Songwriting: Interval Shifting
One thing that is mystifying about writing chord progressions, or melodies, or riffs, is trying to figure out what separates a good one from a bad one. I think it's basically some combination of structure and surprise. If you just play a scale up and down, no one will be interested; if you just play totally random notes, there isn't enough to latch on to. So the key is finding a way to play things that make sense and sound good but still surprise people's ears somehow. One way is to just noodle around until something sounds good; or, you can take the opposite approach and try to work something out based on a theoretical idea. I'm going to write about a concept that I've noticed in a few songs, and used myself, that is somewhere in between these two extremes. It's something I think of as interval shifting.
|F# | Bb | G | B |
Let's look at a few songs that seem (to me) to use this idea and why this way of thinking makes sense in these cases. The first is Lay Lady Lay by Dylan.
Lay Lady Lay
The verse chords to Lay Lady Lay have always seemed pretty hip to me. It's in the key of A, and consists of four chords:
|A | C#m | G | Bm |
This seems simple enough, but, if we look closely, this progression is actually a little bit unusual. Let's look at the roman numerals for this progression (if you don't know what these mean, check this out: https://en.wikipedia.org/wiki/Roman_numeral_analysis):
| I | iii | bVII | ii |
The first thing that jumps out at me when I look at this is that we have chords built on both the major third and minor seven of A major. This means that we already know we aren't dealing with a major or a minor progression (because the major scale has a major seven and the minor scale has a minor third). If you know anything about modes, you'll know that the presence of these two notes is indicative of the Mixolydian mode (and its corresponding chord the dominant seven). So, we're dealing with a modal progression based on Mixolydian, right? That's not too unusual. There's one more wrinkle though. Here is the Mixolydian scale harmonized:
| I | ii | iiio | IV | v | vi | bVII |
Notice that the iii chord is diminished (indicated by the little circle), whereas Dylan just plays it as a regular minor chord. This isn't that weird either really; the diminished chord is very dissonant and it's pretty common to sub in a regular minor for it. Nonetheless, this seemingly simple progressions turns out to be both non-diatonic and modal. Was Dylan thinking along these lines when he wrote it? It's possible, but I doubt it. I think he was thinking in terms of the aforementioned interval shifting. This means he probably wrote it on guitar and not piano, because this technique strikes me as a pretty guitaristic thing to do.
Ok, so what the hell am I talking about? It's very simple: you just take two chords some interval apart, and then move that up or down some other amount and repeat it. If I had to guess, I'd say Dylan started by fiddling around with the I-iii move that the song begins with. This is a pretty common way to start a progression (for instance, The Weight by The Band begins with the exact same move). In rock music, this is less common than a I-IV or I-V move, but still not unusual. Most songs written in this fashion end up meandering through a series of diatonic major scale chords, although Lay Lady Lady, as we saw, does not. Anyway, after playing the I-iii move (on guitar, just move your hands up one string and down a fret), Dylan decided to move the pattern down a whole step (two frets) and do it again. On guitar, this move seems perfectly natural. What's great about this is that the end result is actually kinda weird. You wouldn't necessarily come up with this if you were thinking about music theory, but it's still based on an underlying idea that initially comes from that way of thinking (because the I-iii move is very much based on a traditional songwriting techniques).
One more note before we move on: Notice also that when we move from the second measure to the third measure (c# to G), the root of the chord we're playing moves down a flat fifth, or tritone. The tritone is a very dissonant interval, but also cool and useful. In this case, it's almost unnoticable. I doubt Dylan even noticed it, because I think that he's thinking in terms of two bar phrases. Thus, we're really moving the entire I-iii phrase down two frets; the fact that this means the iii chord will lead to the bVII a tritone away is incidental, but it helps make the progression unique and interesting.
Nirvana
Dylan isn't associated with theory intensive songwriting, but Kurt Cobain is probably even less associated with that way of thinking about music. Still, most of his riffs and progressions do have a certain logic to them, to me at least. A lot of his stuff makes sense if you think in terms of interval shifting. I know he wasn't thinking about theory when he wrote them, but I bet he was seeing various patterns and shapes on the fretboard. As I mentioned, this way of making music makes a lot of sense on guitar, so if you're a guy who writes with a guitar it's a good tool. First, let's look at the ever ubiquitous Smells Like Teen Spirit. Here are the chords:
| F | Bb | G# | C# |
In roman numerals, this corresponds to the following:
| I | IV | bIII | bVI |
The flat three and flat six tell us right away that we're dealing with a minor progression.
(It's worth noting that I've written all the chords as if they were major, in this example and the next one. This song, like most Nirvana songs, is based on ambiguous root-5 power chords, with no third, so you can think of them as either major or minor, depending on the context. If you're were thinking strictly in terms of a the minor scale, it would be more accurate for the one and four chords to be written as if they were minor. I think it makes more sense to think of power chords as always major, since they move in parallel. Strictly speaking, the melody will determine their quality.)
I remember reading an interview where Cobain half-jokingly said that Smells Like Teen Spirit is really not much different than Louis Louis. He's not totally wrong. Louis Louis is based on a I-IV-V-IV progression, so that first two chords are indeed exactly the same (albeit in a different key). As mentioned in the Dylan section, the I-IV move is probably the most common chord change in rock music. Cobain follows it up with something slightly different, though: interval shifting. He moves up three frets (a minor third) and repeats the same interval move (up a fourth). As in Lay, Lady, Lay this gives the riff a nice symmetry. It's also what gives it a unique sound. Moving up a fourth from the bIII brings you to the bVI, which is an evocative chord, especially in a song that starts out with a pretty stock I-IV change. Notice that the bVI is a half step above the V, which was is third chord in Louis, Louis. To me, the key to this riff is that we never get to the five chord, which is what we would expect based on the first two chords and a basic familiarity with rock music. Coban lands on the minor sixth instead, which sounds darker. Instead of moving directly to the bVI, though, we approach from a fourth below, copying the move from the first two measures, and this makes the whole thing seem logical and normal.
There's nothing too weird going on here, but it still gives us an idea of how (I think) Cobain approached songwriting. He would almost certainly roll his eyes at this kind of analysis of his songs, but I'm still convinced that he was thinking in terms of these basic patterns. He's definitely not just jumping around at random. These types of interval shifting moves crop up a lot in Nirvana songs. One of my favorite progressions is in Tourette's, which is built around these four chords:
|F# | Bb | B | G |
If we assume the song is in the key of F#, since it begins on that chord, the roman numerals would be this:
| I | III | IV | bII |
Before we start talking about this in depth, just for the fun of it, it's worth taking a moment to acknowledge that this song begins with the same one to major third change that Lay, Lady, Lay did! Theoretically, we're looking at a progression with a major three and a flat two. I'm sure there's a scale that contains both of these notes, but I don't know what it is; none of the major modes contain both of those notes. So, we can pretty much abandon any sort of scale analysis right away. I mentioned that it began with the same move as Lay, Lady, Lay, so let's try something similar to that song. Instead of moving the major third change down a whole step, however, let's move it up a half step. Here's where that gets us:
|F# | Bb | G | B |
One simple idea, barely any different from the Dylan song, and we're almost all the way there. All we have to do is change the order of the G and B chords and we've got the Tourette's progression. So, basically, Cobain is building this progression around a pair of major third changes a half step apart, with the added wrinkle that the first is up and the second is down. I would be very surprised, personally, if this isn't the way his mind was working when he came up with this progression, because it is weird as all hell if you try to make sense of it in any other way. This is a good example of how any songwriting tool should be used: always be open to changing it around somehow to make things more interesting or better sounding. Just because the first two examples involved shifting an interval and playing it in the same direction doesn't mean that's what you have to do.
Work In Progress
Here's one more example. This one is from a song I'm working with a band I'm in called The Pedants. We haven't presented it to the singer yet, so it's only halfway done, but it originated with this same line of thinking. Here are the chords for the verse, played with power chords in a punk rock fashion:
| E | C | G | Eb Bb | (2/4) B |
(The (2/4) indicates a half measure.) This idea began when I was fiddling with the E-C change (I-bVI). I decided to move that change up a minor third. This left me with E-C-G-Eb. It sounded okay, but I didn't like the Eb-E change that happened on repeat, so I decided to see what happened if I moved the whole thing up a minor third again. This gave me E-C-G-Eb-Bb-F#. I liked this a little better, but it still didn't quite work right. The F#-E move was stronger than the Eb-E one, but still kind of weak, especially for a song with so many unrelated chords.
What to do now? Start breaking the rules of the interval shifting concept we began with. The first thing I did was replace the F# with a B. These two chords are a fourth apart, so that kind of makes since as substitution; and, the B is a half step up from the Bb that precedes it, so that also kind of makes sense as well. Most importantly, now we have a B-E change on repeat. The V-I change is very strong and common, so now, even though the changes are moving through a few different keys and some unrelated chords, on repeat we have something that reestablishes the original key emphatically.
What to do now? Start breaking the rules of the interval shifting concept we began with. The first thing I did was replace the F# with a B. These two chords are a fourth apart, so that kind of makes since as substitution; and, the B is a half step up from the Bb that precedes it, so that also kind of makes sense as well. Most importantly, now we have a B-E change on repeat. The V-I change is very strong and common, so now, even though the changes are moving through a few different keys and some unrelated chords, on repeat we have something that reestablishes the original key emphatically.
I was pretty happy with it, but it still seemed a little bit unfocused. I decided to shorten it up a little bit so we got back to the E a little quicker. The way I did this was to change the harmonic rhythm (a term that refers to the speed at which the chords in a progression change). The first three chords got a full bar of 4/4, but the next three only half a bar. This meant that the last chord got stuck with its own 2/4 half measure. This wasn't my plan originally, but it wasn't a big deal either. The half measure gives the riff another unique characteristic. Not only that, once I sat down with the drummer and showed it to him, it became one of the coolest parts of the riff. He turned the last half measure into an emphatic build up that made the B-E change even more powerful. The end result is, to me, pretty damn cool.
One Last Thing: Coltrane
I'm not going to spend much time on this, because I'm not a jazz player and only somewhat qualified to discuss it, but I think it should be mentioned. John Coltrane famously used a similar technique when he came up with his famous (or infamous, if you have to solo over it) Coltrane changes. The Coltrane changes involve a series of descending major third modulations. Coltrane noticed that if you modulate in this fashion, you will end up back where you started after a mere three changes. Starting in the key of E, for instance, you will go through the keys E | C | G# | E | etc. What Coltrane did was to take these modulations and play them as a series of ii-V-I changes (frequently without playing the I and only implying it). This is really similar to the stuff we discussed above. A ii-V change simply means moving up by a minor third (f#m to A for instance). Take this interval and shift it down a major third twice and you're playing through Coltrane changes. There's a ton of stuff on the internet about Coltrane changes so I won't spend much time on it, but I think it's cool to realize that Coltrane was using a technique that, to me, is quite similar to something that pops up in Nirvana and a lot of other music that otherwise has little in common with it.
Comments
Post a Comment