The Ellington Century

by David Schiff

  Harmonies heat up the melody's seductive moves. Ellington set the initial pitch, B, atop a D dominant-ninth chord to form the interval of a thirteenth above the bass, the highest possible upper addition to a triad. The thirteenth is a double dissonance, one seventh (upward from C to B) stacked on another (from D to C). Music theory terms these combinations of tones dissonances, but they sound sensuous, not harsh. Every note of this melody is either a dissonance, or part of a dissonant chord. In the seven notes of the opening phrase the dissonances are as follows:

  B over a D ninth: a thirteenth

  B: a passing tone forming a diminished thirteenth

  A over a G ninth (with a raised fifth): a ninth

  A: a passing tone forming a diminished ninth

  G: a fifth over a C dominant ninth

  A: an “escape tone” forming a thirteenth

  E: the major seventh of an F major seventh chord

  For the next phrase Ellington transposed the notes and chords just heard down a minor third, seemingly even further from any vestige of C major, but landing on a d minor triad, the first three-note harmony of the song, which, as any jazz pianist can tell you, will easily lead to C major through the progression ii-V-I. Ellington, however, stretched that familiar move out with an upward melodic leap of a ninth on “se-re-na-ding you.”

  All these dissonances may look like early Schoenberg, but they sound more like those of Schoenberg's tennis partner, George Gershwin, just as the lyrics tell us. (Gershwin died in July 1937; perhaps this song was Ellington's musical eulogy.) Indeed the rich chord progression resembles that of “Nice Work If You Can Get It,” written a year earlier for Fred Astaire in A Damsel in Distress. Ellington's melody is more deviously sensual than Gershwin's bromidic glide.

  The eccentricities of “Prelude to a Kiss”—its chromaticism, high dissonance level, and systematic exploitation of the interval of a half step as a thematic idea— mask what we might term its skeletal structure. These good bones become visible once we pare the song down to its two outer lines and remove the many pitches outside the C major scale. Such a reduction reveals how the melody and the bass move steadily (yet evasively) toward the tonic. The melody descends sequentially: B-A-G-E; G-F-E-D. Ellington delays the arrival of the tonic note C so that the melody had to keep moving. When the melody finally completes the move downward to C, it leaps up a ninth so that the descent has to begin all over again. The bass line likewise moves in the time-honored pattern of the circle of fifths, but where another composer might have traced that circle from C to C, Ellington began the circle on D, one note above the tonic (D-G-C-F-B-E-A-D), keeping the harmony in suspense.

  Though Ellington's style was distinctive, his songs employed the same idiom as that used by his Broadway-based colleagues, not a distant jazz world dialect. To better discern the similarities and differences, let's compare “Prelude to a Kiss” to a song that would seem to be its opposite, “My Romance,” by Richard Rodgers and Lorenz Hart. It appeared in their 1935 show Jumbo and was revived as a jazz standard by Bill Evans. “My Romance” appears as plain as “Prelude to a Kiss” seems extravagantly fancy, but perhaps the differences are just skin-deep.

  Celebrating a lack of ornament (“My romance doesn't have to have a moon in the sky”), Rodgers's tune is conspicuously “white.” The entire melody for the chorus uses only the notes of the C major scale. As in “Prelude to a Kiss,” the “chorus” of “My Romance” is thirty-two bars long, but in a different configuration, ABAC, more typical of theater-based songs. ABAC songs save their biggest punch for the ending rather than the bridge. They present two similar statements with different endings, the first tentative, the second assured. Where the overall arc of Ellington's tune moves downward to resolution on the low C, Rodgers's tune is a series of upward scales that reach a climax in the final phrase when the opening three notes, the hook that states the title, blaze forth, transposed up a sixth to state and reiterate the melody's highest note.

  Beneath their surfaces, though, “My Romance” and “Prelude to a Kiss” are nearly identical. Both melodies state an idea then restate it a third lower. Both melodies create tension by moves from the tonic, C, to the submediant, E. Ignoring for a moment the differences between their AABA and ABAC designs, we might even say that Rodgers's diatonic chords and melody represent the harmonic analysis of Ellington's tune. It would be hard to reduce the tune or chords of “My Romance” the way we did for “Prelude to a Kiss” like a shellfish, its skeleton is on the surface.

  If we flip that observation around we can see why the tunes encourage different jazz treatment. To a jazz player, “My Romance” is a blank slate, just waiting for its harmonies to be enriched, its rhythms enlivened. As printed there is not a single jazz rhythm in the entire song apart from its fox-trot feeling, and yet if a player just moves the notes forward half a beat, as Bill Evans does, it becomes effortless jazz. “Prelude to a Kiss,” by contrast, doesn't have to be turned into jazz. Its melody comes already richly ornamented. Its rhythms subtly alternate bars of fox-trot and bars of Charleston so that it swings even played as written. Johnny Hodges doesn't turn it into jazz; he just turns it into Hodges.

  UNDRESSED MELODIES: SEX AND RACE

  The pop tune flourished within a particular time frame, let's say 1911–70, in a particular place, melting pot, mongrel Manhattan, and it served particular social functions, including the definition of national and generational identities. Eighty-five percent of the time (according to Furia) its overt subject, though, was romance (“I love you in thirty-two bars”): not medieval romance, but the American kind where boy meets girl and, if things work out, they check into a small hotel, but, as “There's a Small Hotel” says, only in the “bridal suite.” Romance included sex, love, and marriage; the friction and frustration generated by these categories kept the subject and the songs interesting.

  Popular songs represented romance with four scales: major, minor, blues, and chromatic. The major scale conveyed an ideal of untroubled, “normal” innocence: “Do-re-mi.” Like Mozart, Schubert, and Brahms, however, the tunesmiths knew how to give a song a bittersweet quality by making the major scale sound minor. Here are four mixed-mode examples (of many, many others) you can play through in your head: “My Heart Stood Still” (Rodgers and Hart), “Embraceable You” (the Gershwins), “What'll I Do?” (Berlin), and “Yesterday” (Lennon and McCartney). The first two songs emphasize minor-sounding segments of the major scale. “What'll I Do?” draws its poignancy from just one pitch borrowed from the minor mode, the flat sixth. “Yesterday” hovers between major and relative minor. None of these songs uses blues devices, but the minor mode makes them sound blue. Love, after all, isn't easy, but at least in pop tunes the minor usually gives way to the major.

  Until rhythm and blues crossed over to become rock and roll, the blues scale functioned as a racial marker in pop tunes. The blues scale indicated that a song, like “Stormy Weather,” was intended for black performers because it portrayed emotions that “they” had but “we” could not express, or it portrayed a “mongrel” condition, like that portrayed by mixed-race Julie in Show Boat when she sings “Can't Help Lovin’ That Man of Mine.” The blues scale in particular became a fixture of torch songs, from “The Man I Love” to “The Man That Got Away,” sung by women who had “gone south” and paid the price. The presence of the blues, even in just a passing chord, as in “Somebody Loves Me,” denotes an unsettling, subversive feeling summed up in the phrase “low down.” Although Harold Arlen specialized in songs with a blues tinge, there's not a trace of the blues in his most “American” score, for The Wizard of Oz. In the Emerald City, Dorothy may not be in Kansas anymore, but she is also far from the un-American terrain of Harlem.

  In classical music and popular music alike the chromatic scale often denotes sexuality. Tin Pan Alley's chromatic inhibition therefore may have been more than a question of singability. In a marketplace where songs appeared in the censored media of the movies and radio, full frontal
sexuality was both unsingable and unspeakable. Without sex, though, love songs become, in the words of Dorothy Fields, “as cold as yesterday's mashed potatoes.” The solution appeared in songs like “How Long Has This Been Going On,” which floated a diatonic melody atop a chromatic accompaniment, an upstairs/downstairs scalar divide. This device appears in “The Girl Friend” (Rodgers and Hart, 1926), a paean of “terrible honesty”: “Homely wrecks appeal/When their checks appeal,/But she has sex appeal.” The girlfriend in question is both good and bad: “She's knockout, she's regal,/her beauty's illegal.” Rodgers divides her licit and illicit aspects between the melody and the bass, keeping the erotic element subconsciously below the waist.

  Gershwin pushed the diatonic/chromatic split even farther in “Liza” (1929) and “Boy What Love Has Done to Me” (1930). The title of the second song speaks for itself, but “Liza” is a more complex cocktail. It appeared in “Show Girl,” a short-lived “Dixieland” extravaganza dreamed up by Flo Ziegfeld that featured both the Ellington Orchestra (playing its own music) and a ballet based on Gershwin's An American in Paris.17 “Liza” was Gershwin's response to Ziegfeld's request for a “minstrel number in the second act with one hundred beautiful girls seated on steps that cover the entire stage.”18 The gimmick was that just as Ruby Keller was about to begin the song, Al Jolson, her new husband, would leap out of the audience to serenade her and sing the song himself.19 The song had to be black, white, and blackface all at the same time. To code this racial mélange, Gershwin used three scales. The first phrase (“Liza, Liza, skies are gray”) pits a minstrel-style pentatonic melody against a rising chromatic bass line. Both lines turn diatonic in the second half of the phrase (“But if you'll smile on me / All the clouds'll roll away”). The white-note music mirrors the Kernesque sweetness alluded to in the lyrics; it also allows Keeler and Jolson to sing in their own white voices.

  Once we see how the harmonic language of the Great American Song-book encoded race and sex, we can hear the chromaticism of Ellington's melodies as more than a stylistic refinement. Brandishing its chromaticism rather than hiding it in the bass line, the melody of “Prelude to a Kiss” exposed the code and spoke the unspeakable. The melody made the erotic audible. Consciousness of the codes also shed light on the subversive strategy behind Ellington's own mythology, such as the story of composing “In a Sentimental Mood”: “That was written very spontaneously. One playing—zhwoop!—just like that. The occasion was in Durham, North Carolina…. It was a rather gay party with the exception of two girls and one fellow…. So I had two chicks, one on each side of me, and I said ‘Just listen to this. You girls are too good friends to let anything like this come between you,’ And this is what I played for them. I played it and I remembered it, and then I put it down.”20 It's a lovely story, but a cunning one as well. “In a Sentimental Mood” is one of the most sophisticated songs Ellington wrote, even more than “Prelude to a Kiss.” It also takes direct aim at Gershwin, lifting its first seven notes from “Someone to Watch Over Me,” then exposing the squareness of Gershwin's song through a masterful deployment of two Gershwin devices, the chromatic bass and the passing allusion to the blues, a single A that tints the entire melody. Where Gershwin had built his melody by repeating the same rhythm in a descending melodic sequence (longing to see/I know that he/'s waiting for me), Ellington's tune curls unpredictably, arpeggiating up a ninth when you least expect it. Compared to the Gershwin, it sounds improvised whether or not it actually was. (The performance to hear is by John Coltrane.) Dave Brubeck's nearly as spontaneous-sounding homage, “The Duke,” begins with the melodic inversion of “In a Sentimental Mood,” a technique usually associated with serialism rather than jazz.

  The great popular tunes, particularly the ballads, conveyed the contradictory tenets of urban life: sexual liberation versus “family values,” freedom versus anomie, enlightened sophistication versus abiding prejudice. The closer we look at the songs, the more their craft appears as ways of registering social friction through musical and verbal frissons, dissonances, and uncertainties that, while rarely rising to political protest, absolve them from the Adornian charge of mindless escapism. Their full complexity, though, demanded new styles of performance that would replace literal reproduction with looser reinterpretation. They needed, as we will see, to be sung with a blues voice, as blues.

  “DAY DREAM”: SINGING THE BLUES, BLUEING THE SONG

  Musically, not legally, who owns a melody? If you listen to three tenors—yes, those three tenors—sing, say, “Nessun dorma,” each version will contain the same pitches and pretty much the same rhythms. Even the sounds of their voices will differ only enough so that an experienced opera lover could tell them apart. Puccini still calls almost all the shots. By contrast, when you hear Bessie Smith and Billie Holiday sing the “St. Louis Blues,” their pitches, rhythms, and even some of the words differ sharply from the printed sheet music. W. C. Handy may have received royalties, but both singers felt free to remake his song in their own images. And if you listen to Sarah Vaughan stretch the last phrase of “My Funny Valentine” into infinity, there will be no question of who owns that melody. Unlike concert audiences, listeners to jazz and popular music expect performers to style a song distinctively. Historically, though, the contrast of composers' and performers' prerogatives has not always been so clear-cut. On recordings of Broadway musicals from the 1920s, singers like Gertrude Lawrence perform great Gershwin tunes with (to our ears) a genteel literalism; much of the music termed “jazz” in the 1920s was played as written from mass-produced stock arrangements.21 Opera singers in the baroque and bel canto eras added ornaments and embellishments to arias; concerto soloists were similarly expected to improvise cadenzas.

  The conformity of classical music stems from the power of the printed score as a legal restraint and an aesthetic one, a power asserted by music publishers in the nineteenth century, put into copyright law early in the twentieth, and turned into an aesthetic cause by performers and scholars. Even before Toscanini proclaimed that his performances were come scritto, classical music was getting hooked on literalism. The score was the music. In the face of such textual power the performer's room for interpretation has grown ever narrower.

  Popular singing, by contrast, gradually became ever freer in execution. Exactly how this happened is a complicated story involving many different styles of vocal performance and changing technology,22 but it certainly had a lot to do with the two great waves of blues popularity, first as sheet music in 1914 and then as recordings in 1920. Today we take the link between blues and popular music for granted, but they would have seemed inimical a century ago, and not just racially. Popular songs were designed for mechanical reproduction as sheet music or piano rolls, and they were carefully branded with hooks or novelty subject matter to differentiate themselves from competing projects. Developed in an oral culture, the blues, by contrast, put a premium on the individual singer's bardlike ability to embellish within a common musical and poetic vocabulary. The essence of blues was elaborated repetition, both within a blues stanza where a single line would be sung twice with different inflections, and between songs that shared structure and subject matter. Publication placed the blues at a crossroads. Would the blues now be performed as written, like most popular songs? Actually, there were fairly literal blues performances, at least to judge by Eva Taylor's recording of Joe Oliver's “West End Blues,” which sounds demurely text-bound compared to the famous recording by Louis Armstrong.23

  The blues began to transform the performance style of popular music through recordings by less restrained “blueswomen,” including Ma Rainey and Bessie Smith. Mamie Smith (one of many unrelated blues-singing Smiths) launched the craze with her intensely wailed rendition of Perry Bradford's “Crazy Blues,” which sold seventy-five thousand copies within four weeks of its release in October 1920.24 Anyone who owned a copy of the sheet music would have seen at once that Bradford's song was not being sung come scritto. The blueswomen certainly
employed forms of embellishment and ornamentation not found in popular song performance of the time judging by the few recordings we have of musicals in the 1920s, not on Broadway, where operetta-style singing mingled with cabaret-style Sprechstimme and the minstrel show “mammyisms” of Al Jolson. For conservative listeners, even in the realm of popular music the sounds of blues vocalism were repellant. When Louis Armstrong appeared in London in 1932, Melody Maker dismissed his singing of popular tunes like “Them There Eyes” or “When You're Smiling” as “savage growling.”25 Today these performances sound like classics of coloratura, the fine art of melodic elaboration.

  We can hear contrasting versions of blues coloratura in three wonderful performances of “St. Louis Blues,” the famous January 14, 1925, Bessie Smith recording with Louis Armstrong on cornet and Fred Longshaw on reed organ; Billie Holiday's recording from October 15, 1940, with the Benny Carter Orchestra; and a lesser-known recording from February 9, 1932, by Bing Crosby with the Ellington Orchestra. It's not surprising that Bessie Smith's recording, with its harmonium accompaniment, slow tempo, and trumpet obbligato has become a sacred text in jazz history; it has the weight of a Bach aria. In her first phrase (“I hate to see that evening sun go down”), Smith called attention to the way she sings each note by compressing the notes of the melody, following its four-note motto, within the range of a minor third, between the tonic E and G. We might categorize her embellishments as introductory, a slide to the main pitch; concluding, a fall from the main pitch; or internal, a wavering of the pitch within the note, particularly noticeable on the many Gs, especially on the word “sun.” There are also notes that are sung almost parlando, and others that are sung straight and trumpetlike, matching Armstrong's tone. In the second phrase (“Feelin' tomorrow like I feel today”), she used the two syllables of “mor-row,” which take the place in the melody of the one syllable “see,” to color the pitch F instead of G. In relation to her singing, the term “blue note” has to be applied to a whole family of ornaments that she applied to different steps of the scale, not just the third.