Assume the following rudimentary progression in a major scale:

(Progression 1) I - ii - V7 - I - ii - V7 - I

Progression 1 starts with the tonic. Starting with the tonic, Progression 1 may go anywhere from there, so we choose to use the "ii - V7 - I" progression for the next part of Progression 1, as well as for the final part of Progression 1 (since the previous part ended with the tonic too, we could very well use any valid progression after that, so we chose to use "ii - V7 - I" again). Progression 1 ends with the tonic, so all is fine. Progression 1 is a good progression.

Let us consider tritone substitution and tonic substitution now. Tritone substitution dictates that a dominant chord (like V7 in Progression 1) may be replaced by the dominant chord constructed starting a tritone above (by ♭II7 in our case). V7 and ♭II7 share two notes. Tonic substitution dictates that chords iii and vi may replace chord I. I and iii share two notes. I and vi share two notes.

So, considering tritone substitution and tonic substitution, the following substitutions can take place:

Substitute V7 with ♭II7
Substitute I with iii
Substitute I with vi

Will any combination of the substitutions above, wherever applicable inside Progression 1, result in a valid progression? For instance, are the following progressions valid? If not, what rules are broken?

(Progression 2) I - ii - ♭II7 - vi - ii - V7 - I
(Progression 3) I - ii - ♭II7 - iii - ii - V7 - iii
(Progression 4) iii - ii - ♭II7 - I - ii - V7 - vi
(Progression 5) vi - ii - V7 - I - ii - ♭II7 - I

This is my first question here. I hope I do not ask much.

  • 4
    What do you mean by "valid"? Who decides what is valid? Some academics in a music department of a prestigious university? Justin Bieber fans buying the record featuring a song which incorporates the progression? A proclamation of validity from some notable jazz musician who has "street cred"? I can just about hear the progressions you have there in my head; I go through the same sort of stuff of my guitar and know that it sounds fine. Some of the progressions which neither start nor end on the tonic might not clearly establish tonality, but a) so what, b) in the right context they will.
    – Kaz
    Commented Sep 6, 2016 at 21:34
  • Kaz, thank you for this comment. I would upvote it as an answer too, if you care to submit it as an answer. Maybe "valid" translated to "not bad" in my book, when the question was asked. But it is all about the context, as you said. Because the next question would be about the definition of "not bad".
    – xnakos
    Commented Sep 6, 2016 at 22:02
  • " If not, what rules are broken?" - there is only one rule, and it says "there are no rules". Just use your ears - nothing else matters.
    – user19146
    Commented Sep 7, 2016 at 1:22

5 Answers 5


First off, every progression is valid. Whether it sounds good, makes sense harmonically, or is what you want is entirely different. So every progression you list is valid, however some things should be noted.

When substituting closely related chords the function will be similar to that of what you are substituting, but not serve the exact function. For example on progression 3 and 4 you have deceptive cadences and especially in 3 it may not be what you want. Progression 3 in general is not a very good way to imply I as tonic because you are letting one of the more powerful concepts in functional harmony slip away which is the resolution of the leading tone. In V7->iii and bII7->iii even though the chord can work as a tonic substitution it leaves the leading tone common and if you don't resolve it up to the tonic, like you don't in progression, it really weakens the idea of a cadence being observed at those spots.

In general listen to the progressions and hear the different flavors of each. The effect I just talked about while not working for functional harmony may fit in a piece you are composing. You need to be aware in general of the effects each of the substitutions have on your line and when you can do it and when you shouldn't.


Never mind the 'rules'. They all work, but some could do with careful voicing along the way. It's an interesting angle to work from, but substitution is just that. Some of the notes from a particular chord are the same in another, so sonically, there's a good chance that the sub. chord will sound good. The 'rules' come from good practice - as in does it sound o.k? Then we'll make a 'rule' that says so, more often than not.


As the others have mentioned, the validity of a chord progression is entirely subjective and should essentially be determined by whether or not it sounds good and fulfills your musical needs. I always like to remind people that music theory is not so much a set of rules, as many are often taught, but an explanation of what is happening and a language to discuss it. Theory is basically only rules when you are trying to authentically imitate a specific style, so you will actually find that there are different "rules" in this regard for different genres. There are a few things that I would mention to look for while choosing substitutions though.

The idea of substitution, at least in the traditional sense, is that different chords fulfill a function within your composition and that you can choose other chords that will fulfill the same or a similar function. This is most easily demonstrated with the tritone substitution from your example. The idea here is that the 3 and 7 of the chords are shared, which are the primary notes driving the function of the chord. With these notes in common, the two chords can fulfill the same function but there are differences between these chords, one of which being texture.

When choosing a substitution, it's important to pay attention to what the melody and other important parts of the arrangement are doing. The common tones between two chords may not be enough to justify a substitution. For instance, if you were to choose to substitute iii for I, then you will want to pay attention to whether or not the tonic is in the melody since that is the one note that is not shared between the two chords. The 5 of the iii chord is ^7 (scale degree 7), a half-step below the tonic, so if the melody lands on the tonic during the iii chord, it can cause a lot of tension, possibly beyond what you may deem appropriate or would be considered appropriate for a given genre/style.

Similarly, you can look at the tritone substitution (which I'll call "TT" moving forward) and find that certain notes may cause issues like this. 5 of the TT is ^b6, so if you have ^6 in your melody, you will have a half-step dissonance that may cause issues. Same with 2 of the TT being ^#2 (b3), which could conflict with ^2 or ^3, or 6 of TT being ^#6 (b7) and conflicting with ^6. You also want to mind the 4 of TT, which would usually be replaced with #4, which is ^5. In a Jazz setting, this is often resolved by using an Altered Dominant chord, which would include all the notes from the major scale except the tonic, which isn't usually a melodic note on a dominant chord. The notes of an Altered Dominant in C Major, starting on b2 would be Db (1), D (b9), E(#9), F(3), G(#11), A(#5/b13, depending on who you ask), and B(7), though Jazzers would often spell it enharmonically to line up with the chord tones described in parentheses above (Db, Ebb, E, F, G, A or Bbb (depending on who you ask), Cb). This allows you to use a TT for just about any diatonic melody that appears over a V7 chord without creating major conflicts with the melody, however, it is still a different chord and one that is rather dissonant, so it should be used with care outside of its standard Jazz setting (or even in Jazz).

On the whole, like with all theory, you should be using your ears more than subscribing to some rules. If you follow the rules, you may end up with something that you don't like the sound of but is "right" by academic standards, while breaking those rules intentionally could make you the next Stravinsky. Play something and if it sounds good, try to figure out why it made sense to break those rules and try to find a way to describe it within the framework of theory.

  • 1
    There is a lot of material out there and it can certainly be a lot to digest. Depending on your approach to learning it all, it can be more or less difficult. A lot of people approach the subject by learning all the "rules" so you can break them with proper understanding, however, you can learn concepts and attempt to apply them without having learned the entire repertoire, which is where your ears come in handy. I spent a long time considering theory to be the be all, end all of music when in reality a lot of stuff I wrote that was inspired by theory sounded awful. Your ears won't lie. Commented Sep 7, 2016 at 14:01

Some excellent answers here, but let me also add: depends what your priorities are for "validity". Are you trying to maintain strict counterpoint, just reasonably consistent voice leading, or just keep it sounding fairly consonant behind the the screaming Gilmouresque Strat/bebop saxophone/ambient Frippertronics solo you already recorded over the unmodified progression?

My answers to those would be "probably not", "maybe", and "probably"/"yes"/"it doesn't matter" (screaming Strat/bebop Sax/ambient Frippertronics), respectively.

Remember: Debussy's amazing music is full of stuff that isn't "valid". Ornette Coleman's (or maybe better, Sonny Sharrock's) is almost nothing but. Erik Satie made his name ignoring the rules. Listen to Ligeti's beautiful "Requiem"' or some Meredith Monk stuff... What the hell even is that? If you want a laugh, read Ralph Ellison's contemporary reviews of the "confused" music of Charlie Parker sometime. In the visual arts, read up how the Modernists like Cezanne were received in the 1870s (and by some, still today... Odd Nerdrum, an otherwise great painter if you ask me, still rejects them.)

I think the only composers who were ever usually canonically "correct" in a theoretical sense were JS Bach and maybe Palestrina, of whom I have to cop a fair degree of ignorance. Even Mozart wasn't ("String Quartet in C" that can't seem to stay in C, anyone?) And I think even Bach probably occasionally let a zelcher slip through, once he was well-regarded enough to get away with it.

The point of my half-kidding is this: a better question would be, what are you trying to do, followed immediately by, are you musically astute enough to understand the ways in which what you are doing are valid or invalid steps towards accomplishing that? These are more productive questions. (If the answer to the second question is no, don't worry, that can be fixed with dedication and work.)

Lastly, remember, in the big picture, few truly great artists are remembered for how well they followed the rules. Think about it.


If the tritones in the substitute chords (the F - B in G7, also present in Db7 as F - Cb, which is the reason b5 substitutions work) are the same ones as would have been in the basic version, you'll be OK.

But you've got to dump this idea of "valid". Theory describes what works, sometimes finds reasons for it. It doesn't tell you what you MAY do. If you find a progression that doesn't fit the "rules" of tritone substitution, perhaps you're doing something else instead. That's OK!

  • Laurence, thanks! I do not think that "valid" should be dumped. But you are probably suggesting that one should loosen up, right? "Valid", as I see it, is "sufficient" in order to produce an OK/good/great result. But the consensus is that there is magic not captured by "valid" constructs, i.e. "valid" is not "necessary" in order to produce something nice. Something nice may come from something uncharted. Or from something "invalid". "Valid" is useful, since everyone will need to take some textbook decisions sooner or later. It is also no fun to break a rule that you are unaware of.
    – xnakos
    Commented Sep 7, 2016 at 17:14
  • 1
    More than that. Maybe you've written a progression that uses alternative chords but they're NOT tritone substitutions. That's fine. Tritone substitutions aren't the ONLY substitutions or alternatives. Validity comes from it sounding good, not from it satisfying any particular pre-conceived pattern.
    – Laurence
    Commented Sep 7, 2016 at 17:48

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.