In Debussy's Arabesque #1, mm. 6–7 (among others) are scored as triplets against eighth notes (quavers):

Arabesque #1, mm. 6–7
(Image Source: IMSLP, first edition)

I've always interpreted this rhythm literally, and that is what I've heard in recordings and performances.

However, I came across Debussy's own recording (below), and, short of a painstaking transcription, it sounds more like "right hand as written; left hand arpeggio ad lib." I was left with the impression that Debussy intended this to sound more improvisational, and just notated the "impression" (for lack of a better term) of the sound he imagined.

I'm interested in what scholarship has to say about Debussy's attitude in this regard (literal interpretation vs. improvisation), either specific to this piece or regarding his compositions in general.

Can an improvisational approach to Arabesque #1 specifically, or Debussy generally, be justified by what is known about Debussy's attitudes toward the piece, his music, and/or music generally?

  • If you don't already know it, have a listen to what I think is the definitive (!) version, by Isao Tomita. A few decades old now, but still fresh as a daisy.
    – Tim
    Commented Aug 2, 2023 at 7:16
  • @Tim I just listened to about 20 seconds. I can't tell if you're joking or not.
    – Aaron
    Commented Aug 2, 2023 at 7:19
  • Certainly not joking. There's a whole album dedicated to Debussy, all translated for electronic instruments, quite new at the time. Followed up by a (banned) Holst Planets set, which sister Imogen hated, thus banned. I don't care much for the latter, but the former is, in my opinion, an excellent interpretation. It is, after all, from the Romantic era. Thus lends itself really well, imo. Maybe a Marmite thing...
    – Tim
    Commented Aug 2, 2023 at 7:24
  • Interesting, like the opposite of "Left-hand-strict-right-hand-rubato" Chopin. I've definitely noticed other pieces in which a quirk of performance becomes established as "that's the way everybody does it," but the composer's original recordings differ. Bartok/Szeryng's "Roumanian Folk Dances," Copland's very brusque and jazzy Rhapsody in Blue, etc. Commented Aug 2, 2023 at 13:00
  • @AndyBonner - isn't Rhapsody in Blue Gershwin, not Copland?
    – Tim
    Commented Aug 2, 2023 at 15:55

1 Answer 1


First, note that this is not a recording in the modern sense, but a piano roll. Accomplished pianists are highly adaptive to nuances of tone colour, clarity, attack, balance between different registers, etc, which can vary widely between different instruments. Pedalling effects are even more of a minefield. As a result, many details of the performance may come across differently from how the player originally experienced and intended them.

But also it seems highly doubtful that the performer on this particular track was Debussy himself. Debussy did create some piano rolls, which have been recorded on a restored player-piano and released under the title "Claude Debussy: The Composer as Pianist" on the Pierian label in 2000.

The track you referred to is from a release of more murky provenance on the Favourite Classics label, entitled "Claude Debussy Plays His Finest Works". It appears that several of the tracks on that recording are in fact not performed by Debussy himself. The correct attributions may be the ones listed here, in which case the original performer for the first Arabesque on your track is Lev Pouishnoff.

See the following links for more discussion. It would be nice if someone could clear all this up more authoritatively!:




Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.