Sus une fontayne

Please Stop Playing Bach on the Piano

Vladyslav Nazarchuk

With regards to Baroque music, the experience of the average pianist is this. First, as children, we begin with simple Baroque minuets. Then, we learn the music of J. S. Bach: first his Inventions, then his Preludes and Fugues, then perhaps his Suites, Partitas, and Toccatas. Finally, as adults, we may sometimes perform works by Bach, and very occasionally those by D. Scarlatti, Handel, and Rameau, but nothing else of the Baroque, and certainly nothing earlier.

Quite rightly, we as pianists learn Baroque music, primarily Bach, because it teaches to play multiple voices at once. It also teaches voicing: when simultaneously striking several notes with a single hand, how to play one note louder than the rest. However, why do we never stop to consider, that this music was written not for the modern piano, but for other instruments, primarily harpsichord? Why do we never think to study and perform this music on a historical instrument for which it was actually written?

After all, a piano and a harpsichord are two completely different instruments, with radically different properties. Just because Baroque keyboard music is written on two staves, just like our piano music, and looks playable on our modern piano, gives us no right to play it on a modern instrument so unlike the music’s original intention.

The truth is, Baroque keyboard music sounds objectively better on one of its intended instruments—the harpsichord, for example—than on the modern piano. Most pianists, when hearing their Bach played on a harpsichord, would most likely disagree with this fact; however, most pianists are just not well acquainted with the sound of the harpsichord. When one listens to a fair bit of music on the harpsichord, and especially compares the same piece when played on the two instruments, this fact is plain.

In particular, again taking the harpsichord as an example:

  1. The texture of harpsichord music is too thin, and its range too small, for the piano.

    Notice that in a concert of, say, both Baroque and Romantic music, almost invariably, Baroque music will be played in the beginning, and Romantic music toward the end. This is because we instinctively know that if we play Romantic music first, then the subsequent Baroque pieces will sound small and unimpressive. It is as if to enjoy Baroque music, the audience needs to start from a blank slate, without being primed by music from later centuries. But, if an entire genre cannot stand up on its own, requiring such manipulations of contrast to sound impressive in its own right, surely this suggests that the genre is not being performed in its optimal form, to its highest potential.

    Indeed, when compared with music of later centuries, it is evident that the texture of Baroque music, while perfectly full and sufficient for the harpsichord, is just too thin for the modern piano, and its range is too narrow. There always seems to be this nagging feeling that the piano is not being used to its full potential; that the piano is only half-speaking.

  2. Playing harpsichord music on the piano injects unnecessary dynamics into the music.

    The piano is capable of plenty of fine gradations of volume, and this becomes an inherent part of any music played on it. For example, when playing an upward scale on a piano, you will instinctively add a small crescendo, not to mention the many larger changes in dynamic along the course of the piece.

    Meanwhile, Baroque harpsichord music is complete and sufficient on its own, and does not require the addition of these dynamics. When hearing an upward scale on the harpsichord, the mind recognizes and understands it as an upward line; there is no need for an accompanying crescendo. Except for the occasional forte or piano, if the harpsichord has two manuals, the music sounds perfectly fine on the harpsichord without any dynamics. Adding them, which invariably happens on a piano, is completely unnecessary, and so ultimately is detrimental.

    When playing fugues on the piano, we always play the subject louder than the other voices whenever it appears. Again, this is unnecessary in harpsichord fugues: the mind recognizes when the subject enters without needing additional dynamical contrast. And perhaps the subject does not need to be emphasized every single time it sounds? Sometimes it is there as a compositional tool, to make up the overall musical texture, and is not the most interesting voice.

    If you want to perform some fugues on the piano, note that there are plenty of fugues composed specifically for that instrument which you can choose from, such as those by Czerny, Mendelssohn, and Shostakovich.

  3. The timbre of the piano is too sweet for harpsichord music.

    When compared against each other, the harpsichord has a sharper and more percussive timbre, while that of the piano is smoother and more singing-like. Thus, when a piano is used to play harpsichord music, the sound has too much sweetness—like tea with too much sugar—which ironically makes it ugly.

    Although it doesn’t always carry through in recordings, the sound of the harpsichord is plenty sweet as well, and the harpsichord’s light wooden case gives the sound a beautiful resonance. Listen, for example, to some slow pieces by Rameau or Couperin: the harpsichord can be as mournful as a human voice.

  4. Much of performance practice is lost when transferring harpsichord music to a piano.

    Harpsichord music was historically enhanced by certain performance practices: varied, lush ornamentation, broken chords, notes inégales. But these practices are not as effective on a piano. The same amount and style of Baroque ornamentation, when transferred direcly from a harpsichord to a piano, sounds unconvincing. Breaking chords—a practice that had existed because the sharp percussive attack of the harpsichord can make it difficult to make out all the notes of a chord, but breaking the chord allows each note to present itself more distinctly—is often unnecessary, because the attack of a piano is not as percussive, so all notes in a chord are already more or less audible individually. And notes in a passage on a piano often sound better when played equally, rather than unequally, perhaps because the dynamic interest of a piano compensates for the rhythmic interest that notes inégales create on a harpsichord.

    As a result, where these performance practices were written into the score, they tend to harm the music when played on a piano; while those performance practices that were left implicit are lost to history, depriving us of the musical treasures which they held.

  5. Modern pianos are not tuned using historical temperaments.

    It is unlikely that one would have a piano dedicated exclusively to Baroque music—at that point, one might as well just get a harpsichord instead—so the same piano will be used to play all the piano repertoire. Since it is not easy to change the tuning of a piano, that piano will be always be tuned in equal temperament, thus depriving the listener of hearing Baroque keyboard music in its original meantone temperaments. Meantone temperaments have a significant impact on Baroque music: they make the common keys sweeter and more harmonious, and the uncommon keys harsher and potentially more expressive, and as such were utilized consciously by Baroque composers. But in equal temperament, Baroque music loses this expressive edge, and becomes more bland and mediocre.

    But if one played the music on a harpsichord, the harpsichord would be tuned in a historical meantone temperament, both because it wouldn’t be used to play repertoire which would require equal temperament, and because harpsichords are just easier to tune, and can usually be tuned by the performer himself.

Similar arguments can be made in favor of other historical keyboard instruments against the piano, but all this is only a special case of the more general, blatantly obvious principle, that we should play music on the instruments for which it was written. An a priori proof is this: when a skilled composer is writing music for a given instrument, he writes it to be well-suited for that instrument. If that same music is performed on a different instrument, with significantly different properties, then the probability that the same notes will equally suit the properties of the new instrument, as they did the original instrument, is just very low, mathematically speaking.

Consider this, too, from a listener’s perspective. On a first listen, a piece played on an older, original instrument might seem somewhat ugly or incomprehensible, but this is really only due to a lack of familiarity. As you listen more and more, it is as if a new, wonderful, true world opens up before you: as if walking out of a cave, where you previously had only seen shadows of artificial figures made by the false lights of torches, into the true light of the Sun. And as the man who had finally seen the Sun would not want to return to his torches and artificial light, so you cannot imagine returning to hearing music on modern, unhistorical instruments. This is the case not only with historical instruments, but also historical tunings, musical styles and practices, dances, artistic tastes, languages, and many other treasures of the past.

I do not mean to argue that Bach’s music sounds necessarily bad on the modern piano. Music that is written well and in accord with good musical principles will sound great on any instrument, and in a variety of interpretations. Bach’s music will sound good even on a tin whistle, let alone on a modern piano. But my point is that the music which Bach intended for the harpsichord, among other instruments of his time, sounds much better on the harpsichord than on the modern piano.

I understand, of course, that real life enforces some practical contrains, and that harpsichords are not always available. Few harpsichords existed at all in the 19th and 20th centuries, so it is completely understandable why pianists of that time would perform Baroque music on the piano. Today, however, thanks to the efforts of the “historically-informed performance” movement—which I support wholeheartedly, if that were not obvious—we build historically-accurate keyboard instruments, and know much about historical performance practice. Many music schools and performance spaces have harpsichords available. If you truly do not have access to a harpsichord, it is probably better to play Baroque music on the piano, than to not play it at all. However, we should take any and all opportunities, and try hard to create new ones, to practice and perform Baroque music on Baroque instruments.

Sadly, such a concern does not seem to be at the forefront of most pianists’ minds. Today, you have these famous pianists discussing: how much pedal should be used when playing Bach? Should Bach’s music be played soberly and conservatively, or Romantically and expressively? Should notes be played staccato or legato? And then they take great pride in performing Baroque music at their concerts—which, in fairness, is relatively rare among pianists nowadays—thinking that they are doing well. And not one of these supposed geniuses ever stops to simply think, “Maybe I should perform Baroque music on instruments for which it was actually written? If I like Bach so much, why don’t I perform his music on the harpsichord as he intended, rather than on the piano?” For if we started to perform Baroque music on the harpsichord by default, then all these endless debates of how to perform it on the piano would simply disappear.

Questions and Answers