The process

Learning to dance

When it comes to understanding the dance aspect of Bach’s cello suites, we are at an obvious disadvantage compared with eighteenth-century musicians. For them, la belle danse and the bodily paradigm it represented were simply and naturally an integrated part of their everyday lives and music making. For musicians of today, Baroque dance is at best subject to theoretical knowledge, and the only way of approaching a physical understanding of this dance style is painstakingly to learn it at an adult age. I here give an account of my learning process and the supporting disciplines of analysis and reflection I have used to integrate the experiences from dancing in my musical practice. 


My body 

During my studies I practised yoga; later, I took lessons in both Feldenkreis and Alexander technique. I was therefore convinced I was using my body in a reasonably rational way and that my posture was quite correct. Nevertheless, when I started my dancing lessons I became painfully aware of my lack of coordination and general awkwardness on the dance floor, and I had to admit that my posture left much to be desired. I realised that dancing required an entirely different use of the body than does playing the cello, and that, were I to reach an adequate level in the discipline within the project period, I would need some supporting bodywork.

My co-supervisor Béatrice Massin encouraged me to follow lessons in Body-Mind Centering.1 These sessions helped me liberate my movements from excessive thought control, and taught me to move more in tune with the natural energy flow. Along with the different types of warm-up exercises and bodywork routines offered by my dance teachers, this contributed to develop my body awareness; little by little the necessary conditioning for developing my Baroque dance skills was brought on.


Dance classes

During the entire project period, I followed Ana Yepess Baroque dance classes in Paris as well as the weekend courses organised by Atelier Baroque. Ana Yepess classes are intensive and demanding, but she manages to adapt the lessons to the level of each participant. The essentially oral learning style was difficult for me, especially in the beginning, but learning the Feuillet notation and using it to support my memory allowed me to find a modus operandi that helped me learn more quickly and eventually memorise entire choreographies.


New insights

At different points I had physical and musical revelations or sudden insights that had important consequences for the development of the project. The first of these concerned the difference between the musical upbeat and the sink that precedes each step in Baroque dance. Musicians quite naturally think of the upbeat as an upward movement, a lift preparing a downward movement at the bottom of which you find the first beat. In Baroque dance, the effect is just the opposite: the sink’ (plié), a bend in the knees, prepares a rise’ (élévé), or an upward push of the body on the first beat.2 Tomlinson (1735: 144) describes this as follows: [The] Rise or Beginning of the Step, in Dancing, from a Sink always marks Time to the Tune, as well as the fourth or last Note is the Sink or preparative for the Rise or beating Time to the succeeding Step, which no sooner is performd than the Dancer proceeds to the next.’ This dance upbeat, can be compared to compressing a spring that, when released, propels the phrase physically from bar to bar. Once this feeling is experienced and internalised, it becomes difficult to play heavy downbeats in Baroque dance music (see also Marckx 1999: 39).3

Another revelatory experience came during an intensive course with Béatrice Massin in late February 2011 as I was struggling with the transition to the minuet in the Bourrée d’Achille. Béatrice pointed out that my head was not well aligned with the rest of my body, and went on to touch the fontanelle spot on my head while pressing down. At that moment, I became aware that the central body axis really culminates at this point; whereas, I had always believed that it was situated at the top of the scull (in my case, roughly five centimetres further back). This realisation instantly changed the whole organisation of my body, creating a verticality and springiness that I had never felt before. The immediate effect was that the steps became easier to grasp, and I was able to work through the choreography much more fluently.

When I sat down to play the cello on the following day, it was as if my machinery had been equipped with new springs. My whole body felt more alive – tauter and more concentrated. It seemed like a new world of tactile sensations had opened. I am unable to explain exactly what I had learnt, but this particular feeling has been integrated as a point of reference for both my instrumental technique and my musical thinking.

Music and dance



From the very beginning of the project, I had the intuition there might be a direct connection between certain dance steps and musical gestures in the Bach cello suites, which gave me the idea of making a sort of table of concordances to represent this connection.4 As I was learning the dance steps and arm gestures, I tried to relate them to specific phrase segments, or small musical formulae, in the suites. Of course, there can never be a one-to-one correspondence between these elements, since there are always multiple possibilities for attaching a dance step to a particular musical fragment; in the end, I found it more fruitful to use the elements from la belle danse more freely as physical metaphors for musical gestures and phrases.5 Nevertheless, when demonstrated or experienced, certain dance steps and arm gestures provide striking examples that stimulate new insights and new ideas for performing certain rhythms or elements in the Bach suites, and I used these ideas in the fifth part of the exploration section of the dance performance. They also provided me with a vocabulary of steps when creating a choreography for the Sarabande of the Second Suite.6

Dance tempos and how to play Bach

Whenever the relationship between music and dance is discussed, the question of tempo inevitably comes up; indeed, for most musicians, this seems to be the only useful piece of information that can be gained by consulting specialists in historical dance (Little and Jenne 1991: 19).7 Though I believe character is more important than tempo, it is evident that speed does have important consequences for interpretation, and I spent quite some time in my project dealing with this question. In the remainder of this chapter, I will show how I related to the evidence from historical sources. These include tempo indications such as pendulum markings (LAffilard, Loulié, La Chapelle, and Pajot dOnzembray),8 and the heartbeat-related indications of Quantz (1752: 265, 274–75), as well as descriptions in different textual sources.



It is often asserted that historical dance tempos are too fast to do justice to Bachs music (e.g., Gelking 2010) and that the musical content would suffer if performed at such tempos. In support of this argument, some contemporary sources seem to advocate slower tempos when performing dance music away from the dance floor.9 Nevertheless, these texts are often cited out of context and a closer look at three of the most frequently quoted sources actually yields a more nuanced picture.


‘An allemande for playing and one for dancing are as far removed as heaven and earth’


In his Beschützte Orchestre from 1717, Johann Mattheson describes a species of the Chamber Style, which he calls the Symphoniacus Stylus. It consists of all kinds of dances, or suites for keyboard, lute, viola da gamba, violin, and so forth. This type of music, when played alone, belongs to the Stylus Phantasticus (Sind sie schwach und bestehend in Solis, so gehören sie ad Stylum Phantasticum), along with everything that is improvised (ex tempore gespielet wird’). He continues:

The above-mentioned dance types, which belong to the symphonic style, are elaborated with great art, and may not actually be used for dancing. They only have somehow the tempo of the above-mentioned dances, but are much nobler dances. An allemande for playing and one for dancing are as far removed as heaven and earth, and so on, the sarabandes somewhat excepted. (Mattheson 1717: 137–38, my translation)10

Mattheson obviously feels that the principal difference between danced and non-danced music lies in the character and complexity of the pieces themselves, rather than in the tempo of the performance.11


‘one must give the tempo a bit slower because of the sanctity of the place.’


The second quotation comes from the 1er Livre d’Orgues published in 1688 by André Raison:12

One must observe the sign of the piece that one plays and consider whether it has a relation to a Sarabande, Gigue, Gavotte, Bourrée, Canarie, Passacaille or Chaconne, mouvement de Forgeron, etc. and play it the same way as you would on the harpsichord, only one must give the tempo a bit slower because of the sanctity of the place. (Raison [1688] 1899: 7, my translation)13

Raison here speaks about adapting tempo to circumstance, the solemn setting of a church service demanding slower tempos (he may also be taking resonant church acoustics into consideration). Nervertheless, he does not expressly state that we should modify tempo when playing dances on the harpsichord.


‘there are some minuets for the harpsichord that are ordinarily not played as fast.’


The only source I have encountered that clearly speaks of an instrumental dance being performed slower than the dance tempo is Michel de Saint Lambert’s 1702 Principes de clavecin. In explaining the 3/8 time signature, Saint Lambert (1702: 46–47) writes:

About pieces marked by the sign three for eight … it is the custom to beat this measure only in one, so to speak. … This is how one still beats the dance minuets, even if the measure be three crochets, because one plays them very briskly. I say dance minuets; for there are some minuets for the harpsichord that are ordinarily not played as fast. (My translation)14

This last quotation thus only mentions the minuet (and only some minuets at that); it would be difficult to argue that it is valid for all the other dances.

Thus we see that not even these three oft-cited sources unequivocally indicate that all dances should be slowed down when played outside the context of dancing. We are definitely far from the systematically slower tempos that permeate todays performance tradition.


Testing tempos

This was a good occasion to challenge the Bach tempo consensus’; thus, I decided to try rigorously applying historical dance tempos to selected movements from the cello suites. Since most of these tempos are radically faster than the cellist’s traditional tempos, this required a fair amount of practice; but it was definitely worth the trouble. Table 3 shows each movement, the tempos I used as goals, and my notes. I deliberately chose to test only dances for which historical dance tempos exist.15


As we see, I managed to make five-and-a-half out of nine movements work in historically documented dance tempos. In addition, we can safely disregard the source for the allemande, as the tempo from La Chapelle seems to concern a different type of movement.


A special case: the sarabande

In the course of the project period, I observed that, even in quite complicated movements like the sarabande, my perception of the right tempo started converging with the historical dance tempos. For the sarabande (excluding the sarabande vive, or fast sarabande), the historical sources indicate a tempo of MM 68–86, whereas both normal and historically informed performances of the Bach sarabandes tend to be slower: roughly MM 32–50. My own interpretations, before starting dance practice, were situated close to this range: I used to play the Sarabande of the First Suite at MM 52. After a year of dancing, I made a list of tempo ranges that seemed possible for each of the sarabandes, marking down 50–72 for this particular one (see table 4). One year later, finally able actually to dance a few sarabande choreographies (Sarabande pour une femme, La Bourgogne), I made a new test, this time noting only the tempo that spontaneously felt natural. For the first sarabande, this now turned out to be 72, with no technical discomfort. It is remarkable that the sarabandes now felt natural to perform at a tempo that seemed the fastest possible the year before. When I recorded the suites a year later I had finished the speeding-up process; I could back down a little without feeling I lost the dance character. 


Tempo variation

It is often stated that accompanying dancers limits the liberty of the musician, especially where tempo is concerned; I remember vividly the first time I played with dancers, the choreographer repeatedly having to shout, Dont slow down at the repeats! There does exist a margin for tempo variations in dance music, though, and I have found that whenever musicians and dancers mutually listen to each other, this margin can be quite large. From the musician‘s side the tempo rubato needs to be introduced in a way compatible with the dancers movements on the floor: if the dance in question includes leaps, skips, or suspensions, there will be a gravitational limit to the tempo; however, if the steps are marched, the dancer has more liberty to respond to the musicians tempo nuances. Ideally this interplay goes both ways: the musician should also be attentive to variations in the character of the dance and echo changes in velocity and energy by giving way, or hasten the pace of the music. If there is a mutual physical understanding between the musician and the dancer, even ritardandos before repeats can be introduced with success. To illustrate this, table 5 indicates the tempo ranges observed in the final dance performance of this project.


In the course of my project, working with strictly historical tempos in the dance movements of the cello suites, an important question came up. Bach's music contains a wealth of musical detail normally not found in pure dance music. He also keeps the musician busy compensating for the lack of polyphonic texture by requiring chords, leaps, and arpeggios, all of which make the relatively fast dance tempos technically awkward. Could there be a more practical way of applying and practising historical dance tempos in this music?


I ended up with the solution of making melodic reductions, or simplifications. Stripping the music of its abundant written-out ornaments, diminutions, and arpeggios, I aimed to end up with the simplest possible melody, one that could easily be played in a tempo suitable for dancing. Since the purpose was to look for dance character, these simplifications also needed to be reasonably satisfying as dance tunes. I found the procedure needed to be adapted to each movement; the examples below show different strategies for making such reductions.


I emphasise that these simplifications are meant purely as exercises through which to experience the dance content of the suite movements, providing an opportunity to play the ‘skeleton’ of Bach's dance movements in actual dance tempo without having to deal with the technical difficulties.16 Many of the arpeggios and diminutions in these solo pieces carry essential structural and harmonic functions; I am painfully aware that, by removing them, I have deprived the music of much of its sense, not to mention its beauty. On the other hand, making these reductions helped me liberate myself from my ‘canonical’ listening habits, and made me perceive new structures and phrases in the suite movements.17

Menuet 1, BWV 100718

The first minuet of the First Suite is already quite straightforward in the way Bach wrote it, and a simple version comes close to a reductio ad absurdum. 

Figure 5: Simplified version of Menuet 1 from Suite 1.

Gavotte 1, BWV 1012

The first gavotte from the Sixth Suite also has a clear dance structure, but the omnipresent chords pretty much obstruct the use of a proper dance tempo. The tune is still quite dance-like, and when stripped of its chords it comes through as a piece that would not have been out of place in a French seventeenth-century ballroom. In this version, I have added a simple bass part to compensate for the lacking harmonies.19

Figure 6: Simplified version of Gavotte 1 from Suite 6.

Sarabande BWV 1007

In the Sarabande from the First Suite, the chords also tend to ‘get in the way of the tempo. Bach adds some complex passagework and frequently suggests polyphony by leaping from one voice to another. This makes creating a reduction more challenging, and it certainly opens several possibilities of interpretation. Here, in addition to the bass part, I introduce some graces that make it sound more like a typically French sarabande.

Figure 7: Simplified version of the Sarabande from Suite 1.

Courante, BWV 100

The Courante from the Second Suite is generally classified as an Italian corrente, but the continuous movement of the semiquavers bring associations to a double in the French style (see, e.g., Little and Jenne 1991: 139; Ledbetter 2009: 191). I have imagined a French courante as the missing, hypothetical simple. The goal here was to find a way of applying the more moderate tempo of the French courante in this movement, avoiding the tendency to rush that is often heard in performances of this piece.20

Figure 8: French simple version of the Courante from Suite 2.

Sarabande BWV 1011

Apart from a regular four-bar phrase structure, it is difficult to find standard sarabande characteristics in the Sarabande from the Fifth Suite (Ledbetter 2009: 223; Little and Jenne 1991: 107). Tilman Hoppstock (2009: 114), in his brilliant analysis of the lute version (BWV 995) of this suite, proposes a transcription in Handelian style for guitar, which he calls a danced version. I reproduce it here, in keyboard notation transposed to C minor, as an example of another way of using melodic transformation to help find dance character.

Figure 15: Hoppstock’s ‘danced sarabande’ version of the Sarabande from Suite 5.

Dance character

At this point, I think it is interesting to return to the idea of dance character. What is it that gives dance character to a piece of music? What is it that makes it seem danceable even if we are not moving?

In the course of this project I experienced dance character as tacit knowledge, and the physical movements of the dance are now embedded in my body in a way that makes an ‘undance-like’ performance difficult for me. There is no easy recipe for obtaining this knowledge other than experiencing the movements and gestures of la belle danse for oneself. I can, however, suggest a method that will give a point of departure for expressing dance character when playing the Bach cello suites. The method is based on the ‘analytical’ work already presented. It can be used with (preferably) or without the support of Baroque dance practice.



1. Identify the dance type

This can be as simple as reading the title, but in the case of untitled dances, sonata movements, or vocal music, it may involve looking closely for characteristics of a certain dance (time signature, anacrusis, typical rhythms, phrase structure). If there don’t seem to be any, they may be hiding underneath ornaments and arpeggios.

2. Make a melodic reduction, or ‘skeleton’ melody.

Identify and remove written-out ornaments, diminutions, and so forth. This may also help in finding the dance type if it is obscured by a great number of notes. If there really is no dance there, this exercise is still useful for understanding the structure of the music.

3. Play the reduced melody with the proper dance tempo.

See bibliography for sources. Practise.

4. Keep the tempo steady and identify the phrase structure.

Don‘t be afraid of breathing, or marking periods and cadences: they help the dancer orient him- or herself in the music and form an integral part of the composition. These articulations also help identify the exceptional events’ – that is, when periods are absent or the phrases irregular.

5. Put all the notes back.

Make sure to retain the non-essential character of the written-out ornaments and keep the arpeggios connected to the essential core notes of the skeleton melody. Practise.

6. See whether you need to slow down or speed up the tempo.

This may not be necessary; but, in any case, it should not change the character that you established when practicing the reduced melody.


Next chapter

Table 3: Historical dance tempos applied to Bach suite movements.

Table 4: Sarabande tempos before and during the project period.

Table 5: Tempos in the dance performance.