Handel--another of the greatest Baroque composers--also seized attention when he led his operas from the harpsichord in London, though he wasn't as extreme as Vivaldi. But all harpsichord parts in Baroque operas were improvised, and Handel's improvisations were by all accounts spectacular. Which brings us to what's surely the biggest difference between musical performances in the days before the concept of classical music arose, and classical music performances today. Back in those ancient days--with the audience talking, carrying on, and applauding whenever it heard something it liked--musicians often improvised, whereas now (with the audience sitting in well-behaved silence) they're trained to play only the notes that composers write.

How do we know that they improvised? Because people wrote descriptions of what they did, textbooks on how to do it, and transcriptions into musical notation of what musicians actually played. Corelli, yet another of the great Baroque composers, composed violin music that, as he wrote it down, often looks very simple. But he didn't mean it to be played that way. He himself played it--especially in slow pieces--with cascades of unwritten notes, which he'd make up on the spot, and play differently in each performance. Here's an example, printed in The New Grove Dictionary of Music and Musicians, the leading classical music reference work. It comes from a 1710 edition of Corelli's violin sonatas (reprinted in 1716), which stated, in no uncertain terms, that this is "how Corelli wants one to play these pieces." Not, of course, that any serious musician would simply copy what was printed, but (as the New Grove writer stresses) instead would take the printed score as an example "to be emulated by the performer in his or her own manner." As you can see, the musical notation vividly shows--even if you don't read music!--the difference between what Corelli wrote down, and what he expected to be played. The middle line shows what's written; the top line shows how it should be played. (The music on the bottom is the bass line.)

How could these embellishments change at every performance? Just look at the examples of improvised music published in a famous 1752 book, Essay of a Method for Playing the Transverse Flute, by Johann Joachim Quantz, an important 18th century musician. In one chapter, Quantz gives 20 pages of examples, ideas of what someone might do with simple musical passages. And I do mean simple--the first of them is nothing more than the same note played three times: dah dah dah. Quantz suggests 21 ways to play those three notes, or rather to enhance them. It's as if you got up and took three steps forward, 21 times, sometimes waving your arms, sometimes spinning around at each step, sometimes adding steps sideways, sometimes walking on tiptoes. Quantz's first variation simply adds a flourish to each note: da-da-da-dah, da-da-da-dah, da-da-da-dah. Other variations involve scales running up and down, or complex little twists and turns, but always starting and ending on the same note, and rhythmically dividing into three parts, so the relation to the original idea is always clear.

And of course these examples could be endlessly multiplied, because so many of them have come down to us. Baroque opera, it turns out, was all about vocal embellishment, quite apart from whatever fireworks composers like Vivaldi and Handel could set off in the orchestra. Operas consisted almost exclusively of solo arias. Each aria was divided into three parts, an opening section, something else that contrasted with it, and then a repeat of the opening. In the previous episode, I mentioned a basic misunderstanding of Baroque opera, and it's all about this way of constructing an aria. As we see it now, it's rigid, unyielding, stylized, even undramatic. Once it begins, nothing's going to change. The music always comes back to the place it started from.

But to 18th century audiences--the people who sometimes screamed their approval, and sometimes didn't pay attention to the music at all--this aria form (called the da capo aria, because da capo meant to repeat from the beginning, and that's what these arias did) was nothing less than explosive. The whole point was to see what the singers would do with the repeats. Of course they'd vary them. Of course, if they were very good singers, they'd very them differently each night. So the da capo aria wasn't rigid, wasn't stylized, and above all, wasn't boring. The variations could be wild; purists repeatedly complained that the composer's melody would entirely disappear. Whenever a singer stepped out to sing, anyone listening would wonder, "What's he going to do this time? How wild is he going to get?" Or how touching, how noble, how virile, how yielding, how melancholy, how furiously angry, or how utterly despairing, depending on whatever emotion the music was meant to convey. No wonder these audiences--or at least the people who were paying attention--sometimes screamed. They were ravished, excited, taken by surprise. And no wonder the singers tried so hard to amaze the audience. How else could they make people pay attention?

But there was art in all this, too. Stendhal, the great 19th century French novelist, loved opera, and, to judge from some of his writing, must have gone to the opera every night during a time when he lived in Italy. He wrote a biography of Rossini, the most exciting Italian opera composer of the early 19th century (and, by miles, the most popular composer in all of Europe). This isn't an accurate book, if you're looking for the facts of Rossini's life. But it's a treasure, if you want to know what opera felt like back then. Musical styles had changed, since the time of the Baroque, and nobody wrote da capo arias. But singers still embellished everything composers wrote, and Stendhal thought this was above all necessary because composed music couldn't. by itself, convey the finer nuances of human emotion. He writes a little sadly about all this, because he's afraid that musical styles were changing, and that composers, especially Rossini, were starting to write down exactly what they wanted the singers to sing. Here's what he writes:

In days gone by, the great singers, Babbini, Marchesi, Pacchiarotti, etc., used to compose their own ornamentation [all the italics in this passage are Stendhal's] whenever the musical context required an exceptionally high level of complexity; but in normal circumstances, they were concerned with extempore invention. All the various categories of simpler embellishments (appoggiture, gruppetti, mordenti, and so on) were theirs to dispose and arrange as they thought best, spontaneously, and following the dictates of their art and their inner genius; the whole art of adorning the melody (i vezzi melodici del canto, as Pacchiarotti used to call it, when I met him in Padua in 1816) belonged by right to the performer. For instance, in the aria

Ombra adorata, aspetta .. [from Vaccai's opera Giulietta e Romeo].

Crescentini would suffuse his whole voice and inflexion with a broad and indefinable colouring of satisfaction, because it would strike him, while he was actually standing up and singing, that an impassioned lover about to be reunited with his mistress probably would feel something of the sort. But Velluti, who perceives the situation rather differently, inter­prets the same passage in a vein of melancholy, interspersed with brood­ing reflections upon the common fate of the two lovers. There is no composer on earth, suppose him to be as ingenious as you will, whose score can convey with precision, these and similar infinitely minute nuances of emotional suggestion: yet it is precisely these and similar infinitely minute nuances which form the secret of Crescentini's unique perfection in his interpretation of the aria; furthermore, all this infinitely minute material is itself in a perpetual state of transformation, constantly responding to variations in the physiccal condition of the singer's voice, or to changes in the intensity of the exaltation and ecstasy by which he may happen to be inspired. At one performance, he may tend towards ornaments redolent of indolence and morbidezza; on a different occasion, from the very moment when the sets foot on the stage, he may find himself in a mood for gorgheggi [cascades of notes] instinct with energy and life. Unless he yield to the inspiration of the moment, he can never attain to perfection in his singing.

A book by Manuel Garcia, Jr.,--a top 19th century singing teacher, and son of the man who created the leading tenor role in Rossini's Barber of Seville--gives many examples of 19th century vocal embellishments. In one of them, while the orchestra rests, a tenor holds a long note, trilling on it, and three times swelling his voice to make it louder, then pulling it back to sing softly again; after this (and still, presumably, on the same breath), he launches into gorgheggi. No singer could do that today; the sound must have been astonishing. In other, simpler, examples, singers take the written music, and--sometimes with very small changes--make it wonderfully personal. Giuditta Pasta, who created the title role in Bellini's Norma, takes little groups of four notes from that music, and turns them into unexpected groups of five. Garcia's father, in an excerpt from his Barber role, adds a charming little flourish that bursts out like a delighted smile (it's on the word "stral"):

Sometimes, Garcia shows singers changing the rhythms the composers wrote, as if they were singing jazz, in one case even losing the rhythm in the orchestra, and going off for a moment on their own. This was a well-known expressive effect, which pianists used, too, letting the melody in their right hand diverge from the accompaniment their left hand played. In an 18th century book on singing there's a touching example, reproduced in musical notation. Over a simple bass line, a singer descends from a high note to a lower one, singing a scale downward, but varying the length and loudness of each note, and lingering over the descent, so that she makes the music entirely her own, feeling free to finish a beat or two after the bass line does. As the writer, Pier Francesco Tosi, says (with his 18th century Italian translated into 18th century English),

When on an even and regular Movement of a Bass, which proceeds slowly, a Singer begins with a high Note, dragging it gently down to a low one, with the Forte [loud] and Piano [soft], almost gradually, with Inequality of Motion, that is to say, stopping a little more on some Notes in the Middle, than on those that begin or end. Every good Musician takes it for granted, that in the Art of Singing there is no invention superior, or Execution more apt to touch the Heart than this...

And so it goes. Bach improvised. Mozart improvised. Beethoven improvised, sometimes for an hour at a time, sometimes making people cry, sometimes bringing a performance of his chamber music to a halt:

...he played his Quintet for Pianoforte and Wind Instruments with Ramm as soloist [wrote someone from his time]. In the last Allegro [the final movement, played at a fast tempo] there are several holds before the theme is resumed. At one of these Beethoven suddenly began to improvise [on the piano], took the Rondo for a theme and entertained himself and the others [the audience] for a considerable time, but not the other players. They were displeased and Ramm even very angry. It was really very comical to see them, momentarily expecting the performance to be resumed, put their instruments to their mouths, only to put them down again. At length Beethoven was satisfied and dropped into the Rondo [resumed the final movement]. The whole company was transported with delight. [Though probably not the musicians.]

But then Beethoven wasn't alone. Improvisation was a normal part of piano performances. Pianists in the 19th century commonly improvised preludes to the notated pieces they performed, a practice so widespread that there's even a name for it, "preluding." And even late in that century, when many of these practices had mostly disappeared, a singer, finding a note in one of Brahms's pieces too high to sing comfortably, asked Brahms if he could change it, and Brahms said yes: "As far as I am concerned, a thinking, sensible singer may, without hesitation, change a note which for some reason or other is for the time being out of his compass into one what he can reach with comfort, provided always that the declamation remains correct and the accentuation does not suffer." [emphasis in the original]

So what does all this mean? First, I think we misunderstand the music I've been talking about. Nothing I've written here will come as surprise to musicologists; everything I've mentioned (except perhaps the piano preluding) is very well known. And yet performers today rarely do what would have been done when the music was new. Singers singing Baroque opera, or Bellini or Rossini, barely ornament their music, at least by 18th or 19th century standards. Da capo arias are meticulously staged (the Metropolitan Opera's recent production of Handel's Rodelinda provides a lovely example) to show some kind of motivation for the first-section repeat. The idea, now, is to show how canny a dramatist Handel was; I've never seen a production, or heard one on any recording, that treated the repeats first of all as occasions for vocal display, and let Handel's drama emerge through that. René Jacobs, in his recording of Handel's Rinaldo, at least lets the orchestra improvise (or play written ornaments of the kind that would have been improvised); the result is wonderfully fresh, a great surprise, both terrific art and irresistible entertainment.

And so we lose the spirit of the music, and, I might argue (Christopher Small certainly does), a lot of the spirit of musicmaking itself. Why did this happen? Because of the concept of classical music, which, as we'll see, emerged in the 19th century, and put the composer at the center of the musical universe. The purpose of playing music, at least in the classical world, was to realize the composer's intentions, and that meant playing the notes the composer wrote. And only those notes--even when the composers themselves expected performers to add something of their own. Historical research reveals, beyond any doubt, that this was expected. But research is trumped by the orthodoxy of the present day, and even some scholars get caught up in the contradiction. An important Mozart scholar, Frederick Neumann, starts a discussion of improvised changes in Mozart's works like this: "Mozart lived at a time when composers still gave the performer license to make certain improvisatory additions to the written text." License! Wham! Changes to the written text are, by their very nature, evidently suspect. No wonder Neumann concludes that these changes should be very rare, as opposed to other scholars (the pianist Robert Levin, for instance) who welcome them. Compare that Neumann sentence to Neil Zazlaw, who in his book on Mozart's symphonies happily talks about "Mozart's ornament-loving instrumentalists."

So am I urging us to return to some 18th century (or early 19th century) paradise? Hardly. There were many problems then. Performances, by our standards, were very likely bad. Again by our standards, they were barely rehearsed. Writers of the time complain about singers and instrumentalists who introduced too many ornaments, making the music unrecognizable. (Note Brahms's caution: Singers should only change the music if their changes didn't hurt it.) Singers carried around "baggage arias," as they were called (because their music was, as Stendhal says, "carried around permanently, as it were, like a change of underwear"), which they'd introduce into every opera they sang. 18th century orchestras seem to have improvised, with, sometimes, all the violinists individually--and, you'd think, cacaphonically--adding their own embellishments of the written violin line.

But let me say it again. The spirit of those long-lost days is something we ought to recapture. At the very least, we ought to know that we've lost it. And by losing it--by evolving the concept of classical music, in which improvisation was all but illegal--we may have sown the seeds for classical music's current decline.

This is the last book episode I'm going to post until September. As I've said, I'm taking July off, and (with apologies to everyone) will close this website to comments. If you find you're able to post them, they won't show up on the site. I'll be back at work in August, but I'll delay any further episodes till after Labor Day--when I'll talk about the idea of classical music, and how it evolved. I suspect much of what I've been writing in this episode and the last two will be shortened in the final book, or at least I'll make them a little less academic. Have a terrific summer, everyone, and feel free to make comments during the rest of June. I'll be happy to greet you all again in the fall.

If you'd like to subscribe to this book--which above all means you'll be notified by e-mail when new episodes appear--just click here. Subscribers help me; I feel wonderfully encouraged each time somebody new asks to be put on my subscription list. And feel free to add a note to your e-mail. I'm always curious about who's subscribing, and why you're all interested. That often leads to an e-mail exchange, and often enough to some sharing of ideas (from which I learn a lot). Or let me put it this way: Even if you don't work in the classical music business, you become part of the network of people who've helped me with the book, which means (as I explained in episode six) that the book is partly dedicated to you. I'll also offer special goodies to subscribers--segments I haven't published online, revisions of online episodes, the book proposal I'll eventually send to a publisher, second thoughts on things I've written, special comments, and other things I can't imagine yet..

Please comment on the book, too. Below you'll see where you can post comments, which can either appear with your name attached, or anonymously. Anyone who posts a comment of course also becomes part of the network I'm so happy about. The comments have helped me enormously.

My privacy policy: I'll never share my subscriber list with anyone, for any reason. I send all e-mail to my list myself, without routing it through anyone at ArtsJournal. And I send all e-mail with the names of the recipients hidden. All subscribers have their privacy protected at all times. Further: if you e-mail me about the book, I'll consider your e-mail private. I won't quote it in any public forum without your permission. Comments you post here, though, of course appear in a public forum, and thus can be freely quoted by me or by anyone else who reads them.

Music that got me through this episode:

Comparisons between cut and uncut recordings of three operas, for an article on cuts I'm writing for Opera News. The recordings:

Wagner, Die Walküre, Solti studio recording, with Hans Hotter and Birgit Nilsson, Wotan's monologue in Act 2, uncut. I compared this to a live performance from Buenos Aires in 1940, conducted by Erich Kleiber with Herbert Janssen and Marjorie Lawrence, shockingly cut.

Wagner, Tristan und Isolde, Furtwängler studio recording with Kirsten Flagstad and Ludwig Suthaus, first part of the love duet from Act 2, uncut. Compared it to a 1936 live performance at Covert Garden, conducted by Fritz Reiner with Flagstad and Lauritz Melchior, heavily cut.

Donizetti, Anna Bolena, studio recording with Beverly Sills, Act 1 duet for Giovanna Seymour (Shirley Verrett) and Henry VIII (Paul Plishka), uncut. Compared to live 1957 La Scala performance with Maria Callas (Giulietta Simionato and Nicola Rossi-Lemeni in the duet), not just heavily cut, but virtually sculpted into a new opera written in a different musical style. The most striking musical moment in the duet, in this performance--an abrupt and bracing appearance of a new melody--is actually created by one of the cuts.

Citations:

Richard Taruskin, The Oxford History of Western Music. Oxford, New York: Oxford University Press, 2005. Vol. 2. p.173-4.

"Coping with Life After Juditha," by Alessandro De Marchi ( translated by Charles Johnston), in liner notes to Vivaldi, Orlando finto pazzo, Naïve Records, OP 30392, 2004.

Corelli ornaments: Michael Collins/Robert E. Seletsky: "Improvisation/Western Art Music/Baroque period/(iv) Later italianate embellishments," Grove Music Online ed. L. Macy (Accessed June 11, 2006), <http://www.grovemusic.com>

Johann Joachim Quantz, On Playing the Flute [this is how the title is given in the English version], translated by Edward R. Reilly. New York: The Free Press, 1966, p. 145

Stendhal, Life of Rossini, translated and annotated by Richard N. Coe. London: John Calder; New York: Riverrun Press, 1965, p. 353-4; baggage arias, p. 360.

Pier Francesco Tosi, Observations on the Florid Song, or, Sentiments on the Ancient and Modern Singers, translated by J. E. Galliard. London: J. Wilcox, 1743, p. 179, musical example on plate VI. (Minkoff Reprint, Geneva, 1978)

Manuel Garcia Jr., Garcia's New Treatise on the Art of Singing [Traité complet de l'art du chant]. Boston: Oliver Ditson & Co., no date (the original was published in 1840). Tenor cadenza, p. 63; Pasta, p. 55; Garcia Sr., p. 59; changing the rhythm, p. 55 and especially p. 53.

piano right and left hand: Robert Philip, Early recordings and musical style: changing tastes in instrumental performance, 1900-1950. Cambridge [England]; New York: Cambridge University Press, 1992, page reference to come

Thayer's Life of Beethoven, revised and edited by Elliot Forbes. Princeton: Princeton University Press, 1967. Improvised for an hour, p. 541; made people cry, p. 185; Quintet for Piano and Winds, p. 350. See also index entries for Beethoven, Ludwig Van/as musician/improvises.

Valerie Woodring Goertzen, "By Way of Introduction: Preluding by 18th- and Early 19th-Century Pianists. The Journal of Musicology, Vol. 14, No. 3 (Summer, 1996), pp. 299-337.

Michael Musgrave, A Brahms Reader. New Haven and London: Yale University Press, p. 147.

Frederick Neumann, Ornamentation and Improvisation in Mozart. Princeton: Princeteon University Press, 1986, p. 179.

Robert D. Levin, "Improvised Embellishments in Mozart's Keyboard Music." Early Music, Vol. 20, No. 2, Performing Mozart's Music III. (May, 1992), pp. 221-233.

Neil Zazlaw, Mozart's Symphonies: Context, Performance Practice, Reception. Oxford: Clarendon Press, 1989, p. 482.

John Spitzer and Neil Zazlaw, "Improved Ornamentation in Eighteenth-Century Orchestras." Journal of the American Musicological Society, Vol. 39, No. 3 (Autumn, 1986), pp. 524-577.