This site is not maintained. Click here for the new website of Richard Dawkins.


← Meme Theory, Zahavi's Handicap, and the Baldwin Effect

Zeuglodon's Avatar Jump to comment 8 by Zeuglodon

Comment 3 by Bumpy

ZEUGLODON: So, quite apart from the high mutation rate that this Lamarkian evolution allows, it shows that ideas don't make copies of themselves any more than a paper makes a copy of itself - the photocopier, and in the brain's case the perceptive and motor neurons, have to do all the work. I think this is the biggest objection to memetics, and the main reason I'm sceptical of it.

I don't see why that's problem. Yes the paper doesn't physically forge a copy of itsef - it presupposes a massively more complicated sequence (which the paper didn't create) to do everything for it (i.e. a human has to feed it into, and operate, the photocopier).

Now how is this qualitatively different to what a genetic molecule does? It too presupposes a certain environment - one with electrostatic forces, raw materials to interact with, a certain tempreture range, the contingency that the laws of physics are unchanging etc.

The genetic molecule is such that in the appropiate environment it will be copied. A piece of paper that reads "please photocopy me" is such that in the appropiate environment (one with a photocopier and a very obedient human) it too will be copied.

It is very different, because the exact same argument could be used to buoy up a version of group selection or genome selection for an asexual species. What you are saying is that active replicators and things being replicated both need certain conditions, but this is trivially true whatever process we are talking about. It's the differences between the two that I want to highlight.

The implications cannot be reduced to an unequal assigning of conditions needed for replication. The difference is in the replication mechanism, and to ignore these differences or dismiss them on the grounds that conditions need to be met in either case just sweep the issue aside, not address it.

Genes have autonomy and, when presented with the necessary components of their own replicator molecule, actively create copies of themselves. The machinery involved in cells is actually to refine the process, not to do it for the genes, because a replicator that improved its own replication process would be selected by the very same logic that made replication a viable start-up for evolution in the first place.

The problem is that memetic theory presupposes a similar autonomy for memes which I don't think they have. It's not enough to have a copy made for something - that something must be able, under its own steam, to make copies of itself without intervention by another process. It has to be a genuine replicator before natural selection can act on it. If ever memes have natural-selection like properties, it could only be because they align with genetic interests. The instant there's any divergence, genes win and memetics is revealed as a convincing illusion.

Comment 4 by OHooligan

I'm heartened to see several articles involving "meme" appearing here recently. I've recently re-read The Selfish Gene, especially the New Replicator chapter, and don't really have anything to add to it, I'm just rephrasing things my own way to get the concept comfortable in my own mind. I strongly recommend you (re) read that chapter too, if you haven't recently.

I have it open in front of me right now. I'm fully aware of Richard's own stance on meme theory, which is that he doesn't fully commit to it as much as many suppose he would. But even if he did, my argument still holds and has to be considered on its own grounds, however shaky or solid those grounds turn out to be.

An attractive part for me is that "memes" explain things at a higher level. You don't need to keep returning to the genetic level, any more than molecular biology has to keep on reminding itself of the details of quarks and gluons.

This is an unequal comparison, because invoking the properties of subatomic particles for replicators is, as far as explanation goes, unnecessary. It's an obvious distraction, the equivalent of discussing atoms when telling a driving enthusiast or an apprentice car mechanic how a car works. But with brains built specially by genes, any discussion of biological behaviour needs to refer back to genes at some point because replicator molecules with phenotypes form the core of any biological explanation. It is difficult to make sense of an organism's design without considering genes.

One quick aside: The thing about memes explaining things at a higher level is that they are not higher level explanations. They cannot be a higher level. They have to be autonomous entities, so they have to be a parallel development working alongside genes. That's very different, because then you acknowledge them as two separate replication entities with diverging interests. Group selection would be a higher level explanation because it has to refer back to genes by its very nature (of course, the problems of doing so are what make it an unsound idea).

I think that emphasis on the mechanism of replication is misguided, and overly parochial.

I think it isn't. An asexual species may look like it's body or its genome is replicating, and we might then be fooled into talking about clone selection or genome selection. But it's all genetic however you look at it. The products of genes do not have replicator autonomy, and it's this illusion that makes gene selection a brilliantly successful theory but clone selection and genome selection a waste of time.

I think that comparison between "memes" and complex multi-cellular creatures such as mammals is also a mistake.

I'm not sure I understand why you said this. I was not trying to explain memes with reference to, say, mammals as though the two were comparable. My focus is on what properties memes have and whether real ideas have those properties, not on whether memes are like mammals.

I think that conflating "meme theory" with "group selection" is an attempt to discredit the former by association.

Nothing of the sort, and I would be displeased with myself if this was what it amounted to. Instead, my point is that they both have something in common that should make people stop and think: they look superficially like entities under natural selection, but when you examine the mechanisms behind them, the process is actually not natural selection of replicating entities, but a mechanism that actively mimics it. In the case of group selection, the proposed examples end up being gene-centric explanations like kin selection, herding behaviour for personal benefit, and reciprocal altruism. In memetics, it's that the supposed spontaneous replication is actively orchestrated by genes who have to build brains that make a concerted effort to blindly reproduce the mental patterns hidden in other brains. The implications couldn't be more profound, because this suggests that groups and ideas will never take precedence over genes in the long run. When they do, it's because their interests and those of the individual organisms align.

The genes want brains to be similar to each other. The ideas themselves don't. And I want to know why.

I think that running the idea down is a lot less interesting than running with it.

Running the idea down is only meant to hit it with scientific scepticism and see if we can't come up with a better hypothesis. It's actually bad science to run with a hypothesis without examining it more closely.

Comparing memes with viruses is IMHO much more useful. Neither can replicate by itself, but require a complex host to do the work. Viruses need cells, memes need brains. Not just any old cells either, the viruses that bother us need the cells residing in individuals of our own (or similar) species. It does not damage the analogy to note that all the candidate "memes" detected so far require human brains for their replication.

But even this comparison has its problems. A virus RNA physically enters the cell, creates copies of itself by actively hijacking the machinery, and then physically bursts out of the cell to enter another one. It's a genetic predator. Even on its own, a virus is capable of making a copy of itself, but it so scarcely finds the right ingredients, and in any case there's a bounty of ingredients and of refined copying mechanisms packed away inside cells. Viruses, like planktonic predators, want to find their prey and take full opportunity when they do find it.

By contrast, ideas do not physically leap from brain to brain. An idea in one brain has to be converted by neural mechanisms so that information can be sent to the motor nerves. The motor nerves are where the buck stops for electronic signals. After that, the body moves or does something, which affects the light patterns or the sound patterns around it. These changes fight their way through the noise of the environment, whereupon they meet the sensors of another brain.

The thing is, the sensors themselves receive a flat and uninteresting income of light, and cannot by themselves make anything out of the light source. Brains need to actively reconstruct the environment in something called inverse-optics, which by its very nature requires assumptions to have been already installed in the brain. To get an idea of what a brain has to do, imagine trying to work out whether the 12 you are given is a product of 2 times 6 or 3 times 4.

Now for the interesting part. The sensors send as much information as possible to the perceptive cortices of the brain, which then have to make a serious attempt to interpret the signals and reconstruct the shape of the world outside. They are actively designed, in the case of humans, to reconstruct what's in another brain, and do it blindly, trying to filter noise out of the signal. They never get a chance to check directly whether the result matches the other brain. Moreover, the machinery is doing all the work, but it wouldn't do so if it hadn't been built by genes for a specific purpose. The genes could just as easily dismantle the machinery in evolutionary time, which is why giving the ideas the credit of all the work seems to me so misguided.

This machinery, as shown with all other animals with nervous systems, can occur without memes. Millions of animal species communicate all the time, though when their brains reconstruct the signals, the reconstruction does not have to match that in the sender's brain. There are no mirror neurons in the majority of cases to make an exact match. So the mirror neurons, say, or whatever is the equivalent, are key to this "replication" process. And nothing gets in the brain's head unless genes wanted it to.

To put it in summary: viruses do, as a matter of fact, replicate, but they replicate much faster if they hijack a ready-made host cell and actively use its machinery to pump out copies of itself, which then burst out and physically invade other cells. Ideas never leave their host brains, but nearby machinery send proxy signals which hit other brains. The brains are so designed that they themselves create what seems to them a parallel structure found in the other brain. The idea itself is passive throughout.

For now, memes are only copied via the action of human brains, with our without photocopiers. That need not continue to be the case.

To dismiss the photocopier is to miss the point. My point is that the agency should be granted to the replication process and not to the passive thing that gets replicated. If the two are one and the same, as with genes in RNA or DNA, then we have an autonomous replicator. Memetics is the equivalent of claiming that ideas are both the thing being copied and the thing doing the copying, and these two roles need to be correctly assigned because it's misleading to give credit to the wrong thing.

In any case the analogy appears sound (I hesitate to call it a theory, but maybe it is one). Especially the part that explains how a meme (like a virus) can thrive despite being bad for the hosts that perform the replication.

But how bad is bad for the host? Yes, you can get "epidemics" in which a group holding the same idea kills itself, but this is like when a deer with congenital bad eyesight blunders into a predator in the bush - just a natural byproduct of genes trying their best to implement proxy strategies to meet whatever the environment demands, and having to accept limitations as part of the package. It's reconcilable with the framework of any idea of a genetic feature occasionally causing an error.

It also allows for the field of memetic engineering, in which artificial memes are deliberately constructed and deployed. Scientology might be an example.

In this case, the deliberation comes from the rest of the brain, not the idea itself. Ideas don't come with their own desires - they just sit there. They are manipulated by the rest of the brain using them for the good of the whole organism, or at least insofar as that aligns with genetic interests.

Taking the "memes-eye-view", analagous to the "genes-eye-view" promoted in The Selfish Gene, will (once again IMHO) provide compelling explanations of the way people behave, and provide a framework for developing defenses against the more pernicious "viruses-of-the-mind".

This is another reason why I made the comparison with group selection. The claim that taking another's-eye-view provides extra explanatory assistance may well be true, but we could take the viewpoint of any feature of the natural world - even an electron or a photon - and this by itself doesn't justify assigning it certain qualities. If anything, it can be positively misleading, the equivalent of mistaking the puppet for the puppeteer and wondering what the puppet would think of doing next.

But, being an analogy, I expect it will break down at some point. Finding that point is worthwhile, and I hope discussions like this help that goal.

Once again, thanks for starting this discussion.

No problem. :-) Any attempt to find out what's true needs a checkup every now and again.

Comment 6 by Schrodinger's Cat

You missed the purpose of the thought experiment. The point is that the mutation rate cannot be too high, or it will destroy any serious attempt to call an idea a replicator. The bible example illustrates my point: the mutation rate that we do see in the four gospels is stark. Even if you presuppose intermediate texts between the originals back in the first century, the train always looks like an X-Y-Z-A chain of causality. It's only recently that exact copies of books can be made and mass produced, and even then it only lasts one generation. Genes are literally identical for tens, if not hundreds or thousands of generations before a mutant strain turns up. Ideas vary from generation to generation, even if you account for a shorter time frame.

Even the biological pointlessness of some cultural practices, such as religious duty or art that distracts one from looking for mates, isn't a counter, because genes already have a mechanism by which they prosper, paradoxically, by going against their own interests. Peacock tails would be ironed out if they were under the selection for aerodynamic streamlining or for avoiding predators, but not as a means of advertising genuine phenotypic prowess, as described by the Zahavi-Grafen handicap principle. And even artists succumb to basic needs like eating, toiletry, and sexual lust. The rarity of artists that take it to an extreme is akin to the rarity of peacocks with tails twenty feet long - genes have to strike a balance between self-handicapping and not committing genetic suicide. This is why people, in practice, strike a balance between biologically frivolous activities and biologically necessary ones.

Biologically frivolous practices themselves emerge as byproducts of multiple genetic phenotypes; art, for instance, is a combination of using perception and using manipulation and the brain being intelligently wired to experience pleasure from engaging in both, which in prehistory would have meant intelligently interacting with the environment and spotting something interesting or achieving something interesting.

Fri, 20 Jul 2012 12:04:34 UTC | #949625