Thursday, July 7, 2011

Thoughts on Dawkin's weasel program

In The Blind Watchmaker,  Richard Dawkins made a thought experiment:
I don't know who it was first pointed out that, given enough time, a monkey bashing away at random on a typewriter could produce all the works of Shakespeare. The operative phrase is, of course, given enough time. Let us limit the task facing our monkey somewhat. Suppose that he has to produce, not the complete works of Shakespeare but just the short sentence 'Methinks it is like a weasel', and we shall make it relatively easy by giving him a typewriter with a restricted keyboard, one with just the 26 (capital) letters, and a space bar. How long will he take to write this one little sentence?
More time, by a factor of millions, than the age of the universe.  Dawkins next considers a model of cumulative selection:
...We again use our computer monkey, but with a crucial difference in its program. It again begins by choosing a random sequence of 28 letters, just as before ... it duplicates it repeatedly, but with a certain chance of random error – 'mutation' – in the copying. The computer examines the mutant nonsense phrases, the 'progeny' of the original phrase, and chooses the one which, however slightly, most resembles the target phrase, METHINKS IT IS LIKE A WEASEL....
Dawkins observes that cumulative selection radically accelerates the 'evolution' of the phrase:
The exact time taken by the computer to reach the target doesn't matter. If you want to know, it completed the whole exercise for me, the first time, while I was out to lunch. It took about half an hour. (Computer enthusiasts may think this unduly slow. The reason is that the program was written in BASIC, a sort of computer baby-talk. When I rewrote it in Pascal, it took 11 seconds.) Computers are a bit faster at this kind of thing than monkeys, but the difference really isn't significant. What matters is the difference between the time taken by cumulativeselection, and the time which the same computer, working flat out at the same rate, would take to reach the target phrase if it were forced to use the other procedure of single-step selection: about a million million million million million years. This is more than a million million million times as long as the universe has so far existed.
There's no question that cumulative selection- the retention of close calls to the target- speeds the convergence on the target.

Dawkins' METHINKS IT IS LIKE A WEASEL experiment has given rise to a lot of discussion.

There's no question that real evolution is cumulative, in the sense that adaptive change are collected and propagated in the genome. That's the whole point of heritable variation and natural selection.  Evolution is to a target- adaptation- and in some situations (situations lacking irreducible complexity) partial adaptive improvements are preserved.  Models of evolution need to reflect that fact of nature.

Yet I believe that Dawkins and his defenders misunderstand the weasel program. The weasel model argues strongly for the inference to teleology/design, not for undirected evolution.

My reasons:

1) Dawkins inadvertently and inescapably uses a model of design. The computer itself, the program, the endpoint are all designed by minds. Even the random letter generator is a meticulously designed program. But the essence of Darwin's theory is that  there is no goal-directedness, no design, in evolution.  So why would Dawkins use a model that is entirely designed? The reason is obvious: he has no choice.  Design permeates nature.

2) The teleology in Darkins' weasel program is particularly in the target. The teleology permeates the system, but the crucial teleology is the adaptation- the match between the 'random' process and the environment.

This observation has been made by ID theorists, most notably Bill Dembski. It is not necessary to see the design as put into the system as a (figurative) hand coming down out of the clouds and tinkering with the  genes to make the organism better adapted. The intelligence in nature is in the match between the variation and the environment. The intelligence is in the fit, the adaptation itself, not just in the organism or just in the environment.

Dawkins' weasel program is a nice simulation of intelligent design.


  1. Mike,

    You could save yourself a lot of trouble by actually reading Dawkins's book. Or the Wikipedia entry Weasel program. It turns out that Dawkins was well aware of these limitations. Here is another excerpt from the same book you quoted:

    Although the monkey/Shakespeare model is useful for explaining the distinction between single-step selection and cumulative selection, it is misleading in important ways. One of these is that, in each generation of selective 'breeding', the mutant 'progeny' phrases were judged according to the criterion of resemblance to a distant ideal target, the phrase METHINKS IT IS LIKE A WEASEL. Life isn't like that. Evolution has no long-term goal. There is no long-distance target, no final perfection to serve as a criterion for selection, although human vanity cherishes the absurd notion that our species is the final goal of evolution. In real life, the criterion for selection is always short-term, either simple survival or, more generally, reproductive success.

  2. P.S. Your quote comes from p. 49. Mine is on p. 50.

  3. Oleg,

    I did read the book. It took me days to recover.

    Dawkins' Weasel program is b.s. He and his mignons presented it as evidence for how adaptive functional complexity could arise by a series of small steps despite the 'million monkey' problem, then he claimed 'oh, never mind. evolution isn't really like that'.

    If evolution isn't really like that, why the hell did he present the simulation.

    The fact is that nature is permeated with teleology, and evolution is not exception. Dawkins didn't use a non-teleological model because non-teleological models don't work, and he knew it.

    Dawkins' "Life isn't like that. Evolution has no long-term goal." is an ideological assertion, without scientific evidence to back it.


  4. If evolution isn't really like that, why the hell did he present the simulation.

    To illustrate cumulative selection. The answer is right there, in the book! :)

    There are of course programs that simulate evolution without a specified target. Avida by Richard Lenski. Ev by Thomas Schneider.

  5. Monkeys do not bash away at random, in the mathematical sense, at anything. It is total crap. See if you can get a dripping water faucet to encode these works as an analogy.
    What a stupid misapplication of mathematics. Chaos Theory gets some of its strength from the fact that we cannot find a real world example of truly random behavior.

  6. oleg,

    Avida 'organisms' are 'rewarded' with more rapid replicaton for adaptive success according to a very teleological paradigm- the programmer's choice.

    There are only two choices: non-teleology or teleology.

    Genuine randomness (non-teleology)-- a million monkeys-- can't produce anything resembing life.

    Only teleological processes can produce life. In simulations- methinks, Avida, etc, the teleology is always there, although sometimes hidden from people who don't really want to see it.

    Show me a successful simulation of evolution that lacks teleology utterly.


  7. Mike,

    You are using your own private definitions—again! Unless it's entirely random, it's teleological. Really?

    The word teleology implies that the programmer has a particular outcome in mind. That's true for weasel, not true for avida and ev.

  8. Of course, this is not the first time ID proponents demanded that simulated evolution be a totally random process. Gil Dodgen of Uncommon Descent came up with this gem:

    All computational evolutionary algorithms artificially isolate the effects of random mutation on the underlying machinery: the CPU instruction set, operating system, and algorithmic processes responsible for the replication process.

    If the blind-watchmaker thesis is correct for biological evolution, all of these artificial constraints must be eliminated. Every aspect of the simulation, both hardware and software, must be subject to random errors.

    The mind boggles at such silliness.

  9. Dr. Egnor said:

    Dawkins' weasel program is a nice simulation of intelligent design.

    Here is another example of intelligent design.

    Give a materialist enough rope and he will shoot himself in the foot!

  10. McDonald argument is to poke holes at the irreducible complexity argument, not at ID itself but at one of its main 'rationalization'. In this case as with the weasel program, you are taking an argument to prove or disprove something and applying this argument to another thing. I would hope that you realize that one can prove pretty much anything that way but... I think it'd be too much to hope.

  11. Anybody who stands up to defend the deterministic, will-arrive-at-the-phrase-with-probability-one Weasel program as some sort of illustration of Darwinian evolution simply forfeits all credibility.

    It really is that simple.

  12. The only point of the Weasel program is to demonstrate that the million-monkeys objection (or the aircraft-from-a-junkyard-tornado objection) does not apply to biological evolution. Cumulative changes will transform a system from a starting point to an ending point whether the ending point is known in advance (as in the case of artificial selection) or the species is responding to selection from the environment. Why wouldn't it? Why must the resulting change be teleological?

    It is a very simple point, and actually one of the primary ideas that Darwin put forward himself 150 years ago. You have already admitted that evolution works by artificial selection, now try extending your imagination to the natural world.

  13. Teleology and randomness are independent concepts that can coexist without strain. Example 1: I buy a lottery ticket, and pick random numbers. My goal is to win the lottery, but I am doing it with a tool that incorporates randomness. Example 2: I am bored and I start tapping my pencil rhythmically, without really intending to or not. It's not random, it's cyclic and very predictable. Yet I have no conscious purpose for doing it, and I very well might not even KNOW I am doing it until the annoyed person sitting next to me asks me to stop.

    Example 3: Evolution is not random, yet it does not have an implicit teleologic purpose any more or less than any other natural phenomenon does. Male peacocks don't choose to evolve showy feathers because it helps their mating chances. Giraffes don't have any say in whether or not their offspring have taller necks, to reach high branches. If you propose that the choice exists and that it belongs to a supernatural deity, you're welcome to that belief but don't tell people it's scientific. Or that we need said belief to explain the diversity of life on Earth.

    As for implicit design in demonstrations and experiments: Demonstrations and experiments incorporate design by definition, and that has absolutely nothing to do with the principle one is trying to demonstrate. Allow me to paraphrase your logic:

    1. I put random-shaped objects in a wind tunnel, and observe that none of them fly.
    (More generally, I do an experiment that consists only of random, non-designed steps. Unsurprisingly, I get no interesting results.)

    2. I design a bird-shaped object and put it in a wind tunnel. I observe that this one, unlike my previous objects, has unique aerodynamic properties. (i.e. maybe it flies)
    (I then do an experiment that was carefully designed to demonstrate a principle. Oh look, I see something interesting!)

    3. Therefore, all bird-shaped objects, i.e. living birds, must be designed!
    (The principal I was studying must be designed, because I had to do a designed experiment to test it! Voila!) <- FAIL

    This is my demonstration of the concept known as "Egnorance." (There is a cool meta-corollary here...)

  14. Why must the resulting change be teleological?

    Try driving from New York to Los Angeles with your hands off the steering!

    If you succeed, I will believe in Darwinism...

  15. Pepe, you painfully miss the entire point that was observed by vreejack, or you ignore it.

    You then go on to say, incorrectly, that evolution is as improbable as driving to LA with your hands off the steering wheel.

    A car is not "trying" to get to LA any more than a giraffe is "trying" to evolve a taller neck.

    You find cars in LA for the same reason you find giraffes with tall necks. They made it there through a combination of chance and being subject to basic "rules of the road." Those you see in LA are a minute fraction-- most ended up elsewhere, or crashed along the side of the road.

    Your car analogy is very limited. Yet it sheds some light on your misconceptions about evolution... perhaps you over-focus on the random elements that play into evolution, without acknowledging the mechanistic elements.

    The real point here is that (in the terms of your analogy) Evolution doesn't "care" if it drives to LA, Detroit, New York, or the gutter. However, it's not driving with it's hands off the wheels either (evolution is not random). And, it's not merely one car but a vast number of cars (populations). Each car is driving on a different route (phenotypic variation). A car is not "trying" to get to LA any more than a giraffe is "trying" to evolve a taller neck. Many cars don't make it anywhere (natural selection), let alone the miniscule fraction that make it to LA.

  16. You then go on to say, incorrectly, that evolution is as improbable...

    I never said evolution was improbable, only that the neo-Darwinian explication of it does not hold water. Changes do occur within species (micro-evolution) but Darwin did not prove speciation (macro-evolution), he only extrapolated from micro-evolution and speculated macro-evolution was true. He should have titled his book, The Adaptation of Species because he actually had no idea how species originate.

    ...evolution is not random...

    You are so right! Evolution is intelligently designed.

  17. Pepe- I stand by my interpretation of your original statement. The modern synthesis of evolution, neo-Darwinism, whatever. You are quibbling over irrelevancies, and failing to address the substantial errors in your thinking.

    You also demonstrate that Darwin had a much clearer idea of speciation than you yourself do.

    Your last point is even more foolish than the rest. Just run around and yell, "god did it!" why don't you.

    This conversation has terminally devolved.

  18. ...quibbling over irrelevancies...
    ...substantial errors in your thinking...
    Your last point is even more foolish than the rest...

    This is typical atheist incendiary rhetoric, ad hominem attacks and dispicable sophistry!

    Hope you still remember the post was about Dawkins' weasel program...

  19. I would find the Weasel program to be a perfectly acceptable illustration of an important general principle if the target phrase were replaced with "YOUJUSTDRANKTHEKOOLAID".

  20. If you'd like an example of a genetic algorithm that doesn't have a target phrase you can have a look at my evolutionary computing approach to designing keyboard layouts here: