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 21 by Zeuglodon

Comment 19 by OHooligan

It's nothing like discussing atoms with a car mechanic. I'm trying to check whether an idea ticks all the boxes needed to qualify for replicator. That means asking:

  1. What is the structure, the unit of replication?

  2. Does it make discrete copies of itself?

  3. Is the number of generations large before it hits its first mutation?

To answer one, I zoned in on the structure in one brain, say that could be representing "christianity" or something. Say we represent it by X.

To answer two, you look at two brains. X is in one brain, but not in the other. The replication process occurs. Now, X is in one brain and an X is in the other.

To answer three, you look at more brains in a chain. If it goes only one brain more before X becomes a Y in the next brain, that's hardly enough time for any kind of meaningful spread to occur. Even if it was like one person from generation one told twenty people from generation two, that would still be one generation. If, on the other hand, it goes a hundred or a thousand brains more before X becomes a Y in the next brain, that's plenty of time for a meaningful spread to occur. We'd be able to call it evolution, though not automatically natural selection.

What you call me using the microscope is actually me looking at the right scale - by analogy, watching the moment when an RNA strand makes a copy of itself. And when you look, you notice two things:

  1. I would therefore have to look at nervous system nets such as those in the brain to show one. I have to identify the unit, and that would be a nerve net that represents the idea in one head. To take a broad view is bad policy here because the unit isn't found by assuming it exists and jumping straight to the natural selection metaphor.

  2. I question whether it makes discrete copies of itself on two grounds: the first is the high mutation rate between generation zero (my brain) and generation one (your brain); the second is the stick insect argument.

Genes are crisp, digital, separable from the genome and isolate-able. It's not arbitrarily picked, as you misinterpret it, and it doesn't mutate after only a handful of generations. It's described, in RD's case, with reference to an allele. Any change is not a gradual thing, but an either/or state. A mutation changes a genome in the blink of evolutionary time, and then the resulting genes competing with their alleles follow the same algorithm as blindly as before. Genes are not terms of "convenience". You misunderstood, and subsequently exaggerated, the observation that a gene is not one of a string of beads along the genome, because for all practical purposes it is.

Ideas aren't beads even for practical purposes. When I tell you about christianity, the idea in your head is guaranteed to be a mutation of mine, especially if my explanation is complicated. And if you tell somebody else, another mutation is added. Bits will pass on unscathed, true, but the mutation rate will be so high that it'd be a game of Chinese whispers. The number of generations for this would be too low before the idea morphed beyond recognition compared with the original. And weirder still, it's recursive. Generation one can be fed by generation ten a completely new idea that is, ironically, the result of the idea it passed on.

Any common ground between people - for instance, their understanding of alphabets and language instincts - are the products of genetic phenotypes, just like a stick insect's leg and brain. Damage to these will not be passed on in the next generation. I'd understand what an alphabet is even if my ancestor received a head blow that destroyed his or her language circuits. Those bits of the idea that everyone automatically gets are those bits that didn't replicate, so memetics is as unnecessary as clone selection here.

You also fail to appreciate the significance of the virus RNA. All a virus RNA cares about is meeting nucleotides that, when it is exposed to them, automatically arrange themselves into a copy. The fact that it can exploit the goldmine of stuff in cells follows evolutionary logic: why waste time looking for them elsewhere when you can specialize in parasitism on a nearby and available bounty? In the very early days of replication, simply drifting about bumping into material like plankton do would have to be part of the process. RNA is not a helpless little thing that has to ask organelles to read it and build a copy without it. Give this a moment's thought and it should be apparent that a replicator needed to get life started could never be such a thing. It had to be something that replicated under its own steam. I told you that modern organelles are refinements of that process, with the RNA still hanging around and the DNA acting as a kind of database for it to refer to. RNA has to physically touch the matter needed to make RNA before it can replicate. An idea in your head behaves nothing like this because it is helpless. The machines have to do the dirty work for it. Moreover, they're not doing it for it. The genes have made large-scale copiers for their own purposes.

And this is putting it mildly. Far from the copying mechanisms being spontaneous, it's so mutational and more akin to straightforward causation that the few times an idea lasts more than a few generations without changing at all are rare. Too rare for evolution, never mind for any kind of selection process, to occur.

I maintain you're running with a half-baked idea. That is the last thing you should be doing.

The alien analogy you provide doesn't work. The transmitters don't make more transmitters. They are already set up by the genes long ago. The information "sent" is simply another form of causation, but a transmitter never makes a transmitter any more than a mouth makes a mouth. That's like saying a bit of brain makes another bit of brain just like it. It doesn't - the transmitter and interpreter bit is already set up in the other guy's brain, and the info actually being sent isn't sent: specifications on how to build it have to be translated into code before being reverse-transcribed into info. This is assuming your mechanism doesn't lose anything in transmission.

It doesn't goes like this:

Idea -> Idea -> Idea

Or like this:

Idea (moved around by machinery until it physically touches material) -> Idea (moved around by machinery until it physically touches material) -> Idea etc.

Genes do work like this:

Gene (moved around by machinery until it physically touches material) -> Gene (moved around by machinery until it physically touches material) -> Gene etc.

It goes more like this:

Idea -> Transmitter -> Light/Sound wave -> Sensor -> Reconstructor -> Idea

And even then, it's just as likely to do this:

Idea 1 -> Transmitter -> Light/Sound wave -> Sensor -> Reconstructor -> Idea 2 (mutation of 1)

It's easy enough to say that we can treat it like it's evolving, but that's to confuse any kind of change in general with evolution, which is specific and technical. Star cycles aren't evolution, they're straightforward change. So too is the constant transmission and reconstruction of ideas.

Comment 20 by OHooligan

The trouble is, this doesn't get you out of the mix. A physically identical molecule physically comes away from the original replicating molecule. You could swap them and not tell the difference, and the beauty is that this isn't a coincidence, but will happen over and over again until a mutation, gross or small, occurs. You can't duck it by saying information is passed on via an intermediary or that it's not material. To even declare it's passed on requires a physical replicator in the first place, or else you might as well say that an idea gets passed on when anything happens to anything else - say, a star's explosion disturbs a dust cloud that collapses and forms another star. That's not replication, but under your logic, it would be because stars lead to other stars.

And you commit an awful mistake here:

Evolution is an innate property of Information, independent of the specifics of the processes that encode, replicate, mutate and select it.

Evolution is not an innate property of information. Every non-meme brain is proof of that, because the information in brains doesn't get replicated any more than the bodies of asexual stick insects do. If damage or a change occurs to a stick insect - say she loses a limb - this isn't inherited by its offspring, and the same applies whether it's a change to her leg or to her brain. If she loses an eye or gets damage to a part of her ganglia, this won't get inherited by her children.

Replication is the first step to getting any evolutionary process started. Replication is what we should focus on, because if something doesn't replicate, and do it often enough so that mutation comes only every 100th generation or so, there's a basis for differential survival of replicators. But the first step is to prove replication is happening. Information about the previous star could be held in this new star caused by the previous one's explosion in the form of its molecular content, but again this information is not making an evolutionary process.

My apologies if I come across as a little belligerent, but it's too easy to assume change is evolution, and when it's not I think it's important to get to the heart of the issue and point out these differences. I think genes enable an independent causal chain to operate in parallel, but in a similar sense to how a tool might be passed on outside of genetic generations. The distinction between an active and a passive copying mechanism (self-propelled and set up by something else) is key, and I really want to get that across.

Mon, 23 Jul 2012 15:23:35 UTC | #949897