Naima, from Giant Steps.

While listening, I wondered — what if Bill Evans had played this piece? I couldn’t find recordings. Wynton Kelly was the pianist on the original recording (linked above).

A harmonic analysis of the whole recording. Subtlety abounds.

For voicings, in general, avoid 5ths — sometimes based on the pedal, and sometimes in the chords above (i.e. the ones before the “slash”).

I tuned to A=442.

Bars 1-8 (A section, 0:00)

Bass pedal on the dominant E♭. Even during the tonic A♭ chords, the pedal persists.

Bar 1

Melody starts on the 13 of E♭ sus 13 — a C.

Bar 2

Melody moves (down a 5th) to the 9 of E♭ minor 9 — an F.

It is crucial to avoid the 5th of the chord (the B♭).

Bar 3

Beats 1 and 2

If the pedal weren’t there, this would be an A major 7 chord, with the melody playing the major 7 - a G♯.

But the E♭ pedal is there.

E♭ Locrian scale (based on the 7th degree of the E major scale) contains both the A major 7 and the E♭ pedal. It makes sense that it’s a weird one, because this is a weird sound.

So this ends up being an E♭ half-diminished 7.

Beats 3 and 4

We are closer to resolution — the chord is again built on the pedal E♭, it’s hard to hear the exact voicing, but since a C clashes with the voicing, I hear this is not an E♭ half-whole octatonic scale. Instead, it’s built on the E melodic minor scale.

And so this must be an E♭ dominant 7 alt, with the melody sounding the ♯9 (F♯) and the 1 (E♭).

What makes it really hard to hear is the continued E♭ pedal with the really close piano voicings — just 3 notes each, all using the same m2 M3 voicing:

Bar 3 (beats 1 and 2): G♯ A C♯
Bar 3 (beats 3 and 4): F♯ G B — this voicing is essential to the sound
Bar 4: G A♭ C

Tonal gravity tangent

Another way to think of these first three/four chords is as pedal + major 7 chord with root an interval above the pedal): where the intervals are, respectively m7, m3, TT.

However if we ignore the pedal altogether, what emerges is root motion as two sets of major 7 chords moving with tonal gravity (this is no longer complete support of the melody of Naima on that second half of Bar 3, but it’s close): D♭△ G♭△ A△ D△ — this kind of “wants” to resolve to a D♭△ (not to the A♭△ in Naima).

This also sounds an awful lot like the whole-tone/thirds/Giant Steps situation…

Bar 4

A simple A♭ major 7 — the pedal resolves, and the melody is the 5, an E♭ — incidentally, what was the pedal!

Bar 5 - 8

Repeat bars 1 - 4.

Bars 9-16 (B Section, 0:32)

Bass pedal on the B♭, a 5th up from the first 8 bars — against tonal gravity.

Bar 9

Melody starts on the root 1 of a B♭ sus ♭9. It comes from the III Phrygian mode of G♭ major scale.

Else, it is simply a B major 7 over a B♭ — hence the suggested suspensions — of the ♭9 -> 1 and 4 -> 3 and ♭6 -> 5 — it’s basically the most suspended chord relative to the B♭ major chord whose root is the pedal.

But I think given that this recording is post-_Kind of Blue_ it is safe to assume that Coltrane must have been leaning towards the modal analysis.

Bar 10

G major 6 over the continued pedal B♭. The added 6 makes a tritone with the pedal, which can’t be ignored (the sound isn’t quite Naima without it). It would be hard to find this chord “naturally” — but it’s there, in this piece.

Melody goes — from the root of the pedal in bar 9 — to the root of the “over” chord — the G.

How to analyze it/name the chord (other than as a “slash” chord)? This is a diminished harmony, based on the symmetric octatonic scale — the G half-whole octatonic scale is one such scale. It seems more fruitful to analyze it in that scale than to find a name for it within traditional harmony. But then Coltrane was playing on the edge of traditional harmony and modal harmony during this period. In any case, it’s certainly an off-color sound.

Bars 11 - 12

Repeat bars 9 and 10.

Bar 13

B♭ alt 7 — clearly over the B minor melodic scale — the most “out”, the most “dissonant” sound of this piece — the melody is on the ♯9, the D♭.

Bar 14

Continued B♭ sus ♭9, but melody goes down a minor 3rd to the B♭, either 1) major 7 of the B or 2) root of the pedal B♭ or 3) 3 of the underlying G♭ major scale.

Bar 15

Melody goes down a minor third to the 13 of B♭ sus 13 — a G. Sounds analogous to bar 1, but a 5th up — this is much less dissonant compared to where we came from — almost a “home” sound, except a 5th up.

Bar 16

Melody goes down a major third to E♭ — our friend, the pedal from Bars 1 - 8.

At first I thought this was an E△/B♭, but there is clearly no B in the recording.

What it sounds most like, is the “A” position rootless voicing construction (see Appendix below) that forms a B♭ half-diminished 7, with the melody on the 11 of that chord.

Not only is that 11 the E♭ pedal from Bars 1 - 8, but this half-diminished 7 appears to function as “dominant” to the upcoming E♭ sus 13 — which of course is not the tonic — the eventual tonic that we will get to, only at the last measure (following in the footsteps of Tristan und Isolde), is A♭! And so, this chord can be thought of as the ii half-diminished 7 leading to the V sus 13 leading to the I major seven! Eureka. The scale is the II of the D♭ melodic minor scale.

On another hand, even though there is no B in the sound, the A♭ harmonic minor scale could also work here?

Scales used in B Section

To recap, the scales used in the B section — all with a B♭ pedal:

Bars 9/11: B♭ phrygian (G♭ major)
Bars 10/12: G half-whole octatonic scale
Bar 13: B♭ altered/diminished whole-tone (B melodic minor)
Bar 14: B♭ phrygian (G♭ major)
Bar 15: B♭ mixolydian (E♭ major)
Bar 16: D♭ melodic minor (or A♭ harmonic minor?)

Bars 17 - 20 (A section recap, 1:05)

A shortened recap of the A section, as one 4-bar repetition.

Bars 21 - 40 (Piano solo over AABA, 1:22)

Piano solo over bars 1 - 20.

Bars 41 - 48 (B section, 2:46)

Bars 49 - 52 (A section, 3:21)

A shortened recap of the A section, as one 4-bar repetition.

Some differences? TODO: compare against the Eb sharp 9.. could it be be a G major 7 over Eb earlier (above) ?

Bars 53 - 56 (Pre-coda, 3:38)

Repeat bars 51 and 52 twice. (This would be another A section but the first 2 bars are subbed out for the last 2 bars)

Bars 57 - 61 (Coda, 3:55)

Repeat bars 51 and 52 a third time…

But, instead of ending on the I in bar 58, begin an ascent in the melody where each note lasts 2 beats, up the A♭ major scale: E♭ F G A♭ B♭ C until finally ending on the E♭ in Bar 61.

Over this ascent starting on bar 58, the chords vamp, alternating every two beats between A♭ major 7 and the dominant E♭ sus 13 until finally, in bar 61, the pedal drops out, there is only the A♭ major 7 with the melody holding a fermata on the Eb — the 5 — what had been the pedal.

That vamp at the end is a really, really classic sound. Turns out we can walk up the mixolydian scale, alternating between the I and V:

Scale Degrees	Chord	Chord degrees
`5 7 1 3 5`	`I△`
`6 1 3 4 6`	`Vsus13`	`2 4 6 7 9`
`7 1 3 5 7`	`I△`
`1 3 4 6 1`	`Vsus13`	`4 6 7 9 11`
`2 3 5 7 2`	`I△9`
`3 4 6 1 3`	`Vsus13`	`6 7 9 11 13`
`4 6 1 2 4`	`Vsus`	`7 9 11 5 7`
`5 7 1 3 5`	`I△`

The coda of Naima is essentially this scale, omitting the 4th scale degree chord.

Appendix: “A” and “B” position voicings

Mark Levine, in the The Jazz Piano Book (bottom of p. 46) refers to the following voicings of dominant 7 chords (and this is how I was taught by Anthony Mele):

“A” position: 7 9 3 13
“B” position: 3 13 7 9

Within traditional harmony, these voicings fruitfully apply to at least two other root notes, where the “A” voicing is built on some interval above the root:

“A” voicing m7 above the root forms a dominant 7 — based on the root — using the mixolydian scale — chord tones 7 9 3 13.
“A” voicing TT above the root forms a half-diminished 7 — based on mel. minor scale starting m3 above the root - chord tones ♭5 7 1 11
“A” voicing m3 above the root forms a minor 6 (major 7) — based on mel. minor scale starting on the root — chord tones ♭3 5 6 9
“A” voicing m2 above the root forms a sus flat 9 — based on mel. minor scale M2 below the root — chord tones ♭9 11 5 1

Loose Ends

Coltrane Substitutions

Countdown and Giant Steps sound very similar. See also: Coltrane Substitution.

A Passion for Jazz

A Passion for Jazz maintains an excellent list of jazz quotes.