PZ Myers Does It Again

_{Clive Hayden

August 20, 2009

Darwinism, Education, Evolution, Intelligent Design, Science}

Share: Facebook; Twitter; LinkedIn; Flipboard; Print; Email

PZ Myers has, once again, railed against something that he doesn’t understand at his blog Pharyngula. Hi PZ! Notice that he doesn’t actually address the content of Dr. Dembski and Dr. Marks’ paper, which you can read here: Conservation of Information in Search: Measuring the Cost of Success, published at the IEEE. Given his argument, he doesn’t know how to measure the cost of success, yet claims that Dr. Dembski doesn’t understand selection. A bit of advice PZ, the argument presented by Dr. Dembski and Dr. Marks is very sophisticated PZ, your mud slinging isn’t PZ, you need to step it up PZ. I know this new stuff isn’t ez, but you may want to consider a response that has actual content PZ. Your argument against this peer-reviewed paper is still in its infancy, or, more accurately, still in the pharyngula stage, embryonic in its development.

Since evolution of the kind PZ subscribes to cannot be witnessed, the argument has moved into genetic algorithms with the advent of computational abilities to determine the affair, and the IEEE is an entirely appropriate place to publish on that subject. We’re not going anywhere, we’ll give him time to catch up and educate himself to the tenets of the paper’s actual content. And if/when he does, maybe he’ll write another blog, and possibly write one with active information, that is, actual information, or else his argument will never reach it’s target.

Comments

Alex73#63
Let’s suppose the following: 1. The program chooses a winner only if it has more matching letters than the parent string. (i.e. in case of the best child string(s) having the same number of matches as the parent one the entire generation is discarded) It seems to be perfectly in line with Dawkins’ own words.
Actually, it doesn't. Dawkins' words are:
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.
It is mathematically possible for all progeny to be less fit than the parent. In that case, the Weasel algorithm would pick the best of the bad litter and keep going.
A quick calculation suggests that it would be the case at 1% probability for a letter mutation. BTW it looks like this figure could produce the the results seen in the book.
The site I mentioned above uses a rate of 5% and produces results similar to those in the book. I seem to remember someone running a large number of trials with different population sizes and mutation rates the last time this topic came up, but I can't find it at the moment.DeLurker_{August 21, 2009
August
08
Aug
21
21
2009
03:52 PM
3
03
52
PM
PDT}

R0b, Ahh, I see, "something interesting", like anything that accumulates in any sort of way, even an accumulation of letters from monkeys would produce "something interesting" whatever that means. We have now left the weasel analogy far behind, and are now looking for "something interesting", again, whatever that means. In the respect that things accumulate, is that really "interesting" in virtue of itself? It would of course depend on what accumulated, and if there is no target that is being approximated to in the accumulation, then the accumulation is anything, randomness, and always will be, for there would be nothing to compare the accumulation to, no standard of comparison. Call this "interesting" if you want, for there is no arguing personal taste. Clive Hayden_{August 21, 2009
August
08
Aug
21
21
2009
03:29 PM
3
03
29
PM
PDT}

Clive:
It cannot be common and uncommon at the same time.
Of course it can't. Cumulative selection is common, so if you think it isn't, then you're wrong.
If the phrase is not a target, then any accumulation goes, and the whole weasel endeavor is for naught.
Irrelevant. Just because WEASEL defines fitness in terms of a distant target doesn't mean that every evolutionary algorithm or process has to do so in order to produce something interesting.R0b_{August 21, 2009
August
08
Aug
21
21
2009
03:25 PM
3
03
25
PM
PDT}

R0b, Do you have any biological examples of cumulative selection? And if there isn't a target then "cumulative" is meaningless.Joseph_{August 21, 2009
August
08
Aug
21
21
2009
03:20 PM
3
03
20
PM
PDT}

BillB, Given a large enough population size and a small enough mutation rate you will never see a character reversal. That is ratcheting. Cumulative means to increase by succesive additions. Ratchet means to move in degrees in one direction only. Dawkins clearly meant that cumulative selection is equal to ratcheting.Joseph_{August 21, 2009
August
08
Aug
21
21
2009
03:19 PM
3
03
19
PM
PDT}

R0b, It cannot be common and uncommon at the same time. If the phrase is not a target, then any accumulation goes, and the whole weasel endeavor is for naught.Clive Hayden_{August 21, 2009
August
08
Aug
21
21
2009
03:15 PM
3
03
15
PM
PDT}

Clive:
Because that’s what it says. If there is no target, there is no weasel phrase. Cumulative would mean anything that accumulated, even nonsense phrases.
No, that's not what it says. Biological life and WEASEL have cumulative selection in common, regardless of whether there is a target or not. So you're wrong in acting as if Dawkins thinks that they have nothing in common. And you're wrong in acting as if their differences makes this commonality cease to be common.R0b_{August 21, 2009
August
08
Aug
21
21
2009
03:09 PM
3
03
09
PM
PDT}

SCheesman:
This reflects the quote given on Mark’s and Dembski’s “Evolutionary Informatics” web page at http://www.evolutionaryinformatics.org/ “… no operation performed by a computer can create new information.” – Douglas G. Robertson, “Algorithmic Information Theory, Free Will and the Turing Test,” Complexity, Vol.3, #3 Jan/Feb 1999, pp. 25-34.
And a similar statement by Leon Brillouin is also frequently quoted in EIL work, with the implication that these quotes have something to do with EIL concepts. They don't. This is simply equivocation on the word "information". Robertson was explicitly talking about algorithmic information, not active, endogenous, or exogenous information. Algorithmic information, by definition, is not increased by any computational process. This fact has nothing to do with the EIL's conservation of information. Brilloun is very explicit that computers don't increase information because they are deterministic. Again, this has nothing to do with the EIL conservation of information, which deals with probabilistic searches.R0b_{August 21, 2009
August
08
Aug
21
21
2009
02:59 PM
2
02
59
PM
PDT}

R0b,
Given the context, why in the world do you interpret “that” as referring to the whole of WEASEL? And given the first sentence, why would you think that Dawkins sees WEASEL as being different from life in all respects?
Because that's what it says. If there is no target, there is no weasel phrase. Cumulative would mean anything that accumulated, even nonsense phrases.Clive Hayden_{August 21, 2009
August
08
Aug
21
21
2009
02:58 PM
2
02
58
PM
PDT}

This is more interesting than I initially thought. Staying with the book, anyone can see that the mutation rate is pretty low, because even incorrect letters are likely preserved. Let's suppose the following: 1. The program chooses a winner only if it has more matching letters than the parent string. (i.e. in case of the best child string(s) having the same number of matches as the parent one the entire generation is discarded) It seems to be perfectly in line with Dawkins' own words. 2. The mutation rate is so low that during the simulation there is only a very low chance for any generation to have a child where: a: three mutations happen b: two of these produces the matching letter from previously non-matching ones c: one mutation changes a previous matching letter into a non-matching one then the algorithm will produce results just like a latching algorithm without any explicit latching. A quick calculation suggests that it would be the case at 1% probability for a letter mutation. BTW it looks like this figure could produce the the results seen in the book. Now if a non-latching algorithm is unlikely to produce results different from a latching one, I think it is arguable that it can be evaluated as the latching version.Alex73_{August 21, 2009
August
08
Aug
21
21
2009
02:58 PM
2
02
58
PM
PDT}

Clive:
Dawkins didn’t say “Life is somewhat like that in some respects and not like that in others.” He said “Life’s not like that.” Period. It’s a little strange that it was used at all.
Let's look at the whole paragraph from TBW:
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.
WEASEL illustrates cumulative selection, a fundamental principle of biological life. It is different from life in that it has a long-term goal while life does not. "Life isn't like that" is smack in the middle of a paragraph that discusses that difference. Given the context, why in the world do you interpret "that" as referring to the whole of WEASEL? And given the first sentence, why would you think that Dawkins sees WEASEL as being different from life in all respects?R0b_{August 21, 2009
August
08
Aug
21
21
2009
02:40 PM
2
02
40
PM
PDT}

Clive Hayden#57
DeLurker,
Dawkins himself explains, just before the bit 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.
(Bolding mine, I hope it comes through.) All the Weasel algorithm was intended to do was to explain the difference between random selection and cumulative selection. Nothing more.
I see, one misleading analogy to explain the misleading part of another misleading analogy. I guess I cannot blame Dawkins, it’s hard to get a handle on real life, and we must revert to analogies, and then claim that our analogies have no basis in real life.
Could you please point out exactly what is misleading? The Weasel algorithm is a tool for explaining one aspect of evolutionary theory. Dawkins is very careful in The Blind Watchmaker to specify the limits of the tool. The rest of the book goes considerably beyondthis simple tool. In context, there is nothing misleading about the Weasel algorithm.DeLurker_{August 21, 2009
August
08
Aug
21
21
2009
02:28 PM
2
02
28
PM
PDT}

kibitzer,
“But Dawkins states that targets don’t exist in ToL.” So do we take his word for it?
If there are targets in evolution, then instead of arguing against weasel, Dr. Dembski should use a real example from biology.
If his algorithm is supposed to model evolution, then evolution does involve a targeted search. Dawkins is speaking out of both sides of his mouth.
I didn't think he was talking out of both sides of his mouth when I first read TBW. He presents a simple model which illustrates some aspects of evolution well, but not others. Again, Dawkins explains very clearly the shortcomings of the model, and in the end, I think a reasonable reader would understand the points Dawkins is trying to make.
Where in The Blind Watchmaker does Dawkins explicitly deny ratcheting? In the example he gives (METHINKS…), ratcheting appears to take place. So why not give Dembski the benefit of the doubt?
I don't think we can fault Dawkins for not addressing every possible misinterpretation of the algorithm. Maybe the latching objection didn't occur to him at the time (it makes no sense from an evolutionist's perspective for starters). Also, space is at a premium in a book, and I would imagine he would want to keep his explanation as uncluttered as possible.yakky d_{August 21, 2009
August
08
Aug
21
21
2009
01:50 PM
1
01
50
PM
PDT}

This whole "no latching" objection seems absurd to me. The success of the weasel program relies on some pretty good psuedo-latching at least, to prevent the character strings from drifting away from their past successful matches. It matters very little whether Dawkins explicitly wrote in latching code or whether he chose mutation rates and population sizes to prevent "error catastrophe". Either way, he smuggled in some necessary information (I mean apart from the giant heist of comparing everything to his distant target string). Their constant defending of the WEASEL program as if it meant a damned thing for their case is just one of many reasons why I conclude that Darwinists really have very little idea of what they're talking about. All they've got is a whole bunch of wishful thinking.Matteo_{August 21, 2009
August
08
Aug
21
21
2009
01:40 PM
1
01
40
PM
PDT}

Nakashima:
I do have trouble with the concept of conservation of information which underlies the concept of active information. Suppose I have a program that computes the digits of pi. If I run this program and produce the first million digits, do they have the same information content? If I ran it for the first billion digits, do they have the same information content?
In fact, the number of decimals of pi is not a measure of information at all; it is merely a representation with different degrees of accuracy when expressed in a given numeric base. The true amount of information contained in the program is that required to specify the formula for the generation of pi, and that is independent of the number of decimals generated using the program. This reflects the quote given on Mark's and Dembski's "Evolutionary Informatics" web page at http://www.evolutionaryinformatics.org/
"... no operation performed by a computer can create new information." -- Douglas G. Robertson, "Algorithmic Information Theory, Free Will and the Turing Test," Complexity, Vol.3, #3 Jan/Feb 1999, pp. 25-34.
Put another way, the decimals of pi output by the program add not a single bit to the amount of information already stored in the program necessary to generate them.SCheesman_{August 21, 2009
August
08
Aug
21
21
2009
01:01 PM
1
01
01
PM
PDT}

DeLurker,
Dawkins himself explains, just before the bit 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. (Bolding mine, I hope it comes through.) All the Weasel algorithm was intended to do was to explain the difference between random selection and cumulative selection. Nothing more.
I see, one misleading analogy to explain the misleading part of another misleading analogy. I guess I cannot blame Dawkins, it's hard to get a handle on real life, and we must revert to analogies, and then claim that our analogies have no basis in real life.Clive Hayden_{August 21, 2009
August
08
Aug
21
21
2009
12:55 PM
12
12
55
PM
PDT}

R0b, It's a little more than strange that it would be invented and used if it has no import with actual life. Dawkins didn't say "Life is somewhat like that in some respects and not like that in others." He said "Life's not like that." Period. It's a little strange that it was used at all.Clive Hayden_{August 21, 2009
August
08
Aug
21
21
2009
12:49 PM
12
12
49
PM
PDT}

Mr Kibitzer, You raise an interesting issue, which is obscured by all the predictible response to Dr Dembski's predictible provocation. The paper does go beyond defining active information to asserting a knowledge of the source of the active information. (I do have trouble with the concept of conservation of information which underlies the concept of active information. Suppose I have a program that computes the digits of pi. If I run this program and produce the first million digits, do they have the same information content? If I ran it for the first billion digits, do they have the same information content?)Nakashima_{August 21, 2009
August
08
Aug
21
21
2009
12:48 PM
12
12
48
PM
PDT}

Clive:
Life isn’t like that. In other words, what he demonstrated with the Weasel analogy, is really not an analogy for life, or for anything he’s trying to explain within life.
If Weasel were like Life in every respect, then it would not be an analogy -- it would be Life itself. Weasel is like Life in some respects (selective reproduction with random mutation) but not in others. Everyone including Dawkins agrees on that, so it's a little strange that this criticism has persisted for more than 2 decades.R0b_{August 21, 2009
August
08
Aug
21
21
2009
12:30 PM
12
12
30
PM
PDT}

PS: Another red herring
Another rhetorical dismissal. Please acquire some manners and make an attempt to understand other peoples posts rather than just dismissing them.BillB_{August 21, 2009
August
08
Aug
21
21
2009
12:15 PM
12
12
15
PM
PDT}

In over 200 places where letters in principle could revert, of over 300 letters total, we see no reversions.
Good grief man! That's exactly what will happen when you run the algorithm without a latching mechanism, as you almost admit here:
latching may be explicit or implicit, the latter being based on parameters and selection filter specifications.
Calling it implicit latching is just a semantic device to avoid having to admit that you are wrong. With mutation rates and pop sizes within certain ranges the fittest member of each generation will have an exceedingly low probability of having any correct letter mutated to an incorrect letter. When you 'print off' a small sample of fittest phrases over an entire run, as Dawkins did in his book, it is highly unlikely that you will see any reversions. You can call that implicit latching if you want but it is not an accurate representation of the facts because with a mutation rate anywhere above zero any letter can mutate, even a correct one.BillB_{August 21, 2009
August
08
Aug
21
21
2009
12:10 PM
12
12
10
PM
PDT}

Clive Hayden#49
Life isn’t like that. In other words, what he demonstrated with the Weasel analogy, is really not an analogy for life, or for anything he’s trying to explain within life. Then why offer it at all? What is it supposed to analogize if not life?
Dawkins himself explains, just before the bit 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.
(Bolding mine, I hope it comes through.) All the Weasel algorithm was intended to do was to explain the difference between random selection and cumulative selection. Nothing more. It's just a cartoon version of one evolutionary mechanism, and Dawkins explains that. However, even cartoons can be a useful teaching tool, which is what the Weasel algorithm was developed to be.DeLurker_{August 21, 2009
August
08
Aug
21
21
2009
12:05 PM
12
12
05
PM
PDT}

R0b,
Nor do you for anything biological. How do you do it if you’re not given a target, a lower-level search space, and a higher-level search space?
If we, as human intelligent agents, cannot measure the cost of success, then we cannot say what is successful, for if the cost outweighs the "success", it really isn't successful, in the respect that it is too costly for the attending benefit. This is an argument against evolution if anything.Clive Hayden_{August 21, 2009
August
08
Aug
21
21
2009
11:55 AM
11
11
55
AM
PDT}

DeLurker, Dawkins said "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." Life isn't like that. In other words, what he demonstrated with the Weasel analogy, is really not an analogy for life, or for anything he's trying to explain within life. Then why offer it at all? What is it supposed to analogize if not life? It's not analogous to non-life, or, death. Maybe it was posited to evidence nothing at all, only to confuse people that it should have import with life systems, when in reality it doesn't. I've seen a lot of confusion with this analogy from folks who believe it is like life.Clive Hayden_{August 21, 2009
August
08
Aug
21
21
2009
11:39 AM
11
11
39
AM
PDT}

PS: Another red herring -- change the target half way through. THE of course is the definite article.kairosfocus_{August 21, 2009
August
08
Aug
21
21
2009
11:35 AM
11
11
35
AM
PDT}

Joseph: One adjustment: CRD did not have to use EXPLICIT latching, to get the sort of apparent -- indeed, evident -- latching effect we can see in the 1986 published o/p from "good" runs: ____________________ >> We may conveniently begin by inspecting the published o/p patterns circa 1986, thusly [being derived from Dawkins, R, The Blind Watchmaker , pp 48 ff, and New Scientist, 34, Sept. 25, 1986; p. 34 HT: Dembski, Truman]: 1 WDL*MNLT*DTJBKWIRZREZLMQCO*P 2? WDLTMNLT*DTJBSWIRZREZLMQCO*P 10 MDLDMNLS*ITJISWHRZREZ*MECS*P 20 MELDINLS*IT*ISWPRKE*Z*WECSEL 30 METHINGS*IT*ISWLIKE*B*WECSEL 40 METHINKS*IT*IS*LIKE*I*WEASEL 43 METHINKS*IT*IS*LIKE*A*WEASEL 1 Y*YVMQKZPFJXWVHGLAWFVCHQXYPY 10 Y*YVMQKSPFTXWSHLIKEFV*HQYSPY 20 YETHINKSPITXISHLIKEFA*WQYSEY 30 METHINKS*IT*ISSLIKE*A*WEFSEY 40 METHINKS*IT*ISBLIKE*A*WEASES 50 METHINKS*IT*ISJLIKE*A*WEASEO 60 METHINKS*IT*IS*LIKE*A*WEASEP 64 METHINKS*IT*IS*LIKE*A*WEASEL >> _______________________ In over 200 places where letters in principle could revert, of over 300 letters total, we see no reversions. Add this to the statement by Dawkins in BW: >> It . . . begins by choosing a random sequence of 28 letters ... 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 . . . . What matters is the difference between the time taken by cumulative selection, 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 >> To get that difference, CRD resorts to a proximity based targetted selection process that rewards not functionality -- dismissed as "single step selection" -- but instead mere closeness to the pre-loaded target phrase. Which means that Weasel does not CREATE information, but clumsily replicates already existing information. This is fundamentally unlike what is claimed for evolution by random variations plus natural selection of the fittest sub populations. Also, the runs above and he words about cumulative selection strongly suggest EXPLICIT latching as a natural and legitimate interpretation. We may see such from Dictionary dot com: >> cu·mu·la·tive (kymy-ltv, -y-l-tv) adj. 1. Increasing or enlarging by successive addition. 2. Acquired by or resulting from accumulation. >> That is what was already achieved is locked in and new progress is incremental on top of that. Locked in and latched are synonymous. As my always linked has noted for months, [and as came out in earlier exchanges] latching may be explicit or implicit, the latter being based on parameters and selection filter specifications. But then you and I have been called dishonest and worse -- up to today -- for simply stating the unpopular truth on this: the emperor has no clothes. And, it seems that the cognitive dissonance at work among Darwinists has filled many with a need to deflect inconvenient facts, through red herrings, strawman distortions and ad hominem attacks. Sad, but revealing of the parlous state of Darwinism in the Darwin 200 year. GEM of TKIkairosfocus_{August 21, 2009
August
08
Aug
21
21
2009
11:27 AM
11
11
27
AM
PDT}

And a partitioned search is the same as Dawkins’ cumulative selection.
No it isn't, partitioned search explicitly locks part of the candidates configuration out of the search once has reached a goal. Dawkins algorithm does not. To illustrate the difference and its importance: Partitioned search will fail if you change the target half way through (Unless you explicitly build in a method of un-latching letters), WEASEL will chase whatever target is is given, even if this changes from generation to generation.BillB_{August 21, 2009
August
08
Aug
21
21
2009
11:10 AM
11
11
10
AM
PDT}

Clive:
Given his argument, he doesn’t know how to measure the cost of success, yet claims that Dr. Dembski doesn’t understand selection.
Yes, it's doubtful that PZ has actually read the paper, so he probably doesn't know how to measure the cost of success. Nor do you for anything biological. How do you do it if you're not given a target, a lower-level search space, and a higher-level search space? But the points that PZ makes about the paper are correct. The paper is about searches, with no attempt to tie them to biology. And since Dembski has consistently portrayed WEASEL as using latching rather than selection, in spite of the fact that WEASEL's explicitly stated purpose is to illustrate selection, and in spite of being corrected on this for years, PZ's razzing doesn't seem overly harsh.R0b_{August 21, 2009
August
08
Aug
21
21
2009
11:06 AM
11
11
06
AM
PDT}

kibitzer#41
Where in The Blind Watchmaker does Dawkins explicitly deny ratcheting?
Why would he deny something that isn't a component of the mechanism he was trying to demonstrate? No biologist suggests that natural selection always preserves "correct" genes. Doing so would eliminate the purpose of Dawkins' demonstration. Follow along the example at http://www.softwarematters.org/more-weasel.html and point out, if you can, where anything that Dawkins says could possibly be interpreted to suggest ratcheting.DeLurker_{August 21, 2009
August
08
Aug
21
21
2009
11:04 AM
11
11
04
AM
PDT}

Bill, He didn't have to include latching. The program latches given a high enough population size and a low enough mutation rate. And a partitioned search is the same as Dawkins' cumulative selection. That is once something is found the search for it is over. That was Dawkins' point in the book.Joseph_{August 21, 2009
August
08
Aug
21
21
2009
11:00 AM
11
11
00
AM
PDT}

Prev 1 … 5 6 7 8 9 Next

You must be logged in to post a comment.

Leave a Reply