Tuesday, 13 October 2015

Quotes: Horowitz - His Life and Music (1992) by Harold Schonberg - Preface & Appendices

Harold C. Schonberg

Simon & Schuster, 1992.
8vo. 432 pp. Appendices I–IV The Horowitz Recordings [pp. 317-354].
Discography by Jon Samuels [pp. 357-404]. Index.

First published, 1992.



Near the end of his life Vladimir Horowitz started thinking about writing his memoirs. With his approval, Simon & Schuster engaged me to help him with the project, and we started work in January 1987. He dictated his reminiscences into a tape recorder.

Unfortunately the project broke off about six weeks later. Horowitz was preoccupied with his preparation of the Mozart A major Concerto (K. 488) for his forthcoming recording in Milan with Carlo Maria Giulini and the Scala Orchestra, and he was a little nervous about it. He had never played it in public.

Horowitz left for Milan and I never saw him again. On his return to New York he dropped the project, though in the last months of his life there were indications that he was interested in resuming work on the memoirs. It was not to be. But at least he had passed to me some of the material that would have gone into his memoirs. All of the Horowitz quotations in this book come, unless otherwise specified, from those taped interviews early in 1987.

During most of the 1987 Horowitz sessions his wife, Wanda, the daughter of Arturo Toscanini, was present. She was the moderating force, prompting him, correcting those dates about which he was hazy, and keeping him on line. Horowitz would often tend to digress. Mrs. Horowitz would firmly set him straight. After fifty-four years of married life she knew how to handle him.

I was no stranger to Vladimir and Wanda Horowitz. Since 1954 I had frequently interviewed him for the New York Times. Whatever Horowitz did was, after all, news, and through the years I had a journalistic relationship with him. If we were not exactly close friends – critics on the Times shunned social relations with artists they were in a position to review – at least I believe we understood each other, and, as musicians who had grown up with a set of traditions that had all but vanished, we shared and relished a common musical heritage.

Invariably, when the interviews were over, Horowitz would want to relax, talk, gossip. He never went to many recitals, but his natural curiosity about pianists young and old kept him glued to the three major good-music radio stations of New York. If he did not hear the competition in the concert hall, he could and did hear it broadcast on records.

He did not think too highly of the culture and general musicianship of the pianists he heard (“They don’t sing”), but the fingerwork of a few virtuosos of the younger generation impressed him. “Some of them,” he said in 1987, “have extraordinary techniques. I heard on the radio a Liszt record, the Faust Waltz, played by Jean-Yves Thibaudet. It was amazing, such dexterity, such technique, such articulation, such command. I could not do that today.” Pause. “But I do certain things they can’t do, so that makes us even.”

On the whole, though, he liked very little of what he heard, and his brief, dismissive comments could be devastating. Vladimir Ashkenazy? “Ashkenazy was good once. Not now.” Glenn Gould? “I heard a recording of the Wagner Siegfried Idyll played by Glenn Gould. It was his arrangement. He played like a stupid ass. The tempi were all wrong. He was not normal.” The great Josef Hofmann? “A very good pianist but a second-rate musician.” The legendary Solomon? “Boring” The equally legendary Arturo Benedetti Michelangeli? Here Horowitz expressed admiration with a qualification. “Interesting pianist, but I think he is just a little bit meshuga.” André Watts? “Technically formidable, fantastic fingers, musically horrible.” The new Russian school? “Of the Russian pianists I like only one, Richter. Gilels did some things well, but I did not like his mannerisms, the way he moved around while he was playing.” Claudio Arrau? “I heard his Emperor and it was terrible. He plays so slow, ugh. Also the Waldstein. So slow.”

He would go to the piano to illustrate a point and, once seated, could remain there for several hours. Often I turned pages for him. He was a fabulous reader, taking in five measures at a glance no matter how difficult the music. He had in his fingers much of the active repertoire. In his entire career, for instance, Horowitz never programmed more than about half a dozen of the Beethoven sonatas, but he knew all thirty-two by heart, and at home he would sit at the piano playing one of the early sonatas, or sections of the Hammerklavier, or the beginning of the last movement of Op. 109, wondering why pianists played it merely as chords without getting the melodic line and the chordal balances. Fooling around at the keyboard he might drift from the Saint-Saëns C minor Concerto to sections of Tchaikovsky’s Eugene Onegin, to the Chopin Etude in Thirds (“I am too old to play this in public”), to the Schumann Concerto, to Mozart’s G minor symphony, to some Moszkowski he had learned as a child, to some completely unfamiliar Liszt.

He had in his fingers many of the Russian operas and would play long sections from Tchaikovsky’s Eugene Onegin at the slightest prompt. The Ring cycle he knew by heart and could still play some of the transcriptions he had made as a youth. He also had a passion for Bizet. “When I go to Père Lachaise I put a flower on Bizet’s grave. I think Carmen in unsurpassable. I love Pearl Fishers, a beautiful opera. I love his Symphony in C.” He also loved Saint- Saëns’s Samson and Delilah (but the piano music of Saint-Saëns did not interest him except for the Fourth Concerto).

He liked jazz or, at least, jazz pianists, some of whom almost came up to his own demanding technical standards. Henry Pleasants, who had covered the jazz beat and written with perspicacity about jazz singers, told the story about Horowitz sneaking in to listen to the jazz pianist Art Tatum. He was terribly impressed with Tatum’s technique and his easy, natural way of playing, and on one occasion couldn’t understand how Tatum did what he did in “Tea for Two.” So he introduced himself. Tatum admitted that yeah, he had heard of Vladimir Horowitz. The two men had a pleasant talk, and then Horowitz asked Tatum how long it had taken him to learn “Tea for Two.” Tatum looked at him as though he were crazy. “I just made it up,” he said. Horowitz went home and worked up his own arrangement of “Tea for Two,” which he played as a party piece.

He was to the piano playing of his day what Jascha Heifetz was to the violin. All of the world’s young violinists wanted to be Jascha Heifetz, and all of the world’s young pianists wanted to be Vladimir Horowitz. Especially in the period from 1940 to 1960, the young pianists tried. They tried to get his kind of sonority, they aped his physical mannerisms, they played Horowitz-like programs. If Horowitz revived a seldom-played piece – Schumann’s Kreisleriana, say – almost everybody was programming Kreisleriana the following year. They tried to copy the Horowitz phrasings, the Horowitz agogics. Of course it was an impossible quest. Personality cannot be imitated or transferred. The Horowitz personality was accepted by most pianists with awe and even worship. Horowitz once went backstage after a concert by Alicia de Larrocha, whom he had never met. When she saw him in the green room she nearly fainted. Then she dropped to her knees and kissed his hand. Horowitz had that effect upon pianists.

His virtuosity was, of course, integral to his triumph with the public, as it has always been with any amazing virtuoso since the days of Paganini and Liszt. But sheer virtuosity was only part of the Horowitz gestalt. Not many realized how serious a musician he was. Yes, he sometimes could go overboard. Yes, he could miscalculate in concert. Sometimes his dares did not come off; or a listener could legitimately ask why he took a phrase in this or that manner. But never was there a letdown in interest, tension, or drama, his ability to keep the audience at the edge of the seat.

He had that from the very beginning, all through a professional career that started in 1921 and ended in 1989. When he came to international attention in the late 1920s, he was up against a handful of pianists who had set standards for all time: Rachmaninoff, Hofmann, Lhevinne, Friedman, Cortot, Schnabel, and a few others. But almost all had disappeared by 1950, leaving Horowitz pretty much alone, certainly as an exponent of Romantic music. In the public eye, and to most of the professionals too, he was the world’s greatest pianist for the entire last half of the twentieth century. And when he died there was nobody to replace him.

The literature on Horowitz is scant. There have been one biography and one volume of reminiscences by a friend.[1] Otherwise information about Horowitz has to be gleaned from interviews and from past magazine and newspaper articles and radio interviews. Under the circumstances, a bibliography would be nonexistent; in the running text of this book, I have generally documented the source of quoted material.


The Horowitz Recordings

Appendix I

The most extraordinary thing. I listened to the radio and there was some pianist. Wanda was there. It was the Liszt B minor Ballade. I listened and said to Wanda, it is horrible. I have to know who that is. Tempi too slow and tempi too fast, all like that. It’s probably somebody good and I don’t like it. It turned out to be Vladimir Horowitz. I hated this performance. It was horrible. I would not play one note like that today. So that means the taste changes.[2]
– Vladimir Horowitz on his recordings

Vladimir Horowitz took recordings very seriously. It is true that he fretted about certain built-in factors. Recordings, he said, were like photographs, resemblances of things past. A photograph of a thirty-year-old man differed from a photograph of the same man thirty years later. “Sometimes you recognize the person and sometimes you don’t.” A Horowitz recording of a specific piece made in 1935, he said, would differ from his recording of the same piece made in 1975.

Nor did he think that recordings were necessarily a force for good. Ever since the introduction of magnetic tape and the LP disc in 1948, editing had become so easy that recordings no longer represented what an artist really could do. Through splicing, even a second-rate technician could now sound like a Hofmann or Horowitz. Tape editing created a race of (apparently) perfect technicians of a kind who did not exist in the concert hall and opera house, where fingers and vocal cords could, alas, sometimes show their human imperfections. In addition, the public started to grow up on technically perfect performances that bore no relation to what actually happened in the concert hall.

Horowitz told Elyse Mach, in her Great Pianists Speak for Themselves[3], that too many pianists were influenced by recordings. “They are so used to hearing note-perfect performances on record that they want to duplicate the same note-perfect performance in the concert hall. The result is that instead of projecting the spirit of the music they concern themselves only with the notes…. With recordings today, it is mechanically [Horowitz should have said electronically] possible to do what it took me so many years to develop.”

At the same time Horowitz realized that a musician could live forever only through recordings, and he took full advantage of whatever new techniques were available.

For some sixty years Vladimir Horowitz made records. By the time he had made his last one, only four days before his death, a good part of his repertoire was forever available to posterity. Max Massei, the French expert on Horowitz who made a study of the recordings, estimated that Horowitz recorded about 85 percent of the Schumann works he played in public, 80 of his Chopin and Scriabin, 70 percent of his Scarlatti, 65 percent of his Rachmaninoff, and – surprisingly – only 55 percent of his Liszt.

In the process he consistently won Grammy Awards, far more than any other pianist. The Grammy Awards were initiated in 1959 for records issued the previous year. Horowitz got his first one in 1963 for his Columbia Presents Vladimir Horowitz. In all he got twenty-three Grammy Awards, a Merit Award and, after his death, the Lifetime Achievement Award. This does not take in the many awards he received in Europe, and he continued to receive them after his death. In 1991 his last record won a posthumous Grammy for the best classical-music instrumental recording.

He had prefaced his first series of recordings with a few Welte-Mignon piano rolls (made in Germany) and Duo Art (made in America). He was one of the last of the great piano-roll artists to record; in 1926 and 1928, when he made the rolls, the player-piano industry was almost dead. Electrical recordings and a new thing called radio captured the public’s imagination. Before long, piano rolls, which had been popular since 1904, disappeared.

The Horowitz rolls reflect the faults of the medium. For instance, accurate pedaling was next to impossible to capture. Nor was there much metrical flexibility. And the rolls could be heavily edited, with wrong notes repaired. Piano rolls tended to lack nuance. Hence the expression “He sounds like a player piano,” referring to a pianist who plays metronomically with no color or inflection. The story goes that Schnabel was once approached by a piano-roll company to sign a contract. “We have seven degrees of nuance in our rolls,” the manufacturer bragged. “Sorry,” Schnabel wired back. “I have eight.”

With all that, and keeping in mind the fact that Horowitz himself disavowed the rolls, his playing on them is still recognizably Horowitzian. And there is on the rolls one major work he never played in America, the Mozart-Liszt Figaro, ardent yet controlled, brilliantly unfolded. Another typical example of the young Horowitz is the Schubert-Liszt Liebesbotschaft, which demonstrates a lovely long line. He never played this after his first few seasons in America. There are also the Adagio from the Bach-Busoni Toccata in C, some Chopin mazurkas and etudes with which he later became identified, his own Moment exotique, and some Rachmaninoff.

When Horowitz started to make flat-disc recordings in 1928, the process was new. Only a few years before, the Bell Telephone Laboratories had tested a means of electrical recording in which the frequency response was greatly expanded. Up to then – from 1877, when Edison invented the phonograph, to 1925 – musicians had to record into a horn that had a limited frequency response. Now, with the new electronic process, musicians could play and sing into microphones instead of the acoustic horn, and the advantages were immediately apparent. In 1925 the first commercial electrical recordings were made, and the technique was able to capture sounds, especially of such large groups as symphony orchestras, with then unheard-of fidelity. A complete rethinking of the phonograph as a medium for music resulted.

The first recordings made by Horowitz were, of course, 78-rpm discs. Several things can be said about 78s. There was more in those old grooves than even the manufacturers realized, and a mint or even decent copy of the best of them, played on today’s top-quality audio equipment, could easily be mistaken for a CD. There is startling presence and surprisingly little background noise. The best that a CD transfer of a 78-rpm disc can do is aspire to be as good as the original. Another thing about the 78-rpm discs: they were honest. In those days before tape, there was no way of correcting a mistake. If an artist was not happy with what he did in the recording studio, there was only one option: to record the entire piece over again. What is heard on a 78-rpm disc is what the artist really did, and that cannot be said of LPs or CDs.

Of the many recordings Horowitz made between 1928 and 1953 (the year of his second retirement), by which time the LP record had triumphed and the 78s were beginning to disappear, two stand out – the Rachmaninoff Piano Concert No. 3 with Albert Coates and the London Symphony Orchestra (1930), and the Liszt B minor Sonata (1932).

Both were Horowitz specialties, and neither was much of a repertoire piece before he espoused them. Such a thing is hard to believe, because the Rachmaninoff Third Concerto has become perhaps the most-played piece for piano and orchestra in the international repertoire, and the Liszt B minor Sonata has been undergoing a major reassessment.

But in the 1930s both composers were derided as not worthy of attention by serious musicians. Liszt was equated with flashy, meretricious virtuosity. Rachmaninoff, in the fifth edition of Grove’s Dictionary, was dismissed in only five contemptuous paragraphs, with such statements as: “His music is well constructed and effective, but monotonous in texture, which consists in essence mainly of artificial and gushing tunes accompanied by a variety of figures derived from arpeggios. The enormous popular success some few of Rachmaninoff’s works had in his lifetime is not likely to last, and musicians never regarded it with much favor.” So much for Rachmaninoff in 1935.[4]

Of course he no longer needs defending, and it was Horowitz with his blazing performance of the Third Concerto who did much to establish his real value. Horowitz was to record the Rachmaninoff three times, and many consider his 1930 performance the best. It is propulsive, even savage, and he simply eats up the notes. There are none of the mannerisms of which he was later accused, and the slow movement emerges in a singing, natural kind of poetry. In the last movement he is one of the few pianists (Rachmaninoff himself not included) who, in the two Più mosso sections of the last movement, maintains a steady metric without allowing the patterns to fall into triplets. The prevailing tempos, like those of the composer, are faster than today’s.

Horowitz took some of the cuts sanctioned by Rachmaninoff. So did everybody else in those days. In the last movement there is a tiny ensemble mishap that the participants did not bother to remake; it passes almost unnoticed. This is a massive, thrilling performance that, along with the composer’s own, set the standard for all to follow.

Because of today’s slowdown in tempo, the 1932 Horowitz performance of the Liszt Sonata may sound fast. But it wasn’t in its day. Horowitz’s timing is 26 minutes, 30 seconds. Timings from the 1950s and 1960s show Rubinstein at 25’28’’, Alexander Brailowsky at 26’02’’, Clifford Curzon at 25’35’’, Van Cliburn at 26’50’’, and Emil Gilels at 25’38’’. Today the average is around 30 minutes and getting longer.

The Horowitz performance is typical: raw excitement coupled with moments of delicate lyricism and washes of color. As expected, there are the flashing octaves and awesome technical command. He holds the sprawling piece together beautifully. Even in the last measures he suggests a feeling of motion, and this is hard to do. Old pianists used to say of the Liszt B minor Sonata that he composed it and then decomposed it. Pianists who are too funereal at the end decompose the piece. Horowitz maintains the tension.

It was in 1932, also, that Horowitz took a foray into Classicism with his recording of the last Haydn sonata, No. 52 in E flat (which in those days was the only Haydn sonata in the repertoire, and seldom played at that).[5] Horowitz may have been especially attracted to it by the slow movement, which has a proto-Chopin kind of melisma. He plays the sonata with a minimum of pedal, unflagging rhythm, precise fingerwork, sharp contours – and no repeats, the omission of which is today considered a sin.

But few musicians of his time, or indeed of the nineteenth century, took repeats, and there is every indication that they were not taken even in the composers’ own time. At best they were optional and, as recent research has indicated, not many took the option. In the slow movement Horowitz manfully sings the melodic line without ever getting sloppy; and the grace of his phrasing and precise articulation in the finale, the relationships of the dynamics (never above a forte), the lack of musical trickery or cheap effect, all give the listener a Haydn to treasure. The sheer technical mastery plus the unfaltering rhythm and expressiveness combine in a performance that makes most others sound anemic.

Horowitz was never accepted as a pianist with an affinity for the Classic repertoire, and he certainly could have his problems with some of the Beethoven works he played, but his Haydn and Mozart performances generally have a purity that few associate with his name. In 1947 he recorded his first Mozart sonata – in F (K. 332). It has much the same musical approach as the Haydn. Horowitz was not afraid to use the piano as a piano, which means delicate pedal colorations, a highly expressive slow movement, and glittering passage-work in the finale.

The other concerto recordings, after the Rachmaninoff, that Horowitz made in the period before his 1953 retirement were of the Tchaikovsky B flat minor, the Brahms B flat, and the Beethoven Emperor.

The Brahms came first, in 1940. In 1987 Horowitz said that he did not like this performance with Toscanini and the NBC Symphony. He said it was too fast and too metronomic, and he was right about the speed if he was thinking of the concerto in terms of 1980s performance practice. Horowitz – or, one should say, Toscanini, for the concerto goes according to his idea of the correct tempo – takes 43’21’’. In the 1950s and 1960s, Rubinstein took 48 minutes, Serkin 47, Richter 46, Cliburn 46, and Wilhelm Backhaus 47. In the late 1970s and in the 1980s, such pianists as Vladimir Ashkenazy, Daniel Barenboim, and Vladmir Feltsman were clocked between 48 and 49 minutes in concert performances. Of his collaboration with Toscanini, Horowitz said:

Not long ago I heard a broadcast of the Brahms B flat concerto that I recorded with Toscanini and I asked myself why I ever did it. Whatever its status as music, it is not a concerto for me. I never liked it very much, and I played it so badly, and my ideas about the music were so different from Toscanini’s. Metrically he was so much stricter than I was. I didn’t enjoy rehearing this performance at all.

The Horowitz/Toscanini performance gives the feeling of power rather than poetry. And the slow movement is (uncharacteristically, for Toscanini) dragged, with Horowitz uneasily imitating a metronome. He is happiest in the last movement, the most pianistic of the four, and he hurtles through the tricky double thirds with a nonchalant flourish. Grace and style are here, but this cannot be described as one of the great Horowitz performances.

There were two performances of the Tchaikovsky with Toscanini; the first, a studio recording, in 1941, and the second live from Carnegie Hall at a war bond concert on April 25, 1943. Both performances take about half an hour, roughly four minutes faster than today’s average.

The 1941 performance is not as good as the 1943. In the earlier version Toscanini is absolutely unyielding. Horowitz goes precisely along with him. The performance has its brilliant and even majestic moments, but one can feel Horowitz dying to let loose. Restrained by Toscanini, he did not dare. But the 1943 performance is more relaxed. In the second movement there is a serious disagreement between oboe and piano, and one shudders to think of the Toscanini explosion that must have devastated the backstage area after the performance.

But the recorded performance of the Tchaikovsky that is the most exciting of all never was officially released[6], and is available only on a pirated recording. It was the Horowitz Silver Jubilee Concert on January 12, 1953, and he played the Tchaikovsky with George Szell and the New York Philharmonic in Carnegie Hall.

Wherever the pirated tape came from, it was of top quality, and when Jack Pfeiffer heard it, he thought of brining it out under the RCA label. But too many technical difficulties interfered. Among other things, Szell was under contract to Columbia.

Copies of course inevitably made their way to collectors, first on reel-to-reel tapes. Then, in the middle 1960s, microminiaturization was coming along, and recording machines not much larger than a pack of cigarettes were on the market. Music lovers were, joyously and illegally, taking them into the concert halls and going home with tapes of concerts given by their favorite artists. Now collectors could easily copy something precious and send it to friends. Pirated tapes started making the rounds, and were even sold under the counter in specialty record stores. Thus it was with the Horowitz/Szell Tchaikovsky. Pianophiles had to have a copy, and there was little trouble locating one.

It is a staggering performance, very different from the Toscanini pair. It is taken at about the same speed – 30 minutes, 40 seconds – but does not sound particularly fast except for the two big octave heroics and the last-movement coda, where Horowitz races to the finish line in record speed, skidding off the last note. He employs a rubato and a kind of freedom that he did not attempt with Toscanini. Listening to the performance, it is hard to see what Olin Downes in his review meant by his comment about not everything going as had been planned in rehearsal. The liaison is perfect – no trouble with the oboe in the second movement here! – and Szell anticipates every move.

The story goes – at least, that is how a Philharmonic musician told it to a critic – that before the first rehearsal Szell faced the orchestra and said that the Tchaikovsky B flat minor was a piece of junk. (Only he did not exactly say “junk”.) “So let’s let Horowitz do what he wants. We’ll follow him anywhere.” And Szell and the orchestra did.

Still another off-the-air recording of the Tchaikovsky made the rounds, and in 1990 it was transferred to a commercial compact disc by AS Disc, an Italian company registered in Monaco. The performance, with Bruno Walter and the New York Philharmonic from the broadcast of April 11, 1948, is magnificent, even if the last-movement octaves are little more than a glorious blur. Horowitz was still running the octaves race so decried by Rachmaninoff. The octaves apart, it has a feeling of relaxation and poetry that the other two performances lack. The Horowitz/Szell may have the most animal excitement, but the Horowitz/Walter could well be the single greatest performance ever made available to the public.

Some of the pre-1953 Horowitz recordings present a repertoire that disappeared from his concerts after a few years. In 1946 Victor issued his Mendelssohn album, dominated by a lovely performance of the Variations sérieuses. Here we have simple, flowing, elegant playing very much in the sparkling Mendelssohn style. Using very little pedal, employing a reduced scale of dynamics, Horowitz manages to suggest Mendelssohn’s own kind of Neoclassic playing. (Mendelssohn was considered one of the great pianists, but not a Romantic one; he had been taught in the Mozart-Clementi style by the classicist Ignaz Moscheles.)

1946 also saw Beethoven’s Moonlight Sonata, a low-level recording with noisy surfaces. The first movement is painfully regular except for some artificial, overcalculated dynamic effects that interfere with the natural flow of the music. For most of the movement Horowitz was trying too hard to be “Classic,” and it did not work. The second movement is better, and the finale is an explosion. He was to make a much better recording of the Moonlight ten years later.

The best Beethoven Horowitz ever recorded is probably the 32 Variations in C minor (1934). Rhythmically precise, magnificently fingered, supple in phrase, logical in contour, coherent and natural all the way through, it is a stunning performance. So was his performance of the Schumann Arabesque that same year.  

What was Horowitz’s all-time favorite recording? He was asked that, two years before he died, and immediately replied that it was the Czerny Ricordanza Variations of 1944. One can see why. This unknown piece resurrected by Horowitz is triumphant in every aspect: it is beautifully written for the piano, it has a good deal of period charm, it is played with purling passage-work and exquisite tonal adjustments, and it gives a better idea of the Horowitz tone and color than do most of his records.

Of the large-scale Chopin works that he recorded in those days, one of the most interesting is the Chopin album of 1945 because it contains the Andante Spianato and Polonaise, which he seldom if ever played after that year. A pity. The piece is one of the most effective of Chopin’s early works, and the performance is Horowitz at his best: poetic, flowing, brilliant, sprightly.

The Chopin-Liszt album of 1947 also is a prize, mostly because of the Liszt. It contains Chopin’s G minor Ballade in a performance so thought-out that it lacks spontaneity and ends up a collection of details. Horowitz fought this ballade all his life, constantly playing and recording it, never really making up his mind about how it should go. But the album also contains a ravishing performance of Liszt’s Au bord d’une source and a pulverizing one of the Sixth Hungarian Rhapsody.

Of the three Prokofiev sonatas that Horowitz introduced to America, he recorded only No. 7, in 1945. Horowitz plays it with very little pedal, which accounts for the unusually clear textures. He maintains steady rhythm, takes the toccatalike finale slower than one would have expected, and builds to an enormous climax. No wonder his old friend from Paris, Serge Prokofiev, was carried away when the first copy off the assembly line was sent to him in Moscow.

Then – talking about Russian music – there was Horowitz with the Mussorgsky Pictures at an Exhibition. He made two arrangements, the first of which he recorded in 1947. The second, more elaborate, was recorded from a 1951 Carnegie Hall recital. To read some of the outraged reviews of the Mussorgsky-Horowitz Pictures one gets the impression that mayhem was committed. Not really. In his reconstruction of the familiar piece Horowitz touched up here and there, always with taste and respect for the original. The only significant changes were in the last section, where he added a very effective bell-like counterpoint to the massive chords.

Horowitz gave the piece a pair of transcendent performances, “orchestrating” it on the keyboard and getting a coloristic quality that no pianist, not even the great Sviatoslav Richter, has achieved. The contingent that finds the Mussorgsky Pictures unpianistic and something of a bore think that the Horowitz version is an improvement. Horowitz played the Pictures for only three seasons, the last in 1951, and his version remains unpublished. Thus the chances are that it will never be heard again on the concert stage.

In the pre-LP era of the single 78-rpm disc, Horowitz recorded many short pieces – Chopin mazurkas and etudes, Mendelssohn Songs without Words, works by Debussy, Poulenc, Rachmaninoff, Schumann, and others. Many are very beautiful, and one of the colossal Liszt recorded performances of all time is on one of the early ten-inch singles – Horowitz in 1930 playing the Paganini-Liszt-Busoni Etude in E flat. The articulation of the scale passages is not of this world, nor are those fast, controlled left-hand octaves. This was the etude that Rubinstein heard Horowitz play in Paris in 1926, and Rubinstein was carried away by its “easy elegance.” Horowitz never played it after his first years in America.

Nor did he often play Stravinsky’s Petrushka, but he did record the “Danse russe” in London. Too bad he never recorded the complete three-movement transcription. The recorded performances of the “Danse russe” is terribly exciting, a thrilling example of virtuosity (which one must have for this work) coupled with all kinds of coloristic effects and a special feeling for the Russian element.

Another piece that Horowitz dropped in the 1930s (though he briefly brought it back some three decades later) was the Schumann Toccata, which he recorded in London in 1934. It so happens that his Kiev classmate Simon Barere also recorded the Schumann Toccata, that murderous study in double notes and octaves. The 1937 Barere recording followed the 1934 Horowitz. Barere is not only faster than Horowitz; he plays the fastest version ever recorded. His timing is 4 minutes, 17 seconds. Horowitz took 4 minutes, 37 seconds. Of course in both recordings the repeat was omitted.

The story goes that when Barere was ready to record the Schumann piece, he was asked by Victor what he would use as a filler on the second side. After all, that fabulous technician Josef Lhevinne had made a Victor recording of the Toccata in 1928, and did not get it on one side of the 78-rpm disc. Barere laughed. He said he could easily get it on one side, and he did. Easily. Both Barere and Horowitz play the notes as written, but Horowitz, with his coiled-spring intensity, is much more convincing musically. Barere just shows how fast he can play.

When Horowitz returned to the stage after his retirement in 1953, he no longer ended his programs with the sensational virtuoso tricks he and his public enjoyed so much. But at least he had put nearly all of them on records. His first disc recording of the Carmen Fantasy was made in 1928, followed by another in 1947 and still another in 1968. The Saint-Saëns-Liszt-Horowitz Danse macabre came out in 1942. The Mendelssohn-Liszt-Horowitz Wedding March was issued in 1946; the Liszt-Horowitz Rákoczy March and the Sousa-Horowitz Stars and Stripes Forever in 1950; and the Liszt-Horowitz Hungarian Rhapsody No. 2 in 1953.

All of these are examples of inimitable virtuosity, with the kind of technique – so effortless, so thrown off, so exultant in its easy execution of the impossible – that carries its own esthetic probity. “Is there not a time for cakes and ale?” asked Sir Toby Belch. There is indeed, and as served up by Horowitz in concert it provided a kind of frisson that created hysteria. But Horowitz grew to dislike this kind of exhibitionism; he worried about it and finally discarded it. In an interview with Jacob Siskind of the Montreal Gazette (April 24, 1976) he explained:

When you play these pieces, like I have my variations on Carmen, I find out very funny things, that the people when they hear those pieces for encores, they forget the whole program. It kills the whole effect. They say, Oh my God, did you hear him in Carmen – look what is he doing there, in the Stars and Stripes he is like a band, a band for himself – and you know that is not so good for a program.

But a lot of fun from concerts disappeared when Horowitz dropped his death-defying tightrope act. Incidentally, he considered his arrangement of the Hungarian Rhapsody No. 2 the most difficult he ever created, much more so than Stars and Stripes.

After 1948 all of the future Horowitz records would be LP discs, then stereo discs, then cassettes and finally the compact disc. With the CD, the long, proud reign of the flat, grooved, Berliner disc, from 1895 to the middle 1980s, came to an end. But what a glorious spin it had!

Appendix II
LP and Stereo

Clementi was a great composer. He was admired by Beethoven, he anticipated some of the Romantics, he was a great pianist who created the modern school of playing. Yes, even more than Mozart and Beethoven!
– Vladimir Horowitz on why he recorded Clementi

With the advent of the long-playing record, Horowitz began to put some of the major pieces of his repertoire on disc for RCA Victor. But even there he thought of the LP disc as a mini-recital, which meant that one major piece was generally accompanied by shorter items.

(In the following discussion of the Horowitz’s LPs, the actual names of the discs as they were issued, such as Horowitz Plays Chopin or Horowitz in Concert, are for the most part not given. All those discs have long been out of print. In the discography at the end of this book, information about CD reissues of any Horowitz performance – which means almost everything – can be found.)

Was the LP kinder to him than 78-rpm discs? Yes and no. Yes, because now works of up to almost a half hour in length could be fit on each side of the disc (previously the record had to be turned every four and a half minutes), and because more of the Horowitz dynamics could be captured on LP. No, because the engineer exercised a great deal of control, and not all engineers had in their ears music as actually heard in the concert hall. Too much early LP recordings favored close-up microphoning. Some Horowitz records, especially the early Victor LPs, sounded as though the listener’s ear was right over the strings of the piano. In the concert hall the Horowitz sound was much more subtle. Critics who were familiar with it only on records often heard a misrepresentation, and many of their complaints about a kind of hard clangor were legitimate.

But occasionally an early Victor LP captured the Horowitz essence. Two of the best were issued in 1950 and 1951. The first contained Schumann’s Kinderscenen[7] and seven Chopin mazurkas, and the second was a Liszt disc containing the Funérailles, the Sonetto 104 del Petrarca, the Valse oubliée No. 1, and the Horowitz arrangement of Liszt’s Rákoczy March.

The Kinderscenen represents Horowitz at his most eloquent and elegant. Horowitz was always more comfortable in Schumann than in Chopin. In Chopin he had a tendency to want to “do” something, and the listener can sometimes become more conscious of Horowitz than of the musical architecture. In Schumann he was content to let the music flow in a natural manner, with a fine feeling for tempo fluctuation, and telling dabs of color achieved through imaginative pedaling and the exploitation of inner voices. As for the mazurkas that accompany Kinderscenen, they go with charm, rhythmic snap, and a delicate rubato, as they always did with Horowitz.

In the Funérailles his amazing octaves come into play, and also his ability to voice full chords. This powerful performance is followed by the unabashed virtuosity of the Rákoczy March. The arrangement has little to do with Liszt’s original, which Horowitz largely ignored. Instead, he takes the theme as a point of departure for some giddy acrobatics. Of all the Horowitz transcriptions, this is the most conspicuously showy. He soon dropped it.

His first recording of the Chopin B flat minor Sonata came out in 1950, coupled with the Piano Sonata in E flat minor by Samuel Barber. Barber wrote it for Horowitz, and after the premiere it entered the international repertoire, where it has stayed.

This performance is definitive. Barber and Horowitz worked on it together, and what emerges is propulsive, sharply etched, massive playing. No pianist who has since taken it on has been able to command the Horowitz articulation in which every strand has X-ray definition as the architecture is laid bare.

The performance of Chopin’s Funeral March Sonata is perplexing. It is slow, mannered, and contrived. For some reason Horowitz takes the first-movement repeat; never a good idea. The modulation at the end of the exposition is one of the ugliest that Chopin, the master of modulation, ever conceived. (But Charles Rosen, the pianist-scholar, argued in 1990 that all twentieth-century editions of the sonata are incorrect, and that according to the manuscript the repeat, if taken, should start at the very opening and not the fourth measure.) Naturally there are wonderful moments, especially in the orchestration of the march, but this is not a convincing performance, and Horowitz was to record a much better version a few years later for Columbia.

The first disc of Horowitz “live” performances came in 1951, and it contains some of his most cherishable playing. Especially notable is his performance of Schumann’s Variations on a Theme by Clara Wieck (the third movement of the F minor sonata). Horowitz captures the songlike quality of the music, its sadness close to tears. It is a radiant, intimate, emotionally controlled performance.

On the disc is Chopin’s Polonaise-Fantaisie, a work seldom played fifty years ago (since then, largely thanks to Horowitz, it has been “discovered” and today is very much with us). Horowitz draws together its sprawling organization, makes magic out of the mazurkalike section, and builds to a tremendous climax. He also put Haydn’s E flat Sonata on this disc – the sonata he had recorded in 1932. This time the performance is more Romantic and very fine in its way, but without the Classic logic and elegance displayed in the great 1932 recording. For dessert there is the Stars and Stripes, even more coruscating than in the 1950 studio recording. Audiences always stimulated Horowitz to supreme efforts.

A vintage Horowitz year, 1951 also saw the Horowitz/Fritz Reiner recording of the Rachmaninoff Piano Concerto No. 3 with the RCA Victor Symphony Orchestra. Impulsive, daring, full of raw energy, Horowitz duplicates the feats he had accomplished with Coates in the landmark 1930 recording. He shapes the melodies with more color than before (or does one imagine this because of the more modern sound?) and adopts a looser rhythmic scheme. In the finale he takes the cut sanctioned by the composer. The only time he loses the rhythm is in the Più mosso section, the second time it comes around (Section 65 in the Boosey and Hawkes score). Here he gets excited and lapses into jerky triplets. On the whole this is a monumental performance, a true Horowitz experience. But there are those who insist that of the three Horowitz versions of the Rachmaninoff Third, it is the 1930 version with Coates that has most grandeur and steadiness, the one that most closely matches the splendor of Rachmaninoff himself.

There are, incidentally, early off-the-air recordings of Horowitz in the Rachmaninoff Third held by private collectors. One of them, with Barbirolli and the New York Philharmonic in May 1942, is wild. Horowitz storms through it faster than any player before or since. His timing is 32 minutes, 53 seconds. (Add another two minutes or so for the cuts. The performance even then is about ten minutes faster than the 1990s norm.) One’s first impulse is to say, “Too fast! Ridiculous!” Then one listens again, notices that every note is perfectly articulated, gets caught up in the visceral excitement, responds to the loving shape of the lyrical elements, is floored by a technique of supreme mastery, and surrenders – gladly.

The following year, 1952, Horowitz recorded the Beethoven Emperor Concerto, also with Reiner and the RCA Victor Symphony Orchestra. Horowitz was on his best behavior. This is – one can feel it going through his mind – Beethoven! One could take dictation from the piano part, so accurate is it. Every note is precisely struck, every indication carefully followed. The piano part is logically unfurled, and never is there the thickness one has heard from certain Beethoven specialists. This is an objective, supple, non-Romantic performance that turns out to be a much more Classic Emperor than many of the weightier, thicker, self-indulgent performances that are so often encountered today. It also, at 37’44’’, runs four or five minutes faster than prevailing 1990 tempos. In many respects, the Horowitz/Reiner Emperor is a revelation, even if it has been badmouthed by the purists.

Horowitz’s first big Chopin disc on LP, for Victor in 1952, contains the A flat and F minor Ballades and the B minor Scherzo. There are also the F minor Nocturne, E major Etude, and A flat Impromptu.

For some reason Horowitz seldom played the A flat Ballade [No. 3 – Ed.]. This is his only recording. He is warm and relaxed in this interpretation, and sounds as though he is enjoying himself. On the other hand the B minor Scherzo remained with him almost to the end, and he recorded it five times. The playing here is slashing, nervous, almost febrile. Horowitz swallows it whole, ending with a burst of interlocked octaves (which he always used). Chopin wrote unison scales instead of octaves, and the Horowitz addition bothered some musicians. Yet there was precedent for it. Carl Tausig, Liszt’s phenomenal pupil, used to play the ending of the Chopin E minor Concerto with interlocked octaves, and it produced such a grand effect that audiences swooned and the players in the orchestra stood up and cheered. Yes, the Horowitz double octaves ending the B minor Scherzo are an “effect”. But no notes or harmonies are changed; it is only a reinforcement, and if an artist can do it with panache it does not falsify the music.

The Ballade No. 4 in F minor was a Horowitz staple, and it stayed with him until he felt he was too old to handle its technical demands. He played it differently every time. In this recording he used a highly inflected line with all kinds of agogics. (Agogics involve the temporal variation of notes within a metrical framework; it is a first cousin of rubato.) Horowitz never loses the pulse of the music, no matter how many rhythmic changes he adopts. At the impossibly difficult coda he uses very little pedal in an astonishing display of sheer finger independence. Power and poetry are united in one of the all-time great recorded performances of the piece.

Horowitz gave his twenty-fifth anniversary concerts all over America in 1953, and Victor put most of the Carnegie Hall recital of February 23 on disc. The big piece was Schubert’s B flat Sonata. He had never played it before – he had played almost no Schubert up to then – and his ideas about the work did not get universal approval; his approach ran counter to the received opinion of how Schubert should go. Of course he was accused of Romanticizing the work. Even had he played it à la Serkin or Brendel he automatically would have been accused of Romanticism.

Yet, listening to the record without preconceived opinions, and keeping in mind performances of this work by a dozen or so of the most acclaimed pianists of our day, one wonders about the alleged Romanticisms. Horowitz certainly plays the music in a cleaner manner than any of the others. He maintains a perpetual singing line; the playing at all times moves with a fluent lyricism and transparency of texture. It is also full of striking ideas. The accentuations are different from, say, those of Pollini or Brendel, and are infinitely more exciting. The menacing pianissimo trills in the bass of the first movement, so perfectly calibrated, sound like distant mutters of thunder and are even scary.

As for the “Romanticisms,” Horowitz is, as in the Emperor, again on his best behavior, paying full respect to Schubert’s big form and intensity of utterance. But his rhetoric and inflections are so different from those of the Schubert specialists that one can understand the frightened reaction of many musicians. Perhaps the time has come for a reassessment of Horowitz in such icons as the Beethoven Emperor, the Haydn and Mozart sonatas, and this Schubert B flat. (The Horowitz’s performance of Schubert’s B flat Sonata that Deutsche Grammophon released in 1991 is nowhere near as good. He recorded it in 1986 and never approved it. His instincts were correct; this is labored and artificial playing, full of overdone rubatos that make the music sound like kitsch, and its release did no service to Horowitz’s reputation.)

The other big pieces in the anniversary album are the Chopin B minor Scherzo, Scriabin’s Sonata No. 9, and the Liszt-Horowitz Hungarian Rhapsody No. 2. The Chopin is probably Horowitz’s most successful of the versions he recorded. There is no lingering in the middle section, which can sound interminable at a slow tempo, and the final octaves are a burst of glory. In the Scriabin some details are obscured, but the performance remains one of fire and ice. Horowitz always had a remarkable insight into the dotty but wonderful world of late Scriabin. He was to play the work even more convincingly at his return concert in 1965. In the familiar Liszt Rhapsody No. 2, Horowitz does not tear up the form of the work as he had done in the Rákoczy March. What he did was add countermelodies, insert some new and difficult cadenzas, and change a few harmonies. At the end, the two themes come together à la Godowsky. Horowitz was proud of this arrangement. He said it was one of the most difficult pieces he had ever played.

Horowitz never played this rhapsody after 1953, the year he retired. But although he was not before the public for the next twelve years, he continued to record. His first record after the retirement, released in 1954, was devoted to sonatas by Muzio Clementi. That alone was a story. A supervirtuoso playing Clementi! Clementi, who wrote those Gradus ad Parnassum exercises on which every pianist is weaned!

The three sonatas that Horowitz carefully selected – he had read through every sonata Clementi had written – backed up all of Horowitz’s claims about the importance of Clementi as a creative figure. He played them simply, tenderly, lovingly. And with extraordinary virtuosity, too. In complicated passage-work every finger is equally weighted, and only a phenomenal technician could have achieved such clarity, such transparency of texture.

After the Clementi binge, Horowitz turned to Scriabin, and the recording of Sonata No. 3 and sixteen preludes was issued by Victor in 1956. The sonata is not one of Scriabin’s more arresting works, and not even Horowitz could popularize it. Normally, pianists everywhere rushed to play a work that Horowitz featured. Not this time. Horowitz helped Scriabin along by introducing a short, unwritten repeat in the last section of the second movement. It improves the structure of the music. The preludes are idiomatic, colorful pieces that were seldom played at the time – and are seldom played today, for that matter. Scriabin’s etudes are much more popular with pianists than are the preludes. Horowitz lavished all he could upon the music, and if he could not evoke much interest in it, nobody could.

In 1956, Horowitz also got around to his second Beethoven Moonlight, coupling it with the Waldstein. The Moonlight performance is much better than the 1946: looser, more personal, with much more freedom in the first movement. He burns up the keyboard in the finale and considered it one of the best things he had ever done. The Waldstein is along the lines of the Emperor – objective, clearly contoured, Classic in approach even with the resounding chords and fantastic articulation. He plays the notorious octave passage in the finale presto and staccato. No glissando, as written in the score. A glissando, Horowitz said, could be played on the light-actioned Viennese instrument that Beethoven owned but is not feasible on a modern piano.

The only other Beethoven sonata that Horowitz recorded during his long sabbatical was the D major (Op. 10, No. 3). It was made in 1959 and was the first time Horowitz had been recorded in stereo. The first movement prepares the listener for an elegant example of early-Beethoven playing. It is beautiful: a concert-grand approach with strong basses and impeccable taste. The rhythms are bracing throughout, the fingerwork sparkling, and one gets the feeling that Horowitz is having a wonderful time. But from there on all is downhill. He makes too much of the second movement, and the explosive dynamics are out of proportion. They would work in the big sonatas, such as the Waldstein and Appassionata. Here they make the playing sound overstressed, and what we get is rhetoric rather than poetry. The finale also suffers from undue emphases.

The main interest of Horowitz’s Chopin recital in 1957 is the inclusion of two Chopin scherzos – No. 2 in B flat minor and No. 3 in C sharp minor. Those were not part of his normal repertoire. With them he now had all four of the scherzos on record. In 1936 he had made a brilliant recording of No. 4 in E. No. 1 dated from 1951. Incidentally, Horowitz once said that all four Chopin scherzos should have the same tempo, adding that in No. 3 most pianists were so anxious to show off their octaves that they played too fast. Horowitz followed his own precepts by playing the octaves strictly in time, making them sound powerful and ominous rather than flurried. In the coda he is absolutely demonic.

The B flat minor Scherzo has some incredible articulation in it, along with charm and ravishing color. Horowitz, as always, got a good deal of that color by a deft emphasis of inner voices, and by his canny way of handling the bass. There was always a firm harmonic underpinning to Horowitz’s playing that anchored and reinforced what was going on in the treble.

This Chopin record also includes four nocturnes that Horowitz never again played: in B (Op. 9, No. 3), F (Op. 15, No. 1), C sharp minor (Op. 27, No. 1), and the popular E flat (Op. 9, No. 2). He plays them with appropriate sensuousness and grace. In addition there is the first Horowitz recording of the Barcarolle in a performance that suggests crystals flashing light. Here we experience Horowitz at his best: rhythm is always embedded in rock; fioritura perfectly adjusted; melodies projected in a manner reminiscent of great vocal stylists. The interpretation has poetry, power, and drama and is one of Horowitz’s greatest performances.

Closing Horowitz’s association with Victor was the 1959 recording of the Beethoven Appassionata Sonata. Again, as in most of Horowitz’s Beethoven performances, he holds back, avoiding pronounced Romanticisms. His aim was to deliver a strong, clear Appassionata. His tempos are deliberate (slower than those of Richter, Kempff, and Rubinstein in that period), the outlines are sinewy and well proportioned. It is an impressive achievement that was not universally well received, because everybody of course knew that Horowitz was not really an exponent of Beethoven. But, to repeat, when musicological revisionists get busy on early nineteenth-century performance practice, the Horowitz Appassionata will be high on their list as an example of monumental Beethoven playing.

For the next three years Horowitz made no more recordings; and Victor, unhappy with his record sales – Clementi and Scriabin, indeed! – showed no desire to lure him back to work. Perhaps Victor was relieved when he went to Columbia in 1962. Of course it was a terrible mistake for the company to have let Horowitz go. The first three Columbia records that ensued were best-sellers. Horowitz looked at his new association as a challenge; and he probably had the all-too-human determination to show the Victor executives what idiots they had been.

Columbia also gave Horowitz a richer, more mellow quality of recorded sound than Victor had offered, much closer to what Horowitz actually sounded like in the concert hall. And with a different company Horowitz could start rerecording his repertoire. He also recorded some new music that he had been looking at during his retirement. The first disc contained Chopin’s B flat minor Sonata, the Schumann Arabesque, Rachmaninoff’s Etudes-tableaux in C (Op. 33, No. 2) and E flat minor (Op. 39, No. 5), and his own version of Liszt’s virtually unknown Hungarian Rhapsody No. 19.

In his old age Liszt composed four rhapsodies after No. 15. Nos. 16, 17, and 18 are of no particular interest, but No. 19 as Horowitz arranged it is a dazzler, a throwback to the plangent zigeuner music of such rhapsodies as Nos. 6, 11, and 12. It has no great technical problems, and Horowitz touched it up to make it more virtuosic. As much of Horowitz as of Liszt is in it, and Horowitz threw himself into the performance with abandon. It is one of the most brilliant, exciting transcriptions he ever created, and he plays it like an angel or a devil – take your choice.

The Chopin Sonata is more relaxed and better thought out than the 1950 Victor. Tempos are faster and the line more secure. Horowitz this time avoided the first-movement repeat, and the sonata unfolds with a feeling of logic and inevitability, from the bold initial statement to the sotto-voce mutterings of the finale. Horowitz never played it better.

For his second Columbia record, also in 1962, Horowitz played Schumann’s Kinderscenen and Toccata, some Scarlatti, Schubert’s G flat Impromptu, and several Scriabin pieces – two etudes and the lovely Poème (Op. 32, No. 1). In the Toccata he sets the piano on fire with the velocity and clarity of the double notes and the grace of the octave section.

The Kinderscenen has all the characteristics of Horowitz’s mature style: careful, even sharp-edged phrasing, freedom of tempo, and an unusual feeling for the polyphonic lines embedded into the fabric. Few pianists today know what to do with Schumann’s polyphony, and therefore they ignore it. But in Schumann it is ever-present. Alban Berg once analyzed the apparently simple Träumerei from the Kinderscenen to show how it is, throughout, a piece in the strictest four-part harmony. Horowitz knew exactly how to balance the lines in Schumann, bringing out inner details that not only provide harmonic interest but also an amazing variety of color.

The Schubert Impromptu in G is an anomaly. Schubert wrote it in G flat. For some reason Horowitz, in his 1953 recording of the piece, had used the Bülow arrangement, which transposes it to G major. In this 1962 recording he goes to the original G flat.

For his 1963 Columbia disc, Horowitz made his first recording of Beethoven’s Pathétique Sonata, took another look at his old friend the Chopin B minor Scherzo, added Chopin’s Revolutionary Etude and the C sharp minor (Op. 25, No. 7), and then turned to Debussy for his only recordings of three preludes from Book II – Les fées sont d’exquises danseuses, Bruyères, and General Lavine – eccentric.

In the intimate world of the preludes Horowitz is very much his own man. This is not “authentic” Debussy, as the term is generally understood. For some it may be too hard edged. Yet the color is there, the wit of Lavine and the charm of Les fées are fully captured, and Bruyères has a steady, atmospheric flow. It is a most interesting look at Debussy.

The Pathétique has all the characteristics of the previous Moonlight, Waldstein, Appassionata. It is carefully organized, it is warm, powerful, lyric in the slow movement, and dramatic elsewhere. And the fingerwork has a clarity that supersedes all previous and subsequent recorded performances. Only some pulled-out ritards are questionable. The B minor Scherzo has some great moments and some spasmodic ones. As with Chopin’s G minor Ballade, Horowitz never really came to terms with this piece.

His next Columbia record was issued in 1964, the year before his return to the stage. It contained twelve Scarlatti sonatas, carefully selected for variety and contrast among the 550 or so that Scarlatti composed. It is safe to say that Horowitz had played through every one of the 550. When he developed a passion for a composer, as he had done with Clementi, he went through the entire oeuvre. Horowitz had played Scarlatti sonatas from the beginning. Now that he had retired, he could make a close study of that remarkable composer, one of the most original in music history.

What came out was a labor of love. Horowitz did not attempt to recreate harpsichord sounds. Instead he adapted Scarlatti to the concert grand with amazing success. After all, Scarlatti sonatas were composed for the harpsichord; hence, played on the piano, they amount to a transcription. He played the notes accurately as written, and with his ability to vary touch he was able to suggest the Scarlatti period, though in modern terms, playing always with taste and refinement. One of the most delightful of all Horowitz recordings resulted.

The next Horowitz record commemorated his grand return to the stage. It is as thrilling on records as it was for the listeners in Carnegie Hall on that memorable day, May 9, 1965.

Appendix III

It is not enough to emphasize the upper voice to make the melodic line sing; it must pulsate like the voice of a great singer. The wrist must feel the movement. Rubato can never be a substitute for insufficient dynamic shadings. The moment you know how to color a phrase, the excessive rubato disappears by itself. A rubato in tone has replaced it.
– Vladimir Horowitz to Jan Holcman

When Horowitz returned to the stage on May 9, 1965, the Columbia engineers were in Carnegie Hall to immortalize the event. The album, issued shortly afterward, was advertised as The Return of Horowitz to Carnegie Hall. But, as we know, it was not exactly that. Horowitz, working under extreme pressure, was not his usual infallible technical self, and he went back to the hall to touch up some sections that were not in accord with his legend. This included the one really scary spot in the recital, where he momentarily threatened to lose control – the coda of the second movement of the Schumann Fantasy.

Nevertheless the disc remains a compelling document. And some of the technical mishaps were retained. The famous slip at the beginning of the Bach-Busoni Toccata in C can be heard, for example. It is of no importance except to let listeners know that Vladimir Horowitz could be nervous and hence human. The slow movement is a miracle of color, with chords seemingly suspended in air. With the left-hand octaves of the Fugue, the old Horowitz is talking, and the audience knew that he was really back.

The Schumann Fantasy is interesting. Horowitz takes the first movement at a somewhat slow tempo, playing it powerfully, with moments of deep introspection. Throughout this work he sounds different from all other pianists. He phrases differently, he moves the basses differently, he finds nuances nobody had ever imagined. Some critics greeted this as the “new” Horowitz. It is true that there were dimensions in this performance he had not previously shown. But, then again, Horowitz never played a work twice the same way, and this performance reflected a rethinking rather than a new philosophy.

He had played the weird, wonderful Scriabin Ninth Sonata at his twenty-fifth anniversary concert in Carnegie Hall in 1953, and that performance had been recorded. It was fast, taking 6’17’’. Now, twelve years later, he played it more slowly, at 9’06’’. The result is a better performance. The 1953, exciting as it was, had some scrambled moments. Here everything can be heard, with no loss in power or drive. It is stupendous playing.

The other big piece on the recital was the Chopin G minor Ballade, which was always one of Horowitz’s problem pieces. This time he plays it with certain left-hand accentuations never heard before or since. But Horowitz plays what is written, and if he wanted to bring out things in the music not normally heard, he was being characteristically Horowitzian. If the effects do not always convince, they bring the listener up short and make him rethink the piece.

Two years later, in 1967, Columbia released an album that contained major pieces from two 1966 Carnegie Hall recitals, those of April 17 and November 27. There was Haydn’s Sonata in F, flowing and glittering, full of joy and wit. There was the first Horowitz recorded performance of Mozart’s Sonata in A, the one with the “Turkish March.” He plays it simply, the first two movements with somewhat faster tempos than are today’s norm, shepherding the notes alertly along. The “Turkish March” is, however, played rather slowly, with a steady pulse and a lovely suggestion of cymbals and percussion in the turquerie sections.

Also on the disc was Schumann’s seldom-heard Blumenstück, enchantingly played. Schumann always brought out the best in Horowitz. And there was a simply staggering performance of Liszt’s Vallée d’Obermann, an extraordinary piece. It opens with a plangent melody that could have come right out of Tchaikovsky’s Eugene Onegin, except that the Liszt was composed before 1854, the Tchaikovsky in 1878. In this interpretation Horowitz maintains incredible tension and drama, yet never neglecting the poetry. Or the virtuosity: the controlled left-hand octaves are hair-raising. The result is a sound-world that Horowitz alone could re-create, and a sound-world that also shows how big, innovative, and daring a composer Liszt was. There also is Scriabin’s Tenth Sonata, new to the Horowitz repertoire, with incredible detail and shading despite the exceptionally difficult writing. Again this is Horowitz at his greatest.

The September 22, 1968, Horowitz television broadcast from Carnegie Hall was immediately put on disc by Columbia. The big pieces were Chopin’s F sharp minor Polonaise, G minor Ballade (yet again), the Schumann Arabesque, and the Carmen Fantasy. The F sharp minor Polonaise is played with high drama; the two savage-sounding arpeggiated passages that lead to the reprise are actually terrifying. Here Horowitz, as in the Carmen (which he never again recorded), is at the height of his powers. The G minor Ballade goes spasmodically and unconvincingly.

The series of live Horowitz performances on Columbia records continued with pieces taken from recitals in 1967 and 1968. Some unusual things were present. One was the only Horowitz recording of Beethoven’s Sonata in A (Op. 101) in a problematic performance. Every note is crisp and clear, but the interpretation sounds labored, overaccented (some of the explosive attacks are disturbing), and hard. It is not Horowitz at his best, even if the last movement is articulated with superb maîtrise. Claude Frank, the Beethoven specialist, once said that he would much rather play the entire Hammerklavier than the last movement of Op. 101. The double-note passages, he said, were the most treacherous in the Beethoven canon.

Another oddity in this album is Liszt’s Scherzo and March, a virtually unknown piece. Horowitz played it in public, in his own touched-up version, only once in his life, at a Queens College recital on October 22, 1967. It is a wild work, and Horowitz is at his most awesome technical heights. But there really is not much substance to the music, which is Liszt at his most bombastic, and that may be one reason [why] Horowitz immediately dropped it. The effort involved was not worth the musical reward.

To counteract the high-wire act of the Liszt, there is on this record a lovely performance of Haydn’s two-movement Sonata in C (Hob. XVI:48). It is a work with an introspective first movement, and Horowitz plays it as such, giving a demonstration of how the dynamics of the Haydn fortepiano can be successfully transferred to the modern concert grand. The joyous quality of the last movement is fully captured, with the bouncing rhythm and finger independence that Horowitz alone had. Playing like this makes the early-music specialists of the 1990s sound like amateurs.

In 1969 Horowitz brought into his repertoire Schumann’s Kreisleriana, and of course it was promptly recorded. Most agree that this is one of his greatest records. It encompasses all aspects of Schumann’s wild imagination, from simple lyricism to dramatic outburst. The biggest contribution that the performance makes to the understanding of the piece is Horowitz’s knowledge of its polyphony. He cannily brings out the inner lines of the music, balancing them against bass and melody in a lovely, natural-sounding flow.

Also on the record is Schumann’s Variations on a Theme by Clara Wieck, which he had previously recorded so lovingly for Victor in 1951. Now he has new ideas about the piece, and plays it in a stronger, more assertive manner. It is beautiful, but the earlier recording was even more beautiful in its quieter way.

In his search for repertoire, Horowitz in 1971 came up with Chopin’s Introduction and Rondo in E flat (Op. 16). It is rarely heard. While it has some of Chopin’s most imaginative and brilliant figurations, it is musically lightweight – one of the few Chopin pieces that can be categorized as salon music. Basically it is nothing more than an effective showpiece, and Horowitz cheerfully plays it as such, with supervirtuosity. But he dropped it almost immediately.

A peculiar performance of Chopin’s A major Polonaise (the Military) was made in 1972. For some reason Horowitz seldom played this popular work. Could it be that he did not like it very much? Nothing in this performance suggests love. He plays it doggedly and metronomically, taking all repeats, hitting all the notes, but sounding bored. He takes no ritards, not even at the end. On the disc also are, among other things, mind-blowing performances of Chopin’s Etude in C sharp minor (Op. 10, No. 4) and the Revolutionary Etude.

On another 1972 Columbia disc, Horowitz recorded Beethoven’s Appassionata for the second time and also a peculiar Moonlight Sonata, his third recorded try at that work. The Appassionata is respectful, strong, beautifully organized. The Moonlight is another, sad, story. Its first movement is painfully rigid, and the finale is positively perverse. The forte-subito-piano chords are both played staccato without the loud-soft shift, and the meaning of the passage is lost. Mannerism was beginning to creep into Horowitz’s playing.

But no mannerism can be heard in his 1972 Scriabin disc for Columbia, which contains seven etudes from Opp. 8 and 42, the Tenth Sonata (previously released), two poèmes, Vers la flamme, and other pieces. The etudes, which contain some of Scriabin’s best music, come off brilliantly, especially the great C sharp minor (Op. 42, No. 5). Those left-hand smashes in the climax may seem out of proportion, but they fit the bigness of the piece. Today’s right-hand pianists are much more polite, but they lack the surge and passion that Horowitz achieved. There are other delights on the disc – the unabashed, thrilling virtuosity of the double-note D flat etude, the precision of the tiny, spooky F sharp major Etude (all fifty seconds of it), and the wild, menacing trills of Vers la flamme.

The May 18, 1976, “Concert of the Century” at Carnegie Hall presented Horowitz in an unfamiliar role. He played the first movement of the Tchaikovsky A minor trio with Isaac Stern and Mstislav Rostropovich, partnered the cellist in the slow movement of Rachmaninoff’s Cello Sonata and accompanied Dietrich Fischer-Dieskau is Schumann’s Dichterliebe. The Tchaikovsky went splendidly, with Horowitz listening carefully, playing accompaniment passages in a properly subdued manner, taking command without unduly asserting himself in his solos. The Rachmaninoff saw two brilliant Russian musicians happily playing Russian music, Rostropovich singing away with his endless bow, Horowitz making near-equivalent sounds on the piano.

In the Dichterliebe Horowitz maintained expert liaison with the singer, but there could never be any doubt as to who was at the piano. Horowitz was generally reserved but could not resist some of his usual practices – the sudden agogic changes, the little hesitations and expressive devices that his public adored and that infuriated the anti-affettuoso crowd.

On his return to RCA Victor in 1975, Horowitz was represented by excerpts from actual concerts in his first series of discs. His tours were accompanied by a retinue from the recording company, who took down virtually everything he played. Then a composite of a piece could be made. One movement might have come from New York, another from Chicago, and so on. Or perhaps Horowitz may have liked a single performance well enough to let it go through. His performance of Scriabin’s Fifth Sonata certainly has the feeling of a single performance; it moves with coherence and grand sweep. Sometimes the playing is actually brutal – deliberately so. It far eclipses all other recorded performances of the work.

Less can be said of his performance of Schumann’s Sonata No. 3, the Concerto without Orchestra. The least-played of Schumann’s three piano sonatas, it is a sprawling, thick-textured, terribly difficult piece with only one point of relaxation – the lovely third movement, which is a set of variations on a theme supplied by the composer’s beloved wife Clara and which Horowitz had recorded twice before. Throughout most of the sonata there is a lack of rhythmic steadiness in Horowitz’s playing, which imparts an unsettled feeling. Even technically Horowitz seems ill at ease, and he sounds as though he is working very, very hard. One wonders if he really liked the piece. He dropped it after one season.

In 1978 [sic: 1976 – Ed.] he returned to the Liszt B minor Sonata, which he had last recorded in 1932 and had not played for many years. His basic ideas had not changed very much. The only major difference is a quality of relaxation from a seventy-five-year-old veteran, as opposed to the more impulsive playing of a twenty-nine-years-old. Now Horowitz is more careful, and some of the difficult sections are tough going for him. Yet the playing has all the power needed for the big moments, and all the musical parameters are perfectly commanded. The Liszt B minor was still “his” piece.

But not the Rachmaninoff Third, which he once had owned. His Golden Anniversary concerts of 1978 included a Carnegie Hall performance of the concerto on January 8 with the New York Philharmonic under Eugene Ormandy, and of course it was recorded. This time none of the Rachmaninoff-sanctioned cuts were observed. Horowitz’s fingers were still up to the notes, but otherwise the performance was a near travesty, with Horowitz constantly changing tempo, introducing interminable ritards, and with many accentuations rhythmically incorrect. His timing was 43’18’’, about ten minutes slower than he had been in 1930. He was entering a period where the psychopharmaceutical drugs he was taking were clouding his judgment.

Yet he still could be capable of consummate playing. In an album of pieces from his 1978–79 recitals, he plays the Schumann Humoresque gorgeously. It is a beautiful work that he had not played for many years, and he clearly relished it much more than the previous season’s Schumann F minor sonata. He also takes on the Liszt-Busoni Mephisto Waltz, with some large-scale rewriting of his own. The Horowitz fingers could still create thunder and lightning, and Liszt would have applauded this kind of diablerie. In one of the trickier spots – the fast, big jumps towards the end – Horowitz substituted a series of convoluted maneuvers that are actually more difficult, if less hazardous.

From Horowitz’s 1979–80 recitals Victor put together an album containing a lovely performance of Clementi’s Sonata quasi concerto – poised, elegant, with a witty last movement. There is also a warm and loving Chopin Barcarolle, unfortunately a little mannered toward the end, where Horowitz holds on to certain key notes far too long. On the disc he turned to a Rachmaninoff piece he had never recorded – the coruscating E flat minor Moment Musical [Op. 16 No. 2 – Ed.] in a performance that matches in its finish the composer’s own 1940 recording, though it is completely different in approach. Horowitz is more virtuosic and uses greater dynamic extremes. He also plays Rachmaninoff’s Polka de W. R., that tasty bonbon, in a much freer way than Rachmaninoff did. Rachmaninoff is ever the reserved musical aristocrat; Horowitz is simply having fun camping it up.

More Schumann is on another disc of selections from 1979–80 concerts. Horowitz came up with the seldom-played Fantasiestücke (Op. 111). He also plays the third and fourth pieces from the Nachtstücke (Op. 23), and here we have Horowitz the poet, with simple, tender lines, and delicate dynamics. In Mendelssohn’s Scherzo a Capriccio he gives a demonstration of how fingers can do virtuoso work almost unaided by the pedal. His big piece was Rachmaninoff’s Piano Sonata No. 2 in his own version. Rachmaninoff wrote it in 1913 and revised it in 1931. Horowitz created a montage of both editions. He plays it so brilliantly, with such controlled abandon, that he almost succeeds in convincing the listener that it is an important piece of music.

At his Metropolitan Opera House concert on November 1, 1981, the novelty was Liszt’s B minor Ballade. Severe critics tend to look on this piece as Lisztian-fustian. But granted its abundance of Romantic rhetoric, it also has some of the more startling harmonies of its period and some unusually penetrating melodies. Horowitz loved the piece, and the performance is in a class by itself, even with some stretched-out sections that are more self-indulgent than musical. Some of that indulgence can also be heard in the Chopin F minor Ballade on this disc, which is otherwise a flexible and tonally sumptuous account of one of the greatest piano pieces ever written.

In the live recording of Horowitz’s May 22, 1982, concert in London’s Royal Festival Hall, he plays the British national anthem. Otherwise the program contained no surprises: Chopin’s Polonaise-Fantasy and G minor Ballade, Schumann’s Kinderscenen, and the Rachmaninoff Piano Sonata No. 2. The album came packaged with a floppy disc on which Horowitz reminisced about his life and career. Pianistically the concert went pretty well. Musically it was mannered and, in some spots, actually awkward. Horowitz was not in good shape. He collapsed physically and mentally soon after his tour, and this was to be his last record for several years.

Which did not mean that Horowitz recordings were not released during this time. Almost every piece he ever recorded did double, triple, quadruple duty. Horowitz was a one-man recording trust. Say he recorded a Chopin mazurka. Two years later it might turn up on an LP named My Favorite Chopin. Another year or two after that it might reappear in Horowitz Encores. Then it would appear once more on a digitally remastered disc. When CD became the prime recording medium, nearly all of the Horowitz recordings were promptly transferred to the new format. Not even the beloved Arthur Rubinstein was ever accorded this kind of treatment.

Appendix IV
with a Digression on Horowitz and Mozart

Pablo Casals once told me that Mozart must be played like Chopin, and Chopin should be played like Mozart. He was correct.
– Vladimir Horowitz, talking about Mozart

In the last five years of his life Horowitz concentrated on Mozart as he had never done before. He previously had recorded a few sonatas, but after 1985 he went on an intoxicating Mozartean jag: the Sonatas in C (K. 330), B flat (K. 333) and B flat (K. 281), and the Adagio in B minor [K. 540 – Ed.] and Rondo in D [K. 485 – Ed.], as well as the A major Concerto (K. 488 [No. 23 – Ed.]).[8]

Where does Horowitz the Mozartean stand?

Although a great deal of musicological work has been done on it, Classic performance practice remains a mystery. After all, when a composer named Wolfgang Amadeus Mozart is on record as saying that he never played the same piece twice the same way; that he improvised the solo parts of his concertos and, of course, the cadenzas; that he played adagios as andantes; that he used a good deal of rubato (which is mistakenly supposed to belong only to the Romantics); that the concept of a beautiful sound was integral to his thinking – with all this, we are on ground that nobody today can tread with any great confidence. In a way Mozart was the Rachmaninoff of his day, composing difficult concertos hand tailored to his own abilities, exulting in his virtuosity. “This will make them sweat,” he wrote to his father about a concerto he had just composed.

Every age makes music its own way, and the new style that came in after World War II featured fidelity to the printed note rather than what Bach would have called the Affekt – the emotional meaning of the music. Modern practice meant the use of the Urtext – the authoritative published version of the work, pure and honest, based on the manuscript, purged of editorial insertions and other excrescences of Romanticism. Musicians felt that any artist who did not play from an Urtext edition was an uncultured barbarian.

But the trouble was, as Horowitz well knew, there really can be no such thing as an Urtext, and in recent years performers have started to realize the paradoxical fact that in an attempt to be “authentic” by slavish adherence to approved Urtexts, they got further away from the essential Bach, Mozart, Haydn, Beethoven, whomever. What also resulted, thanks to the urge for “authenticity,” was the negation of personality to a point where everybody began to sound like everybody else.

That was all wrong, artistically and historically. Nearly all composers historically have expected their interpreters to take their notes and make a living organism of them, in a process where the interpreter almost achieves parity with the creator. Music, after all, means nothing on the printed page. Somebody – the performer – has to bring it to life.

In bringing Mozart to life, many exponents of the early-instrument movement have failed completely. They seem to be more interested in translation than interpretation. The fact is that whereas the early-instrument players may give an idea of what classic forces sounded like, they convey little idea of how it actually was played. It is hard to believe that their consistently dull, metronomic Mozart performances served up on the platter of “authenticity” were actually played that way by the brilliant, mercurial, impulsive, ebullient, musically aristocratic Mozart.

Horowitz was familiar with the early-instrument movement and read some of the pronouncements with interest and also disdain. He would have none of it. Horowitz insisted that Mozart himself, in his letters, told pianists how to play Mozart. And what did Mozart say? Horowitz could quote chapter and verse from memory. He would recite Mozart on rubato, Mozart on a singing line, Mozart on empty, meaningless virtuosity, Mozart on tempo and fluctuation of tempo, Mozart on key structure, Mozart on improvisation, Mozart on pianos, Mozart on piano sound, Mozart on singers and singing, Mozart on taste, Mozart on expression.

Mozart, Horowitz believed, must be played so that every note sounds. His music must be played with a singing line, with taste and, above all, with expressiveness. Because Mozart’s textures are so spare and transparent, he needs more color, not less. Horowitz believed that the modern piano could achieve a Mozartean sound by a canny use of the pedal, though the pedals of course should not be used in Mozart as lavishly as in Romantic music. It could well be that Horowitz was closer to the essential Mozart than the modern, early-instrument specialists, who – as such revisionists as Richard Taruskin and Robert Levin have recently been pointing out – play Mozart not in Mozart’s style but in a modern, literal, objective style.


Then there is the problem of rubato, which Mozart wrote about in great detail. Rubato is the delicate displacement of rhythm for expressive purposes. Mozart’s and Chopin’s rubato were alike: the left hand remained steady, the right hand could wander. As early as 1723 Pier Francesco Tosi wrote that good taste in playing included “going from one note to another with singular and unexpected surprises, and stealing the time exactly on the true motion of the bass.” (Rubato in Italian means “stolen”.) All sensitive artists use fluctuation of tempo, otherwise their performances would be metronomic. It is believed that Mozart wrote out some measures of how he played tempo rubato. They can be found in the Adagio of his F major Sonata (K. 332). In this notation, he syncopates the right hand against the left-hand bass.

Horowitz set forth his ideas about Mozart on the piano in his program notes for the Deutsche Grammophone Horowitz at Home recording in 1989. He took the position that “all music is the expression of feelings, and feelings do not change over the centuries.” Therefore he discarded such labels as Classic, Romantic, Modern, Neo-Classic. All music is Romantic. Style and form may change, but not the basic human emotions. “Purists would have us believe that music from the so-called Classical period should be played with emotional restrain, while so-called Romantic music must be played with emotional freedom.” The sad result, according to Horowitz, is exaggeration on both sides: overindulgent, uncontrolled performances of Romantic music, and dry, sterile, dull performances of Classical music.

To play Mozart properly, Horowitz wrote, the requirement is not interpretation but a process of subjective re-creation. He pointed out what every experienced musician knows: “The notation of a composer is a mere skeleton that the performer must endow with flesh and blood.” Expression is more important than an Urtext. “Shouldn’t the performer listen to his heart rather than to intellectual concepts how to play Classical, Romantic, or any other style of music?” True, mastery implies control. “But control that is creative does not limit or restrain feelings or spontaneity. It is rather a setting of standards and boundaries in regard to taste, style, and what is appropriate to each composer.”

Horowitz spoke about Mozart playing for his projected autobiography:

In five days I could memorize and play the Funérailles of Liszt but I could not play a sonata by Mozart. If you give me a Mozart sonata, I promise you that I can learn the notes by heart in one day. That is not hard. No problem. But if I am going to play it the way I think it should be played, with the colors I hear in the music, emotionally, with a singing line, I have to spend a long time working like crazy. In any movement of a Mozart sonata are forty colors, fifty colors, sixty colors. But pianists are no longer receptive to color. Mozart demands a unique kind of color. I try to give him that color. He had the same blood, the same veins as we have. Now in the last years of my life I go to Mozart because he is the most difficult composer. I am familiar with the authentic instrument craze but am not impressed. Mozart is much better on the modern piano. You can do much more. It’s the same with Bach and Scarlatti. You lose more with Chopin on the modern piano than you do with Mozart. Chopin could play his etudes on his light-action Pleyel, but some of the etudes are impossible to play on the modern piano.

Horowitz’s performances of the Classic repertoire adhered to his strictures. In his early Mozart and Haydn performances on records, Horowitz played with a wonderfully bouncing rhythm, very little pedal, some delicate tempo fluctuations, and a perpetually singing line. In the Classic performances of his last years, the rhythm and articulation remained much the same, but the tempos were a shade slower and more delicate color washes were applied.

In the Mozart C major Sonata (K. 330), which he first played in his 1985 Last Romantic film (Deutsche Grammophon immediately put it on CD), and then in Moscow, the musical approach is simple and direct, even with the infusions of color. The dynamics are carefully calculated, all between piano and forte (no pianissimos or fortissimos). The clarity of the playing contradicts some assertions of early-piano specialists who insist that the modern grand piano is not capable of a fortepiano’s subtleties. In the slow movement Horowitz practices what others preach: he actually makes the music sound like an extended operatic aria, and the last four measures are simply melting: Susanna and the Countess in a hitherto unknown Figaro aria. The finale is sweet and graceful.

Mozart’s B flat Sonata (K. 333) as recorded by Horowitz in 1987 has some unusual ornaments. He substitutes turns and other devices for move conventional appoggiaturas and uses them in the best of taste. It is hard to believe that the playing is that of an eighty-four-year-old man. There is life in it, relish and joy, even rapture; Horowitz takes all of the repeats, including the long one in the first movement.

Whether he believed in repeats as important to structure or used them to fill out a not very long compact disc is a moot question. There is every reason to believe that in Mozart’s own day (and Beethoven’s, too) repeats were generally ignored. At best they were optional. The conductor George Smart, who introduced many Haydn, Mozart, and Beethoven symphonies to London in the early 1830s and 1840s, seldom took repeats. Once in a while he entered in his score, “Took repeat.” That did not happen very often.

On this Deutsche Grammophone disc is also the A major Piano Concerto (K. 488), with Giulini and the Scala orchestra. The performance can also be seen on a videodisc (without the sonata). As in all of the Horowitz videodiscs, the cameras concentrate on his hands, and one marvels at his flat-fingered approach to the technical problems. Not that there are any problems for him in this work. The first movement has, if one can use an oxymoron, calm tension. The melodic elements are of course emphasized in this most melodic of concertos, but there is an inexorable buildup to the ingenious Busoni cadenza (which flashes and disappears after seventy-two seconds). Horowitz conceived the second movement as a siciliana (a slow dance in six-eight time) rather than an adagio as marked. This means a faster than normal tempo for the movement, though Horowitz never races. He is intent on keeping the music in constant, elegant flow, and what comes out is an elegy rather than a dirge. The finale is all fun and games. Giulini accompanies faithfully, but one feels that his ideas about the music are much more sober and post–World War II traditional than those of Horowitz.

The last Mozart that Horowitz played appeared in the Deutsche Grammophon Horowitz at Home album. It contained the Sonata in B flat (K. 281), the Adagio in B minor, and the Rondo in D. In the sonata Horowitz again takes all repeats. The first movement is brisk, beautifully articulated, with the controlled flow that Horowitz always brought to his scale passages. In the second movement he starts with a rhythmic error, the right hand coming in with accented upbeats that throw the rhythm off for two measures. He gets it right when he takes the repeat. At measures 53–56 he makes a slight textual change, and in the last movement he inserts a tiny cadenza at measure 70. If one were to be really authentic, it would be mandatory to add many more cadenzas, and to embroider the slow movement much more than anybody today (Horowitz included) would dare to do.

Horowitz responds fully to the piercing sadness of the B minor Adagio’s chromatic harmonies. This is a demonstration not only of expressive playing but also of the way the two hands of a pianist can be in perfect balance. Horowitz uses the left hand to apply just the right combination of harmonic solidity and color without being obvious about it. Yes, it is a “Romantic” performance. But the piece could well be, with the Rondo in A minor, Mozart’s most Romantic writing. The harmonies are far in advance of their day.

Of course Horowitz in his last years recorded other things than Mozart. For Deutsche Grammophon in 1985 he recorded the Schumann Kreisleriana. Could he have had anything further to say about this piece after his great 1969 recording for Columbia? Yes. The tempos are now a bit slower (31’04’’ as against 28’04”), there is a more relaxed feeling, and a simpler, more reflective approach to the music. The 1969 performance swept the listener away with its ardent Romanticism. The new one, less physically exciting, has more warmth and even more color. The technique remains solid; everything is in splendid order. This was the most ambitious work Horowitz recorded in his last five years.

Also on the disc was a rarity: a nocturnelike Liszt Impromptu in F sharp, unknown even to most Liszt specialists. Liszt had Chopin in mind here, as the Horowitz performance makes very clear. The disc contains, in addition, a curiously subdued performance of the oft-recorded (by Horowitz) Scriabin D sharp minor Etude. Schubert’s B flat Impromptu (Op. 142, No. 3) and the Schubert-Tausig Marche militaire fill out the disc. Grace and suppleness, with purling, precisely weighted scale passages in the last variation, mark the B flat Impromptu. It is interesting to compare Horowitz with the recording made by Schnabel around 1950. Schnabel’s is equally great, with overpowering authority and a masculine sense of poetry (and he too does those scale runs with extraordinary finesse). The Horowitz is more feminine, more interesting harmonically (thanks to the inner voices he brings out), more colorful.

As for the Schubert-Tausig, it is great fun. At the turn of the century it was very much in the repertoire, and was recorded by such giants as Hofmann and Godowsky as well as a host of lesser pianists. Horowitz touches up the music a bit, adds a flourish here and there and almost reminds one of the Horowitz from Stars and Stripes or Carmen.

About two years later, in 1988, Deutsche Grammophone brought out a disc containing three of Horowitz’s most enchanting performances.[9] All are Liszt transcriptions of Schubert: the Soirées de Vienne Nos. 6 and 7, and the Ständchen, better known as the celebrated Schubert Serenade. They make the listener wiggle with pleasure, especially the Serenade, with its amazing separation of voices – such control! – and its subtle colorations, together with a melodic line that extends into infinity.

The last Horowitz disc, completed just before his death, is remarkable. It contains music he had never before recorded: Haydn’s Sonata in E flat (No. 49), a Chopin group consisting of the Fantasie-Impromptu, Mazurka in C minor (Op. 56, No. 3), the Nocturnes in E flat (Op. 55, No. 2) and B (Op. 62, No. 1), and the Etudes in A flat (Op. 25, No. 1) and E minor (Op. 25, No. 5). Then there were the Praeludium to the Liszt Weinen, Klagen, Sorgen, Zagen Variations, and the Wagner-Liszt Liebestod.

Every piece on the disc is a spectacular example of great piano playing. The Haydn Sonata, a witty and unconventional work full of charm and unexpected darts, receives a performance that extracts everything in the music. Horowitz uses more dynamics here than he generally did for Classic pieces, but everything he does comes out beautifully without erasing the Haydnesque character. If Haydn is to be played on the modern grand, this is the way to play it. If Haydn is to be played on a period instrument, the player would do well to listen and learn from Horowitz’s expressiveness.

One wonders if Horowitz selected the E flat Nocturne because of the Ignaz Friedman recording of the mid-1930s. Horowitz said many times (and he was so right) that nobody had every played it better. Perhaps he wanted to show that he, Vladimir Horowitz, was no slouch either. There are a few Friedmanesque ideas in the Horowitz performance, and also a great deal of prime Horowitz in the registrations and the way he moves the bass line to get such rich textures. Is he more convincing than Friedman? Let’s call it a standoff.

The B major Nocturne is one of the sexiest things that Horowitz ever did: positively orgiastic, yet never in bad taste. It is a prime example of what the British critic Bryce Morrison called Horowitz’s “Romantic polyphonic style, one which could ‘voice,’ scatter or realign any harmony or texture at will.” The Fantasy-Impromptu is cooler playing. Determined to avoid sentimentality in the middle section, Horowitz alertly moves it along and suddenly it no longer seems interminable. The fingerwork of the fast sections is a Wunder. Of the two etudes, the A flat is supple and singing, but one wonders why at the end of an otherwise deliciously played E minor Etude he holds on to the fortissimo trills so long and so loud. It rather overbalances the piece.

Nobody had ever heard of the Weinen, Klagen Praeludium until this disc came out.[10] There is a big Liszt set of Weinen, Klagen Variations on a theme from a Bach cantata, recorded by Alfred Brendel and others, but that is not what Horowitz plays here. He found among his sheet music a short, separately published work on the Bach theme that Liszt appears to have composed before his big variations. It is a dense, harmonically fascinating example of Liszt’s later period, moody and gripping, and Horowitz plays it with extraordinary richness and intensity. Here again he is in a sound-world that only an old magician like him could bring to life. It was the last piece of music he recorded for this disc, though the Liebestod was, inevitably, placed last. Toward the end of the Liebestod, Horowitz makes a change in the bass to provide a thrilling effect.

It was billed as The Last Recording, and indeed it was. But that does not mean no Horowitz records will be issued in the future. A disc of his last Vienna concert was released early in 1991. So was a CD of the Vienna Kinderscenen and the rejected 1986 Schubert B flat Sonata. More inevitably will be forthcoming. There still is unreleased material in the “icebox” of the various companies that made his records. The Horowitz Archives at Yale contain full concerts he gave in the 1940s and 1950s, and it may be that these will eventually go public. Collectors have pirated cassettes of Horowitz recitals, and those too may be made commercially available. The public will most certainly be deluged with “new” Horowitz recordings for many years to come. And who knows but that in some attic, or hoarded by some reclusive collector, are off-the-air copies of Horowitz in the two Liszt concertos that he was playing in the early 1930s.[11]

It was a huge discography that Horowitz left, and it exemplified what he once said to the German critic Joachim Kaiser: “Piano playing consists of intellect, heart and technique. All should be equally developed. Without intellect you will be a fiasco, without technique an amateur, without heart a machine. The profession has its perils.”[12]

Perils, yes. But also joy, wonderment, and, sometimes, fulfillment. All artists pursue a Grail, that all-but-impossible quest for those rare moments when the light of the music transfigures them to the point where performer, composer, and audience are as one. Vladimir Horowitz consistently achieved that goal, and that is why he was the dominant force in piano playing for the last half of the twentieth century.

[1] Horowitz: A Biography (1983) by Glenn Plaskin and Evenings with Horowitz: A Personal Portrait (1991) by David Dubal. Mr Schonberg seldom if ever uses these volumes in his work. Ed.
[2] The only commercially released recording of Horowitz playing Liszt’s Second Ballade is the one from Metropolitan in 1981 (Live). If, as is very likely, it was this performance that was played on the radio, and if the quote, as is very likely again, comes from the 1987 interviews with Schonberg (see Preface), Horowitz’s taste changed rather quickly. Ed.
[3] Dover, 1991, vol. 1, p. 124-5. 2 vols. bound as one, first published separately in 1980 and 1988. The words in the Dover edition are somewhat different, though their meaning is pretty much the same:
So many people who are studying piano study with recordings, and they are so used to hearing note-perfect performances on record that they want to duplicate the same note-perfect performance in the concert hall. They are not concerned about projecting the spirit of the music because they are concentrating so much on the notes; it becomes an obsession with them. […] With recordings today, it is mechanically possible to do what I worked and sweated so many years to develop.
The correction in the square brackets was apparently Mr Schonberg’s. It is difficult to find out exactly when Horowitz was interviewed by Ms Mach, but it was probably sometime in the late 70s. Ed.
[4] The fifth edition of Grove's Dictionary of Music and Musicians was actually published in 1954. Ed.
[5] This was actually a world premiere recording. Ed.
[6] Since 1992 it has been several times, most charmingly, and incongruously, on a 3-CD set from Andromeda the rest of which is occupied by commercial solo piano recordings made for RCA between 1942 and 1953. Ed.
[7] All throughout the book Mr Schonberg uses this English-German travesty of the title. The correct one is, of course, Kinderszenen. Ed.
[8] Horowitz also recorded the Rondo in A minor, K. 511, but it was first released on The Magic of Horowitz set (DG, 2003, 2CD+DVD). Originally it was supposed to be included in the Horowitz at Home (1989) album, but the pianist felt there was too much Mozart on the disc (it also contained KK. 281, 485 & 540) and decided to save it for a later release. Ed.
[9] This disc is actually the aforementioned, several times, Horowitz at Home, and it was released in 1989. Ed.
[10] This is one of Mr Schonberg’s not untypical dramatizations. The Praeludium was recorded by Leslie Howard in February 1988 for vol. 3 in his series of Liszt’s complete piano music. This disc was released in 1989, a year before Horowitz’s The Last Recording. Ed.
[11] Mr Schonberg’s dream about the Liszt concertos has unfortunately not come true, but otherwise he was prophetic. Some of the Yale stuff was released between 1994 and 2009 in five volumes of the so-called The Private Collection. These include plenty of new pieces to Horowitz’s discography (e.g. Chopin’s F minor Fantaisie, Liszt’s Legende No. 2, Balakirev’s Islamey, and others) as well as fascinating “new” recordings of warhorses (most notably a stunning Liszt Sonata from 1949). Two new recitals, one from 1951 and one from 1967, were included in The Complete Original Jacket Collection (Sony, 2009, 70 CD). Other massive sets like Vladimir Horowitz: Live at Carnegie Hall (Sony, 2013, 41CD+1DVD) and the soon to be released Vladimir Horowitz: The Unreleased Live Recordings 1966-1983 (Sony, 2015, 50 CD) contain loads of previously unreleased material. So Mr Schonberg was spot on. We are deluged with “new” Horowitz recordings. We don’t complain and neither would he. Ed.
[12] Liner notes to the album The Studio Recordings – New York 1985 (DG, 1986). Ed.