| Home | About Edge| Features | Edge Editions | Press | Reality Club | Third Culture | Digerati | Edge:Feed | Edge Search |

THE REALITY CLUB
The Value of Memes:
A Powerful Paradigm or a Poor Metaphor?



Mike Godwin and Jaron Lanier debate the value of memes following Science, Delusion and the Appetite for Wonder, a talk by Richard Dawkins


From: Mike Godwin
Date 12-20-96

Dawkins's powerfully explanatory notion of memes seemed to me at first to have almost casually tossed off in a larger discussion of the dynamics of genetic evolution. Only later did I realize he'd given us a paradigm for understanding how ideas work in cultures, in mass media, and in the growth of knowledge.

It's also a paradigm that gives free-speech advocates some serious social questions to think about. Dawkins's concept of the meme -- that discrete thought that propagates itself, sometimes virulently, through minds and cultures -- forces us to abandon any defense of free speech based on the principle that "words can never hurt you." (Hint: they can hurt you.) Instead, we must defend freedom of expression even though it sometimes allows the spread of *harmful* ideas, because freedom is the only environment that consistently promotes the discovery or creation of the *beneficial* ones.

Together with Karl Popper and Gregory Bateson, whose thinking complements his, Dawkins has done much to shape how I think about the world. He's one scientist who reminds us why we used to call scientists "natural philosophers."


From: Jaron Lanier
To: Mike Godwin
Date: 12-20-96

Hey there Mike,

I just debated Richard Dawkins (it'll appear in Psychology Today, of all places). I'm no fan of memes, though I like Richard, and enjoy other aspects of his thinking. Here's a small part of an article I'm working on that concerns memes and many other ways that evolution is applied outside of genetics.

All the best,

Jaron

Spare me your memes

Biological evolution is a theory that explains the remarkable, creative long term effects of massive numbers of untimely (pre-reproductive) deaths, but it is somewhat immune to variations in the sources of genetic variation from which death culls. The current controversies between scientists studying evolution underline this point. Variation might take place without boundaries or favor, as Dawkins seems to suggest, or might be subject to mathematically predetermined paths, as biologists like Kaufman and Goodwin have proposed. In either case, evolution proceeds, through the mechanism of violence. That the theory of evolution can survive these unresolved controversies shows that it is really the culling and not the sowing that is the key mechanism.

The relative indifference of evolution to the source of variation makes it a poor metaphor for understanding creativity that takes place under the protection of civilization. That is one reason why the idea of the "meme" is misleading. The meme concept, first proposed by Richard Dawkins, is sometimes used to explain how ideas change, but also sometimes as an ideal for how ideas should change. Dennett, in "Darwin's Dangerous Idea" speaks of wishing to extinguish a meme that had infected the physicist Roger Penrose as if it were a freakish individual that should be subject to a eugenics campaign. If it weren't for the romance of evolution, "Memes" would just be a fancy way of pointing out that non-rigorous ideas are often subject to a popularity contest. One danger, however, in the meme idea is an equation of creativity with mental eugenics.

There are so many other things wrong with memes that it's hard to list them succinctly. Equating ideas and genes revives all the worst old wrong ideas about genetics. Ideas do everything genes can't. They can change and effect each other without any concern for species boundaries. They can pass along traits acquired during their "lifespans"- they don't have to wait for some sub-strata of genetic material to be selected for. The long-resolved struggle against these mistaken ideas about genes has been irritated into existence again by a stupid metaphor. It is as if Darwin had never existed.

The notion of memes is an affront to the idea that some ideas can be better than others. Ideas can be rigorous, so the notion of improvement has meaning. Genes, on the other hand, don't improve; they just adapt to local circumstance. And that adaptation is entirely non-intentional and so slow that we learn about it largely from fossils. Many kinds of ideas, on the other hand, can be definitively improved, and this can be done methodically and cumulatively, leading to exponential rates of change. People used to believe God thought the world into existence in just this way, in six days. Darwin's central insight was that genes are not like ideas.

Within civilization, nonetheless, are found pseudo-evolutionary processes, like business and the academic career track, in which competition is harnessed to produce excellence. These should not be understood to be true examples of evolution, though, because the genes of the losers are still passed on without diminution. Even their "memes' are passed on, for those who insist on subscribing to the concept. That is what defines a civilization. If civilization worked like evolution, it would be perfectly ordinary to burn library books that had not been read for a long time. In the real world, when libraries burn, civilizations crumble. Marxism provides a recent example. Ideas are only like memes at the moment when they are extinguished, as happened in the library at Alexandria, or, as might have happened if had he been successful, in Hitler's bonfires.


From: Mike Godwin
To: Jaron Lanier
Date: 12-20-96

Jaron,

As you might expect, I disagree with a number of your arguments. Rather than express my disagreements in great detail, I'll just note some of them here, in a way that perhaps will help you as you further refine your side of the argument. Or perhaps not. It's late.

Biological evolution is a theory that explains the remarkable, creative long term effects of massive numbers of untimely (pre-reproductive) deaths, but it is somewhat immune to variations in the sources of genetic variation from which death culls.

If I understand you correctly here, you're saying that the power of evolutionary theory does not depend on any particular theory as to the source of variation. On that point I agree with you.

So would Karl Popper, I think, were he here to respond to your comment. Popper says something very similar about scientific theories--which might also be called (very loosely) "scientific memes"--in his book CONJECTURES AND REFUTATIONS and elsewhere. In his explanation of the growth of scientific knowledge Popper expressly notes that the *origin* of a theory is irrelevant -- what matters instead is its testability (aka "falsifiability"), which is the indicator of its potential to give us greater knowledge about the world . For example, Kekule's hypothesis about the ringed structure of the benzene molecule originated from a *dream* about a snake eating its tail. But this fact tells us nothing about the value of the the theory, which can only be established empirically.

Thus, dreams, which are arguably the most unstructured and disorded thinking that we ever do, nevertheless can be a source of "variation" as to hypotheses, and ultimately a guidepoint to greater knowledge. Yet even if psychologists were to disagree violently about the relative importance of dreams as a a source of "variation"(read "new ideas), it would not follow from this disagreement that variation itself is relatively unimportant to the growth of knowledge and culture.

Variation might take place without boundaries or favor, as Dawkins seems to suggest, or might be subject to mathematically predetermined paths, as biologists like Kaufman and Goodwin have proposed. In either case, evolution proceeds, through the mechanism of violence. That the theory of evolution can survive these unresolved controversies shows that it is really the culling and not the sowing that is the key mechanism.

I do not believe you have established a syllogism here. I don't see how the robustness of evolutionary theory in the absence of consensus about the sources of genetic variation entails your conclusion that "culling" is more important that "sowing." Both are necessary conditions for Darwin's "origin of species." In fact, Darwin expressly acknowledged that variation was a necessary part of his theory, even though he could articulate no theory as to the source of that variation.

The meme concept, first proposed by Richard Dawkins, is sometimes used to explain how ideas change, but also sometimes as an ideal for how ideas should change.

I think it's unclear to say that "memes" are a notion about "how ideas change." Better to say that they're a notion about how ideas compete with one another, substitute for one another, etc.. (And if "compete" is too telelogical, substitute the verb "interact.") Remember, Dawkins wants us to consider genes as basic units of evolutionary action..

Dennett, in "Darwin's Dangerous Idea" speaks of wishing to extinguish a meme that had infected the physicist Roger Penrose as if it were a freakish individual that should be subject to a eugenics campaign. If it weren't for the romance of evolution, "Memes" would just be a fancy way of pointing out that non-rigorous ideas are often subject to a popularity contest. One danger, however, in the meme idea is an equation of creativity with mental eugenics.

Without going into detail, let me say merely that, in my own experience, thinking about harmful ideas as "bad memes" has been extremely productive for me.

Equating ideas and genes revives all the worst old wrong ideas about genetics.

I think your use of "equating" unfairly dispenses with some of Dawkins's nuance.

Ideas do everything genes can't. They can change and effect each other without any concern for species boundaries. They can pass along traits acquired during their "lifespans"- they don't have to wait for some sub-strata of genetic material to be selected for. The long-resolved struggle against these mistaken ideas about genes has been irritated into existence again by a stupid metaphor. It is as if Darwin had never existed.

It may be that my understanding of genetics has faded since I studied it formally, but much of what you say here about ideas strikes me as self-evidently true about genes as well, .

For one thing, it's not just somatic cells that mutate, but gametic cells as well, and that the latter can pass on their mutations (often but not always deleterious changes). For another, don't ideas require "substrata" as much as genes do?. :Like paper, for example, or air (to transmit sound waves), or a brain?

The notion of memes is an affront to the idea that some ideas can be better than others.

It seems to me to _reinforce_ this very idea. Even we meme-lovers still regard some genes as more harmful than others--harmful either to an organism or to its offspring. Nor does any dispassionate discussion of the dissemination of a meme (a racist meme, say) require that we abandon our opposition that meme. Compare: Does the fact that an epidemiologist can study an epidemic's growth cycle dispassionately entail her abandoning her belief that dying of an infectious disease is a bad thing.

Nothing in Dawkins' metaphor requires us (either as moral actors or as knowledge builders) to think of all ideas as being of equal value *when we are engaged in the process of assessing value*. But the "meme" concept is about understanding the dynamic of the spread of thoughts -- that's where its power as a metaphor lies.,. And the "meme" notion gives us us a way to understand the dynamics of the propagation of ideas that is not clouded by our own assessment of those ideas. In short, thinking about memes allows some of us to see the process more clearly.

Ideas can be rigorous, so the notion of improvement has meaning. Genes, on the other hand, don't improve; they just adapt to local circumstance.

I believe this is both incorrect and a category mistake. Strictly speaking, genes *can* improve (the rare beneficial mutation, for example), and it is not genes but _genotypes_ that adapt.

And that adaptation is entirely non-intentional and so slow that we learn about it largely from fossils.

No problem with your "non-intentional" here, but any bacteriologist, I imagine, can give you what amount to eye-witness accounts of evolutionary adaptation in action. That's one of the nice things about studying the genetics of organisms with short life cycles.

Many kinds of ideas, on the other hand, can be definitively improved, and this can be done methodically and cumulatively, leading to exponential rates of change. People used to believe God thought the world into existence in just this way, in six days. Darwin's central insight was that genes are not like ideas.

I don't recall his saying this. I do recall his recognition that variation is a prerequisite for natural selection. Which to me entails the conclusion that genes are not invariant after all.

Within civilization, nonetheless, are found pseudo-evolutionary processes, like business and the academic career track, in which competition is harnessed to produce excellence.

One can sidestep the road to social Darwinism and still believe that if "pseudo-evolutionary processes" quack just like evolutionary ones, waddle like them, swim and fly like them, why, then we can duck the use of "pseudo." altogether.

These should not be understood to be true examples of evolution, though, because the genes of the losers are still passed on without diminution.

Jaron, I'm not sure I understand your point here, since each of us -- self-evidently the product of our forebears' survival to repductive age -- nevertheless carries in his or her genotype lots of "loser" genes. Unless an allele is lethal to the organism prior to the organism's self-reproduction, the Hardy-Weinberg paradigm more or less still applies, and gene frequencies -- even for ultimately harmful genes! -- in a large population don't change much. (A study of sickle-cell anemia is instructive on this point.)

Commonly it's at the phenotype level that we decide which individuals are "losers" in a particular evolutinary context. -- other individuals who carry the same undesirable allele may well qualify as "winners" in Darwinian terms (they last long enough to reproduce) because their overall phenotype neutralized or minimized the :"loser" effect of that allele. Me, I take Dawkins's argument in THE SELFISH GENE to be in part about transcending this phenotype-centric :"winner/loser" perspect.ive.

I agree of course that one must not *glibly equate* genes and memes. While I still like the notion, I also concede there are countless ways in which this metaphor falls short in representing reality,. Yet isn't this a trivial criticism, given that *all* metaphors -- being comparisons of things that are alike yet also different -- are :necessarily "false" to some degree?.

This irrreducible falsehood of metaphors shouldn't bother us much -- metaphors are meant to be used as tools, not as truths.. And if the tool doesn't work for you, you can abandon it without concluding that it doesn't work for anyone else, either.

Even their "memes' are passed on, for those who insist on subscribing to the concept. That is what defines a civilization. If civilization worked like evolution, it would be perfectly ordinary to burn library books that had not been read for a long time.

As Nicholson Baker has documented, this is in fact perfectly ordinary.

In the real world, when libraries burn, civilizations crumble.

If only this were true. Then book-burning civilizations would invariably die with greater frequency than book-loving ones. But so far as I can tell, all civilizations, including the most literate ones we know of, end up dying, regardless of how nicely they treat their books.

--Mike


From: Jaron Lanier
To: Mike Godwin
Date: 12-20-96

Hello there again,

We do agree on plenty of things. I love Popper's insights on scientific method as much as you do. Alas, no one has yet done such clear work as Popper's to help us choose our metaphors. In examining my criteria for them, and why memes annoy me so, I can propose a starting place: A metaphor ought to inform more than it confuses. Furthermore, it shouldn't unwittingly undermine other notions that one wishes to keep in one's head.

I originally started to dislike memes when I heard students talking about real genes in Lamarkian terms. It turns out they had worked backwards from memes, assuming that ideas must be a reasonable metaphor for genetics in some way. I had to set them straight on that. That set me to wondering if the metaphor worked any better in the forward direction. Since it's very very hard to falsify ideas about ideas, we have to be extra careful about our metaphors for them.

And the "meme" notion gives us us a way to understand the dynamics of the propagation of ideas that is not clouded by our own assessment of those ideas. In short, thinking about memes allows some of us to see the process more clearly.

This I cannot accept. You're making a claim here that you're seeing a process that actually happens, and that you can see it more clearly with the metaphor in mind. First, I worry about the notion of someone becoming a dispassionate observer of ideas, without assessing them. I'm not sure that's possible, and that's a primary problem with the meme metaphor. Can you identify an idea by superficial features, like you can identify an organism? Is it possible to identify an idea without internalizing it? The example I cited in Dennett is not the only one I've seen in which the meme metaphor serves as a tool to help the bearer become somewhat cynical and distanced from the ideas of others.

But I also wonder what process the metaphor of memes can help you observe. Where is the genetic material for an idea?

For another, don't ideas require "substrata" as much as genes do?. :Like paper, for example, or air (to transmit sound waves), or a brain?

You suggest paper and air, but those aren't linked to specific ideas in the way that a particular set of genes are linked to a particular organism. Maybe the metaphor could be lined up in different ways; to the genotype, or wherever, or maybe the idea is like the gene and a behavioral action is like an organism. I've tried to find a way to make the metaphor work! No matter how I try, I can't find a reducible sub-strata in the life of ideas to hang on it. If the meme metaphor informs, it should be possible to name this sub-strata. Can you name it?

Ideas can be rigorous, so the notion of improvement has meaning. Genes, on the other hand, don't improve; they just adapt to local circumstance.

I believe this is both incorrect and a category mistake. Strictly speaking, genes *can* improve (the rare beneficial mutation, for example)

In this case I think you are being confused by putting the meme metaphor into reverse gear, like my Lamarkian students. Surely adaptation is only local, while a mathematical theorem is global. A scientific idea, once falsified, is permanently falsified, while a vanished genetic feature might someday reappear if local circumstances change to once again favor it.

And that adaptation is entirely non-intentional and so slow that we learn about it largely from fossils.

No problem with your "non-intentional" here, but any bacteriologist, I imagine, can give you what amount to eye-witness accounts of evolutionary adaptation in action. That's one of the nice things about studying the genetics of organisms with short life cycles.

You're right on this point. What I meant to say is that the genetic rate of change is far slower than the pace of events in the life of an organism. If the meme metaphor informs, once the "genetic" sub-strata has been named, it ought to change very slowly, relative to the pace of discourse. Or if the metaphor should be lined up differently, and the ideas are the genes, there ought to be a faster moving "organism" equivalent that speeds past our ideas.

Evolution is an evil thing. All your genetic features are the result of the pre-reproductive deaths of your would-be ancestors. They were killed in cold blood by your real ancestors, or by micro-organisms, or by cold or hunger. Your features weren't decided by a nice process. If we really want to understand human discourse by making a metaphor with the heart of cruelty, we ought to have a good reason.

I'm not saying the meme metaphor never works at all. When the last copy of a book concerning non-rigorous ideas is destroyed, I think the metaphor might start to work a bit. You could say the book is like genetic material, slower moving than discourse, with discourse being the organism, and that future discourse on related non-rigorous ideas is shaped a bit by the book's absence. While this does happen, the meme metaphor is most popular in the sciences, where it doesn't fit.

For what it's worth, when I presented my arguments to Dawkins, he agreed with them, and said he thought "memes" had been taken too far. You can read what he says about this in his own words in the Psych Today piece, when it comes out.

All the best,

Jaron


From: Mike Godwin
To: Jaron Lanier
Date: 12-20-96

In examining my criteria for them, and why memes annoy me so, I can propose a starting place: A metaphor ought to inform more than it confuses.

Well, perhaps it says something that I disagree with your "starting place" premise. I'm not sure I can say with precision what it is that metaphors do when they aid in understanding, but I don't think "inform" is the right verb. As I said previously, metaphors are tools, not truths. Kind of like what (as I recall) Wittgenstein said the Tractatus should be considered as -- a sort of ladder to the next level that you can throw away once you're up there.

I originally started to dislike memes when I heard students talking about real genes in Lamarkian terms.

If undergraduate misuse of newly acquired notions is all it takes to generate your initial dislike of those notions, I begin to shudder at the implications. (This is a joke.)

Since it's very very hard to falsify ideas about ideas, we have to be extra careful about our metaphors for them.

I'm inclined to say that Dawkin's "meme" notion is simply a metaphor and not a scientific theory. A very powerful metaphor, true, and perhaps even a harmful one. But not something whose unfalsifiability I'd normally worry much about.

And the "meme" notion gives us us a way to understand the dynamics of the propagation of ideas that is not clouded by our own assessment of those ideas. In short, thinking about memes allows some of us to see the process more clearly.

This I cannot accept. You're making a claim here that you're seeing a process that actually happens, and that you can see it more clearly with the metaphor in mind.

The problem is less my proposition, I think than it is my poor usage. Rather than "see the process more clearly" ( a phrase that connotes actual observation), I should have written something like "think about the process more clearly."

You may still disagree with the amended claim, but I don't mean for it to be taken as a claim about observations.

First, I worry about the notion of someone becoming a dispassionate observer of ideas, without assessing them.

I believe this is a false dichotomy, since (in my view) one can be a *passionate* observer of ideas (and of other human creations) without imposing a value system upon them. Some of my anthropologist friends, for example, seem to me to be doing just this.

Can you identify an idea by superficial features, like you can identify an organism?

I'm not sure what you're getting at with "superficial" here, but I do think ideas can be classified by clearly discernable features. For example, I believe this is what Popper does with his science/nonscience demarcation criterion.

Is it possible to identify an idea without internalizing it?

I think so. For example, I believe I can identify a Marxist proposition without adopting it.

You suggest paper and air, but those aren't linked to specific ideas in the way that a particular set of genes are linked to a particular organism.

When you used the word "substrate," I found myself thinking of nucleic acids, which, of course are no more specific to a particular gene than paper is specific to a particular idea. I'm still not sure I follow your reasoning here.

I believe this is both incorrect and a category mistake. Strictly speaking, genes *can* improve (the rare beneficial mutation, for example)

In this case I think you are being confused by putting the meme metaphor into reverse gear, like my Lamarkian students. Surely adaptation is only local, while a mathematical theorem is global.

Actually, your response suggests a rather different confusion. I don't believe "local" and "global" are terms that represent objective reality.

Popper might have said that a mathematical theorem actually *is* "local" -- it is located in what Popper calls World 3 (the shared domain of human ideas) and it is *not* located under under my bed.

I don't think your local/global distinction is helpful, but you may be reaching for something like the a priori/a posteriori distinction. In any case, once again I have trouble following you.

A scientific idea, once falsified, is permanently falsified, while a vanished genetic feature might someday reappear if local circumstances change to once again favor it.

Popper would say that falsified scientific theories remain in World 3. (They're just reshelved in the "falsified" section.)

I was taught that vanished genetic features *never* simply reappear. E.g., the mammalian species that returns to the sea does not grow scales, even though its long-ago forebears may have had them. Instead, it develops analogous structures or perhaps even arrives at a wholly different solution to the adaptation problem.

You're right on this point. What I meant to say is that the genetic rate of change is far slower than the pace of events in the life of an organism.

This is absolutely right, IMHO, and, incidentally, one of the implications of the Hardy Weinberg equation (or so it seems to me).

If the meme metaphor informs....

Again, I'm uncomfortable with the assumption that metaphors "inform."

Evolution is an evil thing. All your genetic features are the result of the pre-reproductive deaths of your would-be ancestors. They were killed in cold blood by your real ancestors, or by micro-organisms, or by cold or hunger.

Some of them were just too lazy to fuck, Jaron.

I'm not saying the meme metaphor never works at all.

The science of metaphors is never a precise one, I'm thinking.

For what it's worth, when I presented my arguments to Dawkins, he agreed with them, and said he thought "memes" had been taken too far.

Although I disagree with some of what you see, I certainly agree with you and Dawkins (and Danny Hillis) that the notion has been taken too far.

Of course, when my book comes out this spring, you may find that its prolix discussions of memes and media damn me as another culprit in the current meme overload. I'm wincing in anticipation.

Take care.

--Mike

 


Back to Science, Delusion and the Appetite for Wonder by Richard Dawkins



| Home | About Edge| Features | Edge Editions | Press | Reality Club | Third Culture | Digerati | Edge:Feed | Edge Search |

| Top |