>Brian D. Harper continues to be skewer Randy about Denton's use of "pure
>chance" in the quote Randy posted. Here's the latest:
>
><<Good grief, Randy, I'm referring to your quote of Denton, the quote
>that characterizes Darwinism as pure chance, the quote that so
>mis-characterizes Darwinism that it falls just short of being an
>out right lie.>>
>
JB:==
>Brian, I think you're right that "pure chance" is NOT a proper
>characterization of Darwinism as a whole. But in Chapter 13 of Denton's book,
>it becomes clear not only that his use of "pure chance" is related to the
>mutational part of the equation, but that others use the same terminology:
>
I have looked over chpt 13 of Denton briefly and I think you are right.
Actually, it occurred to me when I first read the quotes given by
Randy that those quotes may not be accurately representing Denton's
views. Now, I still believe Denton botches up chpt 13, but not in an
obviously stupid or dishonest way. So I'll retract what I said about
Denton. My complaint now is only against Randy for giving a selective
quote that not only distorts Darwinism but also what Denton said
about Darwinism.
Denton:==
>"According to the central axiom of Darwinian theory, the initial elementary
>mutational changes upon which natural selection acts are entirely random,
>completely blind to whatever effect they may have on the function or structure
>of the organism in which they occur, 'drawn,' in Monod's words, from 'the
>realm of pure chance.' It is only after an innovation has been disclosed by
>chance that it can be seen by natural selection and conserved."
>
JB:==
>Denton, I think, is being held to something he never said. Dawkins does the
>same thing in The Blind Watchmaker (see pp. 306-309). So anxious is he to
>distance himself from pure chance, that he sets up several extreme definitions
>of "random" just so he can knock them down. One of these is the "equally
>likely" argument that Brian has brought up, too. Denton never uses it, so why
>is it used against him?
>
I haven't read all of chpt 13 carefully yet, but skimming through I found
right away an example where Denton does use this assumption:
However, to find by chance, longer words, say seven
letters long such as "English" or "require", would
necessitate a vastly increased search. There are
26^7 or 10^9, that is, one thousand million combinations
of letters seven letters long. As there are certainly
less than ten thousand English words seven letters long,
then to find one by chance we would have to search
through letter strings in the order of one hundred
thousand units long. ... -- Denton, p. 309
Now, let's see how he got those numbers. There are 10^9
possibilities containing 10^4 possible words (a generous est.
I think). The probability of selecting one of the 10^4 possible
words from the total 10^9 possibilities is thus 10^4/10^9 = 10^-5.
Thus one would have to search strings of length 1/10^-5 = 10^5 =
100,000 ...
But the probability equals 10^4/10^9 only if all letters occur
with equal probability. So, Denton is definitely assuming
equal probability whether he is aware of it or not.
Since Denton is using English, let's consider another
example along these lines. A mathematician on
bionet.info-theory helped me with the following math
(about 2 yrs ago if I remember right) the details of
which I will skip. Let's take Dawkins example phrase
methinksitislikeaweasel
[this is all I'm taking from Dawkins, BTW, I hate his
little word game associated with this phrase in TBW]
There are 23 letters in this phrase and thus 26^23
(roughly 10^32) possible combinations of letters.
The probability of selecting that particular phrase
(assuming all letters occur with equal probability)
is thus about 10^-32. Now, I have absolutely no idea
how many phrases with 23 letters convey some meaning.
The probability of selecting a sequence that contains
some message will obviously be a little higher. The
analogy obviously breaks down here anyway as one
wouldn't expect the proportion of phrases containing
a message in English to have any relation to the number
of functional proteins, say.
But now let's suppose the letters don't occur with
equal probability. After all, our example is English
and the letters in English phrases don't occur with
equal probability. So let's suppose we have a totally
random process which now generates letters which
satisfy the statistical laws of English. These
statistical laws were worked out by Shannon (and
subsequently modified somewhat) and have nothing to
do with semantics, grammer, meaning etc. They are
purely statistical, some letters occur more frequently
than others. Letter such and such always follows
letter so and so or this and that. Again, I'm skipping
the math, but there are roughly 10^13 phrases of length
23 that satisfy these statistical laws. Now, all of a
sudden, the probability has increased by 19 orders of
magnitude !!
>Denton does not argue that "anything is possible" at the mutational level.
>Only that the mutations which do occur do so by PURE CHANCE. And they do:
>
>"The mutation process is generally thought to be an uncontrollable chance
>phenomenon." [Volpe, Understanding Evolution, 1970, Wm. Brown & Co., p. 33,
>Jim Bell's college text, dog-eared and highlighted, with nasty notes about the
>professor]
>
>The mutational definition of random Brian uses is:
>
><< In Darwinism, random as in random mutation means
>only that the appearance of a mutation does not anticipate the
>needs of an organism.>>
>
>Dawkins says, "Mutation is not systematically biased in the direction of
>adaptive improvement."
>
>I find this misleading. Mutations not only don't "anticipate needs," they
>arise without ANY predictive preview (except when you zap fruitflies in the
>lab, but that's another story). In nature, no one knows when a mutation will
>arise, or precisely what form it will take (though, yes, there ARE limits. But
>that's a different consideration). Most of the time mutations are negative,
>of course (a point usually ignored in discussions like these), but when one
>pops up that is selected, it is still the result of a FIRST CAUSE that is PURE
>CHANCE.
>
Here I'm going to key in only on the meaning of random mutation. Actually,
your quote of Denton above does not disagree too much with what I
have been saying, depending on how you interpret it. Below is a repeat
of the quote:
Denton:=======================================================
>"According to the central axiom of Darwinian theory, the initial elementary
>mutational changes upon which natural selection acts are entirely random,
>completely blind to whatever effect they may have on the function or structure
>of the organism in which they occur, 'drawn,' in Monod's words, from 'the
>realm of pure chance.' It is only after an innovation has been disclosed by
>chance that it can be seen by natural selection and conserved."
>============================================================
If we look at what follows the comma as a qualifier of "entirely random"
then it seems Denton and I are in at least a partial agreement. i.e.
if we read the above as :
"... entirely random, where entirely random means completely blind
to whatever effect they may have on the function or structure
of the organism in which they occur ..."
Futuyma elaborates a little more on the meaning of random:
Mutations occur at random. This does not mean that all
loci mutate at the same rate, nor that all imaginable
mutations are equally likely. Nor does it mean that
mutations are independent of the effect of the environment;
environmental mutagens increase mutation rates. Mutation
is random in that the chance that a specific mutation
will occur is not affected by how useful that mutation
would be. As Dobzhansky (1970) said, "It may seem a
deplorable imperfection of nature that mutability is
not restricted to changes that enhance the adaptedness
of their carriers. However, only a vitalist Pangloss
could imagine that the genes know how and when it is
good for them to mutate.
-- Futuyma <Evolutionary Biology> 2nd ed. 1986, p. 76
The comment by Dobzhansky is interesting. My own feeling is
that this would not be a deplorable imperfection at all and may
even be the most efficient method of maintaining adaptability
in view of its flexibility. It is not entirely inconceivable that
mutations might occur to the benefit of an organisms adaptability,
in fact Dawkins discusses some recent research which tentatively
suggests that this actually happens in one specific case, though,
if it turns out to be the case it will undoubtedly be an exception.
Of course, if this does turn out to occur occasionly, it couldn't
happen by the extremly reductionistic suggestion of Dobzhansky,
wherein genes are making decisions. Instead, it would probably
occur due to a very complicated and nonlinear feedback mechanism
involving the genes and their environment.
But would such a scheme be an efficient one? I don't think so, not
over the long run. It seems to me pretty clear, given what we now
know about complexity and chaos, that long term predictions about
the environment are impossible. In the short term there will be patterns
that develop which our hypothetical mechanism may be able to anticipate
and take advantage of. But these patterns will eventually break down
and the organism relying on this mechanism will probably go
extinct since it will be producing mutations according to faulty
anticipations. Unless, of course, their is another anticipation
mechanism which anticipates and makes necessary corrections
in the original anticipation mechanism ... and so on ;-).
Thus, I think that an anticipitory mechanism is not the most efficient
way of maintaining adaptedness in view of its inherent inflexibility
to unpredictable long term changes in environment. The occurrance
of mutations which do not anticipate the needs of the organism
may then be the most efficient way to maintain long term adaptability,
especially in view of the fact that long term anticipation of the needs
of an organism is impossible.