The Musical Meme

Bill Benzon (
Thu, 12 Jun 1997 20:50:27 -0500

Message-Id: <>
Date: Thu, 12 Jun 1997 20:50:27 -0500
From: (Bill Benzon)
Subject: The Musical Meme

Tim Perper notes:

>One would have trouble setting the memes of a piece of music or art into a
>*hierarchical* arrangement, if, by that, we mean that _this_ theme in a
>piece of music, say, is more important than _that_ theme. The Western
>musical form of the "theme and variations" might seem to prove me wrong,
>because the initial "theme" in a sense rules the form of its variations.
>If so (I say for the sake of argument) then the "theme and variations" form
>in music is essentially *transgressive* -- the variations acting
>dialectically to challenge the predominance of the theme. However, to be
>honest, I think that interpretation is a *bit* forced, so I want to return
>to a more playful, and less hierarchical, view of the "theme and
>variations" form. Maybe Bill Benzon can comment?

I just put on Beethoven's "Diabelli Variations" to get in the mood. Here

In the theme and variations form the theme serves as a source of material
which is then subjected to a variety of operations. The early variations
in a series tend to be derived from the theme in a simple enough way that
you can hear the resemblance despite some differences. As the series goes
on, the variations tend to be less and less transparently related to
original theme, yet there remains some musical link, some continuing
sameness, between the theme and the variations. The idea is to get as much
interesting music "out of" the theme as you possibly can.

In a way, you're playing games with what Aaron Lynch calls, I believe, the
axiom of abstraction. Two things can be said alike only under the
application of some comparison procedure. If I state a theme and then
simply repeat it, one can "extract" likeness from the two repetitions using
a relatively simple procedure. Once the variation process starts, the
procedure which extracts likeness from the theme and one of its variations
(or from two different variations) gets more complex.

I'm not at all sure how this relates to Tim's suggestion of musical
hierarchy, though it does make sense to talk of one theme being more
important than another in a given piece of music. Nor, for that matter, do
I know where to find a meme in this. It's not difficult to analyze a piece
of music into big units and small units, and to see how components are
combined and recombined in different ways, etc. But figuring out which is
a meme....

Let's take a simple example. Consider George Gershwin's tune "I Got
Rhythm" and put the lyrics aside. Let's think only about the music. This
song is of a fairly standard type; in fact it is a prime example of a type
often referred to as "a Standard." Such songs are typically 32 bars long
and have 4 sections each eight bars long in what is called "AABA" form.

In AABA form a given phrase is stated (A) and the repeated (giving us AA).
Then a contrasting phrase is played (to yield AAB) to be followed by a
repetition of the first phrase (one cycle of the AABA is now complete).
The two 8-bar phrases making up this form generally consist of 2 4-bar
phrases, which in turn consist of 2 2-bar phrases. The A section of "I Got
Rhythm" could be analyzed as:

[m, (i)m] [m, n]

In this analysis m and n are two different 2 bar phrases. (i)m is an
inversion of m; that is one takes phrase m and plays it upside down. The b
section could also be analyzed into its constituent units, though the
analysis would be a bit different.

In performance it is common to perform several "choruses" of the tune (in a
vocal performance, there'll be a different set of lyrics for each chorus),
so a performance becomes, for example:


Now, where's the meme? The chord patterns associated with m & n are
themselves absolutely standard and used in many many other tunes. The same
could be said of the rhythmic and melodic units. Does that make them
memes? My intuition is to say that it doesn't. They are just units from
which the meme, whatever it is, is constructed, just as a gene is itself
constructed from smaller physical units.

I'd be inclined to think of the whole tune as a meme. In performance, one
plays the whole tune, not a part of it. It is the whole tune which takes
part in the social ritual of sharing music (and possibly dance) with other

Note that the tune has been performed in hundreds of different
arrangements. So a memetician would have to exercise some care in asserting
that all these different pieces of music are, in some non-trivial sense,
the same tune (Aaron Lynch's principle again).

Now, let's take this analysis a bit further. Jazz musicians have been
strongly attracted to "I Got Rhythm" and have used it as a vehicle for
countless improvisations. Beyond that, they have used the tune as the
basis for hundreds, if not thousands, of jazz tunes. They do this by
taking the harmonic structure of the tune (the temporal pattern of "legal"
sets of pitches) and writing a new melody for it -- Duke Ellington's
"Cottontail" and Charlie Parker's "Dexterity" are two examples of tunes
based on "rhythm changes" (the vernacular/technical term for the harmonic
structure of "I Got Rhythm").

Again, where's the meme? Jazz musicians have abstracted the harmonic
structure of the tune away from the melody and seem to be treated that
harmonic structure as a meme. This has been done with many tunes, "How
High the Moon" has yielded "Groovin' High," "Georgia Brown" had yielded
"Bright Mississippi" and so forth.

Unless you have listened to a good deal of jazz (or read a little about it)
you wouldn't know that jazz musicians have constructed many jazz tunes by
taking the harmonic structure of a pop tune and providing it with a new
melody. The practice is so common I'm almost inclined to think that, for
jazz musicians, it is harmonic patterns which are memes. But I'm not sure.

Beyond this, the jazz world is full of short fragments called riffs (or
licks) which appear in the improvisations of many different musicians.
Some riffs are thought to have been invented by this or that musician, but
most just seem to be around. A very important subset of riffs are those
used in playing the blues and are called, accordingly, blues riffs. Given
the way musicians talk about riffs and seem to use them I'm inclined to
think riffs might also function as memes (eventhough I want to deny that
musical fragments of similar scope in composed tunes such as "I Got Rhythm"
are not riffs).


There's a good deal of analytical language one can use to describe music.
There's also a growing body of work on the psychology (and even
neuropsychology) of music. AI folks have devoted some attention to getting
computers to compose music. But just where one finds the musical meme is a
bit of a mystery. That is, it is not at all obvious to me how one connects
the concept of a meme with the analytic units currently used to describe

What seems very important to me is that the meme, whatever it is, is a unit
of cultural transmission (& variation), and not a unit in the construction
of the artefacts and behaviors of culture. The 2-bar phrases which
constitute "I Got Rhythm" are units of construction, transmission. For the
most part, the analytic units of music they were created to talk about how
individual pieces of music are constructed; not how they function in
society and culture. It is tunes which are transmitted, and perhaps riffs
in jazz (and related styles). Though even as I type this my mind flashed
on work on motifs in the folktale world and made me want to retract my
speculation that riffs function as memes.

[An aside for post-modern cultural theorists. Not only that, but the terms
of standard music theory were created to describe Western classical music.
They are strongest when deal with harmony and what is called architecture
(large-scale structure), somewhat weaker when dealing with melody, though
still pretty good. And they are lousy when it comes to analyzing rhythm.]

In any event, I can dimly see my way to an argument that, at least within
certain musical styles, the song is the meme unit. But then what does one
do with classical music, with its large scale composed pieces? Is a
Beethoven symphony to be counted as a single meme? If not, then how many
memes is it? The individual movements of a Beethoven symphony can be
performed individually -- and have often been done so. So, are they memes?
What do we do with Wagner's operas, where he constructed them using units
he called leitmotives? Do those individual leitmotives functions as memes?
Probably not, since they have little functional existence outside of
Wagner's operas.

And so on and so forth.


There is, of course, a meta point. This is a crude example of the kind of
analysis we're going to have to do to get a good grip on culture. Most of
what I've said is fairly pedestrian stuff, readily available off the
appropriate shelf. But hooking it up to the dynamics of a Darwinian model
of cultural change is tricky stuff. And I think this kind of analysis has
to be done domain by domain. Working out the memetics of music is one
task. Working out the memetics of stories (such as the myths Levi-Strauss
has analyzed) is a different task. Folks have been working away at that
but it is not obvious to me just how one goes into that literature and
identifies the memic units, much less the meme-complexes, and coadapted
meme-complexes, a memosomes, etc. -- though I have a little more hope of
identifying species. So, music and stories, now we can do dance, and
clothing, a cuisine, and war and its implements, etc.

Given this, as much as I'm unhappy with calling anything and everthing a
meme and then pretending we know what we're talking about, I'm not a great
deal happier with the notion that we gotta define our terms and only then
can we get down to the real stuff. That implies that defining terms is
going to be easy & straightforward, we just have bite the bullet, take a
couple of weeks, and get all the definitions down. I don't think its that
easy. Definining terms in a robust way is hard work and it requires that
we spend a lot of time discussing the objects and events in the domain over
which we want to define our terms. You offer a definition or three, and
try them out on a bunch of cases and see how the definitions work.

It's easy to take the terms and concepts in the biological theory of
evolution and coin words for cultural analogues. But that doesn't give you
a theory of cultural evolution. All that gets you is a bunch of neologisms
in search of a theory.

More later.

Bill B

William L. Benzon 201.217.1010
708 Jersey Ave. Apt. 2A
Jersey City, NJ 07302 USA

This was distributed via the memetics list associated with the
Journal of Memetics - Evolutionary Models of Information Transmission
For information about the journal and the list (e.g. unsubscribing)