Joseph Kerman
Joseph Kerman; drawing by David Levine

Joseph Kerman’s book on the concerto is too short, but it is otherwise splendid, entertaining, original, and often profound. Its excellence is partially disguised by a resolute assumption of modesty. It is also occasionally marred by a style that assumes the all-too-easy, popular approach of a freshman music course. For the most part, however, Kerman speaks directly and informally to a literate and educated public deeply interested in music. These six lectures (the Charles Eliot Norton Lectures of 1997-1998) are characterized by their dissatisfaction with the usual categories of both academic and popular writing about the concerto, although Kerman is always courteous and even grandly generous to his predecessors. His affable manner sometimes makes his approach to the subject seem obvious, the result of common sense, even when it is most innovative.

The first lecture begins at the beginning: how to start a concerto. Flouting the traditional approach, Kerman takes nineteenth-century practice as his model for the form, rather than eighteenth, as most scholars have done. In the eighteenth century, a concerto opened with a presentation of the principal material by the orchestra, a section called the ritornello because it continues to return and punctuate the solo passages. Kerman prefers to consider first the examples after 1800 when the soloist either opens the concerto or enters within a few seconds. The disadvantage of Kerman’s tactic is that it minimizes the extent to which the ritornello scheme continued to function as the normal form: composers who avoided it must, I think, have believed for many decades that they were doing something daring and original even after it had been done dozens of times.

Getting completely rid of the classical ritornello began as an experiment by the radical composers of the 1830s, principally Liszt, Mendelssohn, and Schumann. However, Chopin, whose two concertos were written by the time he was twenty-one, retained the ritornello in its most formal guise, and I suspect that many minor composers of the following age whom no one has any time for today continued to compose the opening, purely orchestral, section, following the established classical recipe.

Kerman thinks that Brahms “was reviving the eighteenth-century concerto ritornello,” and that “of all his many classicizing projects, this was perhaps the most extreme and the most obdurate.” I am not sure that this project was all that extreme, although the seriousness and the symphonic weight of Brahms’s ritornellos were indeed astonishing. For example, in the 1860s an uninteresting Italian composer, Giovanni Sgambati, a protégé and devout disciple of Liszt, and therefore unlikely to be influenced by the only example of Brahms published in 1861, began his piano concerto with a long classical ritornello. Music students at conservatories the world over must have been taught even into the twentieth century that this was the correct way to compose a concerto.

The advantage of Kerman’s approach, however, is that he is able to start by calling attention to the different ways that the relation of soloist to orchestral ensemble can be displayed to catch the ear and the attention of the listeners. This enables him to elucidate the nature of the concerto without discussing purely formal processes. In other words, the composers who abandoned the classical scheme were confronted at the outset by the necessity of inventing a new way of making the position of the solo instrument faced with a large ensemble immediately evident—and of finding an original way of doing this (unless, of course, like Grieg, they merely aped the solution that Schumann found in his only piano concerto).

Kerman seems to begin, in short, as a listener who knows nothing about the history of concertos. This initial and temporary refusal of history is profoundly right; even those of us who know something of the history of the concerto, and have educated expectations about the form and procedures, quickly learn time after time with each individual work that we must abandon those expectations. The familiar theory that composers work mainly by playing on these expectations, frustrating and fulfilling them, is only partially true and has outlived much of its usefulness. We all have to realize with every unfamiliar and familiar work that our expectations are often irrelevant. Kerman addresses the listener who tries to realize the ideal of listening to a concerto with innocent ears. The ideal may be impossible, but any decent judgment of a work requires a striving toward it. The historical imagination, indispensable as it is in the end, should never be permitted to take over and govern the whole of our musical experience.

The abandonment of the opening long orchestral section coincided with the growth of public concerts and the transformation of the concerto into an exclusively public form. Before 1800, there had been almost as much chamber music as public entertainment. And, as Kerman observes, the public does not want to sit down at the beginning and watch the star performer doing nothing for several minutes. The elimination of the ritornello was a rational and timely response to the new public importance of the concerto.


With the second and third lectures, “Particularity and Polarity” and “Reciprocity, Roles, and Relationships,” Kerman substitutes different considerations for the standard analysis of form. What concerns him at first is whether the soloist and the orchestra address the same or different material, whether they are “polarized” by giving contrasted motifs and melodies, or “reciprocate” by endowing the same phrases with a different sound and a different function. As Kerman writes, if the soloist makes the initial entrance after a ritornello by replaying the opening orchestral theme, “we attend less to the melody, which we already know, than to the new sonority.” This makes for a very different effect than the repetition of material in a sonata, and changes the listeners’ focus. More subtly, Kerman seeks to broaden this approach beyond the question of simple material to characterize the relationship of soloist to ensemble—to understand, so to speak, what kind of conversation takes place between solo and orchestra. We might ask, is it an argument or an exchange of views? The concept of conversation was central to the idea of music in the late eighteenth century, when the concerto form came into its own, but the nature of the dialogue shifts from work to work.

What Kerman wishes to set up is a way of viewing concertos that respects and preserves the individuality of each work and evades the simple pigeonholes of the standard academic analysis. His treatment of the opening scales of the pianist’s replaying of the main theme when it enters in the first movement of Beethoven’s C Minor Concerto is particularly cogent and convincing. “The scales,” he observes, “deflect attention from volume to range, from weight to mobility, giving the reciprocal playing of the first theme new vital energy.”

He makes explicit his intention to meditate on the nature of the dramatic relationship between the individual solo and the mass ensemble:

How to develop a criticism that can engage with the diverse qualities of concerto relationships? It will certainly be necessary to get past neutral terms like “accompany” and “answer,” terms that are applied so routinely to musical events that they have lost their original figurative meaning, and might as well be technical terms like “chord” and “antecedent/ consequent.” Fresher tropes are needed, and it seems to me they have to be derived by analogy from human relationships (something commentators have been doing throughout the history of the concerto, as I have said a number of times)….

I also like to think about and attempt to specify the roles that concerto agents assume, roles such as MASTER and SERVANT, or MENTOR and ACOLYTE. (When I was young, analogical or metaphorical efforts of this kind were taboo in academic musical studies. The mandarin, Miltonian [the composer Milton Babbitt, not the poet John Milton] position was to dismiss such discourse as “incorrigible” and “literally meaningless.” In today’s musicology, though, meanings are discovered everywhere.) Among the roles that will be encountered later in these pages are EAVESDROPPER, TEASE, SURVIVOR, VICTIM, MOURNER or PLEURANT, MlNX LOVER CRITIC, EDITOR…some of these, I suppose, will seem fanciful.

This puts Kerman in the center of the present fashion in musicology, alongside younger critics like Scott Burnham, who has studied the relation between metaphorical description and formal analysis, or Robert Hatten, who has written brilliantly about the semantic elements, or the “tropes,” in Beethoven.

The new fashion is a gain when it is controlled, as it is in Kerman’s case, by a grasp of the formal elements of music and an ability to distinguish salient and significant detail from trivia. Otherwise, the hunt for coy and whimsical tropes can become a mere excuse for purple passages or an indulgence in ideological prejudice. Kerman, however, can almost always back up his metaphorical fancies by reference to aspects of the music discovered by the ear, not merely observed on paper.*

As an example of his basic sharpness, I give the only example I found in his book of an analytic mistake, an error that actually addresses and reveals something important about the music, here the opening of Liszt’s Piano Concerto No. 1 in E-flat Major:

The solo enters after an orchestral outburst that modulates at once to a new key. Liszt’s piano does not accept this key; on the contrary, it switches back to the original key.

The mistake is to speak of a modulation: at the end of the four bars of the “orchestral outburst,” the only possible harmony is the dominant seventh of E-flat major, so we are still in the opening key—any other harmony would be an outrage, a solecism. Nevertheless, Kerman’s mistake touches the most important aspect of this opening. The first four orchestral bars are complex, chromatic, and disorienting if not actually ambiguous. The piano’s entrance brings us back to solid ground, reestablishes the hegemony of the opening note, a simple E-flat. Kerman’s description conveys the unsettling dramatic opening and the shock of the solo entrance, but fails to account for the logic of that entrance. Except for this passage, which is more useful in the end than it is misleading, Kerman gets his analyses right as far as I can see, and his observations, even when esoteric, are not insignificant.


The fourth lecture, “Virtuosity/ Virtù,” treats of the virtuosity without which the concerto is not even thinkable. Kerman explains the nature of the virtuosity that is more than mere bravura by the Italian word virtù, above all for its secondary sense of authority. Perhaps one should emphasize an aspect of virtuosity that Kerman only glances at, but that is essential to his consideration of role-playing in the concerto. The virtuosity, the battle against odds, wins the sympathy of the public for the soloist. In any relationship between an individual and a mass, there is bound to be somewhere the sense of a struggle, an agon. It used to be said that Brahms wrote concertos for piano against orchestra; but only when the pianist is downright incompetent do we cheer the orchestra on. It is the soloist that must win for a happy ending, or be defeated in tragedy.

The fifth lecture, “Diffusion: Concerto Textures,” is perhaps the most important. It strengthens the opposing concepts of reciprocity and polarity, adding to them the idea of diffusion: here Kerman treats those striking moments of many concertos where soloist and orchestra coalesce, where there is no longer dialogue or argument but sometimes an exchange of roles. Kerman’s examples are largely from modern examples, like the concertos of Alban Berg, Carl Nielsen, and Igor Stravinsky, yet he notes a beautiful moment in Mendelssohn’s Violin Concerto, where the violin holds a long note that provides the bass for a melody played by the winds. (He mistakenly credits Beethoven with being the first to introduce an extended solo part in a piano concerto for an orchestral wind instrument accompanied by the solo piano; there are already examples by Mozart as early as the Concerto in E-flat Major, K. 271 (with a horn solo in the first movement), and in the finale of the Concerto in C Major, K. 503, where the piano accompanies complete melodies in flute and oboe.)

Kerman’s greatest innovation is perhaps his attempt to come to grips with texture. He remarks:

Musicologists are slow to discuss musical texture because this is notoriously hard to do, and the results are usually unhappy. Of all music’s parameters the most resistant to theory and verbal description is sonority, whether the sonority of single sounds, timbre, or that of multiple combined sounds, texture….

Analysis of concerto textures works well enough when the solo is accompanied by the orchestra, falters when the relationship between them is contrapuntal, and all but collapses when the solo appears to be accompanying the orchestra. One cannot theorize, one can only admire.

There is no reason for Kerman’s despair, and he deals with the different cases well enough. Among the best of his discussions are those of Bach’s Brandenburg Concerto No. 3 and of No. 5 (the latter is the first concerto ever written with a keyboard soloist). In No. 5, Bach treats the harpsichord exhaustively, as Kerman observes:

as a member of a quartet with violin, flute, and continuo…; as a melody instrument…; as a voice in a fugue…; as a virtuoso instrument playing high-speed runs overlaid on trio action as preface to a spectacular written-out cadenza…. He tries out the harpsichord in an even more interesting role, as an ingredient for a composite concerto texture.

Kerman misses only one further role: there is a splendid brief moment (in bars 133-134 of the opening movement) when the harpsichord mimics the string orchestra, with its chordal accompaniment of the flute and violin soloists (an example of what Kerman has called diffusion).

The final lecture, after an examination of the different ways that concertos end, is rounded off by a meditation on the present state of music in general, above all the replacement of the live concert by recordings and videos. This decline of the live performance of classical music today particularly concerns the concerto, since it is the only purely instrumental public genre which is essentially staged, in a for-mat that resembles the dramatic art. Can the effect of the lone individual against the massed forces survive in video? Kerman is cautiously optimistic, but he finds the new reliance of music on recordings and film to be a change comparable in importance to the invention of musical notation a millennium ago. These final pages show Kerman at his most speculative and also at his most impressive.

His discussion inspires further consideration. The invention of notation fixed the text. In so doing it gave Western music a history, with radical changes of style which can be retraced, reexamined, and revived. It also made each piece of music into an independent work of art, detachable from the social context for which it was originally created (this is what makes Western music so difficult to treat by sociologists, as a work that can be performed equally well in church, in the salon, in the academy, and privately for oneself is not easily amenable to social categories). Notation, however, also caused a radical split between composition and performance. The work was more or less fixed permanently as a text: performance, on the other hand, was essentially ephemeral, insubstantial, and, above all, volatile. The work of music was a single, definable object on paper and in the mind, but it could be realized in many different ways, many of them unpredictable by the composer—who did not, in fact, always play the work the same way. Western music was based on a tension between text and performance: behind every performance, we knew, there was a fixed creation, an ideal not yet realized in sound, to which the performer could be more or less faithful.

This situation is transformed by recording. The performance is now fixed, unalterable, each playing of a phonograph record despairingly predictable. A wrong note or a miscalculation in a public performance may be an unimportant witness to the essential humanity of the performer; on a record it can be a major irritant. The cult of the “authentic” performance is, as has often been said, the product of an era of musical reproduction. The recorded performance can now have the authority of a holograph manuscript. Phonograph records are not altered by time and by change of place, only by mistreatment, and it is wicked to mistreat them. The performance of a musical work, however, has always been modified by time and by change of place, as well as by change of performer, and in the last century we have been gradually persuaded to think that these changes are somehow wicked, or at least reprehensible. The relation of text to performance has lost its tension, and has become sclerotic, hardened. (Sometimes, in reaction, we find the contrary nonsense: a performance that willfully ignores the text can appear as the new creation of a romantic free spirit.)

That is why musicologists (including Kerman in this book) are somewhat ill at ease with the question of improvisation in concertos; he does not mention, for example, that in the central solo section of the opening movement, mid-eighteenth-century composers did not always write out the virtuoso figuration for the pianist, but just noted the harmonies and allowed the performer to fill them out. Kerman recognizes the existence of bad taste among Mozart’s contemporaries who improvised ornaments to a Mozart concerto, but he acknowledges the need to ornament certain passages. He also thinks, improbably to my mind, that Mozart improvised his cadenzas in public, full of the carnival spirit, mischievously, in a “bratty mood,” and that this was “one way he held his audience.” We know that Mozart fooled around in private, and we know that he improvised in public. But on January 22, 1783, he wrote to his father, apologizing for not sending the cadenzas to new concertos, “I have not yet altered the introductory cadenzas to the rondo [K. 382 for piano and orchestra] for whenever I perform this concerto, I just play whatever comes into my head.”

Mozart was quite able not only to play but also to writewhatever came into his head, but he was unwilling to do so. Writing required that a more permanent decision be made. When he set down a cadenza on paper for the same concerto on different occasions (probably for other pianists), it is often surprising how close the alternate cadenzas are to each other; evidently at some point in the history of his own performances of his concertos, the conception of what would be the best cadenza gelled.

Little by little, notation encroached upon the rights of the performer. Beethoven, over his career, gradually built into the text whatever improvised cadenzas were traditionally considered necessary to the structure. We know that when he performed his own music (like the Fourth Piano Concerto), he sometimes played more notes than he had written, but we also know that if anyone else performed his music that way, even a favored student of his like Carl Czerny, a public explosion of rage was the result.

Nevertheless, Kerman clearly understands the procedures of improvisation very well. He knows why Liszt’s music sounds like an improvisation:

Brahms doesn’t sound improvisatory. Liszt does, I think, exactly because of his absent-mindedness and his imprecision, and also because of a formulaic quality which, in most true improvisations, is more rather than less in evidence.

This cannily displays some of the reasons that improvisation clashes with written composition, unless—as in the case of Liszt—an extraordinary creative sense of sonority and texture makes up for whatever repetitive formulas come into play.

Concerto Conversations comes with a compact disc that contains selections of many of the works discussed. There is an excellent, deeply musical per-formance of the first movement of Beethoven’s Third Concerto by Robert Levin on a fortepiano, with a cadenza of Levin’s own composition that is very fine, although not as astonishing as the one Beethoven himself composed. The last movement of Mozart’s Concerto in D Minor, K. 466, is played magnificently by Malcolm Bilson with true ferocity. Oddly, the fortepiano in the Beethoven is softer than in the Mozart (the conductor is John Eliot Gardiner in both selections). I had understood that pianos got bigger and louder in the two decades that separated these works. Both Levin and Bilson are passionate advocates of the practice of playing continuo (the keyboard filling out the harmony in the orchestral sections) in Mozart and Beethoven. I have even seen Levin vigorously doing so but have not heard him, as an eighteenth-century piano is not audible during the orchestral sections even to a listener seated in the front rows. It is not audible on Kerman’s CD either, although I presume that Levin and Bilson were both playing along with the ensemble, given their usual practice. I was pleased that the engineers had not miked up the keyboard sound, since these excellent performances demonstrate my claim that it is absolutely authentic to play the continuo in Mozart and Beethoven concertos so long as it is inaudible.

This Issue

January 20, 2000