Freethought & Rationalism Archive

donkey · 12-28-2002, 10:58 AM

This is an interesting attempt to demonstrate that Dawkins Blind Watchmaker simulation was flawed.

http://www.rae.org/MutationProgram.htm

theyeti · 12-28-2002, 11:11 AM

Why creationists/IDists have such a fascination with Dawkins' methinksitsaweasel program is beyond me. The only point of the program is to demonstrate the efficiency of selection over a blind search; it's not meant to be an accurate simulation of evolution. There are plenty of genetic algorithms that are actually intended to simulate biological evolution, in addition to solving real world problems, which they do an excellent job at. The fact that anti-evolutionists spend their time attacking the over-simplified weasel simulation just proves how weak their case is.

theyeti

Principia · 12-28-2002, 11:18 AM

There are more robust evolution simulators, some of which are listed on this thread. If anything, Dawkins' program illustrated the notion that selection reduced the number of trials required to reach a target as compared to pure random guess. For that purpose, it sufficed.

copernicus · 12-28-2002, 11:20 AM

Donkey, I fail to see how Bergman and Sharp's argument has any relevance at all to Dawkins' argument. He was at great pains to emphasize that evolution was not about random mutation, but about environmentally-biased mutation. His simulations imposed "intelligent design" on the process in order to demonstrate how random mutations provided a path to a goal, not a guidance mechanism. The anti-ID element of his argument was simply that environmental pressure, rather than intelligent manipulation, was the guidance mechanism for what we observe in nature. That is, Dawkins argued for the existence of a "watchmaker". His watchmaker just didn't need intelligence, i.e. "sight", in order to make watches.

Dr Rick · 12-28-2002, 12:18 PM

Quote:

In his book, The Blind Watchmaker, Richard Dawkins produced a computer simulation to show that within a relatively small number of 43 generations random mutations can produce a meaningful result. He set as the goal for the program to produce the phrase: �Methinks it is a weasel.� We recreated this simulation, and introduced a number of other factors that Dawkins probably did not consider.

Probably because, unlike these people, Dawkins is smart, educated, and honest...

Quote:

First of all, we make an assumption that when a letter (or gene) matches a desired goal, there is something that prevents it from further mutation, even though there is no biochemical evidence or process to back up this assumption.

The process of natural selection favors conservation of favorable or "desired" mutations over undesirable or inconsequential ones; these guys are setting up a strawman because they don't have a real argument to make. There is nothing that prevents further mutations of a mutated genome, and there is no reason to assume that there should be. Natural selection favors the survival and reproduction of desirable mutations, however, so there is a process that conserves them, contrary to their assertion otherwise.

Quote:

Second, if we presume that assumption to be true, then once the organism achieved the �goal� it is forever prevented from further evolution. His thesis (and therefore evolution) is refuted both ways, for if we take away this �mutation save� function, our simulation goes on for millions of generations without achieving the goal.

This is utter nonsense: these guys are actually trying to argue that "evolution prevents evolution." Mutations can and do continue in an organism that already has mutations with selective environmental pressures favoring the survival and reproduction of those genomes which express "desirable" mutations; there is nothing that prevents "further evolution" in Dawkins's thesis or the theory of evolution.

Quote:

Without the source code to Dawkins� program, we can only run a simulation that gives the benefit of the doubt far beyond what is reasonable, and establish limits within which evolution can occur.

It is neither reasonable nor honest to argue a strawman.

Quote:

In addition to testing the results both with and without the save function...

Both the "save" function and the "unsaved" function ignores natural selection, an essential feature of evolution.

Quote:

our computer simulation adds into the mixture several more factors involved in mutations that serve to increase the number of generations. Our simulation saves the results in a table if the results match a specified number of letters in the goal. When we ran the simulation without the save function for more that 7 million generations, we could only produce 6 matches on 9 letters out of 28.

In other words, "garbage in, garbage out."

Quote:

We also found that there were several ways to do the simulation with the save function. Originally, within one generation, if the random mutation fell on a letter that was �saved,� the mutation simply didn�t occur, and the program went on to the next generation. The results in that case were that it took between 2500 and 4500 generations to produce the goal sentence.

Within one generation? Why would you want to conserve a desirable mutation only within one generation? That does not duplicate evolution: favorable mutations are passed on for generations, not just one.

Quote:

Next, to try to achieve the results Dawkins claimed, we presumed that if a mutation fell upon a �saved� letter, the program went on to try and find another letter to mutate within the same generation. The number of generations was reduced down to 500 � 700. This still does not match the results Dawkins got. The only way I can see for him to get it down to 43 generations would be for him to mutate EVERY gene EVERY generation. This is totally unrealistic.

What's unrealistic is to assert that only one mutation can occur per generation. Reproducing organisms often produce many progeny, not just one, greatly magnifying the chances of mutations. There is no reason to assume that only one mutation can occur with each generation.

Quote:

We distorted the experiment in the following manner to give Dawkins every benefit of the doubt:
1. Mutations are saved if it matches the goal
2. Within the same generation, if a mutation is tried on a letter that matches a goal, mutations continue to occur until one is found that doesn�t match.
3. Within the same generation, every letter is mutated.

We found that if you remove these false assumptions and allow random mutations to occur freely whether they match the goal or not, you can run the simulation for millions of generations without ever getting close to the goal.

The �save function,� therefore, is the overriding factor that might allow Dawkins� conclusions to be possible.

There is a powerful save function in nature called natural selection that favors desirable mutations; evolution is not just a process of random letter generation as these guys imply.

Quote:

In addition, we know the following to be true about mutations:
1. Mutations may be inserted, moving the entire string over one gene.
2. Mutations favor �hotspots� and rarely occur elsewhere.
3. Mutations heavily favor the T gene.
4. Mutations may result in unusable or meaningless code, represented by special characters in our simulation.
5. Mutations may kill the organism with a �poison factor� and stop the process.
We have added these factors in our simulation so that we can give a more realistic representation of what actually happens with mutations. Even with the save function turned on, the addition of these factors causes the mutations to further destabilize. With all of the factors turned on, it either died with the poison factor, or it took 23,000 generations to match the goal.

If the limitations these guys arbitrarily impose on mutations were true, there would be no antibiotic or pesticide resistance.

Even at more complex levels, it only takes a couple of hundred or so mutations to turn a mouse genome into a human one.

Quote:

But the most amazing revelation we had when we were doing this simulation is that if Dawkins� save function were true, it would prevent the organism from ever evolving again, not to mention that implicitly if you try to match a goal, you are introducing intelligent design into the equation.

Wow, two strawmen in one sentence:

Evolved organisms can evolve, there is no reason to assume they can't. Organisms that have evolved continue to evolve as random mutations occur under the influence of natural selection.

There is no goal to match, and that was not the purpose of Dawkins's exercise. Dawkins set out to demonstrate an algorithm that would allow order to arise from random permutations under selective pressures, and these guys set-up several strawman and inserted false assumptions in a futile attempt to discredit his work.

Quote:

This computer simulation serves to quantify the problem in a concrete, demonstrable manner.

The only thing it concretely demonstrates is the intellectual dishonesty of ID'ers.

Rick

Doubting Didymus · 12-29-2002, 02:55 AM

Quote:

First of all, we make an assumption that when a letter (or gene) matches a desired goal, there is something that prevents it from further mutation, even though there is no biochemical evidence or process to back up this assumption.

This gave me a nice little giggle.

Actually, I think this quote demonstrates some of the problems that start appearting when you try to practise a major scientific field (i.e. biology), but need to edit out those findings that contradict ones ideology (i.e. IDists selectively rejecting evolutionary science). It is quite clear to me that the writers of these scentences have not the faintest clue what natural selection even is , and thus they are chasing a basis for the 'assumption' that beneficial mutations stay in the population in the field of biochemistry of all things.

Really? No biocheamical basis for mutation fixation? Boy howdy! And look: no astronomical basis for the assumption that organisms are comprised of cells, no basis in quantum physics for the silly idea that germs cause disease, and worse! No basis in geololgy for the 'growth of human toenails' hypothesis.

Whats the difference between an IDist and an evolutionist? Remedial biology.

idiom · 12-30-2002, 02:41 AM

Isnt natural selection supposed to filter out mutations?

If natural selection kept every mutation every time you end up with a garbage sentence first generation.

But of course Natural Selection KNOWS what the final sentence is going to be and is GUIDING and making sure that every mutated sentence it keeps is a VALID sentence unto itself right.

Quote:

The process of natural selection favors conservation of favorable or "desired" mutations over undesirable or inconsequential ones

Cuz its smart right. You may not get a perfect sentence first generation but you know the sentence is coming so you can make do with it right.

Quote:

Mutations can and do continue in an organism that already has mutations with selective environmental pressures favoring the survival and reproduction of those genomes which express "desirable" mutations

And naturally a slightly mutated malfunctioning sentence is more desirable right...

Not saying that natural selection is wrong or that it doesnt exist or anything, it just needs more research. As soon as we contact those aliens they will explain to us how they made the "genetic algorithms" that allow adaption to take place. Then the IDists will really have a mess. They will have to deny the existence of aliens. Such fools.

pz · 12-30-2002, 05:49 AM

Quote:

Originally posted by idiom
Isnt natural selection supposed to filter out mutations?

If natural selection kept every mutation every time you end up with a garbage sentence first generation.

But of course Natural Selection KNOWS what the final sentence is going to be and is GUIDING and making sure that every mutated sentence it keeps is a VALID sentence unto itself right.

Cuz its smart right. You may not get a perfect sentence first generation but you know the sentence is coming so you can make do with it right.

Are you seriously advocating these ideas, or are you trying to parody creationist misconceptions about selection? Everything you've said above is completely wrong to such a degree that it is kind of comical, but some people do actually believe what you've said.

idiom · 12-30-2002, 06:03 AM

I'm not entirely sure.

I think its the constant personification (deification?) of natural selection that confuses me.

I put my head down and study some more yes?

pz · 12-30-2002, 06:18 AM

Quote:

Originally posted by idiom

I put my head down and study some more yes?

Yes.

For instance, this comment, " But of course Natural Selection KNOWS what the final sentence is going to be and is GUIDING and making sure that every mutated sentence it keeps is a VALID sentence unto itself right", is inherently wrong. There is no final sentence. There is no guidance towards a particular destination. The only part that is partially correct is the recognition that every surviving, successful intermediate has to be valid; however, it is not true that every mutant sequence has to be valid.

12-28-2002, 10:58 AM	#1
donkey New Member Join Date: Nov 2002 Location: usa Posts: 1	Argument against the Blind Watchmaker This is an interesting attempt to demonstrate that Dawkins Blind Watchmaker simulation was flawed. http://www.rae.org/MutationProgram.htm

12-28-2002, 11:11 AM	#2
theyeti Veteran Member Join Date: Jun 2001 Location: Denver, CO, USA Posts: 9,747	WTF Why creationists/IDists have such a fascination with Dawkins' methinksitsaweasel program is beyond me. The only point of the program is to demonstrate the efficiency of selection over a blind search; it's not meant to be an accurate simulation of evolution. There are plenty of genetic algorithms that are actually intended to simulate biological evolution, in addition to solving real world problems, which they do an excellent job at. The fact that anti-evolutionists spend their time attacking the over-simplified weasel simulation just proves how weak their case is. theyeti

Thread Tools	Search this Thread
Show Printable Version	Search this Thread: Advanced Search

12-28-2002, 11:18 AM	#3
Principia Veteran Member Join Date: Mar 2002 Location: anywhere Posts: 1,976	There are more robust evolution simulators, some of which are listed on this thread. If anything, Dawkins' program illustrated the notion that selection reduced the number of trials required to reach a target as compared to pure random guess. For that purpose, it sufficed.

12-28-2002, 11:20 AM	#4
copernicus Veteran Member Join Date: Jul 2001 Location: Bellevue, WA Posts: 1,531	Donkey, I fail to see how Bergman and Sharp's argument has any relevance at all to Dawkins' argument. He was at great pains to emphasize that evolution was not about random mutation, but about environmentally-biased mutation. His simulations imposed "intelligent design" on the process in order to demonstrate how random mutations provided a path to a goal, not a guidance mechanism. The anti-ID element of his argument was simply that environmental pressure, rather than intelligent manipulation, was the guidance mechanism for what we observe in nature. That is, Dawkins argued for the existence of a "watchmaker". His watchmaker just didn't need intelligence, i.e. "sight", in order to make watches.

12-30-2002, 06:03 AM	#9
idiom Junior Member Join Date: Dec 2002 Location: New Zealand Posts: 49	I'm not entirely sure. I think its the constant personification (deification?) of natural selection that confuses me. I put my head down and study some more yes?

Freethought & Rationalism Archive

The archives are read only.