Wednesday, May 26, 2010

Cultural Evolution 3: Performances and Memes

n the first post in this series I showed that, while Peter Richerson and Joseph Carroll apparently differ on the importance of cultural evolution in relation to biological, their common acceptance of gene-cultural evolution gives them no “sharp” way of defining their difference – a problem that’s worse for Richerson than Carroll, as his is the more subtle position. My second post expressed dissatisfaction with Richerson’s position on informal intuitive grounds – I called it a phenomenological gut-check – similar to those Carroll used to dismiss Miller and Pinker on the adaptiveness of the arts, and Dawkins used to dismiss a purely biological approach to human behavior. Now I’m ready to begin laying out my own views on cultural evolution, views which are very much in flux.

But first, some preliminary throat-clearing.

Thoughts on where Dawkins went wrong

While I don’t by any means dismiss biology as being relevant to the study of human behavior – I’ve got too heavy an investment in brain-based approaches – I do think that the evolution of culture needs to be approached as a phenomenon more or less “within” culture itself, much as Martindale or Dawkins approach it. Though on the whole he seems a bit muddled on this point, Dawkins’ key insight is that, in the cultural evolutionary process, selection operates on cultural entities and not on human phenotypes. That is to say, the evolutionary costs and benefits accrue directly to cultural entities, not to the human beings who create and consume them. There are cases where cultural entities seem to thrive at the expense of humans, but this is a secondary matter and, in the large scope, not worth the attention that’s been lavished on them in popular discussions.

In the biological realm, of course, it is phenotypes, bodies, that are the object of selection. But Dawkins says otherwise with respect to culture (The Selfish Gene, p. 199, using his standard anthropomorphic language): “Selection favours memes that exploit their cultural environment to their own advantage. This cultural environment consists of other memes which are also being selected.” Some confusion arises because Dawkins thinks of memes as the cultural analog to genes, and genes are not the direct target of biological selection. Selection works directly on phenotypes, and only indirectly on genes. It is the phenotypes that thrive or die through interaction with the environment. Dawkins has no clear conception of what the cultural analogy to the phenotype might be – nor do his followers. Better, perhaps, they have no clear conception of how to draw a distinction between a cultural phenotype and cultural gene and so tend to use memes in both roles as seems locally expeditious. In fact, Dawkins often abandons the meme-gene analogy and thinks of the meme as a virus, thus obviating the need for a phenotype analog.

I suspect one thing that’s throwing Dawkins off is his (and everyone else’s it would seem) implicit theoretical imagery. Considered as physical entities, phenotypes are relatively large while genes are relatively small. Further, genes are physically contained within phenotypes, not vice versa. When we transfer this implicit imagery into the cultural realm we are led to think of memes as relatively little things that must be enclosed in some relatively large thing, such as the human brain. Hence, we have the common notion of memes as quasi-autonomous agents hopping around from brain to brain, taking over mental real estate, and often driving their “hosts” to irrational acts, like believe in God, psychoanalysis, or Marxism.

This kind of thinking has led nowhere. It’s time to hit the resent button. We need cultural analogs for both the gene and the phenotype, and individual human phenotypes are not going to play either role. How can we do this?

Musical performance as a cultural analog to the phenotype

What’s important are the causal relations among our theoretical entities. As reconceived in the 20th century, the basic Darwinian process is one of random variation among gene entities and selective retention of phenotype entities. Thus it is the phenotypical entity that interacts with the environment. The gene entity plays a different role.

On this point Martindale (1990) has it right. People like certain cultural things and not others. Those they prefer get passed around while the others are forgotten. So the cultural entities (practices or objects) must be our phenotypical entities and the social group is the environment in which they must survive. Why do I say this? Because this seems to me the most obvious fit between cultural process as we observe it and the causal roles available in the Darwinian paradigm. That doesn’t make it correct, however, but why not start there and see where it leads?

If that is the case, then where do we look for the meme entities? The meme entities are somehow constitutive of the phenotype entities. Genes do something like “catalyze” the process of organismic development. What entities do that for cultural entities? Since culture is shared among individuals in the group, we might ask: what allows for the sharability of cultural processes and artifacts? Where do we find the lynchpins of intersubjective agreement?

At this point I’m going start talking of a specific set of cases rather than talking at the most general level. I’m going to talk about music, which I’ve investigated more intensively – at least in these terms – than any other set of cultural objects and processes. Let me start by recasting the basic account of the musical phenotype I gave in Beethoven’s Anvil. There I assert (p. 192): “The phenotypic role in music’s evolution is played by performance level attractors.” For the purposes of my discussion here I’m going to drop talk of attractors – a mathematical notion – and talk of performances. It is performances that thrive or die.

There is one qualification, however. The entity which actually plays the phenotype role is the perceived and experienced performance (in the langauge of attractors, it is a trajectory in neural state space), the performance as it unfolds in the mind and brain. It’s a neural (and hormonal) event. For the most part, that’s what I mean when I talk of a performance, the inner event. Performances thus conceived are properties of brains-in-process, whether a single music-making brain or a group of brains coupled (see my post on the busy bee brain) through music-making.

While musical performances are not solid compact physical objects like biological organisms (phenotypes), they are certainly bounded in space and time, as are organisms. They come into being at a certain time and in a certain place and continue on for a finite period of time. Their constituents, however we may choose to characterize them, can persist (be reconstructed) from one performance to another just as genes survive from one generation to the next. Further, individual constituents can be used in distinctly different performances just as individual genes can be in the genotypes for different organisms. Thus the analogy seems to work in certain basic ways.

Most importantly, a performance can be successful or not, just as organisms can. Success is the subjective experience of pleasure while the lack of pleasure is failure. People will be motivated to repeat pleasurable performances, but not lack-luster ones, much less performances engendering anxiety — excluding, of course, the case of performances considered to be practice or rehearsal for “real” performances. Any constituent contributing to a pleasurable performance is likely to be repeated in other performances while constituents contributing only to unpleasant performances will drop out of the repertoire.

Performances live in the ebb and flow of human groups, two by two, ten or twenty people at once, a hundred or three hundred, for a minute, 10 minutes, or an hour, three hours, a day, and so forth. Varying numbers of people, varying periods of time, couplings of all kinds, some happening on a regular basis while others happen opportunistically, these are the stuff of musical culture.

At least some of the constituents I’ve referred to must be memes. But what, specifically, do I mean by a musical meme? What would be examples of musical memes?

Phase and period as musical memes

The argument I develop in Beethoven’s Anvil posits rhythm as the fundamental property of music. So let’s consider the simplest possible example of a rhythmic performance: people making noises simultaneously on a regular repeating basis. Imagine that some people are clapping their hands, others are stomping their feet, while still others are banging on cans or logs, ringing bells, shaking rattles, some may be yelling, whatever. It doesn’t matter how they make the noise; all that matters is that they make their noises together. What’re the memes?

We’re talking about very simple periodic behavior, like the ticking of a clock. Physicists describe such behaviors using two parameters, phase and period. Period is the length of time between “ticks” while phase is the relative position of two or more streams of repeated “ticks” with respect to one another. If, for example, we have two people clapping, they could be making hits at exactly the same time (in phase) or one could be lagged with respect to the other (out of phase). If we have 30 people, they might all be hitting at the same time, or we might have various lags, making for a ragged sound.

Note that, while period and phase are not chunks of physical stuff, like genes, they are physically real entities. We can measure them, and people can copy them. What role do they play in the performance? They constrain people’s noise-making by dictating when to make the constituent noises. These parameters define what it means for people to be doing the same thing. If you make your noise at the same time (phase) and repeat at the same interval (period) as the others, then you are participating in the performance. Otherwise, you aren’t.

It doesn’t matter just how you make your noise, or how loud you make it. All that matters is when you make it, and that is defined by period and phase. Those are the memes that define this particular performance. Memes set the terms by which people share in the music-making.

Now, you might be thinking something like, “but you’re not talking about real music, that’s just some toy example. What about real music?” Be patient.

In the first place I’ll consider a more complex example in a later post – the chord progression to “I Got Rhythm,” known informally as “rhythm changes.” For now I want to stick with this very simple example. If we can’t make this simple example work, there’s no point in considering more complex cases.

In the second place, simple repeatable rhythm is a real phenomenon, whether or not you want to think of it as music in some robust sense. It’s a feature of marching, for example, which William McNeil (1995) has found to be an important factor in social cohesion at various times and places in human history. Beyond that, physicists have studied periodic clapping and analyzed it as a case of coupled oscillation (Néda and Ravasz, et. al. 2000; cf. Benzon 2001, pp. 67-68), so we know something about how it works at the physical level. That’s important.

It’s clear to me that one reason that memetics hasn’t gone anywhere is that no one has taken the time to look at simple cases in some detail. That’s what I’m doing now, and that’s what I did at considerably more length in Beethoven’s Anvil. In this particular case, I’m arguing that memes are psychologically salient physical properties of cultural objects. It’s important that we understand what those properties are. We know quite a bit about these two properties, period and phase, and how they influence dynamical systems.

Which brings me to my third point. Period and phase aren’t the only properties of the sound we’re considering in this basic example. That’s why I asked you to imagine people making sounds in various ways, so you’d imagine a rich and complex sound. Most of that richness and complexity, however, is irrelevant to the question of how people coordinate their motions – at least in the simiple terms of this example. On that one issue, all that matters are period and phase, those are the memes.

One could imagine, however, that other properties are important in some other particular case. Perhaps we’ve got a case where only metallic sounds are allowed. Then the property “metallic” would become memetic, giving this performance three memes: 1) a specific period, say, half a second; 2) in-phase, and 3) metallic timbre. My general point is that a given “wash” of sound can have many properties, perhaps even an unbounded number, but only some of them are memetically active. Those properties have to be detectable by the human sensory system and they have to be producible by the motor system (with suitable supporting apparatus). When I discuss rhythm changes we’ll consider other sound properties as being memetically active.

Where are we?

We’ve now got tentative accounts of cultural phenotypes and genes for one class of phenomena, music, and illustrated with a single simple example, making a periodic noise. The phenotype is a performance. Performances, of course, are physical events. But whether they succeed or not as cultural events, that depends on whether or not they are pleasurable. The arena for musical performance is a sort-of collective brain (a notion which I investigated with some care in chapters two and three of Beethoven’s Anvil). In my view, then, it is the cultural phenotypes that are in the head, not the pesky brain-hopping memes of pop science.

Memes then are conceived of as those physical properties of the musical performance around which the performers coordinate their actions. Some physical properties are mimetically active, some are not. It follows from this conception that memes are not some exotic new theoretical entity. Rather, “memetic thinking” is a way of reconceiving entities we already know quite about through the study of physics, psychophysics, and perceptual and cognitive psychology. It follows from this that any study of cultural evolution must entail a fairly detailed consideration of perceptual and cognitive psychology – matters which have largely been ignored in popular accounts.

It also follows that we need to think about pleasure. If success or failure depends on whether or not a performance is pleasurable, we ought to know what pleasure is. That’s tricky business and, to be honest, I don’t know whether or not I’ll tackle it in this series of posts, though I discuss it in chapter 4 of Beethoven’s Anvil. I note, however, that, in Martindale’s model, novelty is the source of aesthetic pleasure. This is a common conception – make it new, and all that.

The larger point, of course, is that, whatever artistic pleasure is, it’s something that happens in the mind. It’s not about feeding the body, at least not directly. While cultural evolution ultimately depends on the existence of populations of human bodies, the basic dynamics of cultural evolution are not those of the life and death of human populations. They are the dynamics of the life and death of cultural entities.

Bonus Points: Those familiar with anthropology may notice that I’ve define “meme” is a way that’s similar to the notion of “emic,” in contrast to “etic.” Is there a decent brief account of this distinction, or is it a hopeless mess?


William Benzon (2001). Beethoven’s Anvil: Music in Mind and Culture. Basic Books.

Richard Dawkins (1989). The Selfish Gene, New Edition. Oxford University Press.

William McNeil (1995). Keeping Together in Time. Harvard University Press.

Colin Martindale (1990). The Clockwork Muse: The Predictability of Artistic Change. Basic Books, Inc.


  1. Hi, Bill,

    I think the real problem here is with the analogy of meme to organism with some kind of genetic reproducibility (like genes) in the meme. What gets reproduced over and over in the cultural works that survive are EXPERIENCES. You are on the right track when you speak of pleasure. But I think you get thrown off because you are a jazz musician. Your pleasure is performance and that repeats for you. But I'm not a performer and what repeats for me and for most people is the experience of your jazz. And, understood that way, what repeats for you is your pleasurable EXPERIENCE of your performance. (You will note the reader-response bias in my comment.)

    All the best, Norm

  2. Ah, but Norm, your experience of jazz, whether listening to live music in a club or a concert hall, or listening to recordings in the comfort of your home, or an iPod while jogging through the park, that experience is a performance as far as I'm concerned. Your nervous system is doing something; it is performing.

    I should have made that clear. Thanks for calling me on it. And, yes, the performer is also in the audience for the music they create.

  3. Hi Bill,
    You don't mention it here, but I know you're aware that language works pretty much like music as a metaphor platform for evolution, and vice versa. I agree with Norm that there's a giant dichotomy between performing and perceiving as experiences; but I haven't done much comparison with music as a metaphor.

    (I like music, but I don't perform, and music is not a big mpart of my life. My son, however, is a working musician and I watched it happen, so I have some second-hand experience.)

    My own use of evolutionary metaphors is, as you might expect, in language evolution, and the way it works there (to the extent it can be said to 'work') is to note that, whatever else may be true about it, any real evolution in language (and this goes pretty much for any other cultural phenomena as well), is that evolution of language (and culture) does not follow Mendelian rules, but rather Lamarckian. That makes a big difference in how the system works, since biological evolution depends on just two kinds of selection -- natural and sexual (and sexual selection is rare compared to natural selection). Adding (for example) economic selection, stylistic selection, and social class selection makes the theoretical structure top-heavy with pseudomemes. And not particularly predictable, either.

  4. As usual you, like Dawkins & Co., overlook the real problem with biological determinism as it applies to cultural artifacts such as literature--which is to say, determinism. You can flesh out Darwinism however you want, add the genetic factors, memetics, etc. but you're still affirming determinism, and in a sense making a philosophical claim, at least implicitly--though like most of the Valve boys, or the frauds at Berube co, you're not really capable of defending your views (or at least haven't).

    Writing and language seem to suggest something like...Mind, and the uniqueness of human thinking, and intention for lack of a better term (and I wager Professor Holland, occasional Chomsky-ite agrees, or did at one time, like pre-PoMo). Of course humans share much in common in primates, and most likely descended from a common ancestor, but primates they are not. Aeschylus was not a monkey, or neanderthal--he created something, and that creation cannot be reduced to mechanical or biological terms. That needn't imply substance dualism of the traditional platonic or theological sort, but a strict Darwinian view won't suffice as an explanation for language, or philosophy, mathematics, music (and memetics itself merely adaptationism applied to all of human history). Even Marx realized that, and while accepting Darwinism, rejected the mechanistic implications for human thinking.

    Review Kant, or Marx's critique of Feuerbach's mechanism, or even Chomsky's rips of Skinner (and his crony Quine)--all about Mind. And that doesn't mean (as most Valvesters or Berube-bots would hastily conclude) marxism's correct across the board, or that evolution doesn't hold, or that...souls exist (but that's a possible world, modally speaking).

    Darwinian accounts of evolution work, for finches and horses and neanderthals. But it won't be of much help in understanding Moby Dick or Mozart (and some jazz if Bill Evans was just following bio-chemical orders which were pre-set at his conception. nyet. He was a person, not a bot).

  5. The emic/etic distinction is rather like the extensional/intensional distinction in philosophy, basically the meaning (emic/intensional) of some cultural item, versus the external observation of its dynamics. Likewise the functional/morphological distinction in biology. I wonder if we can make a general categorical claim here?

    Determinism is a red herring. If things are caused by something, then they are determined just to the degree they are caused. So giving any causal account of cultural or biological evolution is deterministic. Likewise, both forms of evolution are, at the populational level, a matter of some stochasticity.

    I think a lot of people have a strawman version of evolution, usually under the useless rubric of "Darwinism", when they consider cultural evolution. The real problem is the question whether one needs something that "codes for" "phenotypes", which I think is not required in either biology or in culture, although in biology a large proportion of evolution occurs of codical entities of a kind. In culture very little is codical. But this is a difference of degree, not kind.

  6. This comment has been removed by the author.

  7. Lots of folks seem to think that cultural “evolution” is Lamarckian, John (Lawler). What I think is that the appropriate conceptualization of the process is still pretty much up in the air, including the question of whether or not an evolutionary model is useful. Still, if we’re going to think about evolutionary models, I don’t find Lamarckian evolution at all compelling.

    Consider such things as books, movies, and sound recordings. Buy and large the people who make such things and (attempt to) sell them, intend to make a profit on each and every title. At the same time, they know that they won’t; they know, in fact, that they’ll loose money on the majority of titles and they hope that the money they make on a minority of titles will cover their losses. (Art De Vany’s book, Hollywood Economics, is the best study I know of this, though it’s rather technical). Why do they loose money? Because their intentions are rather useless when it comes to predicting what audiences will respond to. On this or that occasion, they’ll really go for something by an unknown while, on some other occasion, ignoring the latest offering from some superstar. The audiences decide. That seems best modeled by some kind of selection model. That’s what Colin Martindale’s got.

    Consider the music that evolved in the US over the last four centuries or so. The single most important dynamic has been the mixing of African sources and European sources – which I discuss in the final chapter of Beethoven’s Anvil, and in this paper. I find it easy to think about this, informally at least, as a mixing of memes from Africa-derived and Europe-derived populations. Certainly, whatever individuals intended on this or that occasion, what has survived over the long haul is not something anyone planned or designed. I just happened – Ishmael Reed calls it Jes Grew.

    Or consider the English language. The Norman French invaded England in the middle of the 11th century CE and four centuries later we’ve got Shakespeare’s English. Over that period we’ve got millions of people having billions of conversations – and some writing in there as well. At the level of the individual interaction, they’re just trying to communicate with one another. And whatever works in one conversation, well, maybe it’ll work in another. Multiply that by a billion and spread it over centuries and you’ve got a new language emerging. That’s the sort of thing I’m after in talking about cultural evolution.

  8. I agree with John Wilkins that determinism is a red herring. But I don’t think the emic/etic distinction maps well onto the extensional/intensional distinction, though perhaps that’s how it is used here and there in the literature. Trouble is, I simply don’t have a clear sense of how the distinction is used in the anthropological literature, though I’ve read a reasonable hunk of anthropology over the years.

    But the distinction originates in linguistics, in the distinction between phonetics and phonemics. The former is about the psychophsics of speech sound while the latter is about phoneme systems. These are obviously very closely related matters, but they aren’t the same. We tend to perceive the speech stream as consisting of discrete sound entities, syllables and phonemes; this is the domain of phonemics. But the speech signal is, in fact, continuous. If you look at a sonogram of some chunk of speech, you don’t draw a series of vertical lines through it separating one phoneme from another; nor can you snip a tape recording into phoneme-long or syllable-long segments and reassemble it into something that sounds like natural speech. The aspects of the speech stream which are phonemically active differ from one language to another, which is why foreign languages all sound like “Greek.” Independently of the fact that you don’t know what the words mean or how the syntax works, you can’t even hear the phonemes in the speech stream.

    Now, that’s the distinction I’m after, between phonemes and the raw speech stream. And that doesn’t quite map onto the distinction between intension (emics) and extension (etics). The intension/extension distinction doesn’t focus our conceptual imagination on the right things. This is not about sets and how they’re defined. It’s a bit different.

    I know that the linguist Kenneth Pike elaborated the emic/etic distinction into a whole theory of culture. I’ve not read that stuff and my impression is that it’s something of a cult item, though he published some of that work in the best of places (e.g. Mouton).

  9. It's a red herring to reductionists who deny Mind exists, not to rational humans.

  10. The intension/extension distinction

    From a traditional POV that distinction itself a fairly good indicator of something like rationality, or the limits of a purely empirical/associationist approach, at least to human activities (ie, evo-psych). Let's grant that maybe other higher mammals could do something like extension as in denotation. They bark or stomp their hooves at the right times, and win an apple, or he's taught to. Mr Ed. might understand referentiality in some sense. Mr Ed doesn't understand intension, though, whatsoever; it has no language and thus it has no meanings, in terms of connotations or synonymies. Sort of obvious, but in effect the evo-psych person begins by denying meaning; ie you can't apply the i.-e. distinction to the non-human, since it doesn't exist according to your own assumptions.

  11. The workshop will hopefully identify empirical ways to solve conflicts and inconsistencies which may inform future research and collaborations.