The Ellington Century

by David Schiff

  “THE CLOTHED WOMAN,” OR WHAT IS THIS THING CALLED ATONALITY?

  After the Second World War the history of European art music in the twentieth century was rewritten, first in Paris, then at Princeton, as an inevitable evolution from tonality to “free” atonality to serialism. Somewhere along the line this tale of harmonic progress became interwoven with T. W. Adorno's jeremiads against popular music and jazz, which branded these genres as infantile empty-calorie musical junk food manufactured by the culture industry. These views began to change (at least outside France) as early as the 1960s, when rock music seemed at least as interesting, both musically and politically, as much of the new art music. With the rise of the civil rights movement, the ‘60s were also a fertile and contentious period for African American music in a wide range of styles, from the gospel-inspired soul music of Aretha Franklin to the free jazz of John Coltrane. Ellington's Far East Suite and Sacred Concerts were very much part of that politically charged panorama.

  Free jazz, often played without reference to harmonic progressions, provoked a new debate about atonality. When the style first appeared, with Ornette Coleman Quartet's gig at the Five Spot in late 1959, some critics heard the new jazz as a belated response to the European avant-garde, while others emphasized its political radicalism. Some heard it as a response to Stockhausen, others as a resurgence of primitive blues untainted by popular music. LeRoi Jones summed up the debate in Blues People: “[Cecil] Taylor and [Ornette] Coleman know the music of Anton Webern and are responsible to it intellectually, as they would be to any stimulating art form. But they are not responsible to it emotionally, as an extramusical catalytic form. The emotional significance of most Negro music has been its separation from the emotional and philosophical attitudes of classical music.” Jones advised that avant-garde techniques “be used not canonized” by African American musicians.6

  Ellington anticipated this controversy by more than a decade. At his December 27, 1947, Carnegie Hall concert he announced, without much ado, that he would now play “The Clothed Woman,” sat down at the keyboard, and let loose the most outside piano solo he or anyone else had imagined up to that time. Actually only the opening and ending sections of the piece are so far out, and Ellington never revealed if their eccentric sounds were meant to describe a woman getting clothed or unclothed. Ellington recorded the piece twice in 1947, and the music was later published. The sheet music rather oddly claims to be Marian McPartland's arrangement of the piece, yet “like the Ellington original involves little improvisation.” Perhaps the sheet music is an arrangement in that it transforms the original, which was scored for piano and band, into a piano solo. The printed music, lacking indications of tempo and dynamics, is useful only if you listen first to the Ellington recordings. Its outer sections look like a random collage of arpeggios, blues riffs, and dense polychords sprinkled with seemingly incongruous F major triads. You wonder if Ellington had composed it Cage-style by consulting the I Ching. When you listen to Ellington's performance, however, these fragmented, discordant gestures reveal themselves as a slow and slinky blues in F, the apt prelude and postlude to the central stomp in B, as harmonically mainstream as the framing episodes are not. Ellington's musical cocktail complicates any sense of distinction we might have between atonal and tonal, European and African American, out and in, extraordinary and everyday, pop and art. In its suave subversion of “category” it leapfrogs over the historical determinism of the decades that followed its premiere to a later eclecticism that didn't take shape as an artistic movement until after Ellington's death. But even though postmodernists treat atonality as an option rather than an imperative, it remains ill defined technically and culturally, and it is still very much tied to the works and ideas of Arnold Schoenberg, even though there is a lot of atonal music, by Scriabin, Debussy, Ives, or Ellington, that has little to do with Viennese expressionism and its aftermath. A century after its first appearances atonality still sounds out—and that remains its attraction.

  Critics and theorists have applied the slippery term atonality to music ranging from the late works of Liszt to art rock, from music that departs ever so slightly from nineteenth-century textbook rules to music that doesn't even employ pitches. For the general public it denotes music that is both nasty sounding and deliberately obscure, the product of some diabolical system. To be sure, many twentieth-century composers from Scriabin on propounded harmonic systems or generated speculative compositional procedures. Evaluating the results of systems as different as those of Hindemith and Cage is complicated by the fact that ideas can sometimes be fascinating in their own right. Some of the most admired music of the later twentieth century, like the works of Lutosławski and Steve Reich, sprang from the ideas of John Cage rather than the sound of his music. I have always found the music of Xenakis compelling at the gut level, though I don't understand the mathematics behind it at all.

  Perhaps we can grasp atonality better if we go through the back door and briefly reexamine tonality. Theorists usually apply the term tonal harmony to European music written between, say, 1675 and 1900. The style of that music is characterized by:

  1. A tonal center or tonic defined for an entire piece in terms of major or minor scales. Modulations to other keys serve to confirm tonal unity (called “monotonality” by Schoenberg).

  2. Harmonic vocabulary of triads and seventh chords.

  3. Chordal progressions defined by the bass line and harmonic function.

  4. Voice leading (part writing) connecting chords smoothly while avoiding certain kinds of parallelism (no parallel octaves or perfect fifths).

  5. Nonchord (nonharmonic) tones prepared and resolved with specified sanctioned formulas.

  Although the influential theorist Heinrich Schenker took an all-or-nothing view of tonality, and viewed any music that did not meet all the conditions listed above as aberrant, many composers, not surprisingly, disagreed. Both Bartók and Stravinsky claimed that their music had a tonal center, though Schenker would have been hard-pressed to find much evidence of tonality in Music for Strings, Percussion and Celesta, or Stravinsky's Serenade for Piano, both works considered to be “in A” by their composers. Even twelve-tone works such as Schoenberg's Piano Concerto and Berg's Violin Concerto can sound tonal at times, and more so on repeated listening. Theorists sometimes use the terms “centric” or “polar” for such nontraditional but tonal-sounding harmonies, but perhaps we can better understand a wide range of harmonic idioms, diatonic and chromatic, sweet and harsh, functional and disjunct, as part of an evolving, expanding tonality. I'll reserve the term atonality for works—and there are plenty of fine ones—that don't use pitches at all or that aim to be “out,” though, as “The Clothed Woman” demonstrates, even the “out” can sound tonal on repeated hearing.

  We can test the proposition of expanded tonality on pieces by Debussy, Stravinsky, Schoenberg that seem to travel far from anything you learned in Harmony 101.

  Let's start with “Et la lune descend sur le temple qui fut,” from Debussy's second book of Images, composed in 1907. Its opening phrase breaks all the tonal commandments but one. In the first three bars we hear a melody of chords, but the chords are not triads built in thirds but three different trichords with either a perfect or diminished fifth between their outer notes and a major or minor second between their two top notes. We might hear the first chord (E-A-B, from bottom to top) as what jazz players call a “sus” chord, where the fourth takes the place of the usual third, and we might hear the last chord (F-A-B) as implying a dominant seventh, but they don't resolve in any way that would confirm these readings. There is also no bass line. The music appears nontonal in chord vocabulary, indication of harmonic function, and also voice leading, since the three voices most of the time move in forbidden parallel fifths. Despite the mysterious inconsistency of the chords, the melody is relatively simple. It outlines an octave from B to B using pitches from the e minor scale and even sounds a bit like “I Didn't Know What Time It Was” (Rodge
rs and Hart). If we listen to the harmonies from the top down rather than the usual way, upward from the bass, we can hear the lower voices as forming dissonances against the melody, which Debussy resolves in bar 5 by contracting the three voices to a unison B. As the piece unfolds, with nary a traditional chordal progression, it outlines a double tonality centered on E if we give the bass notes priority, and on B if we give greater weight to the melody in the upper voices.

  The large structure becomes clear if we imagine that Debussy, anticipating audio editing, spliced together two pieces. Piece I, centered on the pitch B, is heard in bars 1-5, 12-19, 25-26, 41-42, 46-52, and 56-57. Piece II enters in bars 6-11 with parallel triads over a pedal harmony (E-B) in the bass, both textural effects absent from Piece I. These textures return in bars 20-24, 27-28, 39-40, 43-46, and 54-55. Unlike Piece I, which always affirms the pitch B, Piece II uses the device of the sequence, repetition at a changed pitch level, to create a sense of harmonic motion. We might say that Piece I functions as the tonic, Piece II as the dominant. At the close of the piece the bass line resolves Piece II on a low E, but the upper voice brings the piece to rest on three octaves of B, a resolution that the pianist can make clear by raising the pedal to release the E in the bass and then lowering it again to emphasize the B, or “faites vibrer,” as Debussy suggests in the score. The word atonal seems out of place here, even though Debussy has upturned the usual way of hearing harmony and also redefined harmonic function as alternation rather than cause and effect.

  For some listeners Debussy's groundbreaking piece may sound like orientalist mood music no matter how transgressive it might appear to the harmonic analyst, so let's look at a grittier piece, the “Pas d'action” from Stravinsky's 1947 ballet Orpheus. The score describes the action: “The Bacchantes attack Orpheus, seize him and tear him to pieces.” Stravinsky, following his trademark practice, defined the movement with a single chord that sounds like it was discovered at the keyboard: G#s in octaves in the bass, the pitches A and C, doubled in octaves, in the treble. Even though there are just three pitches the chord sounds harshly dissonant. By omitting the interval of the fifth, Stravinsky complicated interpretation of the chord. It could be heard either as implying an A minor ninth, or as an inverted minor seventh chord, with the major seventh, G#, in the bass. Either way, it has something of the blues about it. Like jazz pianists, and like Ravel, Stravinsky required that every harmony in this movement include the interval of the seventh, or its inversions, the second and ninth.

  The initial and inscrutable “objet sonore,” however, becomes the harmonic building block for the piece. Theorists would define it as a pitch class set [0,1,4] that contains the intervals of minor second, minor third, and major third. In this movement its structure reappears transposed, inverted, and revoiced in well over 90 percent of the chords. Through repetition it becomes the harmonic norm, a substitute triad, and departures from its sound stand out clearly. The remaining harmonies are similarly derived mainly from two compressed variants of the “tonic' chord: [0,1,3], containing a minor second, a major second, and a major third; and [0,1,2], containing only minor and major seconds. In analysis I'll call the main chord X and these two others Y and Z, respectively.

  Throughout the movement the notes in the bass sound like dissonances while the tonal center appears in the treble. I call this “clothesline tonality”; we have already encountered it in “Et la lune descend,” and it is a characteristic that links Stravinsky's music to Debussy's. Looking at the score of Orpheus, it is hard to explain this inverted center of gravity acoustically except by noting that the upper line moves in stepwise fashion while the bass always jumps. The melodic notes (beginning with a motive that sounds, appropriately, like another Rodgers and Hart tune, “This Can't Be Love”) use the pitches of the a minor scale, while the bass line outlines triads of A major and c minor. Stravinsky deployed this harmonic clash as part of a carefully designed plan of harmonic motion based on the polarity of A and E. In place of the extended ostinatos of his Russian works, Stravinsky composed this movement by alternating static harmonic plateaus with more unstable phrases of harmonic motion.

  The movement closely resembles classical sonata form, albeit in a tightly coiled compression, and uses sonata-style strategies to achieve tonal contrast and tension. The poignant affect of the coda depends on our hearing its single harmony as a subdominant in relation to the tonic harmony (centered on A) announced at the beginning and affirmed climactically at the recapitulation. In its expressive use of harmonic relations the music could be by Beethoven, except for the fact that Stravinsky did not use any chords in the movement that Beethoven, or Brahms, or Schenker would have considered harmony. Monk, who wrote most of his tunes around the time that Orpheus appeared, would have felt right at home.

  Let's venture still further out to the onset of Stravinsky's serial phase. The Bransle Double comes right in the middle of Agon and is the first movement based on a twelve-tone row. Previous movements used shorter rows, and in fact the Bransle Double combines two six-note rows that have already appeared:

  Hexachord I: C D E F E A

  Hexachord II: G A B B D G

  As Peter van den Toorn points out, both of these hexachords contain recognizable chunks of the octatonic scale (alternating major and minor seconds), which underlies much of Stravinsky's music, especially of his Russian period.7 But the Bransle Double doesn't sound Russian; it sounds like jazz. Hexachord I contains most of the notes you need for a blues in C. The first five pitches of Hexachord II make a nice jazz dominant with a flat fifth. This creates the possibility of a quasi-harmonic progression that Stravinsky gives us three chances to hear. The short movement consists of an eight-bar phrase (in time) written in two-part counterpoint, followed by its repetition with a third voice added in the bass. There follows a contrasting middle section, thirteen bars of , then a repeat of the opening eight bars followed by a fourteen-bar coda in . (If you do the math you'll see that each of the phrases contains about the same number of beats.) It is essentially an extended AABA pop tune form. This highly repetitive structure is not something you find in the music of the Viennese serialists, and neither is Stravinsky's manner of stuttering out the twelve-tone row, repeating two-note segments two or three times before moving on.

  Around the time he wrote Agon Stravinsky stated that “the intervals of my series are attracted by tonality; I compose vertically and that is, in some sense at least, to compose tonally.”8 This music immediately sounds more tonal than most Schoenberg and Berg because Stravinsky used tonally affirming octave doubling of pitches, which the others usually avoided. If Stravinsky had followed serial rather than tonal logic in this Bransle he might have constructed his two-part counterpoint by superimposing the two hexachords to create a twelve-tone aggregate. He did combine a statement of I (in the violins) with II (in the brass), but the versions he used share three common pitches:

  I: C-D-E-F-E-A

  II: B-B-D-D-E-A

  Stravinsky seems to have chosen his pitches not in order to get to twelve but to outline a tonal progression from C to G and back; he even used the inverted retrograde forms of both hexachords to crawl back to C at the cadence (where the trombone adds a Monkish minor second below the C in the violin.

  In the short middle section Stravinsky did oppose two complementary hexachords:

  D-E-F-G-F#-B

  A-B-C-D-E-A

  He voiced them, however, to suggest two jazz chords: G7 (in the piano) and C7 (in the clarinets). Stravinsky himself pointed out the jazz character of this passage, a character perhaps most obvious in its rhythm and timbre, but underscored by its harmonies. He even ended the coda jazz style, with a G7 chord with two statements of Hexachord II in the bassoons:

  A-B-C-D-E-A

  D-E-E-F#-G (C)

  The second hexachord never sounds its final C, so we are left hanging both tonally and serially.

  I am not trying to prove here that all music is in some sense tonal; music also needs a sense of risk and f
ree fall; sometimes, as the old jingle goes, you feel like a nut. Or like Schoenberg. The piano piece Op. 23 no. 2 (written in 1920) is a two-part invention recast as a tantrum. Expressionist music employed two metaphors for portraying psychosis: the roller coaster for bipolar disorders and the assembly line for obsessive compulsion. On its surface Op. 23 no. 2 is a lurching roller coaster, a series of steep rises, gut-churning dips, and sudden swerves separated by deep breaths of anxiety. The meter and tempo constantly change; phrase lengths are unpredictable; nothing is repeated literally. Beneath the surface, though, it is a diabolical mechanism. Schoenberg plundered the first six bars to construct the music for the remaining sixteen. (The piece usually lasts around ninety seconds, though Glenn Gould stretched it almost to two minutes.) The rigor of the piece seems both obsessive and guilt-ridden. Everything comes back, but in unrecognizable forms, a perfect crime with a perfect cover-up.