I didn’t find out about hypermeter until very late in my music theory learning journey. I think it should be part of the basic toolkit, especially for songwriters and improvisers. The explanation that follows might seem abstract, but behind the scenes, hypermeter provides the signposts that orient you in medium-scale musical time.
The term “hypermeter” might be new to you if you aren’t a musicologist, but I guarantee that you already intuitively know what it is. When you feel that a verse or chorus has a front half or a back half, that you can or can’t expect when the next section is going to start, or you sense that things do or don’t align with each other, that is your sense of hypermeter at work. At a club or festival, the crowd can easily feel when a 32-bar section of a repetitive dance groove is coming to an end, not because anyone is counting measures, but because of their orientation in the hypermeter.
Edward T. Cone coined the terms “hypermeter” and “hypermeasure” in his 1968 book Musical Form and Musical Performance. The basic idea is simple: if meter is the organization of beats into measures, then hypermeter is the organization of measures into hypermeasures. Imagine a bar of 4/4 time, comprising four beats: beat one (strong), beat two (weak), beat three (strong), and beat four (weak). By analogy, a standard hypermeasure contains four measures: measure one (strong), measure two (weak), measure three (strong), and measure four (weak). Just like you expect important small-scale events to fall on strong beats in the measure, you expect important mid-scale events to fall on strong measures in the hypermeter. Not that music always has to gratify your expectations, it can be more fun when it doesn’t.
You could think of a whole series of nested structural levels of musical time. Four sixteenth notes make a beat, four beats make a measure, four measures make a hypermeasure, four hypermeasures often make up a song section, and so on. The image below shows subdivisions combining into a beat, beats combining into a measure, measures combining into a hypermeasure, and hypermeasures combining into a section.
Just as you can subdivide beats into units other than four, you can group measures into units other than four too. Hypermeasures can be two bars long, or eight bars, or (less commonly) three or six bars. So how do you know how long the hypermeasures are in a given song? They aren’t notated or spelled out, so the answer is always going to be subjective, whatever grouping feels most like a coherent unit.
But so why should you even care how measures are grouped together? For one thing, it’s an underappreciated aspect of musical organization. If the hypermeter is too predictable, then the song can feel same-y and tedious. If it’s too unpredictable, the song can feel shapeless and meandering. Good songwriters can use hypermeter to create musical interest, especially by making it odd, by disrupting it, or by having it conflict with other aspects of musical structure. This happens often in Beatles songs, which is one of many reasons why I devote so much analytical attention to them.
Let’s start with “Baby’s in Black”. Its production is primitive, and its harmony, rhythm and melody are old-fashioned, but the surprising structure more than makes up for it.
The tune starts with its chorus. I hear the hypermeasures as being four bars long, which is typical. However, the chorus is three hypermeasures long, which is not typical. They break down like so:
- Oh dear, what can I do
- Baby’s in black, and I’m feeling blue, tell me
- oh, what can I do
That last hypermeasure sounds different from the others because its harmonic rhythm is faster. Up until this point, each chord has lasted for two measures, but on “oh, what can I”, the chords only last for one measure each.
The verses are even odder. They are three and a half hypermeasures long, because the fourth one is interrupted halfway through.
- She thinks of him, and
- so she dresses in black, and
- though he’ll never come back, she’s dressed in
- (short!) black
Chord changes usually nest neatly within the hypermeter, but in the “Baby’s in Black” verse, there’s a D chord that spans the end of the second hypermeasure and the beginning of the third one. It feels like the chord should be changing on the word “though” because of the new hypermeasure, but it doesn’t. This is a subtle but effective attention-grabber.
“Lucy in the Sky with Diamonds” is famous for its trippy lyrics and production, but the underlying structure is doing a lot to contribute to the feeling of dream logic. In the chorus, a hypermetrical boundary makes it feel like the key changes, even though the chord doesn’t. It’s wild!
I hear the hypermeasures in this tune as being two measures long, so the chorus contains three and a half of them. As in the “Baby’s In Black” verses, the fourth hypermeasure is cut off halfway through.
- Lucy in the sky with diamonds
- Lucy in the sky with diamonds
- Lucy in the sky with diamonds
- (short!) Aaaaahhhh
The first three hypermeasures all end on a D chord (on the word “diamonds.”) The chorus is in G major, so this D is the V chord. However, the fourth hypermeasure is harmonically connected to the following verse, which is in A Mixolydian. That makes the D chord on “Aaaaahhhh” function as the IV chord in A. Even though the chord stays the same from “diamonds” to “Aaaaahhhh”, it sounds like something is harmonically shifting. It’s the chord’s function, not the chord itself, but the effect is similar. I doubt this was intentional on John Lennon’s part, but it’s a neat trick, and I’m sure he enjoyed the effect once he discovered it.
The verses of “Lucy in the Sky with Diamonds” are hypermetrically strange too. I hear the first half of the verse as being four and a half hypermeasures long. The last hypermeasure isn’t cut off early; it’s the second-to-last one.
- Picture yourself in a
- boat on a river with
- tangerine trees and
- (short!) marmalade
- skies
The second half of the verse compounds the oddness, because it’s not the same as the first half. Instead, it’s five full-length hypermeasures.
- Somebody calls you, you
- answer quite slowly, the
- girl with kaleidoscope
- eyes
- (instrumental)
Speaking of John Lennon songs with dream logic, let’s consider “Julia”.
In this tune, I hear the hypermeasures as being four bars long. The first section (I call it the A section) is two hypermeasures long, which is conventionally symmetrical, but unconventionally short.
- Half of what I say is meaningless
- But I say it just to reach you, Julia
The second section (I call it the B section) is three hypermeasures long. John loves those odd-length sections.
- Julia, Julia
- Ocean child calls me
- So I sing the song of love, Julia
Then there’s another B section, identical to the first. This is followed by a little interlude, only one measure long, half a hypermeasure. The next part (I call it the C section) is two and a half hypermeasures long, with the half-hypermeasure sandwiched between the two whole ones.
- Her hair of floating sky is shimmering
- (short!) Glimmering
- In the sun
Maybe you would rather hear this section as a long hypermeasure of three bars followed by a standard hypermeasure, or maybe as a standard hypermeasure followed by a long hypermeasure? No matter how you choose to group the bars, it’s disruptive to the flow. But because the disruption is at this lower structural level, it doesn’t disturb the tune’s placid surface.
You may be wondering if there’s a difference between phrase structure and hypermeter. Melodic phrases do tend to be one hypermeasure long. However, phrases can be more irregular than the underlying hypermeter. In a future post, I’ll look at some tunes where phrase structure and hypermeter conflict in interesting ways. Knowing me, they will probably be more Beatles songs.
Jason Yust’s challenging but fascinating book Organized Time has two chapters on hypermeter. The book deals exclusively with the Western European classical canon, but Yust’s analytical approach has potential for the music that I’m interested in too. He points out that while meter is usually stable and predictable in canonical works, hypermeter is more flexible. He is especially interested in situations where harmonic and hypermetrical structures conflict with each other. Reading his book made me want to seek out such situations in pop.
There’s some music history in Yust’s book that’s beyond the scope of this post, but it’s relevant to my larger interests. Yust points out that before the Baroque era, when the rules of Western tonality fully solidified, cadences in European music felt stronger or weaker due to their metrical placement rather than their intervallic or voice-leading structure. Eventually, harmonic cadences took on a life of their own as a structural element of the music, with a hierarchy of stronger and weaker cadences depending on their voice leading. But how did Europeans come to hear perfect authentic cadences as stronger than imperfect authentic cadences? It is probably because of their conventional metrical placement. Once Europeans got used to hearing stronger cadences in stronger metrical positions, then they came to hear the voice leading itself as stronger. However, metrical placement never stopped being important; Yust argues that even during the canonical era, whenever metrical placement conflicts with harmonic contrast, the metrical placement will have more structural weight.
Why should this matter to someone studying more recent music? A hundred years ago, most Anglo-American popular music followed European classical conventions in its harmonic structure. However, with each passing decade, pop harmony has become less European-sounding and more heavily influenced by blues and other African-descended, groove-oriented music. You can listen to a lot of hip-hop or electronic dance music and never hear a single V-I cadence. Even in top 40 pop, the metrical placement of the chords determines their function more than their intervallic content or voice leading. It is interesting for me to learn that there was a time before Europe’s harmonic conventions solidified, and that the music from that time might be more appealing to me than I expected.