I remember inwardly eye-rolling when I heard a teacher many years ago claim excitedly that her pianist “never played the same thing twice.” Call me cynical, but if you’re a pianist for ballet class, you know that if you play week in week out for the same teacher and you’re playing repertoire, it’s almost impossible not to repeat yourself. You’d have to have the memory of a vaudeville savant to literally not repeat yourself.
On the other hand, if by “never playing the same thing twice,” the teacher meant that the pianist constantly improvised, and never came out with the same thing twice, then—call me cynical a second time—I still have my doubts. Improvising, particularly for ballet class, doesn’t mean producing infinitely random streams of notes, like a lottery ball machine; it involves producing music that sounds as if it’s something that you’ve heard before (see a previous post for more on this and 18th century composition primers).
The Quadrille Melodist
I had been thinking about this in relation to a brief exchange of comments on a previous post about “fear of repetition” when improvising. Then today, by coincidence, I discovered a fascinating article [full text free to download] by Nikita Braguinski on a 19th century quadrille generator, called the Quadrille Melodist, that promised to provide “428 Millions of Quadrilles for 5s. 6d.” Never one to pass up a bargain, I had to read it. It turns out, this is a box of cards, in principle not unlike Mozart’s dice game (Würfelspiel)—you have a selection box of one-bar fragments that you can combine to make an eight-bar phrase.
From the photograph, the “machine” looks a bit like one of those clocking-in card racks for employees. Instead of employees, you have bars of music on cards that you can arrange, sort, move and shuffle, but with constraints on which slot they can be placed in, to ensure the musical logic of each resulting piece.
This seems pretty close to what a lot of ballet class improvisation is (and indeed, the creation of ballet enchaînements too): the amalgamation of a limited range of permissible fragments into a more-or-less logical sounding phrase, with strict rules. Obviously the best improvisers (and ballet teachers) do more than this, but that’s a handful, in my experience; and the chances are they did their best improvisations by shedding, not on the spot.
Incidentally, Braguinski points out that its inventor, John Clinton misrepresented—perhaps on purpose—the possible combinations by several orders of magnitude: the correct figure is in fact 7,400,249,944,258,160,101,211, or just over 7.4 sextillion. If he’d wanted to to achieve the advertised figure of 428 million tunes, he could have done it with far fewer cards.
More of the same?
At the end of the article, Braguinski asks an interesting question.
At the same time, the assumption that original or new music is desirable for a quadrille is debatable. Wouldn’t the constant variation run the risk of confusing the dancers? (p. 98).
Possibly not. It depends a lot on what you mean by “the same” or “original” or “new” music, and whether you’re using the music to cue particular movements, or as accompaniment for a dance that you already know (see previous posts here and there [see under “pedadogical category” heading] ) for more on those topics).
Looking at the possibilities afforded by Clinton’s quadrille machine, the outcomes would be different, to be sure, but they are literally formulaic, and so pretty much indistinguishable from one another, as well as totally unmemorable—which is perhaps the price of endless “novelty,” if novelty is the right word at all.
See: Braguinski, N. (2019). “428 Millions of Quadrilles for 5s. 6d.”: John Clinton’s Combinatorial Music Machine. 19th-Century Music
(2), 86–98. https://doi.org/10.1525/ncm.2019.43.2.86