Misunderstanding in Blue (by Darcy James Argue)


Gunther Schuller is a legendary figure, someone who I’m sure is familiar to Do The Math habituals. I highly recommend Ethan’s extensive two-part [one, two] interview from 2010, in case you missed it the first time. Schuller’s exemplary and wide-ranging career as a horn player, composer, conductor, and scholar have made him A Big Deal in both classical and jazz circles. The Gunther Schuller music that’s been most personally meaningful to me includes his justly-famous 7 Studies on Themes of Paul Klee, his vivid and bracing Symphony for Brass and Percussion, and his masterly chamber arrangements for Joe Lovano’s Rush Hour. I’m also very appreciative of the role he played in reconstructing and posthumously premiering Mingus’s long-form epic, Epitaph.

Schuller’s books Early Jazz (1968) and The Swing Era (1991) are considered indispensable references and landmarks of jazz scholarship. The scope of these books is honestly kind of insane: in the course of researching them, Schuller professes to have listened systematically to essentially every surviving jazz recording (as many of them as he could get his hands on, at least), from the beginning of the recording era up to 1945. They are notable for being among a very small number of works accessible to the general public to offer serious melodic, harmonic, rhythmic, and structural analyses of jazz recordings. Both volumes are underpinned by copious transcriptions from recorded sources, with detailed commentary by Schuller. It’s impossible to come away from these books without admiring Schuller’s deep love of and genuine engagement with jazz. We’d have a healthier musical culture all-around if more classically-trained composers and instrumentalists shared even the tiniest bit of Gunther’s authentic interest in how jazz actually works.

Unfortunately, however much I respect Schuller’s ambition and noble intentions, I think it’s also important to point out that the transcriptions in both of these volumes are often not accurate. And we are not talking about run-of-the-mill transcription errors here — as (expert transcriber) Bill Dobbins noted, when it comes to complex music, even the best transcriptions are going to be least 10 to 15 percent guesswork. That’s understandable. But many of the transcription errors in Schuller’s books are much more fundamental. Worse, they lead to mistaken conclusions.

When I first began to notice this, I was genuinely shocked. Schuller has a reputation for having some of the best ears in the business! And yet here he was making some pretty blatant mistakes. His editors deserve to eat some crow here too, as even a cursory listen to the recorded sources would have revealed many of the problems.

Ethan has had previous occasion to point out some of the rhythmic and metric errors Schuller made in his transcriptions of Art Tatum (in The Swing Era) and Ornette Coleman (in a 1961 anthology of Schuller-transcribed Ornette tunes). I also recently remarked on a particularly glaring howler from Early Jazz, where Schuller’s incorrect transcription of “Mood Indigo” leads him to a mistaken conclusion: that it employs “the kind of parallel motion that a pianist would use,” when in fact “Mood Indigo” is one of the best examples of independent linear voice-leading in all of Ellington.

An even bigger and more perplexing set of mistakes plagues Schuller’s analysis of Ellington’s 1937 “Diminuendo and Crescendo in Blue,” which he discusses at length in The Swing Era. This is a piece that is familiar even to those who don’t know much Ellington — the 1956 Newport recording, featuring Paul Gonsalves’ pandemonium-inducing, 27-chorus “wailing interlude,” remains Duke’s best-selling album. But despite all of the great stories about this legendary performance, discussion of the composition itself tends to get short shrift. “Diminuendo and Crescendo in Blue” actually dates back to 1937 — the original recording is available here, or, for completists, on the Mosaic 1932-1940 Brunswick & Columbia box set. The most successful and viscerally satisfying of Duke’s various forays into “extended works” during the 1930’s, “Diminuendo in Blue” was the A-side and the “Crescendo in Blue” was the B-side of a 78 RPM record; the two halves comprise a single, tightly structured piece. It’s a stunning achievement, six minutes of sustained genius, shot through with radical new ideas. (Jim McNeely once told me that every time he thinks he’s come up with a really hip arranging idea, it usually turns out that Duke did it first, 70-80 years ago. This is probably one of the pieces he had in mind!)

I was thrilled when Ethan made the first two pages of Duke’s original “Diminuendo” manuscript available as part of his Reverential Gesture post, and I begged him for the whole thing. Here it is for everyone to check out:

Dim 1

Dim 2

Dim 3

Dim 4

Dim 5

Dim 6

Dim 7

Having the manuscript also inspired me to look at Schuller’s discussion of “Diminuendo and Crescendo in Blue” from The Swing Era. Here’s how Gunther describes the work:

Diminuendo and Crescendo in its original 1937 form was, like Reminiscing in Tempo, a full-fledged written composition with virtually no improvisation. For a jazz piece it was relatively demanding in structure and harmonically, technically complex. Though based on blues changes, Diminuendo in Blue featured elongated 14-bar blues choruses with 2 1/2 -bar subdivisions (Fig. 5) and modulated through a maze of five different keys. Moreover the modulations are often abrupt, startling shifts in tonality which were hard to digest, both for the players and the audience. After the initial key of Eb, the work modulates to G for two choruses, then to C, then to a highly chromatically altered Ab, finally coming to rest on Db as the overall diminuendo of the piece reaches its quiescent resting point.

Schuller Diminuendo Analysis

The first two sentences above are true, but almost everything Schuller says after that is incorrect.

To take it point-by-point:

– “Diminuendo in Blue” is built primarily from 12-bar blues choruses. There is one (but only one) chorus that is arguably 14 bars.*

– There is one (but only one) passage which could arguably be analyzed as a “2 1/2-bar subdivision.”

– None of the harmonic progressions Schuller outlines above are used in “Diminuendo.”

– All of the modulations are prepared. None of them are “abrupt.”

– The “highly chromatically altered Ab” chorus Schuller mentions is actually an F minor blues.

* “Crescendo in Blue” begins with a pair of unambiguous 14-bar blues choruses, but none of the “Diminuendo” choruses are like that. Some listeners might be inclined to fold in the 4-bar FALSE START with the truncated, 10-bar CHORUS 2 (see below) and label them as a single 14-bar blues chorus… but even so, that would still be the only 14-bar chorus in all of “Diminuendo.”

Here is the actual formal plan of Diminuendo in Blue (in its original 1937 recording, available here):


[A] CHORUS 1: Blues in Eb (12 bars)

[B] CHORUS 2: Blues in Eb (10 bars)

[C] FALSE START+INTERLUDE: FALSE START (Eb) begins same as CHORUS 2 (4 bars) then INTERLUDE on VII pedal (2 bars)

[D] CHORUS 3: Blues in G (12 bars)

[E] CHORUS 4: Blues in G (12 bars)

[F] CHORUS 5: Blues in C (12 bars)

[G] CHORUS 6: Blues in F minor (12 bars)

[H] CHORUS 7: Blues in Db (12 bars)

[I] CHORUS 8: Blues in Db (12 bars)

[J] CHORUS 9: Blues in Db (12 bars)

[K] CHORUS 10: Blues in Db (12 bars)

[L] “STAGE FADE” CODA: Db7#9 vamp (6 bars)



CHORUS 1: Full band. The IV chord is prolonged by a half-bar (which I guess you could label as a “2 1/2-bar subdivision”) but the V chord comes as expected in the ninth bar of the form — not the tenth bar, as Schuller has it on his diagram.

The ninth bar also brings us this strikingly unusual three-measure turnaround:

| V7    | IV7  bII7 | II7  bII7 |

This causes the tonic chord to be delayed until the twelfth bar. But there’s no mistaking the top of the second chorus on the downbeat of the thirteenth bar!

(Also check out the forward-looking upper-structure voicings Duke uses in the brass on those E7 and F7 chords, which sound a lot closer to 1967 than 1937!)

CHORUS 2 and FALSE START+INTERLUDE: Full band. The beginning of the FALSE START following CHORUS 2 is extremely deceptive — on first listen, we really want to hear the first two bars of the FALSE START as the final two bars of CHORUS 2, which would make CHORUS 2 the usual 12-bar blues. But this is actually masterful Ellingtonian misdirection! The first three bars of the FALSE START are identical to the first thee bars of CHORUS 2, and the fourth bar of each is similar. These measures each make up a single, continuous four-bar phrase, meaning that with the benefit of hindsight, we can identify CHORUS 2 as a shortened, 10-bar blues form (lopping off the final two bars of a 12-bar blues become a hip trick favored by many subsequent composers and arrangers), followed by a 4-bar phrase — the FALSE START — which begins as a literal repeat of CHORUS 2. But then instead of going to the IV chord in the fifth bar, the FALSE START transitions into the INTERLUDE: a 2-bar VII7 (D7) pedal, which functions as the dominant of the upcoming key (G).

This stretch of music is extremely deceptive, however so others might not hear it the same way. Sonny Greer doesn’t mark any of this stuff so we are on our own here! As I mentioned above, it’s also possible to hear the FALSE START as a prolongation of CHORUS 2, making it a 14-bar blues chorus followed by a 2-bar interlude. (That interpretation is supported by the placement of the double barlines on Duke’s manuscript, though I’m not sure we should read too much into that — the double bars might easily have been added by a copyist after the fact.)

Ethan and I have talked about this section, and he hears the second chorus as a straight 12-bar blues, followed by a 2-bar false start before the 2-bar interlude. That honestly seems utterly perverse to me — it breaks up the middle of what I’d consider an indivisible 4-bar phrase! But all of these conflicting interpretations really drive home the subtlety and artfulness with which Duke disguises the top of the form.

CHORUS 3: Full band. Here, once again, the top of the form is obscured. This time it’s the continuation of the D pedal into the beginning of CHORUS 3** that leaves us feeling utterly disoriented… and then Sonny Greer goes and throws yet another spanner in the works by setting up the third bar of the form instead of the top! It’s not until the IV7 chord comes in the fifth bar of this chorus that we get our bearings back again. But even after that, Duke continues to mess with us: he cleverly creates the impression of asymmetry by repeating the opening phrase of this chorus a bar earlier than we’re expecting. This surprise entrance in bar eight of the form (instead of bar nine) flips the phrase structure and makes us hear the figure in a new way.

** Different bassists have treated this different ways… Billy Taylor plays the dominant pedal at the top of CHORUS 3 the same way on both the original and alternate 1937 takes. But in a 1945 V-Disc recording, Junior Raglin plays G on the downbeat and walks G7 throughout these two measures instead of continuing the pedal, and most subsequent bassists seem to have followed suit. (Also, Sonny Greer is a bit clearer about the top of the form on the V-Disc version — though he also sets up bar 3!)

CHORUS 4: Full band. The blues tonic I7 (G7) is a pivot that functions as the V7 of the upcoming key (C).

CHORUS 5: Full band. Again, the blues tonic I7 (C7) is a pivot that functions as the V7 of the upcoming key (F minor).

CHORUS 6: Trombones with unison sax counterline. The final measure has a III7 chord (Ab7) on the third beat, which also functions as the dominant of the upcoming key (Db).

CHORUS 7: Sax soli.

CHORUS 8: Cootie Williams solo (composed) over saxophones.

CHORUS 9: Harry Carney solo (composed) over trombones.

CHORUS 10: Duke solo (composed) with bass only (drums drop out). All of the harmonies have #9ths added to them — if there’s an earlier instance of anyone doing this on a blues, I’d love to hear it! Also, the IV7(#9) voicing is planed into from a parallel voicing a whole step above, and the left-hand bass line Ellington plays at the end of the chorus emphasizes the b9th of both V and I. All of this points towards a modal approach to the blues using the entire diminished scale, something that would not really take hold in jazz until Coltrane and McCoy Tyner started doing it in the 1960’s.

“STAGE FADE” CODA: The final phrase of CHORUS 10 becomes a bass-and-piano vamp, fading to nothing. Vamping on I7(#9) is over-familiar to us now (James Brown, Jimi Hendrix, etc etc ad infinitum) but this could easily be Patient Zero of that syndrome.


What’s most frustrating about all these errors is that they cause Schuller to overlook almost everything that makes “Diminuendo in Blue” such a badass tour-de-force. He does not address any of the (actual) harmonic variations to the blues structure Ellington employs: the I-VI-II-V “Rhythm” changes turnaround at the top of the second chorus, the chromatically descending II7 – bII7 figure in the second bar of the fourth chorus, those prophetic sharp-ninth chords in Duke’s solo… not even that jaw-dropping three-bar turnaround at the end of the first chorus. Nor does he point out how the top of the form is at times deliberately obscured, another startlingly innovative technique that remained essentially unused in jazz until Herbie, Ron, and Tony started doing it in Miles’s band in the mid-1960s. Schuller doesn’t outline how the dense, overlapping, asymmetrical figures of the first few choruses gradually resolve into a more standard call-and-responnse pattern by the fifth chorus (the last full band chorus). This is key to the narrative arc of the piece, which involves not just a gradual reduction in volume and instrumental forces, but also a journey from deceptiveness, ambiguity, asymmetry, and instability towards eventual clarity. He doesn’t explore the sophisticated — almost Mahlerian! — relationship between each of the keys, nor the brilliant symmetry of the harmonic plan: we begin with Eb up a major third to G and end with F minor down a major third to Db, separated by two modulations along the circle of fifths.*** Schuller laments the “thrice-familiar riff clichés” as “not striking or strong enough to support or justify such complexity,” but meanwhile he’s missed all of the varied and subtle ways Duke uses our familiarity with “riff clichés” and 12-bar blues structure to fuck with us. I mean, it seems to me that this is the entire point of the piece — to both embody and explode the familiar blues form!

***  This is doubtless coincidental, but that final modulation to Db occurs awfully close to the golden ratio

I’m a bit worried that all this analytical talk might make “Diminuendo” sound like a purely intellectual exercise, but if so, there’s a simple cure: just close your eyes and listen to the damn thing. It’s one of the most purely thrilling recordings in Duke’s entire discography, and the band swings like demons throughout. Amazingly, though, Schuller even chastises their execution, lamenting the “two-beat rhythmic feeling and rather stiffly played figures.” Damn, Gunther, we should all be so stiff! He also claims that “both critics and audiences were confounded and reacted either apathetically or negatively” — but, as Terry Teachout notes, even before the famous Newport version, “Diminuendo and Crescendo in Blue” was a showstopper: following a performance at Randall’s Island in May 1938, three thousand audience members stormed the stage and extra police needed to be called in to restore order.

It seems to me that if Schuller had made comparable errors about, say, Schoenberg, people would notice. Possibly these errors were widely discussed by academics when The Swing Era was first published — though if so, those much-needed correctives don’t seem to have left much of a trail on Google. Honestly, Schuller’s mistakes bother me far more than any number of ignorant think-pieces by writers like Maria Popova, Adam Gopnik, etc. They aren’t musicians or scholars. But Schuller is. He’s a tremendously respected figure whose word carries weight. So it’s dispiriting to see him get a landmark Ellington work so wrong, and to see those mistakes go unchallenged.

This is also why we desperately need a proper critical edition of Ellington scores, one that draws on his original manuscripts, the surviving instrumental parts, and all other relevant sources. This is what’s routinely done for classical composers of any significance, and it’s deeply shameful that we haven’t managed to do the same for Duke. I’m glad Ellington’s papers are being preserved by the Smithsonian, but preservation is not enough. They need to be accessible. Surely the Smithsonian has digitized the manuscripts (or is in the process of doing so) — why not make at least some of them available online? This coming May 24, it will be 40 years since Ellington’s death, and you still can’t buy a score that isn’t a transcription. (I understand Loren Schoenberg has been trying for decades to get the manuscripts published — people need to get behind him on this!)

We all know that what makes Duke Duke can’t be entirely captured by what’s on the page. The Ellington Effect is, and always will be, ineffable, mysterious, unknowable. But to amplify an important theme from Ethan’s Reverential Gesture post: if we’re going to have a proper discussion of Ellington’s body of work and artistic legacy, it’s vitally important to know what he actually put to paper.

Darcy James Argue’s Secret Society.