SOURCE: “On the Verge of Opera,” in Opera Quarterly, Vol. 6, No. 3, Spring, 1989, pp. 76–85.

[In the following essay, Martin discusses the similarities and differences between Sondheim's work and opera, focusing on character, musical structure, orchestration, amplification, and musical style.]

Has Stephen Sondheim, the composer of such Broadway musicals as Follies (1971), A Little Night Music (1973), Pacific Overtures (1976), and Sweeney Todd (1979), been masquerading all these years and actually composing operas? The question is raised vigorously by the new and greatly expanded edition of Craig Zadan's excellent book Sondheim & Co.,¹ and it hovers above the many recent revivals of the works by opera companies that have staged them, without apology, as operas—emphasizing musical values.

The revivals are a phenomenon of the last five years. In the United States they began in 1984 when the Houston Grand Opera and the New York City Opera shared a production of Sweeney Todd, a Victorian melodrama; the Michigan Opera Theatre the next year mounted another. In addition, the Michigan company, the Lyric Opera of Cleveland, and the Opera Ensemble of New York all have produced A Little Night Music, a nostalgic love story in three-quarter time. In England, starting also in 1984, Sweeney Todd was put on by several small companies, and Pacific Overtures, a series of episodes about the opening of Japan to the West, by the English National Opera. The latter event led to an article in the English magazine Opera analyzing some of Sondheim's musical techniques and proposing that the question “Why doesn't Sondheim write an opera?” be answered: “He's been writing them for years.”²

Meanwhile, Follies, a sour summation of theatrical illusion, marital discord, and aging, after a long sleep following a problematic initial run, had two notable revivals, pulling in opposite directions. In 1985 the New York Philharmonic, with an extraordinary collection of Broadway's best performers, presented it for two nights in concert, preserving in a remarkable recording much of the work's astonishing bitterness. Then in 1987, in London's Shaftesbury Theatre, it reappeared as a West End musical with its book rewritten to provide a happier ending and with four numbers replaced, including the famous nervous-breakdown finale. And in this form it has had a triumph, causing the editor of Opera, who regrets the shift in tone and lost numbers, to lament: “Perhaps if Sondheim were to write operas, and admit it, he would need to pay less heed to popular taste.”³

Sondheim's three works following Sweeney Todd have added to the debate without in any way resolving it. The first, Merrily We Roll Along (1981), a story of theatrical people whose friendship is spoiled by success, was a failure on Broadway, though recent revivals in California and England suggest that it was dismissed too quickly.⁴ Next came Sunday in the Park with George (1984), whose first act is a fictionalized account of the painting by Georges Seurat of his masterpiece Sunday Afternoon on the Island of La Grande Jatte, and the second, a repetition of the situation and emotions in the career of his great-grandson, also an artist. Many persons felt that the two acts were very artificially joined, and that the first, if allowed to stand alone, would make a fine one-act opera.⁵ Lastly, there is his current Broadway hit Into the Woods (1987), in which three fairy tales twine into a message that we must go into the woods and face the mysteries of life in order to grow and that in doing so we will not be alone. It won the Antoinette Perry Award for the best original score in the 1987–1988 season, and yet, as usual with Sondheim, there is no agreement about its worth, or its form. Some reject it as sentimental Broadway slop with music to match; others rank it somewhere above Hänsel und Gretel.

The argument over what to call Sondheim's works, how to stage them and where—whether or not in an opera house—is not a semantic quibble in which mean-spirited conservatives attempt to deny an innovator the credit of serious composition. At issue are fundamental questions about what an opera is and what sort of works opera companies should mount.

Further, the questions imply that we may have arrived today at one of those grand shifts in cultural perceptions that put out the door the assumptions of one era while bringing in others that seem quite contrary or at least wholly new. But before examining the differences between Sondheim's musical techniques and those of a typical opera composer, let us consider the similarities, for it is they that have brought his works into the opera houses, unsettling the very definition of opera.

Take Sweeney Todd, generally acknowledged to be the most conventionally operatic of Sondheim's works and the one thus far most often staged by opera companies. It tells a story of revenge that, on being thwarted, turns into an obsession to murder. A London barber, Benjamin Barker, convicted and shipped to Australia on charges trumped up by a judge, returns to the city after fifteen years, seeking his wife, Lucy, his child, Johanna, and vengeance. Calling himself Sweeney Todd, he visits his former landlady Mrs. Lovett, a widow who makes meat pies, and hears from her that the judge had lusted for the wife, raped her most brutally, and after she had killed herself by poison, in remorse had adopted the child. Recognizing the barber, Mrs. Lovett produces his razors, which she has preserved, and the two enter a partnership to trap the judge—though if truth be told, Mrs. Lovett is more interested in trapping Barker/Todd into marriage.

They soon manage to lure the judge into Todd's barber chair, in the room above the pie shop, and Todd has the razor at the man's throat, dawdling with pleasure, when an interruption saves the judge, and he leaves unharmed. Todd, maddened by his lost opportunity, swears to kill all his clients, and Mrs. Lovett proposes that they put them into the meat pies. “Man devouring man,” is London's rule, and “who are we to deny it?”

The tasty pies attract a crowd, and Todd and Mrs. Lovett prosper. Once again they maneuver the judge into the chair, and Todd kills him; but in his haste, eager to get a crazy beggar woman out of the way, he kills her, too, discovering later, in his moment of triumph, that she was his wife, Lucy, who had not died as reported. Mrs. Lovett had lied. Apparently accepting her excuse that a wife without wits was of no use to him, he waltzes Mrs. Lovett around, nearer and nearer the oven, and then pops her in, only to have his own throat slit by her assistant who, having learned the reason for the pies’ success, has gone insane. At the end the only survivors in sound mind are the daughter, Johanna, and the sailor Anthony, who loves her.

In telling this story Sondheim has built character and musical structure, as any opera composer might, by giving the principals distinguishing motifs. A crazed beggar woman, one of the London crowd, asks for “Alms! Alms!” to a falling semitone, and when Todd, finally recognizing her, cries her name, “Lucy” it is to the same motif. The sailor's excitement on arriving in London merges neatly into his excitement on meeting Johanna, and thereafter he sings her name and expresses his emotion to a rising minor third and a whole tone. And the same motif of her name, though with different coloring, runs beneath the judge's song of lust, now transferred to Johanna, and Todd's of parental yearning. Mrs. Lovett has a four-note motif that begins two of her solos and is sung by Todd when he addresses her by name. There are other such motifs, some becoming clear only on repeated hearings, but one that is obvious, which most in any audience will recognize at once, underlies the cry of the London crowd, “Swing your razor high, Sweeney.” Most fittingly, it is based on the opening phrases of the Dies Irae.

Sondheim combines these motifs in duets, quartets, sextets—characters sharing, exchanging, and ignoring them—and in the choruses, of which the work has many. He also offers ballads and music hall and Victorian parlor songs for atmosphere, as well as a parody of mid-nineteenth-century Italian opera. No one, I think, will deny that the music is imaginatively used to reinforce the text, making music drama of a high order, much higher than is usual on Broadway. Why then the hesitation to call it opera? What in its forms and techniques seems incompatible?

Consider the work's orchestration. From Monteverdi onward part of the definition of an opera composer is that he does his own orchestrating. Mozart, Wagner, and Verdi speak with individual voices partly because each has clothed his drama in his own special sound. But this is not true of Sondheim, as the chapter on orchestration in Zadan's book makes clear.

Sondheim composes a vocal score with piano accompaniment that he turns over to an orchestrator, usually, but not always, Jonathan Tunick, and it is the orchestrator who chooses what sound will give meaning to the song. Thus in the Follies song “In Buddy's Eyes,” in which a wife sings ironically of her supposedly happy marriage, whenever she refers to her husband the timbre is “dry, all woodwinds,” and whenever to herself, “all strings.”⁶ The characterization is neat, but it is Tunick's, not Sondheim's so who is the composer?

The question, however, doesn't stop there. When Sunday in the Park with George, for which Michael Starobin was the orchestrator, came to be recorded, the recording producer, Thomas Z. Shephard, reconceptualized it: “I wanted to turn it into a dramatic cantata rather than a collection of numbers. … We had to create an opening where the orchestra would grow, the way the stage filled with light—as George kept adding color. We used three different-sized orchestras. The regular pit orchestra was used for most of the score; the midsized orchestra—with twice as many strings—was used for more sweeping and romantic numbers like ‘Move On’; and the biggest orchestra—with three times as many strings—was used for the finales of both acts.”⁷

It is a fascinating concept and makes a beautiful recording, but it has little relation musically to what happened in the theater. Again, though Sondheim had a hand in this reformation, who is the composer?

Further, if Sunday in the Park with George someday is produced in an opera house, which orchestration should be used: the one for Broadway or for the recording studio? Surely the second, if those who know the recording are not to be disappointed.

Then there is amplification, another question more complicated than at first it may seem. On Broadway the singing actors in Sweeney Todd wore body microphones; only so could their voices fill the huge and acoustically wretched Uris Theatre and be heard above the orchestra. There were, of course, all the usual problems. Their voices were harshened, made hollow and metallic, and forced to share a common electronic quality; and they came not from the bodies onstage but from boxes hanging on the auditorium walls and ceiling. If the players turned their backs to the audience, or started to leave the stage, their voices continued just as loud as if they remained facing the audience at stage front. All sense of direction, space, and distance was obliterated.

When the New York City Opera staged the work, its general director, Beverly Sills, at first hoped to do without microphones but then was persuaded that even opera singers such as Rosalind Elias and Timothy Nolen needed them if their words were to be heard—though none are necessary when the company sings Gilbert and Sullivan.⁸ Why the difference? Because Tunick's orchestration was conceived with amplification in mind, and so, unlike Sullivan, he did not thin the texture of his orchestration when he wanted words to be heard.⁹

How then should an opera company stage Sweeney Todd? With or without amplification? Or should the orchestration be revised, raising once more the question of who is the composer?

In the past most persons would have said that part of the definition of opera is unamplified sound, a human voice unenlarged, undiminished, and not colored electronically. Yet I wonder if the premise still holds. In May 1988 Philip Glass's Fall of the House of Usher had its world premiere at the Loeb Drama Center in Cambridge, Massachusetts. It had been commissioned jointly by the resident American Repertory Theatre and the Kentucky Opera, a recognition on the part of the latter, according to Richard Dyer of the Boston Globe, “that opera is not simply a musical but a theatrical form.”¹⁰ The emphasis is his, but it does not clarify for me his meaning. When has opera not been a mixture of music and theater?

Though the Loeb theater is not large, seating perhaps 1,200, the voices and twelve-piece orchestra, which included a synthesizer and an electric guitar, were amplified. It so happened that a few months earlier I had heard one of the three principals, the baritone David Trombley, in an unamplified performance of Sondheim's A Little Night Music and the quality of his voice then had been considerably finer. But in their reviews neither Dyer of the Globe nor Michael Feingold of the New York Village Voice bothered to mention, even in passing, the fact of amplification, and John Rockwell of the New York Times merely stated that the performers had been “amplified in an undistorted but not always very helpful way.”¹¹

I found the distortion bad, but I must report that it seemed not to trouble the audience, and the critics evidently thought it irrelevant. Today we may have reached an era when most producers, composers, performers, critics, and audiences no longer count an unamplified voice a premise of opera.

Finally, in an area much less precise than amplification or orchestration, there is Sondheim's musical style, which despite any operatic trimmings is unabashedly Broadway, and to many ears is simply not strong enough musically to be treated as opera. Of the City Opera's production of Sweeney Todd, Donal Henahan remarked in the New York Times, “The operatic setting served chiefly to focus more attention on Mr. Sondheim's sing-song score than was good for it.”¹² Peter G. Davis of New York magazine wrote more bluntly, “The main problem is the drab, crabbed, and short-winded melodic invention. … The musical content is so impoverished that I fail to see why the special resources of an opera company should be wasted on it.”¹³

Possibly I can demonstrate a part of what the men mean by “sing-song” and “short-winded melodic invention.” Sondheim writes tunes of limited range because, for the most part, he works with limited voices, and he keeps his phrases short because Broadway voices cannot sustain a long line. Probably his most famous song is “Send in the Clowns” from A Little Night Music. Here is the first verse written out with places indicated where most singers breathe, and with an [a] showing where additional breaths sometimes are taken:

Isn't it [a] rich (breath)
Are we [a] a pair? (breath)
Me here at last on the ground, (breath)
You in (breath) midair … (breath)
Send in [a] the clowns. (breath)

The tempo is lento, but even in a song at a faster pace, such as “The Right Girl” from Follies, to be sung agitato, the phrases are short. And all but operatic singers, and one or two of the very best on Broadway, tend to speaksing them, often only approximating the designated pitch of the notes. Dignify the singing style, if you wish, by calling it Sprechstimme, but when done to an English text set to Broadway rhythms it sounds like cabaret singing: the “chant-oose” speaking the words in an effort to fake the emotion she physically cannot sing.

Cabaret style certainly has a lot of fans, but it also has some difficulties in carrying an evening of what hopes to be opera. The greatest of these is that it does not lend itself to building finales that are musically bigger, longer, more spacious than anything that has preceded. Typically in a Sondheim show the finale of an act is the weakest part. In the score of Sweeney Todd the moment in which Todd, having failed to kill the judge, swears to turn his vengeance on all humanity is grandly titled “Epiphany,” but the music, hobbled by the many short phrases, cannot match it. The scene is the natural finale to act 1, but Sondheim, no doubt aware of its musical weakness, carries us on to a duet between Todd and Mrs. Lovett about what flavor the victims of different professions will give to the meat pies. The duet is enormously clever and humorous, but a letdown. The singers, or more likely those who control the amplification, merely turn up the volume with each verse until—curtain.

Similarly, as the work nears its end, the music thins out, and the story is told increasingly in spoken dialogue or by musical reprise. This is a common form of ending on Broadway, and an unfortunate one, for the audience begins to feel that the piece is over before the last note has sounded. In Sweeney Todd for the grand finale the chorus and principals repeat “The Ballad of Sweeney Todd,” with its refrain based on the Dies Irae. The singers make an impressive amount of noise, but there is no new music for the ending, such as Verdi gave to Aida, or Mozart to The Magic Flute. A reprise, of course, is an old form of ending, exquisitely used by Verdi in Otello; but Sondheim, not at all to his discredit, has not yet equaled that.

Where then are we left with the questions: Are Sondheim's musicals actually operas? Should they be produced by opera companies? And how? With amplification? Rodney Milnes, the editor of Opera, suggests that we call them “light” music or operettas and, if necessary, amplify discreetly.¹⁴ Disliking amplification, I would try first reorchestration. Milnes also urges the opera companies take the works into their repertories as occasional substitutes for the old chestnuts that need a rest. And I agree. For the next few years I would prefer to hear Follies to Fledermaus, A Little Night Music to The Merry Widow, and Pacific Overtures to The Mikado.

Whether Sondheim is capable of composing an opera, as traditionalists understand the form, I am not sure. He apparently distrusts opera, even perhaps does not like it, and though Beverly Sills for some time has been trying to persuade him to compose for the New York City company, thus far he has refused. He is comfortable with the conventions of Broadway, and he does very well in expanding their limitations in ways that are natural to him. That seems to be his role in music, and he may be wise not to push it further than he feels he can sustain—though I, for one, wish he would try.

Notes

1. Craig Zadan, Sondheim & Co., 2d ed. (New York: Harper & Row, 1986). For this second edition Zadan added eight chapters and revised, updated, and expanded the other twenty-one—in effect, a new book. With extended examples and quotations, he recounts not only Sondheim's career but also how Broadway producers, directors, composers, orchestrators, choreographers, and others create a musical and regard themselves.

2. Jeremy Sams, “Sondheim's Operatic Overtures,” Opera 38, no. 9 (September 1987): 1007.

3. Rodney Milnes, A review of Follies, Opera 38, no. 9 (September 1987): 1094.

4. Sams, “Sondheim's Operatic Overtures,” p. 1007; Zadan, Sondheim & Co., pp. 283–85.

5. Zadan, Sondheim & Co., p. 307. For what it is worth, I shared this opinion and, even after listening to the recording, still do.

6. Ibid., p. 157.

7. Quoted in ibid., pp. 177–78.

8. Ibid., p. 347.

9. This point was raised by Patrick J. Smith in his review of the New York City Opera production of Sweeney Todd, Opera 36, no. 1 (January 1985): 47. So far as I have discovered, no other critic has bothered to dig even this deeply into the reasons for and difficulties of amplification. The stunning silence of the critical community on these questions has contributed to the quick acceptance of amplification in many theaters and opera houses.

10. Richard Dyer, “House of Usher Is Due A Salute,” Boston Globe, 19 May 1988.

11. Ibid.; Michael Feingold, “Through Glass, Darkly,” Village Voice, 7 June 1988; John Rockwell, “Glass and Poe Combine In Gothic Goings-On,” New York Times, 21 May 1988.

12. Quoted in Zadan, Sondheim & Co., p. 347.

13. Ibid.

14. Rodney Milnes, review of Pacific Overtures, Opera 38, no. 11 (November 1987): 1317–21. He approved of amplification in this production, stating that it was “made necessary by the fact that Sondheim does not write with conventional operatic range in mind.” It is a curious statement, for about the only thing that amplification at the moment cannot do is extend a singer's range.