Suppose you're playing a melody and your ear tells you that the chord needs to be changed. Why? And then you go ahead and try every chord in that key until one of them sounds right. Great. But why does it?

Normally, the chord changes when an emphasized note (is there an official term for it?) does not "speak to" the chord you're playing. What makes a note "speak to" the chord?

Sometimes this note is the first note of the three-note chord. Other times, it's the last note of the three-note chord. And then there are times when this note isn't part of the three-note chord at all, and yet your ear tells you that it "fits" (or not: I'm always confused by this - how do you pluck out a chord that does not contain the note you're playing).

Why? Is there a technical explanation for it?

For instance, while most popular songs have three or four chords, Don McLean's American Pie has a ton. What is so different about McLean's melody?


The opening of Vissi d'arte goes like this:

enter image description here

Note that it's in D-minor. Note also that the final two notes of the second phrase, "viva!" are the same pitch: it's an E.

The chord there is A (major). Which is the dominant. It is said that the dominant leans towards the tonic and wants to be resolved.

Does this mean that if, playing in D-minor, you end a phrase in E, the chord would ALWAYS, INVARIABLY be A? And not, say, A7, or Em, for that matter?

  • 1
    Consider, for examples, barnesandnoble.com/w/… and barnesandnoble.com/w/… – Carl Witthoft Aug 7 '17 at 13:09
  • 1
    I think that Carl is trying to point out that the "rules" of music composition are too vast to include as an answer on this site. The rules change depending on the period and style. Studying the theory of the style you are composing in will be much more effective than asking someone else to explain it in its vastness on here. – Aric Aug 7 '17 at 13:22
  • 2
    The answer is not that chords "fit" because theory says they should. Chords and melody go together because they sound "right" together. Theory comes after sound; it has always been this way. – ex nihilo Aug 7 '17 at 13:50
  • 1
    @AricFowler: Well, "There isn't" is at least an answer I can sort of accept. As opposed to "Go study theory." – Ricky Aug 7 '17 at 13:51
  • 1
    @Ricky - what's a 'minor second' chord? – Tim Aug 7 '17 at 14:28

Here's one small part of an answer. I expect you're acquainted with the idea of the tonic chord (I) of a key being 'home', the dominant (V or V7) having a tension towards that home chord? Well, sometimes a melody note needs one ov those under it, sometimes another.

There are other tensions between other chords. 'Secondary dominants' if you like. A knowledge of these, and other 'tension-resolution' progressions, can help you decide on a harmony.

But that just describes ONE approach to picking chords. "Theory describes, it does not command". Theory will often help you describe (even justify) something that works, but it will rarely TELL you what to do.

If you're interested in the nuts-and-bolts of music, first get fluent in playing from notation. Read and play LOTS of music. Don't just look at the chord symbols, look at the piano part and see how the notes connect, one chord to another. You'll find that neither melody or harmony always stick to 'the chords in the key', even when the music hasn't fully modulated to another key. The choice is FAR more wide open than you might have imagined!

You don't need an ever more complicated set of theory rules instructing you what to do, you need an ever-widening experience of good-sounding things people HAVE done! Amuse yourself by working out a theoretical justification for the more weird ones if you like!

  • Thank you, this sheds some light on the issue. However, The choice is FAR more wide open than you might have imagined!" - This isn't entirely accurate, since if you look at Va Pensiero from Verdi's Nabucco (to pick a piece at random), you'll know EXACTLY which chords are to be played where; nothing particularly "wide open" about that! The American Pie, which I mentioned in the OP, has a ton of chords, and not just three - there isn't anything wide open about that either. You see what I'm getting at. – Ricky Aug 7 '17 at 14:02
  • 2
    The mark of a great composer is that once heard, you couldn't imagine the piece going ANY other way! But I could easily concoct a re-harmonisation of 'Va Pensiero' or 'American Pie', and so could you. It wouldn't be 'wrong', just unfamiliar. – Laurence Payne Aug 7 '17 at 14:09
  • 2
    @Ricky "you'll know EXACTLY which chords are to be played where" simply isn't true. I once wrote a piece based on the bass line of Pachelbel's Canon - which is just about as solidly nailed down in one key and one chord progression as any music could possibly be, right? WRONG - I repeated the bass line about 40 times without changing a single note of it, and none of the repeats had the same chord sequence except the first and last ones. Most of them weren't in D major either - many of them were what many people would describe as "atonal." But the whole thing worked fine as a piece of music. – user19146 Aug 7 '17 at 14:24
  • @Ricky, for a more prominent example, listen to the Vince Guaraldi version of "O Christmas Tree" and tell me how similar it is to a typical harmonization of that carol. – Dekkadeci Aug 8 '17 at 9:20

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