Towards a theory of balletic movements as non-metrical analogues of acoustic envelopes

©Jonathan Still with Daniel Jones, October 2002. Last updated: Monday, October 28, 2002 23:42

What little literature there is about the accompaniment of ballet classes frequently makes reference to the importance of metre or time signature to different exercises. Discussions with dancers will often elicit preferences on the basis of metre such as 'I prefer pliés on a three, because it gives you more time', or as Suki Schorer reports in Schorer (1999), Balanchine himself preferred triple to duple metre for pliés because duple metre tended to divide the movement - inappropriately - into two sections. In a debate about whether duple or triple time was preferable for ronds de jambe, one dancer struggled to find a reason why 3/4 was better, and then was delighted to have found what seemed to be a reasonable explanation "we don't like doing round movements to square music".

Metre is thus one of the first questions to arise in the decision making process of accompanying ballet classes, and some teachers' courses make it a criterion for assessment that the student teacher should be able to use a variety of appropriate metres and 'dance rhythms' during the course of a class. This has led to a number of orally transmitted conventions which die hard - pliés, ronds de jambes, adage are traditionally constructed on music of triple metre, for example, and grand allegro likewise, albeit faster.

Dancers usually count in hypermetrical phrases which ignore the lower levels of metrical hierarchy (unless there are small steps to be fitted in here and there), and thus nearly everything that happens in a ballet class is effectively duple in nature. Why, then, should it matter what the lower levels of metrical organisation are? I have tested this theory hundreds of times in classes, playing music in 4/4 for adage where the teacher marked the exercise in 3/4, trusting that what really matters is the metre of the music at higher levels of the metrical hierarchy. Even if this were a punishable offence, there is plenty of mitigating evidence: the Nocturne from Les Sylphides, for example, is binary on top, and ternary below, likewise the famous penchée sequence from Kingdom of the Shades (La Bayadère).

No teacher has ever complained - and why should they? For one thing, dividing beats into three at slow tempi is easier in practical terms, since there is less effort involved in maintaining the tempo (it's harder to count undivided beats slowly and accurately) - if this is the only reason to count in threes, then it won't matter if the actual music does not do this; secondly, counting in threes does not eliminate the possibility that what one is referring to is the accompaniment (i.e. lower level of the metrical hierarchy, in a typical 19th century ballet adagio) not the melody. I suspect that if we had all understood this years ago, we might not hear so many inappropriate waltzes for adagio, where the cantabile line is broken up as the melody coincides with the ternary nature of the metre, rather than soars across it.

But above all, if you look at pliés and adage, there is very little 'metre' in evidence at all. Rhythmically, the common combination of demi-demi-full in a plié exercise gives you - in movement terms - an anapaestic rhythm (in a duple metre) at a very slow pulse. Whatever the musician chooses to fill this rhythm with - 3s, 2s, 6s, 9s or 25ths, the metrical and rhythmic structure of the movement does not change. In fact, these movements are probably difficult to perform precisely because they do not have any metric or rhythmic implications within the movement itself, only when you see the movements in a series.

We then arrive at the following question: If a movement is not metric or rhythmic, what is it then?

By chance, at the same time as pondering such questions as these, I happened to be working with a dancer from English National Ballet - Daniel Jones, who apart from being a musical dancer with an organic sense of rhythm, is also a musician (piano, guitar) who is also well acquainted with music technology. Over a number of sessions, working with midi equipment and sampled sound effects, we had begun to build up electronic accompaniments for some of the most troublesome of balletic movements, those for which the 'right' music seemed elusive.

We approached this task by building up sounds in layers, based on a number of characteristics as we thought of them. A rond de jambe, for example, could be thought to be analagous to the opening of a door, so we sampled a door creaking, and placed it on the beat.. To get a movement on the beat, you need to anticipate it, so we put a rather annoying, high pitched sound slightly before the beat, so as to trigger the movement early. Since concentration and contact with the floor seemed important, we then put an impressionistic wash of sound with a sense of pulse created by wave motion rather than beats.

The aim of this experiment had nothing to do with discovering anything about metre and movement, particularly as we made metrical decisions about accompaniment before we started. Our objectives were to find - if possible - musical analogues of what it felt like to be performing a particular movement, as opposed to musical accompaniments to which the accomplished dancer could fit their movements.

However, as we worked on this project , I began to wonder if our work was perhaps providing an answer to the question I posed earlier - if a movement is not rhythmic or metric, what is it? I realised, for example, that the sounds we were creating - and the way in which we constructed them - were more analagous to movements than metre is, Put simply, if you place a movement and a working definition of metre, rhythm or an acoustic envelope together, and look for correlations, there are more correspondences between a movement and a model of a single tone, than between a movement and any conception of metre or rhythm, as illustrated in the table below.

Once you have established the 'envelope' for the movement, this can be mapped on to a metrical grid (and in effect, this is what dancers do when they perform the same movements to different music) but there is nothing implicitly metrical about the movement in the first place, unless the music has been chosen first. In other words, it is easier to correlate a movement such as a plié with an acoustic envelope (ADSR: Attack, sustain, delay, release) than with any model of metre. Indeed, Suki Schorer's description of how Balanchine wished dancers to perform adage movements such as a developpé approximates closely to an ADSR envelope.

This hypothesis seems to provide a possible answer to a number of questions that have concerned me about ballet training (and I mean ballet, rather than dance generally, since ballet is so closely related to the western art music tradition).

Firstly, I have long been unconvinced that the piano is any longer an effective instrument for aiding musical understanding in dance classes, except at professional level, where dancers have had the experience of dancing to orchestral music and can thus construe what comes out of the piano as an aide memoire. Secondly, I have often wondered what the effect might have been of taking ballet class to music played on the violin, where:

1. it is easier for dancers to perceive and empathize with the effort involved in the production of tones
2. it is possible to play sustained notes in adagio music, where the tone of the note can change and grow over a long period;
3. the difference between upbow and downbow can be clearly seen, and thus one of the most important distinctions in music - between upbeat and downbeat, anacrusis and 'crusis' have a true visual analogue. (1)

Thirdly, I have many times had the opportunity to see the effect of the human singing voice (live) on dancer's performance in class, which I would attribute to (1) and (2) above (and [3] also, if you include the necessity to breathe in before breathing out as a model of anacrusis). Finally, I have also had the opportunity to see how percussion (cymbals in particular) can be a much greater incentive to jump than anything the piano can do.

These four experiences, coupled with the theoretical basis for using acoustic envelopes rather than metre as an analogue of movement, lead me to think that any attempts to improve the teaching of ballet technique - or that elusive thing 'musicality' through music, should place considerable emphasis on the production of tone, as well as an understanding of metre and rhythm (2).

Daniel Jones and I hope, in further experimentation with sound, to find ways in which this might be achieved.

To be continued.

Interested? Disagree? Suggestions? Email us at envelopes@jsmusic.org.uk

(1) With regard to the last point, it is interesting that the words 'levé' and 'frappé' in French are used for 'upbeat and 'downbeat' (arsis/thesis); can one assume that these words, when applied to dance steps, have more levels of meaning for French dancers than for English ones?

(2) Although I would argue further that the Great Nineteenth Century Rhythm Problem outlined by William Nathan Rothstein is nowhere more prevalent than in the ballet class. As illustrated above, ballet classes tend inexorably to duple organization, no matter what might be happening on lower levels of metrical hierarchy, and thus we delude ourselves if we think that using a 3/4 today and a 4/4 tomorrow is educating anyone about metre or rhythm. By the same token, a class conducted to music entirely in duple metre, which nevertheless required the dancer to make metrical or rhythmic alterations themselves, might do more than the class containing a multiplicity of low-level metrical alterations in the music which were rendered duple by the hypermetrical structure of the exercises.

© Jonathan Still/Daniel Jones 2002