From musical theory I know of no reason not to use any of the chords of the key in any order. Yet actual music tends to be strongly restricted both to a limited set of chords and to a limited set of orderings of those such that it is suprising every time if there are more of them used.

So what does determine whether a chord progression makes sense? For example, what if I went with I ii V iii I? What determines that a particular chord usually doesn't follow in the progression? When does it make sense to have chords that are out of key?

  • 2
    If it sounds good, it is good. Or maybe, if it creates a feeling in the listener, then it's a valid way to communicate that feeling. That's the only determination that matters. Commented Apr 12, 2015 at 8:35
  • @ToddWilcox if "If it sounds good, it is good" is all we can offer, isn't that to say that theory is dead? (Not that I would necessarily be able to argue with that...) Commented Apr 12, 2015 at 11:10
  • 2
    Personally I see theory as both interesting to think about and almost useless for composition in a conscious way. Maybe it's like influences: sometimes long after I've written something I realize where some ideas came from but I wasn't thinking of those things at all when I was writing. So maybe theory isn't dead but it's also not The Law and we shouldn't get hung up on it for any reason, especially not when writing. Commented Apr 12, 2015 at 13:04

7 Answers 7


You can think of music theory in two parts :

  • The 'scientific' part : Physics and psychoacoustics
  • The 'stylistic advice' part : Patterns and practices that have been observed to be common to certain musical styles - i.e. if you want this kind of sound, you should do this.

From the scientific point of view, maybe we can observe that

  • Some notes have fundamental frequencies that are part of the harmonic series of other notes (or have an octave relationship therewith), giving them a close relationship and making movement between them easy.

  • (obviously) some notes are close in pitch, giving a natural ascending or descending motion when you move from one to the other.

  • We tend to expect pieces of music (or sections thereof) to move towards a root note that the harmony is based around.

The diatonic scale, which is what you are using when you are writing music in a key, has been designed to take those things into account , and so if you are using it, you will have lots of opportunities for good, scientifically-sound harmony :). However, the diatonic scale (and the whole idea of keys) is a stylistic choice. It's certainly not the only way to write good-sounding harmony - for example, the 12-tone equal tempered scale allows many harmonic possibilities that work well but go outside the bounds of a particular key.

So, to ask When does it make sense to have chords that are out of key? to me is slightly backwards thinking. If you are restricting yourself to the notes (and therefore chords) in a key, it's because you've already made a stylistic decision to do so, because you want that particular sound. If you haven't made that decision, then you're not in a key (in the sense of sticking to the notes of the diatonic scale), and therefore question you asked doesn't arise.

Put another way - you can always do whatever you want, and 'making sense' doesn't come into it.

However, if we consider the 'stylistic advice' that music theory gives us about particular styles : to say From musical theory I know of no reason not to use any of the chords of the key in any order indicates to me that you could usefully read around the subject a bit more! A lot of music theory is concerned with ways to choose your harmonic motions. Some starting points to look at would be:

  • Functional Harmony
    • How to set up modulations
    • more advanced ideas such as secondary dominants
  • Jazz harmony and substitutions
  • Modes
  • The blues scale
  • "standard" chord progressions:

A lot of modern music comes from a mixture of classical and common practice harmony, and blues and folk scales and modes. Personally I think this sudden fusion of styles has caught the music theory field on the hop a little, and I've not read anything that really deals with the ways in which the various musical traditions are blended together to create the common effects we see in popular song - unfortunately you often hear some suggestion that 'chord borrowing' is going on, and not much more. Maybe someone better-informed can suggest something! Still, once you've learned about the various traditions I've mentioned, you probably have most of the tools you need to make sense of songs in your own way.

  • My question is not which progressions do exist but which ones can exist and where, i.e. why some progressions would be preferred and some rejected and which contexts determine this. It is not I who has made a stylistic choice but others make their stylistic choices and I do not understand why they make them and what alternatives are possible and which ones aren't.
    – user66554
    Commented Apr 12, 2015 at 11:22
  • @user66554 All progressions can exist. It's that simple. In your question you said actual music tends to be strongly restricted both to a limited set of chords and to a limited set of orderings of those but I don't think that's really true at all - there's a lot of variety out there. However, there are certainly some progressions that are much more common than others. Have you read about functional harmony as I suggested in the answer? If so you will have one perspective to understand a lot of these common progressions from. Commented Apr 12, 2015 at 11:37
  • That they can exist is exactly what I mean by From musical theory I know of no reason not to use any of the chords of the key in any order. Functional harmony is a set of restrictions, with some ad hoc explanations for itself but none for harmony overall.
    – user66554
    Commented Apr 12, 2015 at 12:18
  • The only parts of musical theory that are relevant to all harmony are the mathematical things I mentioned - frequency relationships of notes to other notes - and the physiology of the ear. understanding that will certainly help you understand why IV and V are important chords, for example. It's just that to fill in all the details of why people make the choices they do, you need to add the cultural/stylistic knowledge to the more objective mathematical stuff. Commented Apr 12, 2015 at 12:57
  • Ha, I hadn't actually seen you linked to the hook theory as well before I answered myself. It is a great tool.
    – Pif
    Commented Apr 14, 2015 at 7:25

In answer to your penultimate question - Probably the most common chord change is up a fourth, as in Am - Dm, or C - F. It occurs in just about any tune in existence in the Western world. It was found to be something that worked, and sounded good and natural, so it gets used. Any diatonic chord can and does follow any other, without sounding 'out of tune' to most people, musos or not. Put any other chord in a sequence, and folks could easily say it sounds out of tune. Not that it will be, but it's not the expected change, which can be any out of the 7 usuals.

Trying to answer your final question, it helps if the melody contains notes from the underlying harmony, and vice versa. So if a tune takes a turn away from the diatonic notes, but not necessarily, another chord could fit under. Let's say a song in C has a long note, for a bar, that's B. A B major chord, or in fact any chord containing a B may well fit. That's a big choice - C#7 or Abm are two 'odd' options, but would work or not depending on the chords either side.

Todd's answer is good, and I agree with his recommended reading.


A useful way to look at chord progressions is to treat them as multiple melodic lines played or sung on top of each other.

For example, let's say you have the progression C-Am-Dm-G and you have 3 singers to render it. As displayed below, the first singer (or first finger on a piano) can sing G-A-A-B and the second E-E-F-G and the third C-C-D-D. This visualization can help to see the three melodic lines:

    B                B
    A        A   A 
    G    G           G
    F            F 
    E    E   E 
    D            D   D
    C    C   C

Even just visually, one can see the three voices progressing gradually, with small changes. For example, progressing from C (C-E-G) to Am (A-C-E) can be done by changing just one note, G into A.

In the end, what matters is that something sounds good to you. But if you're composing, and you're not sure where to go next, or if you need to improve a particular change, try to look at the chords as the progression of a set of individual voices, and you'll probably get some useful insights.


Topo gave an excellent answer. I think simply put, your choice of chords (and therefore, your ultimate choice of harmonic progression and overall "harmony" of a bar/phrase/section/piece) should be used in a way that is stylistic (in other words "this is the sound I like/want") but also have a goal.

Remember that when composing, it is not just about how one chord sounds compared to the chord right before or after it. It is about having a goal with harmonic progression, and deciding (composing) on how you want to reach that goal. What goal you set, and how you reach that goal, is the framework for movement within a piece.

For an example: You want to modulate to the dominant key. There are the standard ways to do this (let us say something simple like I-vi-II7-V(which is now the I in your dominant key due to raising the third note of the supertonic). However, you can incorporate other chords, even non-diatonic chords (hello diminished fifth!) and other chromatic alterations within that I-vi-II7-V(now I in dominant key) progression.

By having a goal (tonic modulation to dominant), you can now choose a variety of ways to reach that goal, whether it be simple as described above, or venturing off into more elaborate harmonies before reaching your goal. Of course this is an extremely simple way of explaining this.

You should analyze some pieces in the traditional classical music repertoire. Assuming that you are not very comfortable with classical music, you should find some later Beethoven piano pieces. If you are comfortable with classical music, try analyzing some late romantic or even modern pieces, such as Rachmaninov.

Hope this helps.


Tricky question.

From musical theory I know of no reason not to use any of the chords of the key in any order. Yet actual music tends to be strongly restricted both to a limited set of chords and to a limited set of orderings of those such that it is suprising every time if there are more of them used.

A chord progression makes sense if it works out for you or for most listeners, depending on what you mean.

You're right that you have on the one hand the innumerable possible chord combinations, and on the other hand the ones that are actually used.

Music theory will tell you how tension and release work, mostly of tension can resolve, or falsely resolve, and what tensions are used in a given context (spoiler : it's always the tritone. The tritone is the music's butler that did it), and style will give you a context. If you were to look at all chord progressions for a given style, I guess it would look like a nebula, very dense around standard progressions, with less frequent variations around it, and even less secondary variations. The frequency of chord progressions and chord usage is not random, it's a culture, it's what we're more or less familiar with. A bit like in a language, some words are only found together in an expression, which might not make that much sense for us in itself, but is familiar and internally consistent, and they would never commonly be used out of this context. Nothing prevents you from taking a bit of one expression and trying it with a word it's never been combined with. It might or might not work, or make sense. There's no absolute reason why it's the bee's knees and not the bee's ankle, or the bee's peas, but you can rely on people expecting it to be the knees.

As an example : a turnaround originates in the classic candenza in a blusey context, and there is a tradition of turnarounds, and patterns to play over, and you could eschew expectations or reward the listener with familiarity at any point, until it's an atonal mess.

Or take the Giant Steps grid. It doesn't make any sense to me on paper, but press play and it works.


Classical harmony revolves around the concept that the music is leading somewhere. So, very generally, if you can enhance the sense that the music is going somewhere, it will work well. The simplest way to get a strong sense of direction is using the progression V-I.

You can use the trick for other chords as well. If you know you want one chord at a particular place, try putting the corresponding V or v chord before it. This gives rise to progressions like II-V or III-vi, or even I7-IV, which are all technically outside of the mode, but sometimes work well.

This is of course just one of many principles you can use. Every case must also be judged aesthetically.


Without going into a more technically detailed answer than I'm equipped to give, I think what you're asking about is "voice leading". Even though we don't typically think of them that way, the idea behind a lot of chordal music, at least for a large chunk of western history, was for the chords to be able to be analyzed as a set of individual instrumental voices running concurrently, interplaying with each other. The very, very abbreviated answer is the more consistently and logically your chord sequence can be analyzed like that, the "better" it sounds.

Now, a LOT of knowledgable people might want to yell and scream at this answer, and they'd be right. Go listen to John Coltrane's "Giant Steps" to see a perfect example of how you can make historically great art—almost magic, really, if you ask me—by pointedly ignoring what I just said.

But in a lot of cases, it holds. Maybe google some information about voice leading in modern music and see if it isn't the technical point that underlies what you're asking about.

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.