In a studio at École supérieure de ballet du Québec, Joysanne Sidimus is staging the first movement of Balanchine’s Concerto Barocco. A student drifts toward the phrase shape of Bach’s Double Violin Concerto in D minor, following the melody the way a dancer trained on musicality might. Sidimus stops her. “Don’t listen to the music. Count!” The instruction is not a correction of taste. It is a correction of where the score lives. The student has been reading Bach. Sidimus wants her reading Balanchine — and Balanchine, in this room, is a person standing at the front, carrying counts in her body and in a marked-up score, and deciding in real time what the choreography is.
That scene, reported by Elaine Gaertner in a 2022 interview with Sidimus for Vancouver Ballet Society, is the anchor for this essay. It is a small moment — a coaching note, a redirect of attention — but it exposes the structural oddity of the répétiteur’s role. A répétiteur is a person who functions as a score: a transmissible, authoritative record of a choreographic work, consulted in the studio the way a conductor consults a part. The work exists in her, and she is asked to make it exist in others. What kind of document is that? What does it preserve, and what does it quietly rewrite?
What the répétiteur actually does
The word répétiteur comes from the French for one who repeats — the person who drills and rehearses a work into a company. In the Balanchine repertoire, the role is more specific than generic coaching. Sidimus began staging Balanchine’s Serenade in 1969 while a principal dancer with the newly formed Pennsylvania Ballet, having performed the ballet as a corps member at New York City Ballet and as a principal at National Ballet of Canada. When she asked Balanchine in 1970 for permission to stage his works elsewhere, he agreed on one condition: that she first observe every company class he taught, along with rehearsals and performances of his ballets. That observation period lasted nine months.
The condition is worth sitting with. Balanchine did not hand her a score, a video, or a manual. He handed her time in the room. The transmission was to happen through watching, and the authority to stage would follow from having watched. This is a different model of documentation than the one we usually assume when we say a work has been “preserved.” It treats the observing body as the recording medium, and it treats duration of attention as the qualification.
What Sidimus carries into the studio is a composite apparatus. She describes muscle memory as reliable — she danced these roles — but she does not rely on it alone. She keeps copious notes of every step written beside each count of music, and within the musical score she adds rhythmic detail beyond what the printed page carries. The annotated score and the trained body are used together. Neither is treated as sufficient.
She also has access to archival video, and she is explicit about its status: helpful as a reminder, studied before an engagement begins, never used in the studio. Her reason is precise. “A film is only a record of one particular performance and people make mistakes.” A film is a single reading, frozen. If you teach from it, you teach the mistakes along with the intentions, and you have no way to distinguish them, because the film cannot tell you which is which. The living score can.
The annotated count as a notation system
To see what Sidimus is doing technically, it helps to set her practice beside an actual notation system built for Balanchine’s choreography. Kara Yoo Leaman’s choreomusical notation, developed in her 2016 Yale dissertation and explained on balanchine.org, transcribes dance onto a modified musical staff. Spatial height runs on the vertical axis — full or demi-pointe on the top line, standing or walking in the middle, demi-plié on the bottom, with jumps and lifts above the staff and kneeling below. Left and right legs are distinguished by stem direction. Rhythm runs on the horizontal axis, aligned with the music.
The system’s design choice is the revealing part. It maps dance onto music isomorphically, so that patterns in the choreography can be compared directly with patterns in the score. And it prioritizes weight shifts — up, down, left, right — because, as the primer states, shifts of weight are usually the greatest contributors to the sense of choreographic rhythm in classical ballet. Weight placement, rhythm, and vertical position are the parameters dancers themselves mark with their hands when learning steps.
Now put Sidimus’s annotated counts beside that. She writes each step next to its count, and she adds rhythmic detail inside the musical score. She is doing manually, for one repertoire, what choreomusical notation does systematically for analysis: binding a body position and a weight shift to a moment in the music, so that the two can be read together. Her counts are not the phrase structure of Bach. They are a counter-line — the choreographic rhythm laid over the musical one, which is exactly why she tells the student to stop listening and start counting. Listening follows the music’s phrases. Counting follows the dance’s.
The difference between the two systems is what each is for. Leaman’s primer is candid: choreomusical notation is intended for analysis rather than dance preservation. It is a tool for seeing relationships, not for reconstituting a ballet in a studio. Sidimus’s annotated score is the inverse. It is built for preservation and staging, not for analysis; it is not published, not standardized, and not legible to anyone who has not danced the role. Each form of transmission carries something the other cannot. The notation can be audited, compared, and argued with by a third party. The répétiteur’s score can read the room.
The score that reads back
This is the philosophical crux. A written score is inert. It waits. A film is inert in the same way, with the added disadvantage that it records errors without flagging them. A répétiteur is not inert. She watches the dancers in front of her and adjusts the transmission to what she sees. With students, Sidimus says, “we start from a different point than with company dancers.” The same work is delivered differently depending on who is receiving it.
That adaptivity is the living score’s great advantage and its great liability. It is more responsive than notation — it can meet a body where that body is. It is also less auditable. When a staging is disputed, there is no fixed text to appeal to. There is a person, her notes, her memory, and her judgment. The authority of the transmission rests on the claim that interpretation is more faithful than documentation. That claim deserves examination rather than assent.
Consider what Sidimus says the film gets wrong. It records one performance, and people make mistakes. True. But the living score also records one lineage — one dancer’s experience of the role, filtered through her own body, her own teachers, her own decades. The difference is not that the person is neutral and the film is not. The difference is that the person can tell you which parts she is sure of and which she is reconstructing. The film cannot. The living score’s authority comes from its capacity for self-description, not from its freedom from interpretation.
The gatekeeper’s body
There is a political layer here that the aesthetic one tends to obscure. The Balanchine Trust licenses stagings of the works, and according to Sidimus, the technical level of the dancers dictates the Trust’s decision to allow a staging. The répétiteur is the point at which that judgment becomes concrete. She is in the studio; she sees what the dancers can do; her assessment feeds the decision about whether the work will be staged at all.
So the répétiteur occupies a double position. She is the archive — the work lives in her — and she is the gate. Access to a choreographic work is mediated by one person’s embodied judgment of other people’s bodies. This is not a scandal; it is how an oral and embodied tradition necessarily works. But it is worth naming, because it means that the question “what is this ballet?” and the question “who gets to dance it?” are answered by the same person, in the same room, often in the same sentence.
Sidimus describes her job as respecting the dancers while protecting the authenticity and legacy of the repertoire. Those two obligations can pull against each other. Respecting a dancer may mean adapting a count to what her body can do. Protecting the legacy may mean refusing the adaptation. The répétiteur is the site where that conflict is resolved, case by case, without a rulebook. “Remaining positive, offering encouragement, but also being totally honest,” as she puts it — the honesty is the enforcement, and the encouragement is the diplomacy.
What the living score cannot carry
If the répétiteur is a score, she is a score with a specific set of capacities and a specific set of blind spots. She can carry intention, correction, and the difference between a mistake and a choice. She can carry the reason a step is what it is, which no notation system records. She can carry the musical counterpoint Balanchine wanted dancers to hear — the counts that follow the steps rather than the phrases — because she heard him say it.
What she cannot carry is independence from her own reading. Every staging she does is a reading of the work, and every reading is, in some measure, a revision. Not because she is careless, but because transmission through a body is always interpretation. The written score can be photocopied identically a thousand times. The living score cannot. It is reproduced by being performed, and performance changes the performer.
This is not a defect to be fixed by better notation or more video. It is the condition of an art form that lives in bodies. The interesting question is not whether the répétiteur’s transmission is faithful — it never is, exactly — but what kind of faithfulness it offers. It offers a faithfulness that can explain itself, that can distinguish its certainties from its reconstructions, and that can meet a new dancer in a new studio and decide, in the moment, what this work needs to be today. That is a strange kind of document. It is also, for a repertoire like Balanchine’s, the only kind that has ever really worked.
Questions readers ask
Is a répétiteur the same as a rehearsal director? Not exactly. A rehearsal director typically maintains a company’s existing productions across a season. A répétiteur is usually brought in to stage a specific work from a specific repertoire, often from outside the company, and carries authority over how that work is set. The roles can overlap in practice.
Why not just use video to stage ballets? Sidimus’s objection is that a film records one performance, mistakes included, and cannot tell you which is which. Video is useful as a reminder for someone who already knows the work. It is a poor primary source for someone who does not.
What is choreomusical notation for, if not preservation? Analysis. It maps dance onto a musical staff so that choreographic and musical patterns can be compared directly. It is a scholarly tool, not a staging manual.
Who decides whether a Balanchine ballet can be staged? The Balanchine Trust licenses the works. According to Sidimus, the technical level of the dancers is a factor in that decision, and the répétiteur’s assessment in the studio informs it.