It's very common in jazz education to encourage a chord-oriented perspective on improvisation. In many ways, this makes a lot of sense and can be a valuable way of analyzing the music. However, it runs the risk of one failing to see the forest for the trees when applied to many tonal jazz standards, where a key-oriented approach provides a better perspective of how everything really fits together. It can also lead to more lyrical, melodic lines that can provide a contrast to lines that outline the chord changes more. Ultimately, I think both perspectives have their place. Obviously, in tunes that are less tonal, the key-center perspective becomes much less useful.
I was thinking about this today and thought it would be interesting to look at the diatonic chords in a key, and compare what notes in the key worked where, to get a global view of the relationships. The results are interesting, I think:
|Green: works |
Orange: mildly dissonant
Red: strong dissonance
What is interesting here is how nearly every note in the key "works" against nearly every chord. In particular, that scale degrees 2, 3, 5, and 6 can be used anywhere. Note that this is 4/5 of the major pentatonic scale. The 7th degree works nearly everywhere (and even the dissonance against the ii chord is fairly mild). If you take 5, 6, 7, 2, 3, then you have the major pentatonic on the 5th degree as being a pretty much failsafe option against any diatonic chord.
• Nothing conflicts with the IV chord (this is just the scale-oriented version of the observation from the standard chord-oriented perspective that the Lydian mode has no "avoid" note).
• The 1st and 4th degrees are the only ones that are seriously problematic. The first degree clashes with the dominant functioning chords (V and vii°) and the 4th degree clashes with the tonic-functioning chords (I and iii, and to a lesser extent, vi).
While I do not advocate mindlessly improvising according to general rules, I do think it is interesting and perhaps helpful to be aware of the relationships in the big picture. And of course "works" and "doesn't work" is a extremely simplistic framework for understanding relationships, one could conceivably do a much more nuanced chart of how every scale degree and chord relate to one another (and I think that every accomplished improviser does have a personal sense of this, whether or not it is conscious or articulated).
For instance, a very slightly more nuanced chart might be:
|Same as above, but add: |
Yellow: subtle tension
The fourth degree on a minor chord has a slight tension to it in normal tonal progressions, so one could further refine the chart with this information. This is moving us a bit in the direction of a chord-oriented approach. However personally useful this is, it does make the big picture less clear, which is what I meant earlier about missing the forest for the trees.
Then I thought, what about common chromatic chords? So I made the following chart:
Obviously quite a bit more red here, as one might expect in restricting melody choices to diatonic scale degrees while adding in chromatic chords: at least one note in each chord is outside the key.
Here are my observations:
• The tonic works against all of these common chromatic chords.
• II7, III7 and VI7 only have one conflict, the 4th of each chord (2, 5, and 6, respectively). This results in an interesting tetrachord that will work over all three: 1 3 4 7 (the two diatonic half-step pairs). Although a chord-oriented approach also advises against the 4th degree on a dominant chord, it is interesting to analyze these three dominants according to a chord-oriented approach to observe the differences: II7 (7, 1, 9, +9, 4, 5, 13), III7 (b13/+5, 7, 1, b9, +9, 4, 5), VI7 (+9, 4, 5, b13/+5, 7, 1, 9)—using notes from the key actually encourages the use of altered 9ths and 13ths with respect to the chordal harmony.
• bVII7, a common chord in bebop, works great with the first 5 degrees of the scale (and results in 3, 5, 9, #11, 13 against the chord).
• #iv diminished 7, a troubling chord for a lot of students, works great with the simple 1 2 4 6 7 (rearranged as 6 1 4 7 2 it becomes 3, b5, 7, 11, b13 relative to the chord)
One of the things that comes from adopting the perspective is that you start to see the key as stable, and the chords as merely temporary events that are happening in the context of the key. While it's important to be able to zero in on the notes that change (i.e., playing the "changes") it's also effective to emphasize the stable notes that aren't changing. The simple fact that the 5th degree works over every diatonic chord, for example, could be used as a unifying central element in an improvisation using many other elements.
I frequently make the observation that any note in a piece of music is simultaneously functioning with respect to the key and with respect to the current chord. That means that a chord can be stable/tense on two different levels at any given time, and they can agree or conflict (there are 4 different possibilities, obviously with many subtle gradations: key stable/chord stable, key stable/chord tense, key tense/chord stable, key tense/chord tense). It can be valuable to consider both ways that any note is working. Here is a chart with a few examples of how things can line up. In this case, I think it is helpful to consider more subtle shadings of tension, so here I only included unambiguous cases:
Note that any note that is chromatic to the key would be fairly tense—this is an issue often that confuses students when they are told things like "natural 9 is consonant on a diminished chord", which is true (chord-stable), but in many key contexts, the natural 9 of the dim. chord is extremely key-tense. For example, the common #iv°7 would have #5 of the key, a very tense note, as the 9. It's not unusable, but it requires extra care to use it musically. Each particular combination of notes/chords is different, but you can see that there would be at least 3 basic categories of tension: stable (key-stable/chord-stable), mixed tension (key-stable/chord-tense, key-tense/chord-stable) and very tense (key-tense, chord-tense).
For a practical exercise that has some relationship to all this analysis, see the following post:
This isn't really a methodical theory approach as much as it is just some thoughts and observations that I thought might be helpful. I'll be curious to hear any thoughts you might have in response to these ideas.