When I was learning guitar, I did a lot of studying and memorizing chord progressions. I did even more thinking about chords when I was learning to play jazz. When I shifted over to mainly producing electronic music, all my focus went to thinking about groove and timbre and I stopped thinking about chords completely. But now that I’m teaching music theory, I’m back to thinking about chords, and as I prep examples for class, I am in particular thinking about chords in rock songs in a serious way for the first time since my 20s.
You can’t generalize about chord progressions in Anglo-American pop across the board, because there’s too much stylistic diversity between metal, country, hip-hop, R&B, dance music and so on. However, rock has stabilized into a canon, and it’s possible to get your arms around the entire thing.
So given all that, here’s my explanation of rock harmony in one sentence: put major chords on the notes in the natural minor scale, and put minor chords on the notes in the major scale. That doesn’t explain every chord you’ll find in a rock song, but it does explain a lot of them.
Let’s think about the key of E. The notes in the E natural minor scale are E, F-sharp, G, A, B, C, and D. If you use each scale degree as the root of a major triad, you get E, F#, G, A, B, C, and D chords. These are all plausible chords for a rock song in E, in just about any order and any combination.
Now let’s consider the E major scale: E, F-sharp, G-sharp, A, B, C-sharp, and D-sharp. We’re going to ignore the seventh, D-sharp, so we’re really thinking about the diatonic hexachord. We’re going to use the first six scale degrees as roots of minor triads. That gives you Em, F#m, G#m, Am, Bm, and C#m chords. These are also all plausible chords to use in a rock song in E.
Combining the lists, we get:
- E and/or Em
- F# and/or F#m
- G
- G#m
- A and/or Am
- B and/or Bm
- C
- C#m
- D
Like I said, this is not every chord you would find in a rock song in E, but does include all the common ones and several less-common ones too. This set of roots also aligns well with the global pitch set used in most rock melodies, which David Temperley calls “the supermode”.
My one-sentence theory has some serious problems!
- It says nothing about melody.
- It says nothing about musical time or harmonic rhythm.
- It doesn’t tell you anything about how the chords function, what order you should use them in, or what combinations sound good.
- It doesn’t tell you about how often the chords are used in rock. The bIII major chord is much more common than the iii minor chord, for example, and the minor ii chord is much more common than the major II chord.
- It only includes major and minor triads, and doesn’t say anything about seventh chords. That means that it can’t make sense of the blues. You could plausibly add a flat seventh to almost all of the chords on my list, but not quite (G7 sounds goofy and D7 sounds too jazzy.)
- It doesn’t tell you about other kinds of Anglo-American pop, or about more extreme kinds of metal.
If you want to go deeper into these questions, I recommend David Temperley’s book The Musical Language of Rock. His chapter on harmony generally supports my theory. He bases his explanation on a study of the 500 greatest rock songs compiled by Rolling Stone magazine in 2004. He and his research partners transcribed all 500 songs from this corpus and did statistical analysis on them. You could debate the validity of this corpus as a representation of the rock music canon, but it seems to me as good a place to start as any. Temperley argues that rock melody and harmony both draw from a pitch collection that he calls the “supermode”, the union of the major and natural minor scales. In E, the supermode contains every note except for F and A-sharp.
Temperley presents the supermode as a section of the circle of fifths. If you start with E and go sharpwards, you get E, B, F-sharp, C-sharp, G-sharp and D-sharp. If you start with E and go flatwards, you get E, A, D, G and C.
Temperley observes that in rock, chord roots usually stick close to the tonic on the circle. So for songs in E, you mostly see chords with roots on E, A and B; you commonly see chords with roots are on D, G, F-sharp and C-sharp; and you occasionally see chords with roots further from E. Furthermore, the other notes comprising the chords will almost all come from within the supermode. In the key of E, F#m is more common than F#. Temperley believes that this is because F#m includes the notes F-sharp, A and C-sharp, all within the E supermode, whereas F# includes A-sharp, which is outside the E supermode.
Temperley finds some patterns in which chords appear together. In general, chords from the flat side of the circle of fifths don’t appear with chords from the sharp side. So if a song in E uses D, G or C chords, then it is unlikely to use F#m, C#m or G#m chords, and vice versa. (Chords in the middle, the E, A and B chords, mix well with both sides.) Temperley observes that songs with more sharp-side chords sound more pop, while flat-side songs sound harder. It’s very striking when a songwriter mixes chords from both sides. The Grateful Dead frequently mix sharp-side and flat-side chords, which makes their songs sound quirky and colorful.
What about progressions? Temperley finds that rock songwriters have clear preferences about which chords they use, but they don’t seem to care what order the chords are in. So for a song in E, you can have E come before A or after it; before D or after it; before G or after it. You move from A to B as often as you move from B to A. This is very different from Western European classical tradition, where you would hardly ever move from B to A in a piece in E major.
Temperley’s book is an extraordinarily thorough and rigorous analysis of the Rolling Stone corpus, but you might have some doubts about how well that corpus represents rock as a music. There isn’t much metal in there, certainly not any of the more extreme varieties. A corpus centered on metal would look very different; there would hardly be any sharp-side harmony and few full chords. Flat two would be a pretty common chord root, and sharp four wouldn’t be unheard of. But if a consensus exists around a canon of metal, I am not aware of it. Among the metal fans I know, the only consensus is that there is no consensus. None of this is meant as a criticism of David Temperley! He is as rigorous as musicologists get. My silly theory is something I came up with out of the blue. It’s gratifying that Temperley’s research mostly validates my instincts, but they are still my instincts.