Re: Probability or Predictability?

From: Peter Ruest (
Date: Mon Dec 30 2002 - 00:53:34 EST

  • Next message: Peter Ruest: "Re: Quality of 'ID Science'"

    Dick Fischer wrote:
    > Peter Ruest wrote:
    > > I think it will not be possible to compute the probability of formation
    > >of any complex system. The best we can hope for
    > >is the estimation of some very very rough upper bounds for such a
    > >probability.
    > I think probability is misunderstood. The probability of my father meeting
    > my mother multiplied by the probabilities of my grandparents meeting, my
    > great grandparents, my great, great grandparents, etc. for hundreds of
    > generations is so low that the probability that I'm here is so negligible
    > that either I'm not here or - it must be God!


    I assume that you are convinced that it must be God who is responsible
    that you are here...

    I am really astonished and frustrated that, after I have explained
    several times in PSCF and other publications - and numerous times on
    this list - what I mean by probability in this context, I am still being
    misunderstood. I feel like a christian accepting evolution who is being
    charged with being a YEC, just because he talks about creation. Just as
    all that believe in any kind of creation are thrown into the same pot by
    non-christians, all that talk of any kind of probability in a biological
    context are thrown into the same pot.

    I am sorry that apparently I have not - once more - made the
    qualifications of probability calculations plain. I just thought,
    mistakenly, that this point is by now obvious. For perhaps 30 years, I
    have clearly distinguished between the naive probability calculations
    you are talking about and qualified estimates of a maximum probability
    taking into account things like code degeneracy, irrelevance of some
    amino acid positions and variability of others, evolutionary
    convergence, functional synonymity of distinct sequences and folds,
    minimal functionality, possible exaptations, different sequences of
    emergence of different functions, etc. It's not just for handwaving that
    I am talking of the extreme difficulties of proposing reasonable
    probability estimates of functional emergence.

    But the trouble is that some just follow Dawkins and company in
    _assuming_ that the emergence of _any_ biological function by random
    walk mutagenesis is no problem at all. The _assumption_ is that the
    configurational space of proteins is densely populated with a fully
    connected network of functional sequences. These assumptions have never
    been critically looked at. Not even the question of whether such a
    checking would be feasible has been asked. Everyone is just caught up in
    the circular argument of "we are here - therefore the fully accidental
    emergence of all biological functions must have happened - therefore any
    bothering about probabilities is anti-science".

    > Where do we go wrong? It is because we specify a known result. Then we
    > say the likelihood of such a result from random processes is impossibly
    > low. Essentially, this is wrongheaded.

    That's right. It's not what I am proposing.

    > It's like looking at Sagittarius the teapot and saying, "What is the
    > likelihood that a handful of stars could arrange themselves into what looks
    > just like a teapot? An intelligent designer must have arranged the
    > stars." This ignores the billions of stars all over the sky that don't
    > resemble anything, and a teapot is only one of thousands of possible objects.
    > Another example. Place a teacup in a field, and throw a rock
    > blindfolded. What is the probability it will land in the cup. Low. But
    > throw the rock first, and then place the cup over the rock. Can't
    > miss. That, basically, is what is going on. We know already where the
    > rock will land, and then say, "Gee, how did it land there"? A low
    > probability only exists if we specify where the rock will land before we
    > throw the rock.
    > Random variation produces things. Some things are useful. Some are
    > complex. Random variation also produces things that are simple and
    > useless. Natural selection sorts it out.
    > If you looked at a continuum of results you could grade them out on the
    > basis of usefulness and complexity, with useless, simple stuff at one end,
    > and useful, complex stuff at the other. To be fair we should look at the
    > entire continuum. But what IDers do instead is to look at only the tip end
    > of the continuum, the useful, complex end, and say, random variation
    > couldn't produce it.
    > Go back to the rock and teacup example. What is the likelihood you could
    > throw a rock into a teacup blindfolded. You might think that it is
    > low. But what if you threw 10,000 rocks 10,000 times with 10,000 teacups
    > scattered all around? The probability certainly is low if the rock is
    > specified, the throw is specified, and the teacup is specified
    > beforehand. But that is predictability, not probability.
    > Dick Fischer - Genesis Proclaimed Association
    > Finding Harmony in Bible, Science, and History

    I agree that you are talking about predictability. But I am not. All
    your examples illustrate exactly the same case, and that is the claim
    that the probability of emergence of a specific gene of 300 nucleotides
    (and therefore of a protein of 100 amino acids) is 4^300. Of course this
    is a nonsense claim. Think through the example I gave, instead.

    Or discuss the qualifications I gave above. Each one of them modifies
    the probability of accidental emergence. But by how much?

    By the way, the ID people don't work with this predictability mistake,
    either (at least some - I don't think Behe, Dembski, or Meyer can be
    charged with this). They are talking about specification, which is
    something entirely different. It's a functional specification, which
    necessarily must take into consideration all the qualifications. I think
    the idea behind this is sound - it's just extremely difficult to produce
    such a specification in a meaningful and useful way.

    The problem is not the "natural selection of the fittest" but the origin
    of a minimal selectable functionality out of _non-selectable_ molecules
    by random walks!


    Dr. Peter Ruest, CH-3148 Lanzenhaeusern, Switzerland
    <> - Biochemistry - Creation and evolution
    "..the work which God created to evolve it" (Genesis 2:3)

    This archive was generated by hypermail 2.1.4 : Mon Dec 30 2002 - 19:33:48 EST