To the Editors:

Charles Rosen’s article on the Early Music Movement [“The Shock of the Old”] in the July 19 issue is one of the most insightful treatments of the subject I have seen. While Rosen has admittedly always had little sympathy for the use of earlier instruments, his intelligent mind and deep, astute musicality always seem to lead him to brilliant insights and conclusions.

I would like to address what I see as two main problems with Rosen’s assessment of the present situation regarding instruments and performance practices. The first is his conception of what we (and I think I can speak as a fair representative of “Early Instrument” types) seem to think we are doing. Rosen’s article is liberally sprinkled with the word “authentic.” Towards the beginning of the article we find:

Fidelity is no longer enough: a performance must be authentic. The new rallying cry, authenticity, represents a goal simpler and grander than fidelity…. Fidelity demanded of performers a genuine sympathy with the composer’s style. Authenticity dispenses with all this guesswork and uncertainty. It does not ask what the composer wanted, but only what he got.

Near the end of the article we read:

It is clear that I think that the basic philosophy of Early Music is indefensible, above all in its abstraction of original sound from everything that gave it meaning….

I have been playing on old pianos for over twenty years; I have played with the orchestras of John Eliot Gardiner, Christopher Hogwood, Nicholas McCegan and Roger Norrington. On my series in New York “On Original Instruments” in Merkin Hall, I have tried to bring most of the important players of Early Music (capitals not from me; Rosen uses them, although I don’t think the practice originates with him) from both sides of the Atlantic to New York for concerts, singers as well as instrumentalists. I don’t recall that in my dealings with any of these musicians the word “authentic” was ever used to recommend a musical idea or consideration of performance, and I certainly don’t know a single serious Early Music player who espouses the “philosophy of…abstraction of original sound from everything that gave it meaning.” I certainly feel that what I am trying to do is in many ways different from what I was taught by virtually all of my teachers, for using new instruments and new setups between instruments stimulates one’s musical imagination in fresh ways, and I would be happy to find a convenient term to describe this new approach; it seems a difficult task. The whole matter is really quite simple: I play on pianos of the period because I think they serve the repertoire better.

To give a simple example:

The single measure slurs over each bar in this opening passage (indeed, over almost every bar in the movement) are specific expressive instructions on Mozart’s part; all 18th century sources are unequivocal in their meaning; in a two-note figure under a slur, the first note is longer and louder, the second is shorter and weaker. These were commonly referred to as “sighs.” (Note that in bar 4, the “sigh” is quickened.) In my opinion these sighs are the main stuff of this theme. Such a figure is simply not playable on a modern piano, for the modern piano, designed in the 1860s and 70s, has a very slowly developing tone designed for long, uninterrupted singing lines. On a modern piano, when one gets to the second note of the two-note figure, the first note has barely got going; if one plays the second note softer than the first by the necessary amount to make the “sighs,” it will be inaudible—so nobody does it. I have yet to hear anyone play this passage on a modern piano with anything but a “long, singing line,” in which the entire rhythmic lilt of the movement is undermined. As this piece is usually played today it would doubtless seem as strange to Mozart as The Blue Danube Waltz would seem to Strauss if played with no Viennese after-beat. In K. 332/i the rhythmic lilt is a basic part of the musical message, and for that reason I find it not very rewarding to play this movement on a modern piano, where expressivity of this type is restricted.

My second objection to Rosen’s opinions concerns his views on pianos and their development. Rosen seems to know a great deal about historic pianos; his factual assessment of their various qualities shows this. But his conclusions as to what they do for the music seem to me far from the mark. On page 48, for example, regarding Mozart piano concertos, Rosen states:

The problem of balance between piano and orchestra already existed in the eighteenth century. The violins and other string instruments were reduced to one on a part whenever they accompanied the piano, a traditional practice, deriving from the Baroque concerto grosso where a small group of solo instruments were contrasted with an orchestra. The real solution to the problem of balancing piano and orchestra was to build bigger and louder pianos. Piano construction changed constantly during Mozart’s, Beethoven’s, and Liszt’s lifetimes in response to the use of large concert halls, and above all to meet the demands of the music…. It is a familiar mistake to think that a composer writes only for the instruments available to him, a mistake basic to the hard-line Early Music dogma.

This reasoning is faulty. One need only read Mozart’s three famous letters to his father from Augsburg in October, 1777, to see that he considered Stein’s pianos to be exemplary. Nowhere does he complain of their being not loud enough (to be sure, Beethoven does later on). And while it is true that pianos changed considerably after around 1800, it is not true that they changed much during the last quarter of the 18th century. Mozart was not writing for a later, louder piano any more than a composer today writes a flute concerto for some imaginary flute of the future that may be as loud as a trumpet! Indeed, Mozart’s skill in utilizing the solo fortepiano in a way to make it balance with the orchestra is one of the most ingenious and brilliant aspects of these works. Let me cite just a few examples.


In K. 175, Mozart’s first original concerto, Mozart gives the piano the same forte tune at its solo entrance the orchestra had at the opening, and it doesn’t work very well: the piano sounds puny. He never did this again. Every first movement having a piano theme at the beginning of the opening orchestral ritornello has the fortepiano state that theme at its solo entrance. In later concertos the louder and more splendid the opening bars of the first ritornello, the smaller and almost more timid will be the solo piano entrance. In the two grand C Major concertos, for example (K. 467 & K. 503), the piano comes in “in the wrong place,” as if by accident, and of course piano. In the two minorkey concertos, K. 466 and K. 491, the piano states its own, new theme, again piano. The fortepiano that Mozart knew, a not very loud instrument, had to develop its brilliancy as it went along. This is one of the most important psychological aspects of these works.

It is true, as Rosen says, that we have sets of parts from the late 18th century showing that when the solo piano plays the strings are reduced. That may well have started out as a balancing measure, but by the time we get to the mature concertos, Mozart’s genius turns this to great advantage. What we get then, is the grand symphonious sound of the tuttis, vs. the chamber music sound of the solo passages, with piano, individual winds and individual strings trading off roles. And the ubiquitous scale and trill at the end of every exposition and every reprise in every first movement is in part a signal: HERE COMES THE BIG SOUND AGAIN! For even more essential than the sound quality of Mozart’s instrument in these concertos is the actual quantity of volume! To use a “bigger, louder piano,” as Rosen suggests, is to do considerable damage to the balance so carefully worked out by Mozart for those instruments he knew. (In this regard one can only deplore the now virtually ubiquitous practice of using a Mozart-era size orchestra of 25 or 30 in combination with a modern piano. When using a modern piano, it seems to me one should use a fairly hefty string section to compensate for the much louder solo instrument.)

Now in Rosen’s defense I must admit to having played Mozart concertos with original instruments in a way in which the solo piano was somewhat submerged. There are two problems Rosen touches on, and both exist. One is the use of an 18th century fortepiano of limited volume in a large hall; the other is the balance with the orchestra. I have played solo fortepiano concerts in some large halls where it seems to be quite satisfactory, and I have played in small halls that were unsatisfactory; a great deal depends on the acoustics of the hall in question, not merely the size. I have played Mozart concertos in situations where I think the piano has been covered. I think I finally have solved the problem, curiously enough by going back to what I believe Mozart to have done in most cases: sit in the middle of the band, facing out towards the audience, with the lid of the piano removed.1 It took me a long time to realize the advantages of this setup, for it would seem only logical that the piano, out in front with the lid on, would project into the hall better than from the middle of the orchestra with the lid removed. And so it would. What I had not taken into account was the simple fact that with the piano in front (even if the instrument is a modern piano, incidentally) the winds can hear almost nothing and can thus hoot away to their heart’s content, virtually swamping the fortepiano. With the piano seated in the middle of the orchestra, everybody hears everybody else, and several things seem to happen at once:


In the “chamber music” sections there is genuine chamber music, for everyone can really hear each other. During the orchestral tuttis the fortepiano continuo has real significance, for I sit right next to the cellos and we really have the feeling of “jamming” together. In my Mozart Concerto recordings with John Eliot Gardiner and the English Baroque Soloists the piano was in the orchestra with the lid off, and I know that Steven Lubin’s Mozart recordings followed the same procedure.

But balance between any individuals or groups must be worked out on the basis of what is desired to be heard as forefront and background, and what is feasible among the playing forces. If one attends a violin-piano recital, on conventional modern instruments, the question of balance can exist on several levels. Firstly, one can only hope that those days will soon disappear forever where a famous violinist brings along an “accompanist” to play a program of sonatas by Beethoven, Brahms and César Franck—works designated by their respective composers for Piano and Violin. All of us have heard recitals of this kind, and often the pianist is praised the following day by the review for “not being too obtrusive.” I have also heard concerts by pianist and violinist of equal standing where the violinist was drowned out by the pianist who had failed to learn, not to “accompany,” but to reduce the sound of his instrument to make it an equal partner of the violin.

In the case of the Mozart piano concertos, years of hearing them played with (yes I will dare use the word) inappropriate instruments has given us a false sense of what should be heard. It may be that the balance was not ideal in some of the concerts in which I and others on early instruments have participated, but in most of the Mozart piano concertos with modern instruments I hear in public and on recording, especially with the reduced orchestras, the balance is just plain awful, with exciting orchestral parts relegated to the background and pianistic figures, many of which are really written-out continuo, put into the foreground. The comparison to a Brahms piano-violin sonata with the piano “accompanying” in the background is, it seems to me, quite apposite.

I think that no pianist used mainly to modern pianos can have an adequate understanding of the bass register as Mozart and Beethoven conceived it. To quote Rosen:

The heavy thick-sounding bass of the modern piano makes it very difficult if not impossible to achieve the light, detached sonority demanded by many passages of Mozart and Beethoven. On the other hand, the rapid decay of sound of the old pianos would not have allowed them to sustain the longer melodic notes in Beethoven and Schubert, as even modern instruments are barely adequate here…. See, for example, bars 28-29 of the slow movement of Schubert’s last sonata.

Rosen understands that the bass of the modern piano is thicker than that of the old ones but the conclusion he draws, namely that this makes it difficult to render the “light, detached sonorities,” is quite different from my own perceptions in this regard. The bass register of every piano from Stein to Steinway is where the power is, and every composer has used the bass for “guts” (look at Liszt, Bartók or Rachmaninoff’s treatment of the bass register in late 19th and early 20th century pianos). One of the aspects of Mozart on the modern piano that I find the most disturbing is the fact that what is invariably heard when going down into the bass is a decrescendo, in my opinion just the opposite of what the music requires. When Mozart, like Liszt and Rachmaninoff later, uses the bass register in brilliant passages it is more power he wants, not less (cf. for example the E flat Major Concerto, K. 449/iii, bars 57–59).

But the most widespread argument against the use of older pianos is that the tone decays much faster than on modern ones, and the basic presumption is that longer is better. I believe this to be a serious misunderstanding of what the expressivity of pianos was (and to an extent still is). The so-called decay of sound is the only expressive trait the piano can really call its own, and it has been used as such by sensitive composers and players since the early days of its existence. One of the most telling characteristics of the early Viennese instrument was a pronounced attack followed by a rapid decay before the actual singing part began, and lightning-quick leather dampers to stop the tone (faster and more audible than modern felt dampers). One could almost say, although this is a great oversimplification, that the Viennese piano was built to emphasize consonants rather than vowels, and that is their great advantage in passages such as the beginning of the F Major sonata cited above. What astonishes me in the choice of the passage from the Schubert B flat Sonata cited by Rosen [see above] is the fact that the long chord that doesn’t “sustain adequately” even on the modern piano is marked by Schubert with a diminuendo hairpin on the first chord and a decrescendo on the long second chord! (Can a modern piano diminuendo quickly enough for the adequate expressivity required here?—if the tempo is slow enough it probably can.) My opinion would be that the greater the decrescendo, the more expressive the passage.

And where Beethoven, in Opus 90, asks for a successive crescendo and diminuendo on a single sustained note, the instrument that can realize this has not yet been invented.

On the contrary, it has. Both Yamaha and Baldwin have built standard concert grands with transducers instead of soundboards, and there is a pedal that would clearly produce the swell apparently asked for by Beethoven. And of course these pianos are much louder than modern pianos, and could produce, in front of the New York Philharmonic 100 strong, the same kind of balance in a Rachmaninoff concerto that standard modern pianos bring forth in a Mozart concerto with a band of 25 or 30. Whenever I bring up these pianos, most pianists sputter “but that’s not a piano.” It is, of course, very much a piano, but our horror at the idea of electronics being applied to a musical instrument may not be too different from the horror many felt in the nineteenth century when a piano, a beautiful wooden instrument, was to be given a cast-iron frame!

But pianists who tell me that “that’s not a piano” are right in one important aspect; what they are really saying is “I have spent my whole life learning to make music through the limitations of the instrument I know; indeed I have transcended those limitations and that’s what makes me an artist—this new ‘improved’ instrument will rob me of my artistry.” And artistry is what we want from a great interpreter, based on years of developing one’s style through study of the sources, one’s own intellectual and emotional capacities through the instrument (or voice) through which one learns to turn all these facets into palpable musical reality. To assert that “A recording on a ‘Schubert’ pianoforte would add little to our appreciation of the work [the B flat Sonata, Opus Posthumous]” is to deny the possibilities of new artistries gained in this manner. I would hope, playing Schubert on a Graf piano, not to displace Rosen’s, Richter’s or Schnabel’s interpretations, but rather, by a somewhat different route, to enrich the existing knowledge and appreciation of those masterpieces.

Malcolm Bilson
Cornell University
Ithaca, New York

Charles Rosen replies:

It is odd that Bilson should never have heard anyone play Mozart’s two-note figures at the opening of the F major sonata correctly on a modern piano, with the second note softer and shorter than the first, but this is an indication of the restricted world of Early Music in which Bilson exists. This is not a reproach: Why should he listen to anyone perform on a modern piano when he can execute it so happily on his replica of an ancient piano?
Nonetheless, a disturbing note of extravagance appears in his exposition when he claims that, correctly played on a modern piano, the second note “will be inaudible.” This is so patently untrue that one might suspect bad faith on Bilson’s part—but clearly Bilson believes what he says. His inability to hear what all of us can hear needs some explanation.

It is true that Mozart’s indications of phrasing are largely disregarded today—even by some Early Pianists. A similar series of two-note groups is found in the Concerto in E flat major, K. 271, played first in bars 8 to 12 of the first movement by the orchestra and later in the movement by the solo piano. I once performed this work with a well-known conductor and asked him at rehearsal if he would have the violins phrase this theme as I would do later. He replied firmly, “Those are not phrase marks, Charles; those are bowing marks.” It did not help the cordial understanding that should exist between conductor and soloist to reply that if they were bowing marks, they ought not to be in my part; as a pianist, I could not bow, only phrase.

Bilson is right about the neglect of Mozart’s indications. Why, however, is he unable to hear it done correctly on a modern instrument—or, rather, why is he even unable to hear himself doing it when he tries it on a modern instrument? We need not suspect anything wrong with his hearing, or surmise that the second note of the two-note “sigh” is inaudible because Bilson is drowning it out by playing the left-hand accompaniment too loud in order to get at the “guts” of the piano. He finds the second note inaudible on the modern piano because he does not want to hear it. It does not matter if it is softer and shorter than the first: until it sounds exactly as soft as it does on Bilson’s replica, it is not right.

This is the fundamental fallacy of Early Music: the absolute identification of a piece by, say, Mozart with the sound the composer would get, or with the sound he thought he would get—or, better, with the sound that we think he thought he would get (since the most clearheaded of Early Musicians admit that we are not certain of just exactly how those pianos would have sounded). The Early Musicians do not think that Mozart’s intentions can be realized by a modern instrument because they simply equate Mozart’s intentions with a specific sound. It does not do any good to repeat my arguments that such an equation is foreign to eighteenth-century thinking. They are deaf to such arguments because their intentions are not the realization of Mozart’s ideas but the reproduction of an eighteenth-century sound. This makes life easier for them: ideas are nebulous, but a sound is reassuringly material. They try to determine as closely as possible what Mozart’s sound was, and bring it to life again.

What is the matter with this basically noble project can be seen from Bilson’s further reflections on the opening of the F major Sonata. It “is simply not playable on a modern piano, for the modern piano, designed in the 1860s and 70s, has a very slowly developing tone designed for long, uninterrupted singing lines.” It seems to me that “very slowly developing” is a very large exaggeration, but other issues are more important.

When Beethoven was asked what he thought of Mozart’s playing, he replied that it was “too choppy.” This means that Mozart’s highly articulated style of phrasing, with a single slur over almost every bar in the movement, implying a very slight pause or breath at the end of the bar, was distasteful to Beethoven, and one must assume that he himself played Mozart with a longer, singing line, overriding many of the short articulations by which Mozart divided up his phrases. This, no doubt, meant a performance less “authentic” and even less faithful to the composer’s intentions. We must emphasize, however, that the creation of such a new way of playing Mozart is not an irrelevant or perverse imposition by a musical taste at odds with the original, but a genuine response to the text. The performance of Mozart with “a long uninterrupted singing line” is a part of the history of Mozart’s music which cannot be written off and forgotten.

The issue is a complex one for the performer today. Like Bilson, I believe the first few bars should be executed with Mozart’s articulations carefully observed, and in my article I protested against the unthinking conservatory teaching which makes today’s pianists perform everything with a beautiful singing tone—the long, uninterrupted line of Bilson, in short. But what about bars 5 to 8, also quoted above? Mozart has again slurred every bar separately. Unlike the Early Musician, I do not care if Mozart really played it that way: I find such phrasing “choppy,” and an “authentic” rendition here would be pedantic: the more sustained lyricism of these bars requires a tone that contrasts with the articulated opening and needs a longer line.

Furthermore, Bilson’s contention that the single measure slurs “over almost every bar in the movement…are specific expressive instructions on Mozart’s part” is not always tenable; some of them are only perfunctory, as one finds the same phrase written sometimes mechanically with the single slur over each bar and at other times with an indication that it is to be carried over until the next bar. Notation is always more conservative than performance, and composers must have used the old articulations even when a more continuous line was already welcome. Bilson and the Early Musicians may attempt to carry out all of Mozart’s indications literally: Beethoven and I prefer a more flexible approach.

If Mozart is inaudible to Bilson on the modern piano, with Schubert he rejoices in this inaudibility on an old one. A glance at the passage obligingly excerpted by Bilson will show a long, sustained chord: the top note, B, is released at last and descends through A to G sharp. My contention was that even on a modern piano the B has generally decayed almost into inaudibility before the descent can begin. The notation, however, implies that one should hear it move legato downward. On an ancient piano in a small hall, the B can be heard at the end only by the pianist himself, if at all. Even on the largest modern concert grand, it can barely be detected. In any case, the smooth descent implied by the notation, the A coming out of the B to make a continuous decrescendo, is not attainable. The pianist must stimulate the listeners’ imagination into thinking that is what they hear. This effort by the performer to circumvent by rhythm and touch the limitations of his instrument is a normal part of the experience of playing music in public.

Bilson will have none of this. There can be no inadequacies in Mozart’s or Schubert’s pianos. Whatever was, was right. Is Bilson implying that the final inaudibility of the long principal melody note that occurs in most performances is a desirable effect, intended by the composer? To defend this, for the bars of Schubert in question, he supplies an argument which misrepresents Schubert’s notation and, in fact, musical notation in general. Triumphantly, with an exclamation mark, he points out a diminuendo sign and a decrescendo in the passage.

Here Bilson falls into a trap that he digs for himself. He does not take Schubert’s indications as directions to the performer: he thinks they are descriptions of what happens. However, no composer in his right mind directs the performer to do what he cannot avoid doing under any circumstances. He may ask for the impossible (as we have seen Beethoven do to indicate the expressive character) but not for the inevitable. The decrescendo in bar 31 does not refer to the sustained chord in the right hand, as Bilson thinks, and does not direct the pianist to diminish the sound of the chord, because that is what is going to happen anyway, no matter what piano, ancient or modern, is used. The decrescendo directs the pianist to diminish the left-hand figure: insofar as it relates to the right hand at all, it asks for the last note of the bar to be played more softly than the long B of the chord that precedes it—and this is, strictly, asking, like Beethoven, for the impossible, as the B has decayed into literal or near inaudibility and the A will in fact be necessarily louder.2

I am sorry if I appear discourteous to a musician of Bilson’s distinction, but what disturbs me about his approach is its complacency. It seems to me that an appreciation of Schubert’s conception in this passage demands a sense that any keyboard instrument—old or new—is inadequate to render Schubert’s beautiful phrase, that the score implies a chord slowly and gradually diminishing but still sustained long enough to achieve a continuous melodic line—as if by violins or flutes, and that the pianist’s task is to cheat the audience into a belief that the impossible is what they hear. Bilson, on the other hand, is fully satisfied with the sound he produces and the lack of sustaining power does not disturb him.

That is because—like all Early Musicians—Bilson makes a composer’s musical idea absolutely identical with the way it sounded on the old instruments played with the traditions of performance current at the time. He does not ask himself how a passage might be realized: he tries it out on his replica of an old piano and then proves to his own satisfaction that the sound he gets is specifically demanded by the notation, and that any other sound is a deviation from the true way. For him, a work of music is not a flexible structure, but a rigid mold.

This approach seeks to destroy one of the essential conditions of Western music, the tension between conception and realization, the way in which a work of Mozart or Schubert can survive across history and not only tolerate, but be enriched by, changes of instrument and traditions of performance, the way, for example, in which a Bach keyboard fugue can survive, essentially unaltered, its transcription by Mozart for string quartet.

The Early Music movement has enriched musical life today, but its basic philosophy as it is lucidly exposed in every paragraph of Bilson’s letter is an attempt to impoverish it, to stimulate a sense of deep satisfaction with a single kind of sonority, and a single style of interpretation.

This Issue

November 8, 1990