A Neutral Theory of Memetics

i-4f4e40ca05c55fa1569d69cc450aeb64-The_Selfish_Gene3.jpg

In his classic book from 1976, The Selfish Gene, Richard Dawkins introduced the concept of the meme. Basically, a meme can be thought of as a cultural gene -- an idea that is transmitted in a population.

This being Dawkins, memetics has a certain adaptationist flavor to it. The Selfish Gene introduces evolution from the gene's eye view. But Dawkins is a staunch advocate of natural selection as the primary driving force behind evolution. The debate over the role of natural selection in evolution has been carried out using empirical data, theoretical modeling, and philosophical arguments. I won't get into any details of it here.

What I am interested in is the adaptationist nature of memetics. From what I have seen, memeticists argue that memes survive or fail based on their ability to propagate through a culture. Memes that are good at infecting new individuals persists, while those that are not go extinct. The probability of extinction depends on some properties inherent to the memes and the culture they inhabit.

i-0b9d25b7e63d529fe135b8496da6d3ce-dennett_pss_sm.JPG

Daniel Dennett has taken a prominent role as an advocate for memetics. I spoke with him a few weeks ago and asked if there was a neutral theory for memetics. In his reply, he suggested that neutral evolution is pretty irrelevant. With such a position, it's hard to imagine that he doesn't spend much time thinking about neutral memes.

Memetics itself is a controversial field of study -- in the sense that some people don't think it's worth studying at all. But, assuming the idea of the meme is a valid one, does it make sense to consider neutral dynamics of memes? That is, propagation of ideas merely because of their historical contingencies, and not because they are good at propagating themselves. For example, consider a cultural practice that is common because it happened to spread merely by chance. Perhaps it hitchhiked along with an adaptive meme (prayer with religion?). Or maybe it just got lucky.

On the other hand, there may be no need for a neutral theory. Perhaps coming along at the right point in time is the adaptive feature of that meme. So, are there neutral memes? Or does it not make sense to think of neutral memes?

More like this

You're probably right that a lack of interest in a neutral theory of memeflow is related to adaptationist presuppositions. There's a problem with such presuppositions, though, since fitness for memes can't simply be "read off" demographic statistics any more than the fitness of genes can. We need at least a "plausible story" about how particular ecological/environmental factors interact with the trait in questions to determine its fitness in that environment.

By bob koepp (not verified) on 27 May 2008 #permalink

see boyd & richerson or feldman & cavalli-sforza's work using population genetic models to explore the theoretical possibilities of cultural transmission. they even talk about analogs to selective sweeps and the hitchhiking of characters around them.

A quick question: WHY would we want to assume that the idea of the meme is valid?

Why not, say, just assume the concept of an all-powerful creator God is also true while we're at it?

Or how about the trans? What about that new entity (the teme) Susan Blackmore introduced recently at TED?

How is the meme anything but a contemptible elision of the difficulties of explaining human cultural phenomena?

(On the trans, See Martin Gardner's review of Susan Blackmore: "In all human cultures, even in chimp society, objects not connected to the body are shifted from place to place. Call every such move a "tran," short for translocate or transfer. Moving our shoes when we walk, run or dance is not trans because the objects are attached to our body. Nor are the movements of things in cars, trains, ships, planes and elevators examples of trans because the propelling forces are independent of us even though we may direct such movements.
Examples of genuine trans abound. The motions of pitched and batted baseballs are obvious trans, as are the movements of
objects in dozens of other sports: football, basketball, bowling, tennis, golf, hockey, pool and so on. When a chess player pushes a pawn, it's a tran. Dealing playing cards is a tran. Raking leaves and moving vacuum cleaners and using dust busters are trans. Serving food and washing dishes are trans. Hammering a nail and sawing wood are trans. Eating is a tran because food is moved from plate to mouth, though swallowing it is not because the food becomes joined to the body. ... There are tens of thousands of other examples.
A vexing question arises: How should we distinguish trans from transplexes? The flight of a pitched baseball is a tran, but if the ball is hit, caught and tossed to first base, is that familiar sequence a tran or a transplex? Shall we call an entire inning a tran or a transplex? Should transplex be reserved for a complete game, with its hundreds of trans?
What is gained by introducing the concept of a tran? Nothing. Trans are no more than a bizarre terminology for saying what is better said in ordinary language. We don't need a new science of tranetics to tell us that in every culture, persons move things. ..."

Using genetic or epidemic models for cultural and social phenomena actually goes back to Bartholomew's 1967 book, Stochastic Models for Social Processes -- maybe earlier, this field is obscure before then.

Here's an easy example, which may or may not have been studied already: accent. We pick up -- are infected with -- an accent from the crowd of peers who surround us. If by chance the strains of accent in the environment become small in number (maybe some tribe comes in and wipes out a lot of people), then it's likely that one of the remaining will become standardized. It wouldn't be due to greater ease of acquiring or anything like that.

Actually, consider the sound-meaning pair for a word. There's no reason to pronounce the concept "duck" the way English speakers do -- nothing ducky about the sounds. Why did one string of sounds become standardized in the English-speaking population, while other hopefuls were weeded out? Again, probably chance.

Saussure actually observed the arbitrary relationship between sound and referent in the early 1900s.

But is a duck "duck" because of chance? I doubt that very much.

My guess is that it has something to do with who was more powerful & perstigious at some crucial point in history.

The thing is, we have no reason to believe that duck and alternative-to-duck compete in any way. They're probably just part of the baggage in another, higher-level competitive struggle.

I'm willing to accept that culture is influenced by evolutionary forces.

I am unwilling to grant the existence of memes, the usefulness of a meme's eye view, or even the proposition that memes are anything but a detriment to human inquiry regarding cultural phenomena.

Given that there is no neurological reason to believe that ideas are stored as "memes", "memetics" cannot be taken seriously, adaptive or not. It's a cute analogy, granted, and, yes, it's fun to compare popular ideas that one dislikes to viruses, but what makes the study of the evolution and transfer of genes meaningful is that genes have been demonstrated to exist.

I didn't mean to imply that the sound-meaning pairing is completely arbitrary -- for one thing, it'd be a pain to have a 32-phoneme word for "and." And yeah, there are some ways that sound influences meaning, like small things tend to get high and front vowels, while big things tend to get low and back vowels.

I mean, in general, the reason we say duck and the Spanish-speakers say pato is not a question of survival of the fittest.

Oran -- powerful people generally have little power over how languages evolve. It's a grassroots thing. You haven't studied anything we're talking about, even on a superficial level, so who cares what you think? I know I don't mind losing a lazy thinker.

Given that there is no neurological reason to believe that ideas are stored as "memes",

The use of meme is to model how cultural units are spread, not how they're stored -- the term meme encompasses things beyond ideas, like hairdos, for instance. Scientists know jackshit about neuroscience anyway, at least at the level you're describing, so who cares?

It's a cute analogy, granted, and, yes, it's fun to compare popular ideas that one dislikes to viruses, but what makes the study of the evolution and transfer of genes meaningful is that genes have been demonstrated to exist.

So we're just hallucinating that there are ideas, accents, ways of dressing, words, etc.? Again, if you'd read any of this stuff on even a superficial level, you'd know that epidemic models are pretty good at modeling these things, especially fashion cycles.

Seriously dude, even the science-related blogosphere has become one giant "everyone's a pundit" policy wonk wankfest. Democracy doesn't work, Wikipedia has proven that. There should be a "type in the letters" thingie to comment, only it would test your IQ and background knowledge relevant to the post. If it were a "for fun" post, then anyone could comment, but not on other things.

The use of meme is to model how cultural units are spread, not how they're stored -- the term meme encompasses things beyond ideas, like hairdos, for instance. Scientists know jackshit about neuroscience anyway, at least at the level you're describing, so who cares?

I care, for starters. One of the most dangerous fallacies is reification -- the erroneous treating of abstractions as real concrete things when they aren't at all. For something to "spread" in a real, rather than metaphorical manner, it must be more than an abstraction. What made biology really take off after the discovery of the double helix is that genes were finally demonstrated to not be abstractions open to debate but real physical things open to experiment.

So we're just hallucinating that there are ideas, accents, ways of dressing, words, etc.?

No, but it is safe to say that we don't know the concrete basis (if any) behind these abstractions.

Again, if you'd read any of this stuff on even a superficial level, you'd know that epidemic models are pretty good at modeling these things, especially fashion cycles.

And you can use cellular automata to model the movement of birds. But just like your fashion example, it isn't that interesting unless one can show that something physical in a bird is actually acting like a cellular automatan.

Agnostic:

So the reason why I in Michigan say "duck" rather than the Ottawa word for duck or the Gaelic word for duck has nothing to do with power?

And even if we discount inter-group struggle, precisely how much data do we have on groups without a common language choosing words for things? Very little, I'd suspect.

Lazy? Like not bothering to find out that Saussure is the locus classicus for the basic observation you are making?

I am not worth arguing with if you think you can just browbeat me into submission or pretend to know something you don't. The "bow down before me, I once read a book" routine doesn't cut much ice with me.

And the fact is that I do have a basic grounding in linguistics. But since you apparently know so much more about it than I do, I look forward to learning all kinds of interesting new things about how signifiers compete with one another in language formation.

It's interesting that you are so ready to toss around insults about underqualification to comment. What, precisely, have you contributed to the conversation aside from insults?

Anyhow, memes are a model of cultural dissemination, but they're a bad one. For one thing they are pretty explicitly an attempt to remove or minimize the interests of the human agent from the model.

Why? Because what Dawkins had in mind when he theorized memes was religion (you can read the chapter in Selfish Gene if you like). And one of the things that Dawkins has always wanted to argue is that religion is useless to humans, and one clever way a strong adaptionist can argue that something that is universal and enduring is useless is to introduce the interests of another agent into the equation.

Thus memes.

They're nothing but an argumentative trick, and they're fatally flawed because the very thing the concept is designed to do--minimize the role of human agency--makes for an oversimplification.

Of course, you can model idea dissemination epidemiologically. One part of epidemiology is the study of the dynamics of contact and commerce. This can model the spread of practically anything this way: it doesn't mean that the thing being disseminated has interests of its own.

Seriously, dude, if anyone in this thread is suffering from signal-to-noise problems, it's you.

What made biology really take off after the discovery of the double helix is that genes were finally demonstrated to not be abstractions open to debate but real physical things open to experiment.

What was Thomas Hunt Morgan, chopped liver?

What was Thomas Hunt Morgan, chopped liver?

I was wondering if someone was going to bring him up even as I writing my response :-). Yes, in the case of genes, there were multiple stages from abstract to concrete, and the structure of DNA was merely the ultimate demonstration. Morgan's "beads on a string" were certainly more concrete than Mendel's "factors". And Avery's "transforming agent" was more concrete still.

I still want to quibble with the abstract-to-concrete thing a bit: once it was "one gene one protein" and now a gene is perhaps again something more "abstract". In genetics, there were two different questions: "what is the unit of heredity?" and "how is genetic information stored biochemically?". Not knowing the answer to the second question did not stop people from studying the first question, although it went slower. The meme theorist analogy to the first question is "what is the unit of cultural transmission?" I'm not saying that there's necessarily much to meme theory, but your criticism seems to miss the mark.

Windy:

With Mendel we have work that mapped out a schema of trait transmission, showed us that there were species where this schema worked reasonably consitently and allowed the proposal of a fundamental unit of transmission.

The first question was NOT "what is the fundamental unit of transmission." The first question was "is there a consistent logic to heredity." Only after sketching an answer to that out was it possible to codify Mendel's work as the proposition that there were fundamental units, that they interacted in particular ways (with greater or lesser complexity) blah blah blah.

I don't think anyone's done Mendel's work in this field yet. The question of whether there are fundamental units is simply premature. Positing fundamental units of cultural transmission--whose properties are all inevitably modelled on fundamental units we already DO know something about: genes--is at this point nothing but a way of queering the entire process of investigating cultural transmission.

And it isn't a disinterested mistake: the distortion is precisely to make cultural transmission a version of genetic transmission and for science to essentially colonize the study of society and culture.

Oran: And it isn't a disinterested mistake: the distortion is precisely to make cultural transmission a version of genetic transmission and for science to essentially colonize the study of society and culture.

The problem isn't that science is colonizing sociology and anthropology --- that has been the explicit goal of those fields for over a century. The problem is that memetics is a terrible way to do so, as you pointed out. And it's probably because there are no essential units of transmission of culture - the units are domain specific.

Memetics is like trying to apply genetics to signalling at the membrane. Genetics only applies to one tiny subset of information transmission in organism (an important one, but very domain specific).

There have been a number of attempts at applying the neutral theory to "memetics," including most importantly the Boyd and Richerson cites given above. For a recent example see:

Hahn and Bentley (2003) Drift as a mechanism for cultural change: An example from baby names. Proc Roy Soc B.

By your_confabulator (not verified) on 29 May 2008 #permalink