What are memes made of?

From: Robin Faichney (robin@faichney.demon.co.uk)
Date: Sun Jan 30 2000 - 14:16:23 GMT

  • Next message: Kenneth Van Oost: "Re:Should memetics embrace Sheldrake?"

    Received: by alpheratz.cpm.aca.mmu.ac.uk id OAA20811 (8.6.9/5.3[ref pg@gmsl.co.uk] for cpm.aca.mmu.ac.uk from fmb-majordomo@mmu.ac.uk); Sun, 30 Jan 2000 14:46:10 GMT
    From: Robin Faichney <robin@faichney.demon.co.uk>
    Organization: Reborn Technology
    To: memetics@mmu.ac.uk
    Subject: What are memes made of?
    Date: Sun, 30 Jan 2000 14:16:23 +0000
    X-Mailer: KMail [version 1.0.21]
    Content-Type: text/plain
    Message-Id: <00013014435801.00956@faichney>
    Content-Transfer-Encoding: quoted-printable
    Sender: fmb-majordomo@mmu.ac.uk
    Precedence: bulk
    Reply-To: memetics@mmu.ac.uk
    

    I'd be grateful for any feedback on the following essay, which I hope will
    eventually evolve into an article for the JoM. It's around 2 400 words.

    1

    Information has a fundamental, but quite straightforward, place in
    physics: it is simply the form, or structure, of matter. ("Information"
    and "form" are very closely related.) Physical information is inversely
    proportional to entropy, because as entropy increases, structure breaks
    down.

    Roy Frieden (1998) has recently suggested that the laws of physics are
    generated by the attempt to minimise the difference between an entity or
    system's own physical information, and the information that physicists
    can obtain about it. What follows is based not on Frieden's work, but on
    that distinction, between physical and intentional information. I intend
    to show that memes, like genes, are items of physical information, even
    though, unlike genes, they are not tied to a particular molecular-level
    encoding.

    Physical information is really just the form of matter, but to think in
    terms of information rather than those of matter can be extremely useful,
    even in physics, never mind communications, computation, and so on.
    Following Daniel Dennett (1987), we can call this strategy "the formal
    stance". (To think in terms of matter is to take "the physical stance".)
    As the term suggests, we take this stance not only when we think in terms
    of information, but whenever we focus on form. To focus on the form of
    something is just to generalise from the particular instance, to look at
    what could happen again elsewhere. This is an absolutely crucial ability
    which I think has generally been neglected, or at least taken forgranted.

    When we adopt the formal stance consciously and consistently, an object
    becomes its own description, and a flow of energy becomes a stream of
    information, because any object is the ultimate source of information
    about itself, and every information stream rides on the back of an
    energy flow, being (certain aspects of) its form, as an electrical current
    is enabled and interrupted to form the dots and dashes of Morse code.

    2

    Identity, from the formal stance, ie the identity of items of
    information, is rather interesting: no matter how many instances of
    an item of information exist, it remains just one item. No matter how
    many productions of The Marriage of Figaro have been put on since it was
    written, it remains AN opera. Regardless of the number of performances,
    one production remains just that. Thousands or even millions of CD's
    might be made from the recording of a particular performance, but it's
    still A particular performance. Strictly speaking, this is information's
    "numerical identity", and it is absolutely essential. It might even be
    said to be the defining feature of information, as opposed to physical
    substance. It means, of course, that information is an abstraction,
    but then the same can be said of "physical substance" (though I won't
    argue that here).

    To see something as an example of information processing, just as in
    seeing information anywhere, is the result of taking the formal stance.
    All information processes, assuming they are actually instanciated
    somewhere, are also physical processes. And every physical process, in
    principle at least, can be viewed as an information process, though it is
    obviously more useful to do so in some cases than in others. What goes on
    inside a computer is nearly always best viewed as information processing,
    whereas the processes in the interior of the sun will only rarely,
    if ever, be best understood that way. But one naturally occurring
    process that can best be understood in terms of information processing
    is evolution.

    3

    Forms can (be) reproduce(d), substance cannot. The concept of evolution
    requires the formal stance. Only information can replicate with variation
    and selection.

    In biological evolution, the entities that actually replicate, according
    to Richard Dawkins (1976), are the genes. Replication of other entities,
    such as organs, organisms, species or proteins, is a side-effect of gene
    replication. So what is the gene? There is still some controversy on
    this, fuelled in the main by two factors. People who use the concept
    for different purposes, tend to define it in different ways -- eg,
    evolutionary biologists (like Dawkins) use a hierarchical, recursive
    definition, so that one higher level gene is composed of two or more lower
    level ones, while molecular biologists tend to focus on the lowest level
    only, probably saying something like "higher level units are not *really*
    genes, but mere epiphenomena". Rather more significant, at least when
    more general, theoretical issues are of concern, as here, is this:
    are genes molecular entities, ie bits of DNA, or units of information?

    It was previously stated (section 2) that the "numerical identity" of
    information is crucial. Thus, no matter how many instances of it can
    be said to exist, one item of information remains just one item. Genes
    (often) survive the death of their hosts, because they are *copied*.
    Items made of physical stuff cannot be copied -- or they can, but
    what is actually copied is the *form*, not the substance. Parents and
    children *share* many genes, and a smaller number is shared by the wider
    population. That is, different organisms can carry the *same* gene.
    This is an item of information.

    It was also previously stated that, by taking the formal stance, any
    physical process can be viewed as an information process -- but that to
    do so is much more useful in some cases than in others, and one of the
    most useful cases is evolution. So what, in purely informational terms,
    is going on here?

    The form of the biological organism is determined not by DNA alone, but
    by the interaction between the DNA and its environment, the machinery of
    the cell that carries it (while that, in turn, is affected by the wider
    environment in which the cell finds itself). The gene is *encoded*
    in the DNA, while the cellular machinery is the decoding mechanism.
    (The wider environment can also be considered part of that mechanism.)

    Now, the "heresy" of biological design, is very simply conceived: a
    designer would have to *encode* the gene in the DNA. In the case of a
    theological designer, that is supposed to have happened at some point in
    the distant past. The (discredited) theory of the inheritance of acquired
    characteristics requires some mechanism by which such characteristics
    become encoded in the DNA. The standard current understanding (with which
    I entirely concur) is that in the case of biological evolution, there
    is no encoding process: the genes are as they are due to an extremely
    long series of random events, whose "successes" are concatenated. But I
    continue to talk in terms of genes being "encoded", simply because I know
    of no better term. They are encoded, *in effect*, and are certainly
    decoded, during normal biological operations. But what is the most
    fundamental fact about genes is that, however they came into being, they
    survive, and so life is the survival of encoded items of information.
    So is culture, of course, but there are some further wrinkles there.

    4

    What, from the physical stance, is any physical process, can, from the
    formal stance, ie by focussing on its form, be considered an information
    process. Eg, that box of gubbins that seems designed to shunt electrons
    around in fantastically complex yet extremely precise ways[1], is also
    an information processor, or computer. Similarly, life is a set of
    processes by which the items of encoded information that we call genes
    are perpetuated. Genes are "encoded" in that they are not directly
    instantiated in physical form, but are implicit in the relationship
    between what we might consider the "carrier", the DNA (or RNA) molecule,
    and its environment, the cellular mechanisms which do the decoding.
    DNA on its own is just another molecule, but put it in the right kind
    of cell and between them, they can make a human being. Or whatever.
    You might consider the DNA to embody the program that is then run on
    the cell as processor. But in any case, both elements are essential.

    Generalising from this, any physical transformation can be considered a
    formal transformation, or an en/decoding operation. Where any physical
    entity undergoes a change of any kind, its pre-change form can be thought
    of as the carrier, and the post-change form as the message, while the
    cause of the change is the decoding mechanism. It's important not to
    get mind-boggled by this, and realise it's no big deal. "Encoding" and
    "decoding" are just words we use for particular types of transformation,
    especially where the change is reversible, so that by decoding the result
    of an encoding operation, we end with the form that we started with
    (assuming, of course, the same code is used).

    What's probably the most difficult part of this, is the "implicit" nature
    of the information carried, for instance by DNA. What this actually
    means is that the outcome of the decoding operation is predetermined.
    In principle, a difference either in the DNA or in the cellular mechanism
    would result in a difference in the outcome -- the resulting organism
    -- but we assume that the cellular mechanism is, in effect, constant,
    so we consider the DNA to be the significant factor, the information
    carrier, and in practice, because of the way these things work, this is
    a useful strategy. But in general, it is the relationship between the
    carrier and the decoding mechanism that really matters, and it is because
    their interaction is determined by all the laws of nature, that before it
    takes place, its outcome can be considered encoded in that relationship.
    To encode is to predetermine a transformation.

    [1] That's actually the design stance rather than the physical stance,
    but no matter.

    5

    When Dawkins came up with the concept of the meme (1976), he was trying
    to show that genes need not be the only kind of replicator. Catchy tunes
    replicate too: I hear one on the radio and can't get it out of my head
    all day, humming, whistling or singing it at times. Someone hears me
    and starts humming it herself: it has propagated again.

    With the formal stance, and the understanding of the gene as an encoded
    item of physical information, we have all we need to understand *exactly*
    what memes are made of.

    What we have here are replicating behaviours, but what does that mean?
    One individual observes (hears/whatever) another performing a particular
    behaviour, and as a result performs that same behaviour himself, whether
    immediately or on some later occasion. But what is "a behaviour"?
    It could be a simple physical action, or it could be very much more
    complex (eg setting up a blind trust), but either way, there is at
    some hierarchical level(s) of organisation an identifiable (set of)
    physical process(es) (unless this is happening on the astral plane).
    One particular behaviour, performed by different individuals at
    different times and places, can be identified as such by some form of
    pattern matching, though it might have to be extremely sophisticated
    -- these instances share some elements of form, some formal aspects.
    These common forms are carried, or *encoded*, in however complex a way,
    in physical processes. That encoded information is the meme, in one of
    the two main phases of its lifecycle.

    Between the acting out of actual behaviours, where does the meme reside?
    In the brain. It is perceived, say, one day, and acted out again the
    next, and overnight it is, of course, stored in the brain -- where else?
    Now, it is not hard (I hope) to see that the meme has to be *encoded*
    in a suitable form, for neural storage. Also, that this neural form
    must be very different from the behavioural encoding in which it is
    expressed. So brain storage is, of course, the other main phase of the
    meme's lifecycle.

    The concepts of encoding and decoding are essential to an understanding
    of memetic transmission. Like all other information processes, these
    are aspects of physical processes. Consider any repeated cycle of en-
    and decoding operations, say, between two different languages. (For the
    sake of simplicity we will assume that a perfect translation takes place
    every time.) If we did not know which came first, we could not say which
    was the "clear", and which the encoded form. In fact, the two forms might
    as well be considered different encodings, neither being "clear". What
    they have in common, that allows the cycle to be repeated indefinitely,
    is never explicit. This is the case with memes -- they have two phases of
    existence: encoded in the brain, and encoded in behaviour. A brain that
    carries the meme is thereby motivated, in appropriate circumstances, to
    perform the relevant behaviour. Another brain, observing that behaviour,
    processes the incoming information with the result that the meme takes up
    residence in this brain, and if the action is again triggered, the cycle
    continues, brain to behaviour to brain, and so on, each transformation
    being viewable as encoding from one point of view (brain or behaviour)
    and decoding from the other (behaviour or brain).

    Can we say that one phase of the lifecycle is immature, and the other
    adult, as in many organisms? Well, if reproduction is the criterion,
    then only when it is encoded as behaviour, can the meme be taken up by
    other brains. On the other hand, a certain sort of reproduction also
    occurs in the brain, because the meme generally remains there during and
    after the time it is expressed as behaviour, and it might be expressed
    any number of times. So maybe the maturity/immaturity distinction is
    pushing this metaphor too far.

    I talk of the "main" lifecycle phases, because it can be argued
    that encoding in artefacts, whether deliberately in books and such,
    or incidentally in anything that retains a trace of human behaviour,
    is another (set of) phase(s). I think that's largely if not entirely a
    matter of semantics, but this type of encoding is certainly worth noting,
    in any case.

    Executive Summary

    Memes, like genes, are encoded physical information, but exist at a
    higher level of organisation.

    References

    AUTHOR = "Richard Dawkins",
    TITLE = "The Selfish Gene",
    PUBLISHER = "Oxford University Press",
    ADDRESS = "Oxford",
    YEAR = 1976

    AUTHOR = "Daniel Dennett",
    TITLE = "The Intentional Stance",
    PUBLISHER = "MIT Press/A Bradford Book",
    ADDRESS = "Cambridge, Mass.",
    YEAR = 1987

    AUTHOR = "B.\ Roy Frieden",
    TITLE = "Physics from Fisher Information",
    PUBLISHER = "Cambridge University Press",
    ADDRESS = "Cambridge, UK",
    YEAR = 1998

    --
    Robin Faichney
    

    ===============================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: http://www.cpm.mmu.ac.uk/jom-emit



    This archive was generated by hypermail 2b29 : Sun Jan 30 2000 - 14:46:12 GMT