Re: Precision of replication

From: Chris Taylor (
Date: Fri 20 Jun 2003 - 21:40:07 GMT

  • Next message: Chris Taylor: "Re: Meme definition"

    I think we need to factor in the idea of a quasispecies here - it exists as a cloud in sequence space, and the centre (eidos-ish thing) may not even exist in the extant population. The population spreads over time, but that spread is curtailed by selection for deviation (down the slopes of the fitness peak). Incidentally this gives the mutation thing a whole new twist; there are, for example, at a rough count, about seventy seperate mutations at different sites in just one gene that cause cycstic fibrosis; AIDS is of course a classic qsp., as an RNA virus its copying fidelity is bloody awful - the cloud diffuses rapidly, and is trimmed fiercely - which is why it adapts so quickly. If the landscape has neutrality, various forms of neutral change may accumulate; if peaked, the cloud stays near the peak, but 'individual' trajectories
    [i.e. reproductive chains of static points) in a necessarily imperfect statistical subsample of all the possible poitions in sequence space (= on the fitness landscape) will appear to drift all over the place, and usually downhill, because as Dawkins (inter alia) points out, there are many more ways to get it wrong than there are to get it right.

    So even though we each replicate imperfectly (and I see no other possibility) by reusing whatever bits we each personally have to hand
    (our fragmented, recombined experience) to build things which differ
    'internally', we still stay near to the 'out there' attractor because selection favours anyone who is nearer (or, tbh, just punishes those who stray too far - see the 'life-dinner' principle). We can trust that an idea will be maintained if it resonates (steady teddy) with something in the world. Chinese whispers shows how quick copying can go wrong given semi-imperfect transmission so something else is herding the quasipecies.

    So Lawry is right to assume 'degradation' but only in a sense - you must assume you are somewhere within range of the 'centre' (/peak/basin/cycle even...). AIDS id continually 'degrading' in that sense but it seems to do all right.

    Cheers, Chris.

    Keith's training is a good example of something more like smallpox which is almost invariant, with good (DNA-based) copying fidelity I suppose. The Shannon index view is a good one for random mutation.

    Lawrence DeBivort wrote:
    > Ahhhh, ok, I understand better what you are suggesting: "> Well, a meme gets
    > replicated with 100% fidelity by definition. If it doesn't get replicated
    > exactly then it's a mutation."
    > I have not been using the term meme with the same standard of 100% identical
    > replication that you do. In our view of memetics, we accept that rarely will
    > a meme be replicated 100% identically. Like you, I am much intrigued with
    > the usage of words and the promise that resides within the availability of
    > words, both for precision and for mind-changing potential. But I think the
    > wording of a meme -- that is, the stuff that goes between two people -- is
    > only part of the story of dissemination. There is also the processing that
    > occurs within the transmitter, through which he tries with only partial
    > success to find the words that best convey the idea, and the effort by the
    > receiver, also only partly successful, to conjure from the transmitter's
    > words what the actual idea is that lies behind them, and then we have Wade's
    > prescient observation that the cultural venue also exercises its deleterious
    > effect on the fidelity of the dissemination.
    > At our end, here, we accept, therefore that the dissemination loses
    > fidelity, and we still call the whole thing a meme. You suggest, if I
    > understand correctly, that we should be calling it a mutation, and you may
    > be right, but I think that there is something in between a 'mutation' and a
    > 100% replicated transmission. A mutation to me seems more random than the
    > process of deterioration that I discuss above. If a transmission mutates, as
    > I hear the term, it becomes something different, its similarity to the
    > original is so perverted that it can not be said to be anymore related,
    > substantively. At the other end, if a meme is only what is passed with 100%
    > fidelity, than I think the term becomes relegated to an extremely small
    > number of instances, so I think that I prefer the more robust definition of
    > meme, one in which deterioration happens routinely, yet the pedigree and
    > connection of influence readily and adequately persists. My operational
    > standards for 'memes' have more to do with the needs and dynamics of
    > influence than with a standard of fidelity that is akin to genetic
    > replication.
    > I know that you too are interested in influence and cultural evolution: how
    > can a meme, defined as 100% fidelity, serve the needs of those interests?
    > I hope I have understood your thinking correctly, and that I am not dragging
    > us off into a pointless tangent.
    > Cheers,
    > Lawry
    >>-----Original Message-----
    >>From: []On Behalf
    >>Of Richard Brodie
    >>Sent: Thu, June 19, 2003 12:30 PM
    >>Subject: RE: Precision of replication
    >>Lawry wrote:
    >><<I use the term and concept of memeplex, too, in the sense that
    >>a memeplex
    >>a collection of memes with a common theme at its core, and in which the
    >>participating memes tightly reinforce and amplify each other, or in which
    >>the memes are mutually dependent. IIRC, this is, I think, compatible with
    >>your own use of the term.>>
    >>I don't know what a "common theme" is (other than another meme
    >>perhaps). But
    >>sure, a memeplex is a bundle of memes that tends to get the memes
    >>it replicated with greater frequency than they do on their own.
    >><<Then I think that the comments I am making pertain to both memes and
    >>memeplexes. In terms of precision of replication, I would guess that a
    >>memeplex -- simply because it is more complex -- might replicate with less
    >>precision than a single meme, but as I type this I wonder: might the
    >>interdependence and interaction of the memes within a memeplex
    >>not serve to
    >>_increase_ the fidelity of the transmission? Hmmmmmm.....interesting
    >>Well, a meme gets replicated with 100% fidelity by definition. If
    >>it doesn't
    >>get replicated exactly then it's a mutation. And yes, I agree
    >>100% with your
    >>interesting thought. The memes comprising a memeplex work together to get
    >>the whole bundle passed on.
    >><<Richard, is it your thought that memeplexes and memes behave differently
    >>when it comes to transmission and replication? I have looked at
    >>the case in
    >>which only a subset of the memes in a memeplex are transmitted, and what
    >>effect that has, but short of this I have been proceeding on the
    >>that a memeplex is essentially just a 'big meme' when it comes to
    >>I think you've put your finger on it. It makes no sense to talk
    >>about a meme
    >>being transmitted with less than 100% fidelity unless you are
    >>talking about
    >>mutation. With a memeplex, you might be interested in less than
    >>100%-fidelity transmission as long as the receiving mind exhibits similar
    >>behavior as a result of sharing the memeplex.
    >>Richard Brodie
    >>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 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)
    > see:

      Chris Taylor ( »people»chris
    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 : Fri 20 Jun 2003 - 21:48:44 GMT