Approaching Nonfunctional Harmony
In 1993, I began developing a harmonic analysis course focusing mainly on the unique melodic, harmonic, and rhythmic characteristics of music by contemporary jazz and fusion artists like the Yellowjackets, Wayne Shorter, Pat Metheny, Chick Corea, and others. Music by these composers possesses some harmonic characteristics that are very different from those found in standard tunes. The criteria used to analyze standard tunes were inadequate to analyze the music of the artists mentioned above. Through teaching, a new set of criteria began to emerge that helped me to understand and ultimately to write music in the style of these composers.
As my students learned to use these approaches, their compositions began to sound similar in nature to those by the composers we were studying. The challenge for the students was not so much deciding which techniques to use to create the sound we were after, but rather which ones to avoid. Functional harmonic patterns (like II-V-I patterns, dominant to tonic resolutions, circle of fifths sequences, and line clichés) contribute to harmonic predictability and set up expectations in standard tunes. These are not found as much in contemporary jazz composition and became designated as elements we would avoid.
For this article, I have written a tune using some nonfunctional approaches. The chord changes and melody are not built around an established key center. This means that there is little or no feeling of a tonic center throughout the tune.
The harmonic foundation of the piece grows out of two eight-note series. They are used somewhat like a 12-tone row in classical music, but their use is not as strict. Example one is the series that I used to derive the root motion of the chords in bars 1-11. Example two maps out the root motion for the chords in bars 12-22. Examples three and four show the chord qualities of the structures built over the bass notes in examples one and two. The criteria used to choose each chord quality are based on what I call "adjacent chord relationships." What precedes and follows any chord determines that chord's color. The last two examples (five and six) contain the primary melodic target notes that form the foundation of the melody.
Once I had written the series, the next step was to shape it into a tune by coming up with melodic and harmonic rhythms. The chord changes in the coda (also nonfunctional) are a series of constant-structure major seventh chords.
For developing the melody, I used a technique that I have seen in the music of many contemporary composers. A melodic fragment is built on a given pitch. When that fragment reappears in later bars, the chord has changed and so the melody note's relation to the chord is changed too. I call it "changing melodic function." In measure one, the G note is heard as the seventh of the A-flat chord. A similar phrase appears in bar three, but this time the G note is heard as the ninth on an F minor chord. In measures five and six, the C-sharp is the sixth of the E major chord and becomes the major seventh when played against the D major chord in bar six. Similar melodic function changes happen throughout the tune.
The melody notes that occur on beat four in measures 23, 25, and 26 appear to be incorrect academically speaking. We have all been taught not to harmonize the lowered seventh with a major seventh chord. What makes it work in this case is the horizontal melodic sequence that is established. The horizontal motion created by the melodic sequence is strong enough to cancel out the vertical melody-to-harmony relationship that would make these notes be perceived as incorrect.
I hope these methods for finding unpredictable harmonic and melodic sounds might lead you to some discoveries in your writing. To learn more about this topic, my book Beyond Functional Harmony , gives an in-depth treatment of the subject.