The D Sandwich: What Confirmation's Bridge Taught Me About Leaving a Key
Brando Miranda — September 2026 · ~8 min read
TL;DR. The bridge of Charlie Parker’s Confirmation leaves home twice and comes back once. On alto it reads D, then G, then Bb, then D — a D sandwich. I kept trying to memorize it as a list of chords and it kept sliding off, because I was memorizing the glue instead of the structure: every arrival is announced by a ii–V, so the ii–V’s cannot tell you why those keys. Strip the connectors and there are three moves, only one of which is the ordinary one — one step counter-clockwise on the circle of fifths to the IV, then three steps at once to the bVI, then home by a major third. In circle steps: −1, −3, +4 = 0.
Eight bars that kept sliding off
I’ve been shedding Confirmation on alto. The A section is its own project — that chain of ii–V’s that keeps handing you off, one key to the next, is most of why the tune gets assigned. But the A section stuck and the bridge didn’t. I could play the bridge. I could not say it. Away from the horn I’d get the first four bars and then guess.
Notation first, so nothing is ambiguous: everything below is in alto key (Eb), since that’s the part I read, so concert pitch is a major sixth lower. My D is your F, my G is your Bb, my Bb is your Db.
Written out, the bridge is Am7, D7, then Gmaj7 for two bars; then Cm7, F7, Bbmaj7, and a last bar that turns around and points home. Eight bars, two blocks of four. (Concert, if you’re on a piano part: Cm7 F7 into Bbmaj7, then Ebm7 Ab7 into Dbmaj7, and home to F.) Charts disagree about that eighth bar — a major ii–V home, a minor one, or just holding the new tonic — and nothing below depends on which version you play.
Three key centers counting home: D, G, Bb, back to D. A D sandwich — start at D, stuff two other keys in the middle, end at D. And the question that would not go away while I practiced it: why those two?
The glue is not the structure
My first answer was the one everybody gives, including me, for about a week: he uses ii–V’s. True, and useless. Both halves of the bridge are announced the same way — Am7 D7 into G, Cm7 F7 into Bb — which means the connective tissue is identical in the two halves and therefore cannot explain why the halves go to different places. A ii–V is a ramp. It will take you to any key you point it at. Asking a ii–V why you ended up in Bb is like asking a road why you moved cities.
That was the confusion I had to clear up, and it was mine, not the chart’s: I was memorizing the glue and expecting it to hand me the shape. Take the connectors out and eight bars collapse to three letters — D, G, Bb — and then the interesting question is finally askable.
The first departure is the oldest move there is
D to G. G is the IV. And the way Bird gets there is by turning the tonic into a dominant: the D that has been home for sixteen bars becomes D7, which is not a tonic at all — it’s the V7 of G. One chord changes quality and the floor moves under you.
On the circle of fifths that is one step counter-clockwise, one step toward the flats, the direction of falling fifths, which is the most traveled interval in all of this music. Nothing here is hip and nothing here is a surprise. It’s gravity. If the bridge had continued in that direction — G to C, C to F — you’d have a perfectly good tune that nobody would write a blog post about.
Then there’s the detail that finally made the first half stick, and I’m slightly embarrassed it took as long as it did: this move is already in the A section. Bar 4 of the head is Am7 D7 and bar 5 lands on G (concert: Cm7 F7 into Bb). Same roots, same ii–V–I into the IV, at double speed — there the ii and the V get half a bar each and the arrival gets one; in the bridge they get a bar each and the arrival gets two. What changes is what happens on landing. In the head, G is a waypoint that keeps moving, and some charts write it as a dominant to say exactly that. In the bridge it is a major seventh with two bars to sit in. Bird takes a transition out of the head and promotes it to a key. Once I saw that I stopped memorizing the first half of the bridge — it was already in my hands, I just hadn’t recognized it slowed down.
The second departure does not come from the circle
Here’s the part I actually wanted to understand. If Bird were walking the circle, the next stop after G is C. He doesn’t go to C. He goes to Bb.
And he gets there by doing something that doesn’t feel like a modulation while you’re playing it: he takes the three bars that got him to G and transposes them up a minor third. Am7 D7 into Gmaj7 becomes Cm7 F7 into Bbmaj7 — same shape, same rhythm, three semitones higher. The bridge is one idea said twice at two pitches, and the sequence breaks in exactly one place: the eighth bar, where instead of holding the new tonic the way G got held, the chart turns around and aims at home. That’s why the second half becomes easy the instant you stop treating it as four unrelated new chords.
Two things that the minor third buys him.
It is not a circle move. Up a minor third is three counter-clockwise steps at once: G to C to F to Bb. He skips two stops. The circle of fifths is exactly the system that makes D to G feel inevitable, and the move to Bb is made against that system, in one jump — which is why it lifts instead of settling.
It lands on the bVI. The sixth degree of D major is B; Bb is what you get by borrowing from D minor. That’s modal interchange: a key the tune has no business visiting, which it’s allowed to visit anyway because the parallel minor is always sitting right there. So those three bars sit entirely outside the home key — Cm7, F7 and Bbmaj7 each contain notes D major does not have — and the only thing legitimizing the visit is a relationship to a key we are not in. That’s a strange and wonderful sentence to be able to write about eight bars of a bebop head.
People call the minor-third jump a Coltrane move, and I want to be careful there, because I was sloppy about it in my own head first. Coltrane changes are major thirds — B, G, Eb in Giant Steps — and they’re a system: three keys evenly dividing the octave, cycled. One minor-third lift is not that. What Bird has in common with Coltrane here is the family of the move, not its arithmetic: go by a third instead of a fifth. Bebop is built on fifths, so thirds are the exit door. Coltrane later turned the exit door into a floor plan.
The arithmetic closes
Which leaves the return. From Bb, home is D — up a major third. And that one is a Coltrane interval exactly: concert Db to F, the same interval class Giant Steps cycles through.
Count the whole bridge in circle steps, counter-clockwise negative:
- D to G: one step counter-clockwise, −1.
- G to Bb: three steps counter-clockwise, −3.
- Bb to D: four steps clockwise, +4.
And −1 − 3 + 4 = 0. Of course it does; you end where you started. The point isn’t the sum, it’s the shape of the ledger: the bridge drifts flatward in two unequal moves and then repays the entire debt in one sharpward move, by a major third. Exactly one of the three moves is the ordinary one, and it’s the first — the one my ear predicted before I’d worked any of this out.
(The eighth bar doesn’t change the count, whichever version your chart prints. A turnaround is glue: a ramp pointed at D, not a fourth key center. Same lesson as before, showing up one more time at the end.)
One more thing about that count, which is really a question for a different afternoon. Every interval in the tune has a whole-number address on the circle: up a fourth is one step, up a minor third is three, up a major third is four. That works because a fifth is seven semitones, and seven shares no factor with twelve, so stacking fifths eventually reaches all twelve notes — the circle of fifths is just a relabeling of the twelve pitches by multiples of seven. Whether that’s the reason it sounds correct, or only the reason it’s easy to draw, is the thing I actually want to know.
Why this tune and not Rhythm Changes
There’s a thought I’ve had more than once while working on this: I should probably be spending these hours on Anthropology instead.
Rhythm changes are the utility vehicle of this music. Hundreds of heads sit on that form; learn it once and you can walk into most sessions and be useful. And its bridge is four dominants, two bars each, falling by fifths — D7 G7 C7 F7 in concert Bb — the first three never resolving to anything but the next dominant, until the last one drops you back into the A section. Four chords, one shape. You can run a single idea through all of them and sound like you meant it. It’s a cycle: one rule, applied four times.
Confirmation’s bridge refuses that deal. Two complete ii–V–I’s, both actually resolved, a minor third apart, the second one in a key borrowed from the parallel minor. There is no one rule to run through it. You have to know where you are, in three keys, in eight bars.
For a while I filed that under regret — a suspicion that I’d picked the ornamental tune over the useful one. I don’t think that anymore, and working out the bridge is why. The move to G is the part my brain likes: derivable, falls straight out of the circle, I could have predicted it from the first chord. The jump to Bb is the part I love, and it’s precisely the part that isn’t derivable. You can explain it afterward — bVI, modal interchange, a sequence up a minor third — but no rule generates it. Nothing about bar 4 of the bridge tells you bar 5 is coming.
That’s not a worse thing to practice. A cycle teaches you a pattern. This teaches you a place.
The lesson
The eight bars stopped sliding off the day I stopped storing them as chords. Stored as chords, the bridge is a list of symbols, and I could never keep straight which half came second. Stored as structure, it’s three sentences: take the ii–V–I into the IV that is already sitting across bars 4 and 5 of the head and give it four bars; say it again a minor third higher; turn home by a major third. The ii–V’s then take care of themselves, which is the entire job description of glue.
Two things I want to chase from here. One is that circle arithmetic — whether the seven-and-twelve coincidence is doing real work or is just bookkeeping. The other is the comparison I kept making without earning it: a real side-by-side of rhythm changes against this, template against sculpture, to see how much of the difference is structural and how much is just familiarity talking. I did the same kind of sitting-with-it a few days ago with Maiden Voyage’s sus chords, and that one also ended with me discovering the confusion was a naming problem.
That is enough for a Saturday afternoon.
If you’d like to cite this post:
@misc{miranda2026confirmationdsandwich,
author = {Miranda, Brando},
title = {The D Sandwich: What Confirmation's Bridge Taught Me About Leaving a Key},
year = {2026},
month = {September},
howpublished = {\url{https://cs.stanford.edu/people/brando9/2026/09/12/confirmation-d-sandwich.html}},
note = {Blog post}
}