The melodic-harmonic divorce in pop

This week in pop theory class, we are talking about the melodic-harmonic divorce, where the chords and melody to a song are all from the same major or minor key, but do not necessarily agree with each other at the local level. This is a common feature of current pop. It’s so common, in fact, that my students are having a hard time hearing it. This is not due to any lack of musicality among my students; it has to do with their listening expectations.

 

In Western European tonal tradition, melodies are supposed to relate closely to their underlying chords. (More accurately, those chords are the result of counterpoint between the melody and the bass or other voices.) If you are in C major and you are writing a melody on top of a C chord, that melody should emphasize the chord tones, C, E, and G. These notes will sound stable and resolved. You can use any other notes you want, but they will feel unstable. So you use chord tones to set the listener’s expectation, use non-chord tones to create tension and suspense, and then return to the chord tones to relieve the tension.

In the early part of the 20th century, Anglo-American popular and vernacular music mostly followed the conventions of Western Europe. However, there were also styles of music that followed different conventions, most notably the blues. Jazz, R&B and rock musicians combined ideas from the blues with tonal and modal concepts from Western European tradition, resulting in melodies that sometimes don’t align with the chords in the way that Mozart would expect them to. The music-theoretic term for this “misalignment” is the melodic-harmonic divorce. That term is a problematic one, because there is nothing misaligned about the blues, but it’s the term we have.

With each passing decade, Anglo-American pop has moved further away from Western European convention, to the point where melodic-harmonic divorce is practically the defining sound of the pop mainstream. I’ll talk through some examples, and then discuss why my students might be struggling with this.

Whitney Houston, “How Will I Know”

The tune is in F major. Listen to the very first line, “There’s a boy.” Whitney sings the word “boy” on D, the sixth of the underlying F chord. This note is part of the F major scale, but it isn’t part of the F chord itself. And yet, the D is not an embellishment or a suspension; it’s the main note in this little phrase. The song’s melody mostly agrees with the chords, but that first phrase is striking.

Tracy Chapman, “Fast Car

The song is in A major. The verses are a loop of D, A, F#m and E. Given that this loop starts on D, you would expect to hear the note D at some point, but Chapman’s melody avoids it completely until the chorus. She uses every other note in the A major scale but that one! There is more overt melodic-harmonic divorce at the end of each line, because the phrases mostly end on the note A over the E chord. The A clashes directly with the G-sharp in the chord. It isn’t a suspension, either; it’s the “rest” point of the phrases. That is a lot of harmonic friction for such a quiet and gentle song.

Taylor Swift, “We Are Never Ever Getting Back Together

The song is in G, and the chords are a loop of C, G, Dsus4, Em. In the first line, Taylor sings the word “are” on D on top of the C chord. The words “never ever ever” alternate between B and A over the G chord. The B is a chord tone, but it acts like an upper neighbor to the “resolved” A, even though A isn’t a chord tone. Then Taylor sings the word “back” on A over the Em chord. 

In the second line of the chorus, Taylor sings the word “we-e” on B and D over the C chord, the major seventh and ninth. These notes are arpeggiating the tonic G chord, seemingly without regard for the C chord underneath. The words “talk to” are on G and D, also arpeggiating a G chord on top of a C chord. The notes in the rest of the chorus don’t conflict so conspicuously with the chords, but you still get the sense that Taylor wrote this melody in one room while Max Martin was writing the chord progression in another room.

Carly Rae Jepsen, “Call Me Maybe

This song is in G major, and it uses the same chord loop as “We Are Never Ever Getting Back Together” – C, G, D, Em. The first line of the verse (“I threw a wish in a well”) repeats the note B, along with an A toward the end, over the C chord. On the line “I looked to you as it fell”, the string of B’s ends on a pair of A’s sung over a brief G chord, which undermines the sense of the G as a point of resolution. Here, again, it feels like the melody and harmony were written separately and then superimposed. It sounds to me more like a remix than a traditional composition.

Sabrina Carpenter, “Espresso

The tune is in A minor. It’s a two-bar loop throughout. The first bar is Dm, and the second bar is… complicated. The guitar plays Em through it, but the bassline implies two beats of Em, and then a beat each of Am and C. The first half of the loop is where most of the divorce happens; in the verses and choruses, Carpenter sings on those Dm chords as if she’s singing on Am, always hitting E and C and always avoiding D. In the chorus, the melody over the second half of the loop outlines the Em chord and ignores the nominal Am. 

In the prechorus, the melody over the Dm chord does center around D, so that is more conventional alignment. However, the melody over the Em chord also centers around D. It’s a chord tone if you think of the chord as Em7, but not a strong one. Once again, the overall sense is that the melody and chords just don’t really matter to each other that much, beyond the fact that they both stay on the white-key pitches.

Chappell Roan, “Good Luck Babe!

The song is in D major. The verses and choruses are a loop of G, A, D, Bm. The verse begins with conventional agreement between the chords and melody; that first line literally arpeggiates the G and A chords. However, in the prechorus, the melody implies a D chord while the backing sits on an Em chord. The chorus melody similarly accents chord tones from D, regardless of which chord is going by underneath.

David Temperley’s paper, “The melodic-harmonic ‘divorce’ in rock“, is a good reference for this concept. The scare quotes are important; Temperley is reluctant to treat Western European tonal convention as normative, and to see rock melody as deviating from the rules. He observes that rock songs tend to have more divorce in their verses, and more conventional alignment in the choruses. This is called the “loose-verse/tight-chorus” model. Temperley thinks that in this model, the independence of melody and harmony in the verses expresses individuality, while the more unified melody and harmony in the choruses expresses coming together, coordination, group consciousness. I have not systematically surveyed current pop songs, but I have noticed that they don’t follow the loose-verse/tight-chorus model. Sometimes it’s the opposite! I don’t know how broadly this trend applies, or what it means, but it’s striking.

My students have a hard time understanding the idea of melodic-harmonic divorce. They have grown up in a world where the  divorce is so ubiquitous that they don’t hear anything remarkable about it. They just don’t approach music with the expectation that it will function tonally. They intellectually understand why, say, an accented C in a melody is a conflict with an underlying G chord, but they don’t actually feel that as a conflict intuitively. It’s like the way that I feel about the “rule” that V7 chords resolve to I, I just don’t feel the pull. But I do feel a C on top of a G chord as a tension that should resolve, and my students seem like they genuinely don’t. In composition assignments, they will use C as a bass note for a G chord, and when I tell them they shouldn’t do that, they genuinely don’t see why, since it sounds fine to them.

I’m trying to follow my own injunctions to be descriptive rather than prescriptive and say, okay, if you are writing for an audience of young people, feel free to use all the melodic-harmonic divorce you want, but if you are writing for old people, keep things functional. Or maybe the guideline should go, if you are writing for young people, make sure not to have too much agreement between the chords and the melody or it will sound dated.

The melodic-harmonic divorce may have originated in blues, but blues is not the main influence on current songwriters. Instead, I think classic rock is the more salient influence. For example, Taylor Swift never talks about the blues, but she does talk a lot about Paul McCartney and Bob Dylan. These guys were certainly influenced by the blues, but for Taylor, it’s second-order influence.

It would be easy to say that melodic-harmonic divorce is a function of naive songwriting, people plonking around on the keyboard with their index finger or drawing in the MIDI piano roll without knowing what they’re doing. I do not buy this theory. Some pop songwriters are naive, but not all of them are. Max Martin knows the rules of Western tonal theory very well, but he also knows how to write an idiomatic pop melody. At this point, I think people are avoiding functional harmony deliberately, whether they think explicitly in those terms or not.

My NYU colleague Kevin Laskey sees the melodic-harmonic divorce as possible being a consequence of the “track and hook” songwriting method common to the current pop mainstream. Songwriters and producers will create complete instrumental tracks in a DAW and then hand them over to top-line writers, who add the vocal melodies. This means that the same person is not usually coming up with the chords and melody at the same time; the chords are part of the background texture, which the melody writer can adhere to or not.

When I listen to my aural skills students improvise melodies over instrumental backing in class, some of them spell out the chord tones, but most of them treat the key as a unified whole and shape their melodies without reference to the individual chords (unless I ask them too.) I also see it in my theory students’ composition projects; several of them treat MuseScore like a DAW with semi-non-interacting streams in the bass vs the chords vs the melody. Sometimes this is due to confusion, but it is much more often a simple matter of taste. So what does this mean for music education? For me, it’s a strong incentive to be descriptive rather than prescriptive as much as possible, to point to Western tonal tradition as an important and influential body of knowledge, but not a set of rules to be followed.

Leave a comment

Leave a Reply