To take a harpsichord concerto by Johann Sebastian Bach and arrange it for a four-part chorus, organ, and orchestra would not, for most music lovers today, be considered the proper way to realize the composer’s intentions or even to show decent respect for the score. Yet this is what Bach himself did to his own harpsichord concerto in D minor—which was, incidentally, in its original version a violin concerto of a somewhat simpler cast. The ideal of performing a work as it would have been done during the composer’s lifetime or even by the composer himself gives rise to unexpected considerations, of which this is an extreme case, but by no means a rare one.

The effort to revive ancient instruments and early performance practice is not strictly modern: it can already be found in the first half of the nineteenth century. Early in our own century Arnold Dolmetsch and Wanda Landowska became major public figures with their championship of the harpsichord. It is, however, during the past two decades that the “Early Music” movement has taken on the character of a crusade, above all as it has moved beyond the sphere of medieval and baroque music and into the late eighteenth and the nineteenth centuries. Early Music is still a good name for the movement even now that it has reached Mozart, Beethoven, and Chopin and is looking to Brahms and Debussy: the goal is to make these composers sound more ancient than we had imagined.

The success of the crusading spirit is undeniable: it can be measured by the extent to which it has imposed a new orthodoxy. In the days of our innocence, what we wanted was a performance that was technically perfect, effective, beautiful, moving, and even, for the most idealistic, faithful to the work or to the intentions of the composer. Fidelity is no longer enough: a performance must be authentic.

The new rallying cry, authenticity, represents a goal simpler and grander than fidelity: it is aptly modern in that it transcends the composer’s intentions, or at least circumvents them. The old ideal of fidelity demanded that the performer try to infer the composer’s intentions, and realize them with the least possible distortion. In a faithful interpretation, the performer’s own personality and his need for expression come into play essentially as a medium through which the work can be made public; the performer’s style would be capricious, willful, lyric, or dramatic as the work demanded it. Fidelity has its dangers, as the performer identifies himself only too easily with the composer, convinces himself without difficulty that the composer would have approved such-and-such a cut, been delighted with this accent, made an expressive relaxation of tempo in just that place. Nevertheless, 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. Intentions are irrelevant. (Some performers of Early Music now claim to return to the study of intentions, but the concentration is still on what was actually heard.) We no longer try to infer what Bach would have liked; instead, we ascertain how he was played during his lifetime, in what style, with which instruments, and how many of them there were in his orchestra. This substitutes genuine research for sympathy, and it makes a study of the conditions of old performance more urgent than a study of the text.

The success of the battle for authenticity is well merited, above all when one considers the contempt with which most professional musicians and critics only a generation ago greeted the efforts to revive old instruments and old ways of playing—the present intolerance of modern instruments is a natural reaction. Nevertheless, a new orthodoxy inevitably provokes doubts, inspires heresies. The hostility to Early Music is no longer as significant as the dissension within the ranks. Things are not as simple as they seemed in the earlier days of the movement, and the certainties of some decades ago have evaporated. We are no longer so sure that Scarlatti, wrote his five-hundred-odd sonatas for the harpsichord and not for the pianoforte; there is no agreement on whether Bach’s rhythms are to be executed in French style, with an irregular swing to it. There is fierce controversy about tempo in Mozart and Beethoven and about improvised ornamentation in opera and instrumental music. Above all, as our knowledge has increased of the wide variations in performance practice that coexisted throughout the eighteenth and nineteenth centuries, it has become less and less evident what it means to return to the style of playing current during a composer’s lifetime—what, in short, an authentic interpretation would be.

A collection of essays called Authenticity and Early Music, skillfully put together and edited by Nicholas Kenyon, confronts some of the issues raised by Early Music and glances at others. All of the contributions are intelligent and stimulating. Will Crutchfield of The New York Times writes of the role of the performer, and of the necessity to substitute personal conviction and freedom for the arid realization of simple rules of historical performance. Philip Brett of the University of California at Berkeley discusses the problems of editing, and what it means to prepare an authentic text. He describes persuasively how even the most neutral and apparently inoffensive modernization of old forms of musical notation can alter our conceptions of the music and influence the performance. Even the simple reproduction of the old notation will not lead directly to authenticity; just the act of printing a score may lead to misunderstanding when the work as we have it was intended to be executed only a few times by a special group of singers and musicians and only under special circumstances.


Such circumstances are dealt with by Gary Tomlinson of the University of Pennsylvania, in an essay on Angelo Poliziano’s mythological play Orfeo, written at Mantua probably in 1480. Tomlinson claims that the “authentic meaning” of the music depends on a reconstruction of the historical circumstances and background of the work. He confesses to “some polemical mischief” in his choice of example, since the music to Orfeo has not survived, and may not have been written down in the first place, but only improvised by the singer, Ugolini, who portrayed Orpheus. Tomlinson gives a brilliant exposition of the probable influence of contemporary Platonism on Ugolini, with its idiosyncratic views of music, magic, and poetic frenzy. He does not speculate on the ways in which these ideas might have affected either the music that Ugolini may have improvised or the manner in which he performed, and perhaps Tomlinson is right not to attempt such speculations.

This makes his argument, however, an evasion of everything of major importance to music lovers, to musicians, and even to most historians of music. No one denies the interest of an initial study of the historical conditions in which a musical work was created; the point of difficulty, however, has always been to apply this knowledge to the music itself and to how it was once, and might be again, performed. There is some point to a savage remark by Richard Taruskin of the University of California at Berkeley (in a later essay in the same book) that by Tomlinson’s “lights an ‘authentic’ performance would seem to be a performance accompanied by a good set of programme notes.”

Taruskin writes brilliantly and at the top of his voice, and his most crushing arguments are often reserved for opinions that no one really holds. He asserts: “To presume that the use of historical instruments guarantees a historical result is simply preposterous.” No doubt. Still, Taruskin beats his dead horses with infections enthusiasm, and some of them have occasional twitches of life.

His main thesis, repeated here from earlier articles, is that the Early Music movement is not a genuinely historical crusade at all but a variety of modernism, an attempt to make the music of the past conform to the austere aesthetic that we associate with Stravinsky and his successors, and to make it sound astonishingly different in order to achieve that shock of originality demanded by the modernist ideal. This does, indeed, describe and clarify certain aspects of Early Music with great precision, and Taruskin demonstrates his thesis with easy conviction. As he says:

Changes in performing styles in the twentieth century, no less than in past centuries, have been allied with changes in composing style, and with more general changes in the aesthetic and philosophical outlook of the time.

When he adds, however, that “a multiplicity of styles is always available in any present, of which some are allied more with the past and others with the future,” this seems true but is only too pat. To group styles into reactionary and progressive as he tends to do is not helpful either: they all reach backward and forward. Taruskin’s view of modernism is too narrow, and is fueled by his hostility to many aspects of contemporary art. Modernism in music is not confined to the hard-edged neoclassicism of Stravinsky, but has its neo-Romantic side in Schoenberg and Berg which reaches into the work of Elliott Carter and Karl-Heinz Stockhausen and even into much of Pierre Boulez. In many ways, Fürtwangler, classified by Taruskin as a ghost from the past, was as “allied with the future” as Toscanini was and John Eliot Gardiner is today. In addition, Taruskin is curiously grudging; he does not want to admit that our greater knowledge of performance practice and instruments of the past can have a beneficial effect, and do more than give an unwarranted sense of superiority.


The relation of Early Music to the past is seen in a more sophisticated light by Robert Morgan of Yale University. The successive revolutions of style imposed by modernism in the twentieth century from Schoenberg and Stravinsky to Karl-Heinz Stockhausen and Pierre Boulez have made not only the general public, but many professional musicians, feel that the apparent continuity of tradition from Bach to the present has now been interrupted. Even the most unequivocal successes of modernism have not yet taken deep root in our musical consciousness: only fifteen years ago I still saw Philharmonic subscribers stalk out noisily in protest during Alban Berg’s delicate and enchanting “Post-card” songs. Those who accept Stockhausen or Philip Glass do not generally think of them as sharing a common language with Mozart. The works of the eighteenth and nineteenth century have become museum pieces that we continue to love but no longer think of as playing an active part in the creation of new music. As Morgan observes:

All this suggests that the past is slowly slipping away from us. It is no longer ours to interpret as we wish, but ours only to reconstruct as faithfully as possible.

Historical accuracy for Morgan as for Taruskin is both a product of modernism and a reaction to it: a way of making the past different, both old and new at the same time.

Historical accuracy in performance matters a great deal to Howard Mayer Brown of the University of Chicago, and he sketches the history of attempts to achieve it. His style may be all velvet paws, careful to offend no one, but he approaches his subject with considerable sophistication and sharpness. The nature of his skepticism is sufficiently shown by the title of his contribution: “Pedantry or Liberation?” He holds the balance amusingly between these two tendencies of the Early Music movement.

Brown is the only contributor who sees the fundamental distinction between music that demands a reconstitution of the original style of performance with the original instruments and music that is susceptible to reinterpretation by different styles and sounds.

The truth—although it may seem controversial to say so now—is that it is more acceptable to play Bach’s music on modern instruments than Rameau’s, for it can be argued that authentic sonorities and old playing techniques are less important in the one than the other, and that therefore the essential nature of Bach’s music can emerge in a performance that translates the original into modern terms. In short, it is possible to defend what might be described as a woolly-headed liberal approach to the question of repertory.

It is refreshing to get away from a monolithic view of music, to be reminded that the “essential nature” of one composer’s music is not the same as another’s, and that the possibilities of realization in sound will be more varied in one case, more restricted in others. This flexibility of approach, however, makes life difficult for musicians: it is much more convenient to assume that there is only one ideal sound for each work of music—either the sound the composer imagined in his head as he wrote, or the one he knew he was going to get from current instruments and contemporary practice—and that the goal of the responsible performer should be to renounce the delights of imagination and realize this ideal sound as closely as possible.


What goes largely unrecognized or unstated in these essays is that the basis of the Early Music movement—aesthetic as well as commercial—is recording. Only Nicholas Kenyon in a parenthesis in his introduction seems to notice that there is something interesting about this:

(It is of course ironic, and a comment on the whole “authenticity” business, that most of its artefacts are in the extremely inauthentic form of recordings without audience which sound the same every time one plays them.)

(“Most of its artefacts”? An odd phrase.) Indeed, the recent success of Early Music was originally stimulated by recordings, and it is still today largely sustained by the recording industry.

The expansion of Early Music came when it invaded the symphonic repertoire of Mozart and Beethoven to play it with the original instruments, as they were before the violins had had their necks lengthened to give the strings extra tension, before wind instruments were made considerably more powerful. The trouble was that twenty years ago few musicians knew how to handle these old instruments, which tend to go out of tune much more quickly than modern ones and to emit strange grunts, squeaks, and quacks in inexperienced hands. The only solution was magnetic tape, which could be pieced together in thirty-second segments so that the orchestra could retune every half minute or so and the more objectionable noises could be edited out.

By now there are many performers who have become expert enough with these ancient instruments, and even learned to keep them better in tune, although not, of course, quite as well in tune as modern ones. Tuning had always been a problem, and we must not imagine that no one was troubled by it in the eighteenth century and earlier. “At the beginning of the concerts,” wrote Saint-Euremond in 1684, “we observe the accuracy of the chords; nothing escapes of all the different variety that unites to form the sweetness of harmony; some time after the instruments make a din; the music is for our ears no longer anything but a confused noise, which allows us to distinguish nothing.” Nevertheless, in spite of the more convincing sounds now made in public by orchestras made up of ancient instruments (or replicas of them), the commercial interest of Early Music is largely confined to records.

The reason for this is partly acoustics, partly economics. Eighteenth-century instruments (except, of course, for organs) were largely meant for halls with a capacity of a few hundred people. In typical modern halls seating fifteen hundred to three thousand they sound thin and puny, and their more delicate nuances are lost in space. They can of course be amplified, but that radically alters and in fact, homogenizes their tone colors. With an amplified harpsichord, for example, nuances of registration are obscured, and different sonorities begin to sound alike. The modernization of instrumental construction was inspired largely by the growth of public concerts and the gradually increasing use of larger halls.

Recording takes almost no account of volume of sound: maximum levels (and minimum as well) remain approximately the same for all records, whether of an orchestra and chorus of a thousand or of a solo flute. I have always found, when a radio station follows the playing of a recorded symphony with a record of a single harpsichord, that I must rush to adjust the dials, because the blast of sound from the harpsichord is not only as loud as the symphony but much louder. The range between the softest and loudest points is much smaller with a harpsichord than with an orchestra, and sound engineers generally set their levels as high as is compatible with the system of reproduction, since this gives the greatest possible fidelity. Good microphone placement can make three violins sound like a dozen, and the recording industry allows Early Music to rival the power and the volume of more conventional orchestras.

In public, “historical” orchestras can compete with the standard bands only if they can afford to muster the necessary forces. It is well known that Handel called upon forty oboes and as many bassoons with the other instruments in proportion, but this of course was on a special occasion and outdoors. Nevertheless, Mozart enjoyed doing his C major Symphony, K. 338, at one indoor concert with forty violins, eight oboes, eight bassoons, twelve double basses, and the rest to match. This would unmercifully stretch the purse of any Early Music society.

The other solution—playing in smaller halls—is also not economically workable, not at least without a heavy subsidy. Half a century ago a pianist or a string quartet could make a living playing to audiences of five hundred people: those days are long gone. Travel and advertising expenses have risen disproportionately to any acceptable increase in the price of tickets. It costs more than eight hundred dollars in New York just to move a concert piano across Fifty-Seventh Street from Steinway to Carnegie Hall. Established symphony orchestras and opera houses have their regular sources of fund raising, and often their traditional access to government subsidies. Early Music societies are still somewhat in the position of outlaw raiders: they depend on the aid of the record companies to force their way into the mainstream of the classical music business.

With the success of their recordings, Early Music has come to seem less radically different—objectively so, indeed, since the greater expertise at handling old instruments in recent years, coupled with the ability of the sound engineers to make a chamber group sound like a large symphony orchestra, has brought Early Music performance much closer to conventional sound. In fact, the more professional an Early Music orchestra becomes, the more it sounds like a conventional one; in addition, the most successful conductors of Early Music—Nikolaus Harnoncourt, Christopher Hogwood, Roger Norrington, and John Eliot Gardner, have turned increasingly toward engagements with established orchestral groups. One has to listen carefully now to a recording of a Handel concerto grosso or a Mozart symphony to decide whether it is an execution on old instruments or a regular symphony orchestra that has picked up some of the new ideas on old performance practice from Early Music. (Of the fruitful influence of the authenticity crusade on general performance there can be no doubt.) In large halls, however, the public (resembling, in this respect, the audience at rock concerts) is often forced to rely on its memory of what the record sounded like.

We generally listen to records with the hope of hearing the acoustical effect of a large concert hall: at Early Music concerts we hope to hear what we remember from playing the record in our own living room. This is particularly true when we are dealing with late eighteenth- and early nineteenth-century pianos. In a recent article in this journal (May 18, 1989) on the Mozart piano concertos, Joseph Kerman wrote at length on Malcolm Bilson’s recordings on an eighteenth-century piano, but said not a word about what these performances are like in public. Bilson’s musical qualities remain the same in a public performance as on disc; but when one is using late eighteenth-century instruments (or replicas thereof), it is much easier on disc than in a live performance to make the piano stand out clearly from the orchestra, and to avoid its being overshadowed by the other instruments.

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 a small group of solo instruments were contrasted with an orchestra. 1 The real solution to the problem of balancing piano and orchestra was to build bigger and louder pianos.

No other instrument changed so radically between 1780 and 1850, and from 1790 on a wide variety of instruments were produced, all called pianoforte, fortepiano, or Hammerflügel. They all had a lighter bass and a faster decay of sound than modern instruments, but the difference between a Viennese piano and an English piano in the 1790s was far greater than that between a Steinway and a Bösendorfer, to take the two most commonly called upon today for public use. Piano construction changed constantly during Mozart’s, Beethoven’s and Liszt’s lifetimes in response to the use of larger concert halls, and above all to meet the demands of the music. Not only did pianos become bigger and louder, but they also had enlarged keyboards with higher and lower pitches.

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. In a sense, composers already used the extra high and low notes before they had them. The existence of these notes is implied in the music. There is an irritating or piquant wrong note in the first movement of Beethoven’s first piano concerto, a high F-natural where the melody obviously calls for an F-sharp. Pianos did not yet have that F-sharp when Beethoven composed the concerto, although they did some years later when he wrote the most interesting of his cadenzas for the work and employed the new high register. We know that Beethoven himself would have played the F-sharp, as he once announced his intention of revising his early works in order to make use of the extended range. To the new aesthetic of authenticity, his intentions count for very little: to employ an instrument contemporary with the work would mean not only playing a note evidently wrong but even renouncing the best cadenza.

The heavy thick-sounding bass of modern concert pianos 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.2 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.

From a common sense viewpoint, the most nearly adequate instruments, although still imperfect, might be those whose construction was inspired by the music rather than those that the composer was forced to use. The desire for authenticity, however, is driven by idealism rather than common sense. I have even heard a pianist on a modern Steinway play the clearly wrong F-natural because that is what Beethoven wrote. An authentic wrong note is evidently better than an inauthentic right one. With a piano of the time, Beethoven’s public could easily see that the F-natural was the last note on the keyboard, and sympathetically supply the necessary sharp in imagination. The wrong note has a pathos in this case, but only if one can see it—and of course this pathos is not available on a recording. One of the chief advantages of using an early piano is that the public can appreciate the way Haydn, Mozart, and Beethoven used the upper and lower limits of the keyboards for the most powerful climaxes. The visual aspect of the performance on old instruments may seem a trivial point, but the dramatic effect of striking the highest or lowest note on the keyboard was an essential part of the musical structure—an effect that can be realized neither on a modern instrument nor by the recording of an ancient one.

The forced alliance of Early Music and the recording industry is more than odd: they have a natural affinity for each other. Recording fixes the sound of a single performance and enables one to reproduce it. Early Music seeks to ascertain the original sound of a work of music and to reproduce it. Both are admirable endeavors with what may seem like minor flaws. Combined, these flaws have a pernicious effect, doubly dangerous because largely unconscious. Both of them reduce a work of music to its sound, tend to abstract it from all of its social functions, from any interaction with the rest of culture—all of this with the innocent conviction that nothing essential is lost in this process of abstraction, and without the slightest suspicion of how profoundly the nature of music is altered.


To try to figure out as well as one can how a composer thought his music would sound and do one’s best to realize this in performance—what a simple and reasonable program! Its attraction needs no explanation. What is astonishing is that almost no one ever thought of it before.

Of course, people were interested in how a composer played his own works: both Czerny and Schindler described, accurately or not, the way that Beethoven played, and conveyed his ideas of style and execution. Neither of them ever suggested, however, that we return to the instruments available to Beethoven when he originally wrote his works—just as Beethoven in later life would have found incomprehensible a recommendation to return to the pianos of the 1790s for his early sonatas instead of using the more powerful instruments at hand after 1815. Not just the volume but the tone quality of the pianos had changed.

Czerny described not only how Beethoven played his own works but how he played Bach’s Well-Tempered Keyboard as well. He did not consider Beethoven’s way of playing his own sonatas and concertos more authentic but more authoritative—most of all because it was better, just as he thought Beethoven’s way of playing Bach was better.

This apparent indifference to the actual sound of an old piece when first performed (or at least to certain aspects of the original sound) went along with a dedication to selected details of the original style, and this coupling has been an integral part of Western music since the eighteenth century. It has even deeper roots within that tradition than Robert Morgan suggests when claiming that the old sense of a living and unbroken musical language from past to present has enabled us to translate the sound of the past into the new sounds of contemporary instruments. It came also from the fundamental idea that musical composition is partly divorced from the actual realization in sound.

When Bach transcribed his D minor harpsichord concerto for organ and chorus, he made no attempt either to preserve some of the qualities of the original sonorities or to give the work a more choral or organ-like character. He just rescored an abstract set of pitches and rhythms for a different sound. He did remove the keyboard style ornaments from the melodic line of the slow movement—but ornamentation was considered precisely as belonging not to composition but to the realization of the sound of the notes, and it was often left entirely to the performer. (If Bach’s choruses were performed with only one singer on a part, as Joshua Rifkin has insisted, then some ornamentation might have been restored.) Bach in fact irritated his contemporaries by writing out the embellishments and so constraining the imagination of the performers.

When Mozart arranged his wind serenade in C minor K. 384a as a string quintet (K. 516b), he was only interested in making it possible for five string instruments to execute the notes originally given to eight winds. Considerations of the actual sound played little part in the arrangement. Similarly, when Beethoven arranged his violin concerto for piano and orchestra (a task which had a financial interest for Beethoven and, except for the cadenza, no musical interest for anyone), he gave the violin part to the pianist’s right hand, allowing the left to strum the harmonies. No thought of making the work pianistic entered the composer’s head. Nor, on the other hand, did he think of finding a way for the piano to create a violinistic sonority.

This does not mean that composers did not care how the music sounded, but that composition and realization were not identical, and that while composition was more or less fixed, realization was more or less open. Mozart’s Symphony in G minor, K. 550 exists in two versions, one without clarinets, but it is the same composition. Mozart no doubt preferred the version with clarinets, since he used them elsewhere when he could get them. The clarinet concerto was originally sketched for basset horn, with the structure and details of the solo part fully worked out; reconceiving it for clarinet did not entail any process of recomposition. It is clear that for Mozart as for Bach, composition did not entail fixing all aspects of the sound in advance.

Things changed radically in the middle of the nineteenth century. In 1838, when Liszt arranged the Paganini caprices for the piano, he not only rethought the music to make it sound spectacularly pianistic, he also rewrote the caprices to allow the piano to sound like a violin, and to reproduce effects like spiccato and pizzicato on the keyboard. Other composers (Scarlatti, Handel, and J.S. Bach, for example) had imitated horns, trumpets, tympanis, and guitars with the keyboard, but only with original works: their transcriptions from other instruments never bother to evoke the original sound. Liszt, however, in his transcriptions of the Beethoven symphonies was unbelievably ingenious in imitating the original instrumentation on the piano. Liszt also rewrote his own music all his life: if the two versions of Mozart’s G minor Symphony are the same piece with two different sonorities, that is no longer true of Liszt’s multiple versions. There is almost as much difference between his two piano versions of Sonetto 104 of Petrarch as there is between works with different titles. For Liszt, realization—actual sound, in short—was becoming a more essential and intimate part of composition.

From the eighteenth century to the present, many of those aspects of realization left to the decision of performers (guided by the relatively loose, but never completely free, performance practice of their time) were gradually incorporated into composition. With Beethoven’s work, for example, the traditional forms of embellishment—trills, grace notes—have become part of the basic musical structure; they now figure among the work’s motifs, and nothing more can be added by the performer—Beethoven was known to make a disagreeable scene when a performer dared to add a few trills.

Even what might seem like trivial aspects of realization could become essential to a musical conception later in the nineteenth century. When Brahms made a transcription for piano of Bach’s Chaconne for solo violin, he scored it for the left hand alone, partly as a finger exercise, and partly, as he confessed, because it made him feel like a violinist as he played it, determining all the pitches with his left hand. Not only the sound of an instrument became an essential element of composition, but even the physical experience of playing. Liszt’s transcription of Paganini’s Caprice in G minor also begins the principal melody with the left hand alone for a similar effect. Like the promoters of Early Music, Brahms and Liszt were trying to preserve a part of the original experience of the work that would once have been considered marginal. Brahms, in fact, is the real precursor of Early Music: when he edited the works of Couperin, he insisted on using all the old clefs, by then extremely difficult to read for contemporary musicians, and was only prevented by the refusal of his coeditor, Chrysander. Brahms wanted the text to look archaic: it is only a step from wanting it to sound archaic as well.

Early Music’s narrow concentration on the sound the composer would most probably have heard rather than on the sound that he explicitly said he would like to hear represents a giant step forward and an almost equal step backward. It is an immense progress because every composer writes to a great extent with the performance he is likely to get in mind, even if he is deeply dissatisfied with the performance practice of his day and with the actual forces and instruments that were available. The notation of music can only be understood in terms of how people really played; and the body of knowledge gained about old music in the past two decades and the experience of playing the obsolete instruments for which it was written has been an inestimable boon to musicians and scholars alike.

The authenticity movement is also a regression because many composers write partly with the hope of an ideal performance which transcends the pitiable means and degenerate practice they have to compromise with. New ways of writing create new ways of playing, and composers are often unaware that the music they have just set down on paper demands a new style of performance. If one could resurrect both Mozart’s performance of his own Concerto in D minor, K. 466, and Beethoven’s performance of the same concerto (which he is known to have played, and for which he wrote a set of startlingly innovative cadenzas), which one would seem to us to reveal more adequately the character and originality of the work?

Yet some of the most distinguished practitioners of Early Music have openly professed a policy of forgetting everything that happened to performance after Bach or Haydn or whomever they happen to be playing. Walter Benjamin was right to characterize the origin of this kind of historical outlook as melancholy and bitterness of heart. It springs from a rejection of the contemporary world. For many Early Music enthusiasts, the alienation was triple: a rejection of avant-garde music, of the standard performance of the classics, and of the conditions of professional musical life.3 A large number of the most radical seekers after authenticity as late as the 1950s were amateur musicians, for whom Early Music was part of a style of life that included playing the recorder, eating brown rice and whole wheat bread, and making their own clothes. The movement has now developed into a genuinely professional and profitable one, but some of the prejudice against the modern world remains with it.

Much of the prejudice was, and is, justifiable. The conventional aspects of all standard performances of music from Bach to Debussy—the customary insistence on producing a “beautiful” or “expressive” sound, for example, even when that kind of expression or beauty may be unsuited to the music being played—can become repulsive, doubly offensive when the performers are truly accomplished. When we realize how much of even the finest music making is mechanical rather than spontaneous, it needs only a moderately refined or jaded sensibility to find little to choose between artists as opposed as Toscanini and Furtwängler. The original impulse behind Early Music was a desire for a thoroughgoing renewal. Nevertheless, to refuse to come to terms with the way Bach, for example, has been interpreted and misinterpreted through time, to see how his work carried the seeds of its own future, is to shut oneself off sadly from the real life of music. The search for “authenticity” is often forced to remain content with the ways the composer’s contemporaries understood his music, and this can amount to perpetuating misconceptions rendered obsolete by two centuries.


The true sadness of “authenticity” is that its scholarly and idealistic interest in the sound of music as it once was divorces music and realization (or performance) from the life of which they were, and are, a part. Even the way many practitioners of Early Music conceive the original sound of a work is flawed. The discovery that Haydn’s and Mozart’s symphonies were conducted during the composer’s lifetime by a pianist who played the harmonies to keep the orchestra together gave rise to a series of recordings in which a piano or harpsichord could be heard obtrusively in the foreground, even at times messing up Haydn’s delicately austere, dry moments of two-part counterpoint. (All the evidence points to the orchestral piano or harpsichord being heard in the late eighteenth century largely by the orchestra alone; it was almost inaudible to the public, and it was dropped as musically unnecessary as soon as one had found a better way to beat time.)

On the other hand, no one as far as I know has revived the French eighteenth-century operatic practice of keeping orchestra and singers together by a loud banging on the floor with a big stick. Of course, the practice was deplored even at the time as distracting in soft passages, and it was not strictly part of the music, but surely it was a part of the auditory experience less incidental than the underground rumbling of the subway in Carnegie Hall, or the whine of an overhead plane at open-air concerts. A good thump at the beginning of every bar must have added definition and even induced a little rhythmic excitement.

Much more serious is the refusal to take into account how styles of playing and even instruments are dependent upon social circumstances—not the size of the hall, but the character of the audience and its interests and reasons for being present. Let us start with the zero degree of the question, with no audience at all: music not intended for any form of public or semi-public performance. This includes two of the supreme masterpieces of the eighteenth century: Bach’s Well-Tempered Keyboard and his Art of Fugue, both educational works.

A fundamental distinction must be made in Bach’s output between public and private works—or, better, between music to be played for oneself and one or two pupils or friends and music to be played for others. The organ toccatas and fugues are public, meant to be played in church; the closest that we have to works by Bach for the modern concert hall since their liturgical function is not an important aspect. Performing them in public has never been a problem. The Art of Fugue and The Well-Tempered Keyboard are a different matter: both were intended for completely private performance, by two hands on a keyboard—any keyboard that one had at home, harpsichord, clavichord, small portable organ, early piano. (Bach was interested in early pianos and even sold them.)

Mozart, who arranged fugues from these two educational works for string trio and string quartet to make them available for semi-public performance in the 1780s, once asserted that fugues must always be played slowly or only moderately fast, otherwise one would not hear the entrances of the theme in the different voices. This is not a consideration that would have occurred to Bach, and it shows to what extent the change from private to public music entailed radical changes in performance only thirty years after the composer’s death. The entrances of the theme in Bach’s public fugues for organ are easily heard and appear with dramatic effect. In the private works, however, the theme is often hidden, disguised, its opening note tied to the last note of the previous phrase. There was no need for the performer of these educational fugues to set the theme in relief: he could hear it himself as he knew where it was, and, even more, he could feel its presence in his fingers. If he allowed a pupil or friend to hear him play the work, both of them would be reading the score. The different appearances of the theme needed no illustration or emphasis from the performer.

When the Well-Tempered Keyboard was played for a small or large group of people, this easy intelligibility disappeared. The first public or semi-public performances of the Well-Tempered Keyboard that I know about took place after 1780 either for string ensemble or on the piano. It now became imperative to allow the listeners to perceive what went on in the fugue, to give them an idea how the individual voices moved, where the theme was. Here the Early Music aesthetic cannot help us. What the composer wanted or what he expected to get is no guide unless we are content to leave the public in the dark.

The proper question now is: which instrument will deliver Bach’s original conception to the public most adequately? The piano is clearly more capable of setting details in low relief than either the harpsichord or organ (we must remember that a performance of these private fugues would in Bach’s lifetime rarely have used any change of registration or employed more than one keyboard), and the clavichord, which can make dynamic nuances, is too soft for effective rendition of the larger fugues even on records (unless we turn up the dial to produce an “inauthentic” volume of sound).

It is true that harpsichordists today in performing these once private fugues have exercised their ingenuity to bring out the entrances of the theme. Taruskin, attacking Edward Cone for claiming that neither harpsichord nor organ was capable of “applied accentuation,” observes that “harpsichordists and organists…have invested gallons of sweat and tears learning successfully to belie [this].” The sweat and tears are indeed necessary: in all the performances I have heard of the Well-Tempered Keyboard on the harpsichord, the performer has executed a set of stylistic shenanigans in order to make this music truly public. Is there any evidence, or any reason to believe, that anyone ever played a fugue that way in the first half of the eighteenth century? Nevertheless, to perform this music in public as if it were still private is not only self-defeating but psychologically impossible. (Professional musical training today is directed toward the format of public presentation in large halls.)

On the other hand, most of us have heard pianists produce the egregious style of fugue playing in which the melody is brought out to the detriment of everything else—the standard way of playing Bach on the piano, and still encouraged in all conservatories. In a fugue, however, it is not the theme that is fascinating, but the way it combines with the other voices, into a homogenneous sonority. The ideal is still to play these fugues on the harpsichord for oneself, to see the voices as separate in the score and to hear them come together in a harmonious whole. In public, however, the most satisfying performance in my experience was one, on a piano, by Solomon of the C minor Prelude and Fugue from Book II of the Well-Tempered Keyboard: it was in the immense barn of the Salle Pleyel in Paris; the effect of transparence made one think that Solomon was setting nothing into relief, but in fact he delicately shifted our attention from voice to voice throughout, and it was profoundly moving.

Playing this private music as it would have been played during the composer’s lifetime is not only never done even by the most fanatical seeker after authenticity; it is also neither a practical nor a desirable ideal. When we reach the semi-private music of Mozart, Beethoven, and Schubert, written to be played to small gatherings, often in private houses, the ideal of authenticity is even more treacherous, as we are less conscious of how we are deforming the music. We shall also find that Tomlinson’s view of authenticity as a reconstruction of the historical moment when the work of music came into being produces results of insoluble ambiguity both for performing and understanding the music.

Perhaps more than any other city, Vienna had a very active tradition of semi-private music making, regular gatherings of twenty or thirty friends and guests: it was in this way that the majority of chamber works, piano sonatas, and songs were made known. Beethoven’s piano sonatas are not fully public works: only two of the thirty-two were performed at a public concert in Vienna during his lifetime, while all of his string quartets were presented in public. In spite of the current mythology of the intimacy of the string quartet, written for the delectation of the players and a handful of connoisseurs, the piano sonatas of Beethoven and Schubert were even more clearly designed for semi-private setting.

The force of this tradition is largely neglected today, and we fail to recognize how it influenced the works that were created for it. This matters least for Beethoven: he was a brilliant and famous concert pianist; he had considerable experience of public performance with his symphonies and concertos; and several of his sonatas (the “Waldstein” for piano, the “Kreutzer” for violin and piano) are clearly concert pieces. Most of his sonatas transfer without difficulty to the modern concert hall, but many of them nevertheless show that they were not intended for a large public: the sudden soft return to a strict tempo in the last few seconds of Opus 90, for example, is genuinely poetic, but can seem almost offhand in public, and pianists rarely resist the temptation to lengthen the very short final note to make it more impressive.

Schubert, on the other hand, was only an indifferent pianist, and few of his works were performed in public during his lifetime: his real element was the small gathering of friends at which his songs and chamber music were performed. At relatively informal occasions, attention may be as keen as at the public concert, but it is caught and held in different ways: delicate nuances are more telling, an almost hypnotic uniformity of texture is more effective, and large climaxes do not have to be prepared so long in advance, but can spring out as a surprise. The aspects of Schubert’s style singularly well-adapted to this semi-private occasion came out in an interesting article last year in these pages by the distinguished pianist Alfred Brendel, and a subsequent controversy with Walter Frisch of Columbia University on Schubert’s indications for repeating the long expositions of his sonatas.4

Brendel, whose viewpoint is partly determined by his project of playing Schubert’s last three sonatas in one program, quotes Antonin Dvorák’s remark about the length of Schubert’s symphonies: “If the repeats are omitted, a course of which I thoroughly approve, and which indeed is not generally adopted, they are not too long.” Brendel also adds that Brahms omitted the repeat of the exposition of the Second Symphony at one concert; the composer explained:

Formerly when the piece was new to the audience, the repeat was necessary: today, the work is so well known that I can go on without it.

In sonatas, symphonies, and quartets from Haydn to Beethoven, the expositions were repeated not in order to let the audience familiarize itself with the material, but for reasons of proportion and harmonic equilibrium—the exposition set up the basic harmonic tension of the work, and it was generally much less complex and chromatic than the succeeding development. At the end of his life, Haydn had experimented with leaving out the repeat, and Beethoven early on had continued these experiments, but Schubert was resolutely conservative: not once is the indication omitted.

The remarks of Brahms and Dvorák were made a half century after Schubert’s death when this sense of proportion had largely disappeared, and had become an old-fashioned convention; this was a process that had only begun during Schubert’s lifetime. More interesting for us here is that Brahms and Dvorák were both speaking of the public genre of the symphony, and they were considering the question of repeats with regard to the reactions of the audience. However, the tolerance of a small, intimate group of listeners is much greater than that of a large crowd. A small intimate group will bear a long work with greater stoicism and courtesy. Moriz Rosenthal related that once in Vienna after dinner Ferruccio Busoni was persuaded to play, and he sat down and played the last five sonatas of Beethoven without getting up from the keyboard. (“It had been a very heavy dinner,” was Rosenthal’s comment.)

Walter Frisch wrote in to protest that the indications of repeats in the last two sonatas contain material not otherwise present in the exposition, and to observe that Brendel’s public would be happy to spend the extra half hour that the repeats would add to the program. Brendel replied with a condemnation of the new material in the first ending of the exposition of the B-flat major Sonata: he objected to “the new syncopated, jerky rhythm,” as well as the fortissimo trill, which remains “elsewhere in the movement…remote and mysterious,” and he feels that “its irate dynamic outburst rob[s] the development’s grand dynamic climax of its singularity.” These bars are indeed violent, and contain the most brutal dynamic indications of the entire work. They explode without warning, and for this reason a convincing realization is not easily achieved in a modern recital hall.

Brendel’s point of view can only be appreciated in relation to his ideas on the molto moderato tempo mark of the opening movement. He interprets this as corresponding “to a none-too-dragging allegretto,” and attacks the recent fashion of conceiving the molto moderato as very slow:

While some older pianists played the first movement of the B-flat Sonata in an almost nervous alla breve, two beats to a bar, it is nowadays, in extreme cases, played in eight, with the repeated exposition thrown in for good measure.

Most of the musicians I have spoken to have taken, rightly or wrongly, Brendel’s target here as Sviatoslav Richter. His is indeed an extreme case: Richter’s interpretation is notorious for a first movement tempo that makes it last almost half an hour. As his recording shows, the tempo may be perversely slow, but the performance is a tour de force, never ceasing for a single instant to hold our attention.

The recording tells only part of the story: on occasion, when playing this work in a large hall, Richter has closed the piano—not just half closed, as if to accompany a singer, but with the lid down over the strings and sounding board, damping most of the brilliance and removing much of the volume. It is clear that Richter’s idea of music has been conditioned as much as Brendel’s by the concert stage, but that he has also been infected to some extent with the ideology of Early Music, although he has not remained within its narrow boundaries and tried to reproduce an antique sound in a modern hall: he essays a large-scale illusion of the intimate setting for which Schubert’s sonata was designed and tries further to dramatize this intimacy for a public of two thousand or more.

The Sonata in B-flat major is often considered Schubert’s greatest work for the piano. The first movement has, nevertheless, posed difficult problems. Donald Francis Tovey, who loved most of it, described the exposition as descending from “the sublime to the picturesque and then drifting from the picturesque through prettiness to a garrulous frivolity.” The difficulty lies in an attempt to adapt the work for the exigencies of modern concert life, to make it “go.” Most pianists opt for a reasonably brisk tempo, which does indeed make the latter part of the exposition sound pretty and garrulous, and they abridge the first movement either by cutting the repetitious phrases, as Harold Bauer did, or by omitting, like Brendel, the repeat of the exposition.

Yet it is the unequaled breadth, tranquil and even leisurely, that gives the work its supreme lyricism. It is the only sonata movement ever written to sustain such intense lyricism at such length, and it is the intensity as well as the length that makes the work difficult for a large audience. Unlike the opening movements of the other two late sonatas, this one has no technical display, no apparent virtuosity. The only technical difficulties lie in sustaining the line and balancing the delicate nuances demanded by the texture: astonishingly, more than half the exposition must be played within the range defined by pianissimo. At the end of the exposition, the music breaks into fragments, and erupts without warning into forte and fortissimo; the most dramatic of these fierce explosions is reserved for the disputed first ending.

These bars were originally not a first ending at all but an integral part of the exposition to be played twice according to the first draft. It is the second playing that Schubert revised, and he substituted a magical pianissimo transition into the development, with a harmonic change of the utmost daring, the kind of liberty for which he had been reproached by critics all his life. The dynamic explosion of the first ending and the harmonic outrage of the second complement each other, and the wonderfully poetic effect of the return of the inspired opening theme after the thunderous roar of the trill in the first ending is a counterpart to the opening theme’s return after the end of the development with the long trill now heard twice pianissimo.

This sonata’s creation may have been shaped by the intimate setting for which it was intended, and its tranquil breadth, lack of display, and sudden violence are more easily realized in such a format, but the sonata is not contained by it. I do not know if it was ever played in Vienna before being published a decade after Schubert’s death, but it would have destroyed the gemütlich, Biedermaier intimacy of the semiprivate musical evening, just as the songs of Winterreise made Schubert’s friends ill at ease and uncomfortable. An “authentic” performance today is an absurdity. A recording on a “Schubert” pianoforte would add little to our appreciation of the work. It is also possible to maintain that Brendel’s practical, modern concert approach distorts the music less than Richter’s theatrical illusion of intimacy in a large space. Just as the sonata would have burst beyond the frame of the semi-private concert, so today in the modern concert hall, even though it must be played in a way that Schubert would never have imagined, it makes the traditional brilliance and dramatic effect that the recital demands seem irrelevant. Every performance today is a translation; a reconstruction of the original sound is the most misleading translation because it pretends to be the original, while the significance of the old sounds have irrevocably changed.

Style of performance may be affected by aspects of presentation that seem irrelevant at first glance. Sir Charles Mackerras pointed out to me that a great deal of evidence tells us that in the early nineteenth century, the first three movements of a symphony of Mozart or Beethoven were most often played faster, and the last movement slightly slower, that performances today. This is easy to understand if we reflect that at the time, the public clapped and cheered after each movement (often forcing a repeat of the slow movement), and not just at the end of the symphony, so that the musicians were not forced into extra brilliance in the finale in order to get their meed of applause. To make sense of an authentic restoration of the original tempi, we should have to encourage applause between the movements (by holding up a sign perhaps?); yet the modern tradition of keeping back the applause until the end is a justified tribute to the unity and integrity we now perceive in (or attribute to) the symphony as a whole. An unthinking restoration of the original tempi is not a wholly rational project—and if recordings ever become not just a mimesis of public performance but the normal musical experience, our concepts of musical excitement and intensity will have changed and will alter our ideas of tempo once again.

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: authentic sound is not only insufficient, as most people would grant, and not only often an illusion, as many now realize, but sometimes positively disastrous. Sound is dependent on function: Bach’s harpsichord concerto became a concerto for organ, chorus, and orchestra when it was used for a church service.

It remains to acknowledge that Early Music has been and is a remarkably beneficial movement. It has made us realize how constricting are the demands of concert life, which force us to play Bach, Mozart, Beethoven, Chopin, Brahms, Debussy, and Boulez with the same instruments, the same kind of sound—looked at rationally, an insane result, even if the facile solutions of Early Music are equally odd and even more constricting. Early Music has laid bare the deadening uniformity of today’s conventional concert world, where we find the same phrasing for Mozart and Beethoven, the same vibrato for Mendelssohn and Tchaikovsky, the same pedaling for Beethoven and Chopin. It has also had extraordinary successes, from Roger Norrington’s work with Beethoven and Sir Charles Mackerras’s performances of Mozart and Schubert symphonies to John Eliot Gardiner’s reconception of Monteverdi.

One could say that the successes have been achieved in spite of the philosophy. This would not be true: it has been by taking the indefensible ideal of authenticity seriously that our knowledge has been increased and our musical life enriched.

This Issue

July 19, 1990