33

So, I've been playing guitar for about a year, and I want to learn some chord theory. I crack open a book, I learn about I ii iii IV V vi viio and why it works and how to construct it, then I learn how to apply it to I IV V being a simple easy medley and I V vi IV being the go-to pop progression. It all makes sense on paper. And the book is very adamant about seeing these chords all around me and having that realization like Neo in The Matrix that everything is built around these chord structures.

The problem is that every time I try to be clever and apply it to a song I know, these rules don't help. I can give some examples:

  • Knockin' on Heaven's Door is the most basic-sounding song I can think of. It goes G D C, G D Am. Four chords, I would expect it to fit somewhere into I V vi IV or something. But it doesn't. The closest I can get it to fit is I ii IV V, which the books have never prepared me for. I have no idea if this is the correct way to interpret it or if it's just coincidence that I got it to fit that way.

  • Jenny Jenny, the poppiest pop hit I can come up with off the top of my head, four chords, I would have put money on it being what the pop progression was built for. But it's not. I've tried numerous ways to get F♯m D A B to fit any kind of chord progression based off the above, but nothing exactly fits. By my reasoning, A B would have to be the IV V, but in that case F♯m and D don't fit.

  • Karma Police, simple chords in a simple consistent order, and I accidentally stumbled on a song that needs a manual to explain why it works.

  • Where Is My Mind, four chords, E C♯m G♯ A it's so tantalizingly close to I V vi I, but it's not. It's still off slightly, which again leads me to believe it might just be a fluke that it's so close to the system I'm trying to force it into.

Frankly, this it's disheartening to be learning all of these rules and theory without being able to apply it anywhere. I've confirmed that my understanding of the rules are correct in the books trivial nursery rhyme and lullaby examples, so I don't think the issue is with my application. I understand music can sound fantastic without following the rules, and at times sounds good because it doesn't follow the rules, but I wish I could stumble upon a song at some point that follows the rules that I've spent the last month learning. I don't feel like I'm learning anything useful.

Am I going at this the wrong way? Or am I applying the rules incorrectly?

15
  • 19
    you should definitely check out the four chord song by Axis of Awesome, you will find lots and lots of examples for I V vi IV progressions in pop
    – Chen Huang
    Jul 10, 2018 at 13:15
  • 12
    There are no rules ... only guidelines ;-)
    – Time4Tea
    Jul 10, 2018 at 13:19
  • 2
    Closely related question. Jul 10, 2018 at 13:41
  • 5
    It's like you're going to a random shelf in a bookstore and expecting to find a cookbook. The Pixies song isn't any other progression than I vi III IV. You can talk about how "close" that is to other sequences of chords, but why? Jul 10, 2018 at 15:31
  • 11
    You've stumbled onto one of the keys of creative success, whether it's music, art, writing, or anything else: First, learn all the rules and conventions of your field and how to work within them. Then, learn how to break them and do something new and unexpected. Jul 10, 2018 at 15:34

11 Answers 11

29

As ggcg said, those rules are mainly the rules of western common-practice harmony, i.e. the harmony that was used in baroque, classical and early romantic music.

It has, for sure, also had a lot of influence on pop music, but it also has a lot of differences. A crucial one is what I would sum up with the following statement, which may sound a bit crazy:

Classical music has no chords.

That is, chords in the sense of lead-sheet chords, where the composition basically specifies which pitch-classes you're allowed to use at any given time, but doesn't really specify any particular voicing.

Instead, classical music is all about voices. Each voice has a melodic role by itself, and many of the common-practice “harmony” rules are actually rather corollaries of melodic rules applying to the single voices.

This aspect just doesn't make sense for most rock/pop music, because the harmonies aren't usually layed out in a melodic way at all, instead you typically have some chord which is strummed or arpeggiated in some rhythmic pattern, then a switch to another chord with a similar pattern but no clear connection which notes in the former chord lead on to a particular note in the next one. Thus a large number of the classical rules don't apply.

In particular, the rule that a dominant should resolve to the tonic (or its relative) is most of all a rule about the leading tones in the dominant. These are notes that, if you think about them melodically, strongly “demand” making a particular step. The ⅶ note (third of the 7) strongly wants to go upwards a step, whereas the ⅳ note (seventh of the 7) wants to go down. Well, if you fulfill that, there's only so much harmonies you can end up on.

But when you don't even consider the notes in the chords as belonging to individual voices that would need to be continued, then this whole issue doesn't arise. In that light it's not that remarkable to find - in a song like Knocking on Heaven's Door.

Voice leading still is relevant for lines that actually are treated as melodies. In rock/pop this is usually most of the time a single voice, which has much more freedom; not much rules to be found here (they aren't needed because the soloist can just improvise whatever they feel fitting; this gives a nice lot of expressive freedom). It is often at least one part of the chord progression too, though, Knocking on Heaven's Door happens to a strong example: the guitar has quite prominently the notes G-F♯-E on top, which in many version is also doubled by choir or a lead instrument:

X:1
L:1/2
M:C
K:G
%%score T1 T2 A B
V:T1           clef=treble-8
% 1
[V:T1] [G,B,DGBg] [DAdf] | [A,EAce]2  | [G,B,DGBg] [DAdf] | [CEGce]2

Common-practice would have you think that the F♯ should lead rigorously upwards, but here it doesn't. Why not? Well, it certainly helps that the D is not played as a dominant-seventh chord but just as a major chord, which doesn't include the tritone between F♯ and C (the dissonance which is to quite some part responsible for the resolution-demand). Then again, you regularly find 7-7 in blues, so that doesn't seem to be the make-or-break either (rather, this song's very simplistic folkyness just eschews all chords that aren't simple triads).

Rather, I would suggest that it has most to do with the intended mood/expression. Common-practice music is very “forward-idealistic” – in baroque it's the drive towards perfection and god (or, for a more down-to-earth explanation, pleasing the absolute king), in classical and romantic music the yearning for some goal that may be reachable (major-resolution-happy-nice) or not (desparate buildup towards the inevitable disaster). Not so blues and descended genres: here resignation is a main theme. On the risk of over-interpreting, one might say it's the subversion of European ideas into a failed version of the American Dream: “there is a clear path upwards, but we don't get to actually take it”. Knocking on Heaven's Door is not bluesy, but it too very strongly expresses such a resignation mood.

And that's just one example of a wider theme, which is perhaps the best reason why you should care about theory: theory tells you what mood you will cause when using such and such progressions. What effect you want is your own decision, but if you know what then a firm grasp on theory makes it much easier to figure out how to actually achieve it.


I'm sure somebody will remark at this spot “but what about figured bass?” Isn't figured bass essentially the same thing as a lead sheet? — Indeed it kind of is, however unlike with pop lead sheets a baroque continuo player is expected to actually infer the intended voicing, not just the pitch classes it contains.

Classical music also has “chord arpeggiation patterns”, such as Alberti bass. But they're not really the default, more of an extra embellishment (and relying too much on such devices would be considered somewhat lazy).

6
  • 3
    Just to comment that figured bass might be "essentially" the same as lead sheet (except you get the bass line written down but not the tune!) - but the baroque concept of "a chord" was very different from the modern one. For example there was no consideration of which scale degree a chord was on, and no concept of "Inversions" - i.e. for a baroque composer or performer C, C/E and C/G were three "completely different" chords, not three inversions of the same C major chord.
    – user19146
    Jul 10, 2018 at 14:28
  • 1
    Thank you for this explanation about how chords (or something like them) in classical music really differ from the concept of chords in rock/pop music. It helps me understand why the square peg doesn't fit so well in the round hole.
    – LarsH
    Jul 10, 2018 at 19:36
  • To be fair, a large percentage of "classical" music doesn't follow the rules either. That may have been the case back in the Baroque era of J.S.Bach, but as early as his sons, C.P.E. and J.C.Bach, you can already see the rules being stretched and broken all the time. By the time you get up to Beethoven, then Chopin, Debussey, etc., it's all basically out the window. The rules are a starting point, but they are by no means universally applied, particularly in modern day music. Jul 12, 2018 at 18:11
  • @DarrelHoffman well, it was J.S.Bach who invented many of the commonly accepted rules in the first place. Beethoven, Chopin and Debussy simply added extra rules, which would occasionally override the older ones. (Actually many of Bach's rules also aren't in the standard books.) Jul 12, 2018 at 19:17
  • 1
    @TasosPapastylianou music.meta.stackexchange.com/questions/2782/… Jul 13, 2018 at 9:14
15

You are going the right and wrong way!

The right way is to be aware of what works - the theory.

The wrong way is to expect this theory to be spelled out in each and every song.

Firstly, having I, ii, iii, IV, V, vi, viio available will work (does work) in a lot of songs. Heaven's Door is one. I, V, IV, I, V, ii. Nothing wrong with that. All the chords are diatonic (from the key). The order they come in is different - that's the way the song is!

Secondly, there's no rule which says only chords from that bank can be used. In fact, there are really no rules at all.

Apart from if it sounds good, it probably is. The mistake you (and countless others) have made is believing that the theory you found out is far reaching. To a degree it is - things that work and keep on working get noted (sic) and because they're known to work, they'll keep on recurring in usage. That does not preclude any other ideas that may or may not work in other situations.

It's then the job of theorists to come up with reasons why the other ideas work. Note - ideas, not rules. When something can be explained in a way that people understand, it makes it easier to put the idea into use - almost like giving it credence or acceptable accessibility.

One other 'theory' you need next, which will put most of your confusion at bay, is that chords from parallel keys can be used. (Heck, any chords, anywhere can be used, but this narrows it down and makes it acceptable!). So, in, say, C, any chords from Cm may also turn up. And bear in mind that Cm has more than just the 7 chords available...

Another is that modes from the same root will spawn other usable chords.

You might want to look at RNA - Roman Numeral Analysis - which makes more sense, theory-wise, of what chords are actually doing. It'll introduce you to secondary dominants, backdoor dominants, TTS and a whole new world of reasons why stuff works. It's actually quite interesting!!

8
  • 1
    If it makes you more comfortable... I downvoted because this answer doesn't really give any advice. Theory is not just a list of things that can be done – clearly the list of things that can be done is actually unbounded. Rather it is a list of a) building blocks that tell you what even to start with, b) a guideline as to which elements will achieve which effect. That's perhaps the most crucial, and particularly when you don't want to settle with just “making it sound good” fleetingly, but actually give it some depth/meaning/context. Also I'd argue that ideas are rules. Jul 10, 2018 at 21:11
  • 1
    I hope you don't mind that I bolded "there are really no rules at all", because I think that's a critical bit of information the asker should get from these answers, and you have it stated best in yours, so I wanted to emphasize it. I have a history of feeling lukewarm about a lot of your answers, but this one I think is right on the money. I also don't quite understand the downvotes. I don't feel like a good answer to this question should necessarily have advice. Jul 10, 2018 at 21:54
  • @leftaroundabout - firstly, I do appreciate honesty. Secondly, ideas cannot be rules - they may be a starting point for rules to be built on, there again, it's music theory, not rules. Thirdly, there is advice on what to look at next in theory. Fourthly, if misinformation is contained in an answer, it's a good reason to downvote.
    – Tim
    Jul 11, 2018 at 7:18
  • 1
    @ToddWilcox the thing is, even the most anarchic punk or wild jazz very much does follow rules. If you say “there really are no rules” then what you probably mean is “stick to a uniform time-grid of bars which are subdivided into low powers of 2 and 3, use any pitches from the 12-edo tuning system, and find a progression over a few bars that repeats a couple of times”. Which is on one hand quite permissive, but on the other hand already precludes a giant chunk of ideas one could have (as well as, perhaps ironically, the original Blues tunes). And pop clearly follows much tighter rules. Jul 11, 2018 at 11:09
  • 1
    @ToddWilcox I don't suppose you mean that punk was actually deliberately composed on any non-regular grid, it was merely played with strong deviations. Whereas dubstep tends to be build around a very exact grid, but then adds crazy wobbles to the mix. In either case, the effect of the music is very much dominated by the breaking of the standard rules. But that in itself is a rule too. Jul 11, 2018 at 12:28
13

"Play by the rules". Western music theory is, first of all, a description of the music of western Europe. This is not the only music out there. These "rules" came about after centuries of ethnic musical development and later analysis of what is popular and in some sense common among the various musical trends. While it is true that certain patterns are more common than other and that some patterns are generally considered "not good", the fact is that there are no rules that forbid certain patterns, chord progressions, melody lines, etc. What we have is a set of guidelines that describe how "acceptable" patterns work and what patterns are generally pleasing to the west European ear. Time changes things, exposure to other cultures changes things (for the better, hopefully). My classical harmony book pretty much cites that any chord can proceed any other chord (I'm embellishing slightly). The pattern emerges when there is a cycle, a pattern that repeats. Of course it strongly enforces the V-->I, V7-->, IV-->I, etc, as endings to a cycle but there are examples of ii-->I, etc.

Don't fall into the trap that if you can't find a progression in a text book that it is somehow not allowed. The laws of physics are seldom broken, but these "laws" of music theory are not the same. They do describe why certain patterns work better than others but they don't forbid anything.

Songs modulate key, they move to their relative minor (which is introduces a major chord on the iii to create resolution, if you need a theoretical explanation). A common cadence is IV-->iv (minor 4) -->I, which clearly brings in an "out of key" note.

Two things may emerge from your studies.

  1. As you learn more you will find "reasons", or "explanations" that do justify the progressions that aren't playing by the rules. In other words, you haven't learned all the rules yet.

  2. You may decide that the rules are interesting in their own right but not reason to allow or forbid anything. In my case, after decades of studying music from many different cultures I began to not like the term "accidental". For me, no note is an accident unless it doesn't match the note you intended to play.

Keep studying music theory, it is a wonderful discipline, but keep it in context.

2
  • 2
    I downvoted this because it ignored some important points the OP made, primarily that they already understand that songs don't necessarily need to follow the rules and don't think the rules are broken. Their question was "why does this thing the book told me to do not work?" to which this is perilously close to "it doesn't, don't do it." Jul 10, 2018 at 21:52
  • I believe I did address the question. You are entitled to your opinion.
    – user50691
    Jul 10, 2018 at 22:28
6

If you analyze classical music in terms of chords, you find that the most common progressions are the authentic cadence V-I and its expanded version IV-V-I. So naturally, these cadences are the center piece of classical music theory.

Some pop music and much of jazz can be analyzed from this starting point. However, there is an important musical genre which is decidedly different harmonically speaking: the Blues. If you look at the standard 12-bar progression || I I I I | IV IV I I | V IV I V || you see, that most transitions to the tonic are plagal, i.e. IV-I.

Since the blues had a big impact upon pop music, it is a better starting point for understanding what's going on in many songs. Let's have a look at your songs (I don't know Jenny Jenny and couldn't find it, so no comment on this song).

Knocking on Heaven's Door
| I V IIm IIm | I V IV IV |
-> IIm has subdominant chord quality and can replace the subdominant IV. So it's two times the expanded plagal cadence I-V-IV-I which is part of the Blues progression.

Where Is My Mind
| I VIm III IV |
-> I, VIm and III all have tonic chord quality, so there's little harmonic movement until the last chord - which is the IV, so we have the plagal cadence again. III isn't actually part of the major scale - it "should" be IIIm. Besides the use of plagal progressions, this is a second principle which is inspired by the Blues: use major chords instead of minor chords. (The most common scale for Blues melody is the Blues scale. It consists of the minor pentatonic with the addition of the so-called blue notes. Yet all the chords in the Blues progression are major chords. So the Blues transcends the speparation of minor and major.)

Karma police
This song includes a lot of interesting harmony. I haven't listened to it for long, so I give only a quick and error-prone sketch of an analysis. The key is A minor in the verse and G major in the chorus. The verse uses bVII-Im (G-Am) and IV-Im (D-Am) cadences to return to the tonic. bVII has subdominant chord quality like the IV. So both are our plagal cadences again. The minor scale "should" have a IVm chord, so again a major chord replaces a minor chord. The chorus features classical harmony with a IV-V-I progression (C-D-G). A lot more could be said, but the two principles of plagal progressions and using major chords instead of minor chords already give a pretty good picture about what's happening in the verse. (If you want to dig deeper, have a look at modes. Using the mode terminology, the verse of Karma police oscillates between A dorian and A aeolian.)

So to sum it up: Classical pieces, some pop songs and most of jazz progress harmonically mostly by (chains of) V-I and IV-V-I with various substitution chords of similar tonal quality. In other pop songs and in much of rock the Blues has left strong traces: they progress harmonically mostly by (chains of) IV-I and V-IV-I with various substitution chords of similar tonal quality. There's a tendency to replace the minor chords of the scale by major chords and to transcend the major-minor dualism.

Let's have a look at a final example to illustrate this.

Hey Joe
| C G | D A | E E | E E |
| bVI bIII | bVII IV | I I | I I |
The interval between the root notes of two subsequent chords is in all cases a perfect fourth (except for the transition from the last chord to the first chord again). So this is a chain of plagal IV-I progressions. If we are looking for a key, the root notes of the chords suggest the E minor scale. Yet all chords, including the tonic, are major chords. Jimi Hendrix plays mostly the E (minor) blues scale.

3
  • In A minor, A is 'i', and G is 'VII'. Also, there exists a major blues scale.
    – Tim
    Jul 11, 2018 at 11:44
  • Thanks for the input. I corrected the tonic chord in Karma Police. The bVII, however, was intentional. It depends on the convention one uses. Did you mention the major blues scale because of completeness or because you think I don't give it the prominence it deserves? I think it isn't used very often in pop and rock music.
    – Marc
    Jul 11, 2018 at 15:36
  • If you're looking for Jenny Jenny, you could always call her: 867-5309
    – Anthony
    Jul 11, 2018 at 21:35
4

Jenny Jenny is in A major. A is the relative major, F#m is the relative minor, D is the IV chord. Where it gets trickier is the B chord, which consists of B, D# and F#. B and F# appear in your A major scale, but D#...?

If you're restricting your set of notes to only those in your A scale, then a II chord must be minor, because we only have access to B, D and F#. That's the theory you've learnt so far, and it's perfectly correct - if you restrict yourself to only those notes. However if you add a blue note then your A major scale A, B, C#, D, E, F#, G#, A becomes A, B, C#, D, D#, E, F#, G#, A. By expanding the notes in our scale, we now have access to a B chord as well, and away we go. Often you'll find that blue notes are used as passing notes which are resolved by the next chord playing an adjacent note, and that's what we get here.

This appears in other songs as well. The chords for the first line of House of the Rising Sun are Am, C, D, F - and that D chord (D, F#, A) is using the F# which is a blue note in the key of C. Again it's a passing note from the G in the C chord, to the F# in the D chord, to the F tonic in the F chord.

1
  • That D chord could also easily be explained by the 'use the parallel key chords' theory. It belongs to key A major.
    – Tim
    Dec 7, 2022 at 11:20
1

Personally, I think Knockin' on Heaven's Door is supposed to be bittersweet. G, D, C = I-V-IV (sweet). G, D, Am = I-V-ii (hint of the bitter). It sounds simple, because of the lack of notes within the triad falling outside the key of G. Harmony is a journey - following a key prevents uneasy feelings (discord), but this too can be a useful tool for expression. Good luck on your musical journey.

1
  • Welcome to the Music Stack exchange. Please check out the tour and read the help center page. Jul 11, 2018 at 14:55
1

This question is adressed by Théodore Dubois in his famous Traité d'harmonie (1921). Let me freely translate a quote from that book:

Students will notice that the masters do not always follow these rules and actually quite liberally ignore them. When they do ignore them, they do this as accomplished artists, sure in their own aesthetic judgements. Until their artistic taste has fully maturated, students should fully observe these rules and consider them as safeguards against the most obvious blunders.

(I am quoting from memory, as I was unable to find quickly the reference again.)

Even if the context is slightly different, I would like to draw a parallel. The “rules” you are referring to are actually a systematic description of what seems to be considered aesthetically pleasing. It does imply that following these rules should result in a pleasing composition, but not that any pleasing composition must fit in these rules!

1

Short answer:

There aren't any rules (none), but there are some arrangements which have gained popularity and perhpas become conventional.

Example: 12-bar Blues is a straightforward progression, very popular, but it still only covers a tiny percentage of the music written. It establishes a convention, which other musicians then delight in breaking :-)

I'd say the book is wrong to call these common progressions "rules" (if it uses that terminology). That'd be like saying everyone has to wear green because a lot of people do.

1

I've confirmed that my understanding of the rules are correct in the books trivial nursery rhyme and lullaby examples, so I don't think the issue is with my application. I understand music can sound fantastic without following the rules, and at times sounds good because it doesn't follow the rules, but I wish I could stumble upon a song at some point that follows the rules that I've spent the last month learning.

You are learning the rules to the language, but there are exceptions and edge cases. There are different dialects, and slang. This is like learning about grammar even though you have been speaking for years before getting to the 2nd grade. You know what a noun is, you know what a verb is but you haven't learned about adjectives yet. You use them, but you haven't learned what they are called or the formal "rules" around them. You can't pick up Shakespeare and analyse it yet because you haven't learned enough rules, and sometimes he doesn't follow the rules, or the "rules" have changed.

The rules describe what has been done before, and have been done a lot. Not every creative decision is going to be put into the rules. Some things are common enough to get a rule, somethings are just creative choices. Don't try to break down every song you hear based on what you have learned so far. keep learning and try to hear the things you have learned in the songs you hear. Oh, a I IV V progression, I get that...huh they went to bVII (flat/minor VII) I haven't learning about why that works yet....let's keep learning, maybe use what you hear to search for why something works.

Stick with it, it is worth the effort.

0

With most music theory you learn, it probably isn't helpful to think in terms of 'rules' (at least when you are thinking about music in general). The only things that are really 'rules' in music are the things that have one foot in physics and psychoacoustics - the way that the ear hears pitches an octave apart as equivalent, or the way the harmonic series comes about, for example.

Most other things you will learn about in music theory are patterns that come about partly because of those deeper physical/psychoacoustic truths, but also simply because of the preferences and cultural experiences of groups of individuals. There are many different groups of people who have made many different types of patterns (in different styles of music), and therefore you have to learn a lot of music theory before you actually learn enough to describe most patterns that occur in the variety of music that's out there.

So you are right - learning just a little music theory from a general perspective can indeed be a waste of time, because the knowledge you gain may be too thin to do anything useful with.

Possible approaches you could take are:

  • Learn a lot more theory
  • Try to focus on the theory that relates to particular types of music you are interested in
  • Not worry too much about "someone else's theory", but also allow yourself to come up with your own mental model of how music works.

There are pros and cons to all these approaches, and ultimately the solution is probably to find your own balance between these.

10
  • 3
    "Learn a lot more theory" - yes, but the OP might have another 50 years to do that, so it's not the top priority, which IMO is listen to a lot more music, until it starts to make sense. That's how you learned your native spoken language, so you've done it once already!
    – user19146
    Jul 10, 2018 at 18:51
  • 2
    @Dom "Stopping after an intro class of any subject will yield poor results. You can't expect people to be able to understand how to apply a topic if they don't progress out of the basics" - I think we're pretty much in agreement TBH! I'm just trying to make the point that someone who has just done a first book or course in music theory is still at that beginning stage - that's something that might be obvious to some, not so much to others. And of course on realising that there's much more to learn, you have to do the cost/benefit analysis on what the best use of your time is. Jul 10, 2018 at 21:01
  • 1
    Also the word "waste of time" is rather strong. Being able to understand Roman Numerals is pretty significant for communication between musicians that know it.
    – Dom
    Jul 10, 2018 at 21:18
  • 1
    @Dom "Any subject will take some time to get spun up on " - agree totally; that's what I'm saying here. I'm just 'replying in kind' to the tone of the OP; sometimes introductory music theory isn't presented as introductory, but as fundamental, which can lead to frustrations when it doesn't tie up with 'real life'. Jul 10, 2018 at 21:56
  • 2
    @Dom yep, or "A little knowledge is a dangerous thing". Most subjects have some 'rules' or ideas that are quite generally-applicable (at least within certain bounds or contexts), and other ideas that are perhaps fuzzier and represent techniques or perspectives that you might use. And in every subject, I think it's often possible to be clear which is which without complicating the subject. From Tim's answer: "The mistake you (and countless others) have made is believing that the theory you found out is far reaching". That kind of mistake stems from lack of clarity about this distinction. Jul 11, 2018 at 7:38
0

All very good answers here. I would like to add that perhaps you are making things a little too complicated for yourself by applying too much theory to what are aesthetic choices by the composers of the pop tunes you are studying.

Try just using the following rules of Western music theory:

  • Scales: Each tune uses one or more scales. If a tune modulates to another scale then we call these scales key centers.
  • Chords and arpeggios: A chord is every other note of a scale played at once and an arpeggio is the chord notes played one after the other.
  • Harmonized scale: A harmonized scale takes the notes of the scale and stacks chords on each note. For example a C harmonized scale is C Dm Em F G Am B half-dim. You can use these chords in any inversion that supports the melody.

Learn the melody and then start listening where the chords fit using the harmonized scale. I usually start with finding a chord that will fit the notes of each bar and listen for unintentional tensions. Just for fun you can sometimes play the entire melody using the harmonized scale, which will give you an excellent idea of the harmonic space your melody occupies but is complete overkill for everything except Xmas carols.

Try to make the chords move in a direction and tension that supports the melody. This movement is why I-IV-V7 progressions are so common. The IV wants to resolve to the root I and the V7 really wants to resolve so it builds a good tension that is released when you resolve to the root.

All other notes and chords that aren't part of the harmonized scale of the key center are for color or for adding tension. These tensions are what make the music breathe and are the melodic equivalent of rhythmic syncopation--they create tensions that the listener wants to hear released. Sometimes they are, sometimes they aren't.

Use playing, listening and theory as the legs of the tripod that supports your playing and composing. Monopods aren't very stable. :-)

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