structural tones of melody

That's a good question. I have a feeling you're not going to like the answer :)

It's really a meat-grinder problem. You can put a cow into a meat grinder, look at the result and say "that's my cow" after you're done. It's a lot harder to start with ground up meat and come to the same conclusion. Compositions are cows in this analogy; structural tones are ground beef.

I'm aware of exactly two composition books that talk about "structural tones" - the Complete Idiot's Guide, and (oddly enough) Music Composition for Dummies. It's not that they invented the term; it's just not part of the vocabulary of composition... it's borrowed from music analysis.

In music analysis, you take some piece that's already been composed and try to dissect what happened, and why it works. An orchestral score has a lot of instruments playing a lot of notes, so the first step in understanding what's going on is called reduction - separating the key elements from the less important stuff that fills out the piece.

A theorist named Heinrich Schenker came up with the idea that some tones are structural, and others aren't. Schenker's work is used in lots of conservatories today, because it's seen as an improvement on the older model, which analyzed tones as being harmonic or not... in Schenkerian analysis, a tone may be non-structural (which he called a prolongation) without being non-harmonic. Simply using a chord tone doesn't guarantee you'll end up with a structural tone.

For a composer, the idea of structural tones presents three problems: first, you don't know what they are in advance - because the concept is analytical, you don't know what they are until after you've cast your melody. Second, because a prolongation tone can still be harmonic, there's no easy way to see what they are. And third, because it's not really a compositional model, you can't take the application of structural tones in one piece and duplicate the concept in another (although you can do the reverse, after something has been composed).

Since structural tones aren't a standard compositional tool, I'll try to work backwards from the tonal analysis point of view and give you a framework to use in composition.

What they are:

Structural tones will appear on accented beats. If you're in 4/4, you'll want to focus on beats 1 & 3. If you're in 3/4, they'll only fall on beat 1. So those are your starting points. (But you don't have to have a structural tone on every strong beat.)

Structural tones in a harmonized piece always make up part of the implied harmony. If you're working with a G chord, the tones you have available are G, B, and D - they're part of the explicit harmonization. But you may also have other tones available - if you want a tension at that point in the melody, an F note will create the implied harmony of G7. But you ignore any possibilities outside the implied chord of the moment - it doesn't matter if the tone falls within the key, but it must be part of the implied harmony at the moment. (For example, you could use F# if you want to imply Gmaj7)

Structural tones always give a sense of being a 'destination'. This doesn't mean they're a resting point in the melody; it means you can't do without them. That F note (or F#) must be critical in resolving to an E (or G) note that follows.

Not much help, I know. You've ended up with a bunch of half notes, whole notes, or dotted halves, and they don't say much melodically. But from Schenker you can find a few tricks for...

Applications in composition

Appearing on an accented beat doesn't mean they start on the beat. You can shift a tone forward in time, starting on beat 4 and carrying it through beat 1. (You could also work the other way; instead of putting the structural tone on beat 3, you could put it on the 'and' after 3, and keep your listener waiting).

You can move from one structural tone to another in three basic ways:

- you can leave the structural tone by half-step or whole step, and return to it. That's called a neighbor note.

- you can move along the chord of the moment through an arpeggio. If you have a G chord with G as a structural tone, you can play G-B-D-G-D-G. Although these are ALL harmonic tones, only the G is melodically important (this is where Schenker really deviates from what came before him).

- you can move from one structural tone to another in a direct line of any makeup. If you have G as a structural tone, and the next one is E, you can go G-B-C-D-D#-E, or G-B-A-C#-B-D#-E, or any other plan that makes sense (the first is an ascending line, the second is made up of ascending thirds - both include notes outside the basic scale, but you don't have to do that - I only put them in the examples to show you're not limited to diatonic tones). Schenker would consider all the stuff in between 'passing tones'; in analysis before Schenker, passing tones are those that occur directly between chord tones, so you'd get somewhat different labels for the in-between stuff.

OR - and I think this is probably the best use of Schenker for a composer - you can first cast a melody without worrying about what's structural, then apply a Schenkerian view to figure out what's important in what you wrote, and then you can make a new melody that keeps the same structural tones in the same rhythmic places. Schenker would consider the two melodies inflections of each other (but here again you have to be careful in talking about what tools you used; there's a different set of compositional tools that are also called 'inflections'!)

Hope that helps a bit. And I've simplified Schenker's views (a lot!). If you want to know about it in more depth, the best start would probably be Felix Salzer's book "Structural Hearing". Salzer explains Schenker's approach, and extends it historically, to music before and after the "common practice" period. It's not a beginner's read, but if you've got a bit of classical theory it's a good book.