Re: Definition of meme

From: William Benzon (
Date: Thu 22 May 2003 - 20:36:09 GMT

  • Next message: "Re: Definition of meme"

    on 5/19/03 1:26 PM, Keith Henson at wrote:


    > Do you have any doubt that the physical representation of a meme in a brain
    > (where memes exist as a class of memory) can be found? (Given fine enough
    > tools of course.) Here is a thought experiment on how to do it:
    > Take a snapshot of the places and states of every atom in a
    > brain. Immediately have the brain learn a new a new phone number, snapshot
    > again. Subtract. Ignoring the (eventually solvable) engineering problems,
    > what is left is the physical representation of the encoded phone number or
    > meme or whatever. Might be hard to figure out how it is encoded, but
    > that's just detail.

    I don't doubt that physical changes happen in the brain as a result of learning this that and the other thing. That's not an issue. But there is no good reason to regard any of those changes as evidence for memes-in-the-brain.

    In any event, you have more faith in our engineering capacity than I do. I haven't got the faintest idea whether or not we'll ever be able to do what you propose. But, assuming we can do that, there's another problem with your thought experiment. As others have said, the brain is a very fluid world, cells are always dying, etc. So, when we get your difference image, not only does that difference image reflect the results of learning the phone number, but it reflects other things as well. And we have no way to sort those things out.

    But let's assume we can solve that problem as well, for the sake of your thought experiment. Let's go through that procedure with Jack. Now we snapshot Jill's brain and Jack gives her a sheet of paper with the phone number on it and asks her to memorize it. When she's done that we snapshot her brain and make the difference image.

    Let's compare the difference image for the two brains. We have no particularly good reason be believe that they will be very much like one another. What is even worse, we don't even have any principled way to make a comparison. One of the interesting characteristics of the nervous systems of small invertebrates, like C. elegans (and, I suspect, your slug) is that each phenotype has the same numbers of neurons in the same arrangement. Thus each individual nematode (C. elegans) has 302 neurons in its nervous system, no more, no less. This means that it is easy to put the neurons in one nematode into 1-to-1 correspondence with the neurons in another.

    But vertebrate nervous systems are not like that. We have many more neurons and we have no reason whatever to believe that one person has the same number of neurons as another. Given the range of sizes from one brain to another it is obvious that different brains have different numbers of neurons. So we can identify some neuron in Jack's brain but there is no way, not even in principle, to identify the corresponding neuron in Jill's brain. And if we can't establish a 1-to-1 mapping between neurons, we can't do it for synapses either. So how can compare the difference images to see if the meme has "replicated" from one brain to the other?

    With that in mind, let's add a step to our experiment. When we've made our
    "before" image of Jill's brain we're going to use our "before" and "after" images of Jack's brain to predict the "after" image of Jill's brain. If we really understand how the encoding works, and if replication really is involved, then we should be able to come up with a pretty good prediction of what Jill's "after" image will be like.

    But how could we make a prediction? If we don't know what neurons in Jill's brain correspond to the neuron's in Jack's, then how can we make a prediction? It's quite meaningless.

    > We know enough about how sea slugs form memory (with only a few giant
    > neurons) to strongly suspect this physical/chemical change encoding
    > memories will be in synapses. (I am ignoring the possibility that memory
    > may be in circulating electrical patterns for a while before it gets
    > committed to physical structure, but we *know* long term memory is retained
    > over complete electrical shutdown of brains--much like a computer disk
    > retains information when the power is off.)

    Not if a complete shutdown means no electrical activity at all. When that happens a neuron is dead. You simply cannot compare a living brain, even a sleeping or comatose one, with a computer disk that has been powered down.

    What happens in computers is that bit patterns are moved from one place to another and new bit patterns are formed by operations on old bit patterns, and so forth. There is no reason to believe that brains operate by moving bit patterns from one place to another.

    > If you don't buy this model, then are you making a claim that memory has no
    > physical encoding or that it is outside the physical world?
    >> These mentalist memes are purely hypothetical entities. Now, if these
    >> hypothetical entities had given rise to a good research program with
    >> substantial empirical results, that would be a different matter. But that
    >> research program has not come into being. Memetics doesn't have a Gregor
    >> Mendel.
    > I don't think that's the problem with research on memetics. The problem is
    > like the baseball-island experiment, the outcome of such experiments are
    > just *too* obvious to get funding.

    Well, if the outcomes really are so very obvious, then they aren't worth doing and memetics really isn't worth much. It's just a dialect of quasi-scientific speech with no unique conceptual content.

    William L. Benzon
    708 Jersey Avenue, Apt. 2A
    Jersey City, NJ 07302
    201 217-1010
    "You won't get a wild heroic ride to heaven on pretty little
    sounds."--George Ives
    Mind-Culture Coevolution: 
    This was distributed via the memetics list associated with the
    Journal of Memetics - Evolutionary Models of Information Transmission
    For information about the journal and the list (e.g. unsubscribing)

    This archive was generated by hypermail 2.1.5 : Thu 22 May 2003 - 20:41:18 GMT