This is one of those jazz theory ideas that gets explained endlessly online and in texts and is relatively rare in a typical American’s listening experience. But when you do hear it, it does sound cool. I made an interactive explainer on Noteflight, because as with so many jazz theory concepts, tritone substitutions make more sense when you hear them than when you see them represented symbolically.
Here’s the verbal explanation, for what it’s worth. Say you have a V7-I cadence in C major, that is, G7 resolving to C. The active ingredient in G7 is the tritone between the third, B, and the flat seventh, F. This same tritone is also present in Db7; its third is F and its flat seventh is C-flat (the same pitch as B.) For jazz purposes, this means that you can substitute Db7 for G7 and it will function the same way, but with an edgier and more chromatic sound. This is called tritone substitution because the substitute chord’s root is a tritone away from the original chord. The practical consequence is that you can precede any chord with the dominant seventh chord whose root is a half step higher and it will create a nice resolution.
The three most common use cases for substitute dominants are to lead into the I chord, to lead into the IV chord, and to lead into the V chord. Let’s take them in order.
SubV7/I
The substitute dominant of the I chord has its root on the flat second. In C, that’s Db7. It resolves to C or Cm.
The very first chord of “I Can’t Help It” by Michael Jackson (written by Stevie Wonder) is a substitute dominant. The tune is in Ab major. The intro is a loop of A7 and Abmaj7. (I guess technically it’s Bbb7 to Abmaj7 but no one wants to see double flats in a chord symbol.)
“Doing It To Death Parts One and Two” by the JBs is a funk groove in F. It’s a loop of F9 (the I chord) and Gb9 (the subV7/I chord.) Later it shifts down to a loop of D9 and Eb9.
Substitute dominants are a very guitaristic sound generally, because moving a chord shape up and down chromatically is so easy on the fretboard. Many guitarists discover them accidentally without ever getting into the theory behind them.
SubV7/IV
The substitute dominant of the IV chord has its root on the flat fifth. In C, that’s Gb7. It resolves to F or Fm.
In the blues, it’s common to put a substitute dominant in front of the IV chord. Listen to “Stormy Monday” by T-Bone Walker. It’s a blues in G. At 0:08, Walker leads into C7, the IV chord, with Db7, its substitute dominant. (He also plays lots of Ab7 chords before the G7 chords, the subV7/V leading into the V.)
John Coltrane’s “Cousin Mary” is mostly a twelve bar blues in Ab, but in measure nine, the chord is D7, the subV/IV chord. This resolves to Db7, the IV chord.
Outside the blues, you can hear subV7/IV in Beyoncé’s song “Love On Top”. The tune is in C, and the first chord in the intro is Cmaj7. The second chord is Gb7, the subV7/IV. The third chord is Fmaj7, the IV chord. This same progression begins each verse and chorus.
SubV7/V
The substitute dominant of the V chord has its root on the flat sixth. In C, that’s Ab7. It resolves to G or Gm.
Many blues songs (including “Stormy Monday”) put subV7/V in front of the V7 chord. I have also heard that chord called the minor blues subdominant, because it so often appears in, you guessed it, minor blues. The archetypal minor blues is “The Thrill Is Gone” by BB King. It’s in B minor. At 0:23, there’s a G7 chord, the subV7/V. This is followed by F#7sus4 and F#7, the V7sus4 and V7 chords, which in turn lead back to Bm, the i chord.
Edge cases and ambiguities
People sometimes describe “Toxic” by Britney Spears as an example of tritone substitutions. The tune is in C minor, and the chorus includes the chords Eb, D, Db and Cm. Is this a chain of tritone subs? I don’t think so. The chords are triads, not dominant seventh chords. You could play them as dominant seventh chords and it would sound cool, but since they aren’t dominants in “Toxic”, I think this is just fake-exotic chromatic planing.
“The Payback” by James Brown is another tune that uses chromatic planing of non-dominant chords. The main groove is a loop of Bbm13 and Bm13. The effect is similar to a substitute dominant, but not exactly the same.
Improvising over substitute dominants
The conventional scale for improvising over a substitute dominant chord is Lydian dominant mode. It’s Mixolydian with a raised fourth, or Lydian with a flat seventh. So if you have a Db7 acting as a substitute dominant for C, you would use Db Lydian dominant for your melodies. You can also add extensions to the Db7 chord using notes from Lydian dominant, giving you the ninth, sharp eleventh, and thirteenth.
As with secondary dominants, there is a whole universe of jazz theory surrounding tritone substitutions. Jazz musicians mostly use tritone subs as an improvisational tool: whenever you see a G7 written, you can swap in Db7, whether you are soloing or comping. My class is only concerned about tritone subs that are written into the tune. Advanced improvisation is beyond the scope of this post (and my class.)
Fantastic overview! I agonise over how much time to spend on subject this with my mainly pop song writer theory class. I will point them to this as the conversation starter from now on as it’s the perfect ice breaker. Also, spent a bit of time on the “Love on Top” Tritone sub couple of years back with my perfect pitch enabled students and the voicing of the second chord is a Gb7(b5). With the recurring A natural in the motif above that vertical stack is can be heard as Gb altered. All the online sheet music I have ever seen is incorrect in this matter.
Online sheet music is such a mixed bag. I use it to sanity-check my transcriptions sometimes, but I assume that whoever made the chart doesn’t have any special insight that I don’t have. (It’s not like Beyoncé herself is writing those charts, you know?) It’s helpful to have someone with perfect pitch around to check things! The closest I can come to that is the spectrogram.
I’m enjoying your blog. Thank you. Is there any way to see a list of all the topics you have covered and a way to click onto those topics directly? Thank you, Roman Lohyn ( relohyn@gmail.com )
At the bottom of the page, you can see tags and categories, just click any one.
I’m curious as to the theoretical explanation of why Lydian dominant is the go-to scale here. As always, thanks for the super helpful general primer on another fascinating feature of music.
That’s a good question, and I don’t have a good answer. One guess is that it’s a mode of the altered scale, and that’s the thing that really sounds good. But why does the altered scale sound good? Who knows. Another guess is that it’s a variant on Mixolydian that feels more searching and unsettled. Or that the sharp eleventh is usually a diatonic pitch, so it’s less jarring. Really, though, your guess is as good as mine.