Evolution in One Dimension, and Some of its Implications

From: Chris Cogan (ccogan@telepath.com)
Date: Sun Sep 24 2000 - 18:23:12 EDT

  • Next message: Chris Cogan: "Re: WWYD - What Would You Do to make evolution work??"

    Evolution in One Dimension

    As odd as it may seem, evolution is possible, in an
    abstract sense, in a one dimensional "world." That is,
    if we start with a number-line and a zero-point and
    any other point marked, and then replicate the marked
    value while allowing for small random variations,
    then, to any finite degree of accuracy, any number on
    the number line can be reached in a finite number of
    evolutionary steps (i.e., replications with occasional
    variations, but with no selection, no exclusion of any
    values).

    What is even more interesting is that we can map
    numbers on the line to numbers in a three-dimensional
    "world" in such a way that any point in 3-space that
    we might mark relative to X-Y-Z axes will also be
    reached, with whatever finite degree of accuracy we
    wish. This principle can be extended to *any* finite
    number of dimensions, thus allowing us to represent,
    solely by numbers on a line, any finite-length genome,
    no matter how complex. This mapping can be
    achieved in almost any number of ways, with different
    results in 3-space depending on the mapping used
    (and what happens in 1-space, of course).

    What about selection? Yes, we can prevent some
    values that appear on the number line from
    replicating, thus excluding them and any values that
    would depend on them. Thus, if the maximum size of
    a random variation is, let us say, .5 of a unit of length,
    and we block off a 1-unit length, the evolutionary
    process will never get from one side of that blockage
    to the other side.

    What this means is that, in a 1-dimensional "universe"
    with these kinds of restrictions, *if* a "species"
    appears on the other side of the blockage, it got there
    by some outside intervention of some sort. Could it be
    design? Yes, in principle.

    Further, since blocks to evolution can be of various
    sizes, and since we could move them around, and
    since selective mechanisms could be of weird types
    (excluding only values close to prime numbers, etc.),
    we can produce, even in this impoverished "world," a
    number of bizarre and counter-intuitive results, some
    of which might *look* like design (at least to Jones
    and Bertvan), but which in fact could be the result of
    randomly generated numeric variations starting from
    an initial randomly chosen small number.

    Since the blockages can come and go (depending on
    how *they* are determined), we lose one key method
    of determining design. We can no longer say that a
    "species" that is *now* "too far" from other "species"
    to have evolved from them or any of their ancestors
    was *always* too far. The blockage may have
    occurred *after* its ancestors were already on that
    side of the area where the blockage is now placed. In
    this case, in order to exclude random generation, we
    have to have a fairly good knowledge of the *history*
    of at least significant local span from this "world." If
    the history is long and the data is sparse, we may be
    *stuck* with presuming that the "species" arose
    "naturalistically" (i.e., via the random variation and
    repetition process).

    In three or more dimensions, the problem becomes
    even more nearly intractable.

    Since most genomes can only be reasonably
    represented in thousands or *millions* of dimensions
    (without "folding" them into arbitrary numbering
    schemes that use fewer dimensions), the process of
    tracing a current genome back and ensuring that at
    *some* time there was a sufficiently large genetic
    "gap" and/or spatial-temporal gap between its first
    ancestor and all other previously-occurring organisms
    could be effectively impossible.

    Yet, this is what ID theory depends on. It must show
    that some organisms exist that could not have evolved
    naturalistically from any prior organisms, or from
    replicating autocatalytic sets of molecules, etc. It must
    do this by showing that the gaps involved have always
    been too large to bridge naturalistically. I'd like to be
    able to offer some help, but I'm afraid the ID folks are
    on their own on this one, because I doubt that there
    ever *have* been gaps too large for existing
    organisms to evolve across naturalistically, gaps that
    have been present over a sufficiently long and
    appropriate period of time, etc. To show that early
    hominids did *not* evolve into modern man, for
    example, they'd have to show that the spatio-temporal-
    genetic gap was too large in one dimension or another
    (or that modern man appeared on the scene *before*
    any other suitable hominid ancestor-prospects, etc.).

    It's interesting that, although ID claims to be a
    positive theory, it can only be meaningfully validated
    by showing that naturalistic evolution theories are
    false.

    This is because of one simple mathematical fact:
    Random variation on a replicating genetic population
    can eventually reach *any* possible genome and thus
    any physically possible biologically viable organism.
    It's conceivable, of course, that some such organism
    could be too far, in all possible "directions" from all
    earlier-appearing organisms to come into existence
    naturalistically if there is a "zone of non-viability"
    around it. But, this seems unlikely in the case of real-
    world species. We can *imagine* such creatures, of
    course (such as a pig-fish-tree-like animal with
    feathered wings and a tendency to write books on
    mathematics; such a creature might be to far from any
    other creature we know of to be considered a likely
    *natural* descendant of any *other* creature in
    Earth's history).

    Hmmm. This started out as merely an exposition of
    the interesting features of one-dimensional evolution.
    Sorry.

    --Chris



    This archive was generated by hypermail 2b29 : Sun Sep 24 2000 - 18:27:30 EDT