Dawkins’ WEASEL: Proximity Search With or Without Locking?

_{William Dembski
March 16, 2009

Darwinism, Evolution, Informatics

3}_{Categories
Darwinism
Evolution
Informatics}

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

On pp. 47-48 of THE BLIND WATCHMAKER, Richard Dawkins gives two runs of his WEASEL program (note that there were typos in both initial seeds — one had 27 characters, the other 29 whereas they should have 28; I’ve corrected that). Here are the two runs using the Courier typeface, which assigns equal width to each character (spaces are represented by asterisks):

WDL*MNLT*DTJBKWIRZREZLMQCO*P
WDLTMNLT*DTJBSWIRZREZLMQCO*P
MDLDMNLS*ITJISWHRZREZ*MECS*P
MELDINLS*IT*ISWPRKE*Z*WECSEL
METHINGS*IT*ISWLIKE*B*WECSEL
METHINKS*IT*IS*LIKE*I*WEASEL
METHINKS*IT*IS*LIKE*A*WEASEL

Y*YVMQKZPFJXWVHGLAWFVCHQXYPY
Y*YVMQKSPFTXWSHLIKEFV*HQYSPY
YETHINKSPITXISHLIKEFA*WQYSEY
METHINKS*IT*ISSLIKE*A*WEFSEY
METHINKS*IT*ISBLIKE*A*WEASES
METHINKS*IT*ISJLIKE*A*WEASEO
METHINKS*IT*IS*LIKE*A*WEASEP
METHINKS*IT*IS*LIKE*A*WEASEL

These runs are incomplete. The first, according to Dawkins, required 43 iterations to converge, the second 64 (Dawkins omitted the other iterates to save space).

As you can see, by using the Courier font, one can read up from the target sequence METHINKS*IT*IS*LIKE*A*WEASEL, as it were column by column, over each letter of the target sequence. From this it’s clear that once the right letter in the target sequence is latched on to, it locks on and never changes. In other words, in these examples of Dawkins’ WEASEL program as given in his book THE BLIND WATCHMAKER, it never happens (as far as we can tell) that some intermediate sequences achieves the corresponding letter in the target sequence, then loses it, and in the end regains it.

Thus, since Dawkins does not make explicit in THE BLIND WATCHMAKER just how his algorithm works, it is natural to conclude that it is a proximity search with locking (i.e., it locks on characters in the target sequence and never lets go).

Interestingly, when Dawkins did his 1987 BBC Horizons takeoff on his book, he ran the program in front of the film camera:

www.youtube.com/watch?v=5sUQIpFajsg (go to 6:15)

There you see that his WEASEL program does a proximity search without locking (letters in the target sequence appear, disappear, and then reappear).

That leads one to wonder whether the WEASEL program, as Dawkins had programmed and described it in his book, is the same as in the BBC Horizons documentary.

In any case, our chief programmer at the Evolutionary Informatics Lab (www.evoinfo.org) is expanding our WEASEL WARE software to model both these possibilities. Stay tuned.

Comments

Arthur Smith (#81): "Zachriel is about the most polite commenter in the blogosphere that I have ever come across. Why is he banned here?" Can I join in asking that Zachriel may post here? (If he wants, obviously). I had a long exchange with him on another blog, and I agree that he is polite, but the most important thing is that he is very intelligent. And there is nothing better than an intelligent adversary. Regarding the discussion here about the weasel algorithm, I must admit that I cannot find it really interesting. Basically, I agree with Gil: an algorithm which already knows its target is of no interest for our debate here. I have always found Dawkins'Weasel a silly issue, however the algorithm may work in the details. And anyway, if we really have to discuss the algorithm, we should know it exactly (I can't see why Dawkins has not made it publicly available). Frankly, trying to recreate an unknown algorithm which anyway does not demonstrate anything seems not a very fruitful task (although I must commend Atom for having the patience and goodwill to try that just the same). Zachriel's word games are another thing. Unfortunately, I have never had the time to study them well, but I am sure that they demostrate something. They probably demonstrate that it is possible to easily explore a semantic space with an algorithm well designed to explore semantic spaces, but I am not really sure of that. But I will give only two comments here about that: 1) The protein functional space is not a semantic space: function, in proteins, is not linked to a symbolic meaning, but to biochemical activities. 2) Although Zachriel's algorithm can explore the space of words and even phrases, it still cannot really generate original language and original meanings. Language and meaning are still the product of conscious beings. Machines can only copy them, remix them, and recognize them only passively. In other words, a program can recognize that a phrase has a meaning only if it already "knows" (IOW, it has been inputted in it) that it has. The program, being a machine, has no concept of meaning. It can only match something with what it already "knows" in its code. Conscious intelligent beings are all another thing (although AI enthusiasts very much like to believe differently). So, after all, even Zachriel's programs are probably in Dawkin's category (they already know the answer they are searching), even if they certainly are much smarter.gpuccio_{March 18, 2009
March
03
Mar
18
18
2009
02:12 PM
2
02
12
PM
PDT}

Tribune7 (#82): My post in 83 deals with your question. There is no real target as such (at least not one in the distant future). For natural selection there are current selective pressures.Hoki_{March 18, 2009
March
03
Mar
18
18
2009
01:17 PM
1
01
17
PM
PDT}

StephenB: Of course. The problem is natural selection knows nothing of targets; therefore, the whole exercise is invalid. See my question @65—which choice would you make? Neither obviously. ... Oh, you want a more exhaustive answer??? OK, you're too teleological. How about this for an example that BETTER represents natural selection: Look at the results from the weasel program in the opposite order. I.e. start with the string "METHINKS IT IS LIKE WEASEL" (the modern "evolved" string we have found) and work your way towards a (to us) totally random string (it's distant ancestor), as such: 1. METHINKS IT IS LIKE A WEASEL 2. METHINKS IT WS LIKE A WEASEL 3. METHINKS IT WS WIKE A WEASEL ... x. HKADHFKLUWHEKJLHWEHUREKJEWHR Each string represents a sequence of letters that were more fit than the sequence below it under it's CURRENT selective pressure. The trick is to realize that "METHINKS IT IS LIKE A WEASEL" was only the target (i.e. more fit) when the sequence was on the second line. The target when on the third line was "METHINKS IT WS LIKE A WEASEL" and the target when on the fourth line was "METHINKS IT WS WIKE A WEASEL". In this way, the original random sequence never strived for a distant "METHINKS IT IS LIKE A WEASEL" Note: this is an ANALOGY, so it will be far from a perfect representation of natural selection (and evolution in general). We could always quibble about whether or not natural selection (or other evolutionary processes) actually could create a, e.g,. a human from something prokaryotish, but that is missing the point: you created a strawman version of how natural selection supposedly works.Hoki_{March 18, 2009
March
03
Mar
18
18
2009
01:14 PM
1
01
14
PM
PDT}

Hoki--The purpose of the program is simply to show that iterative randomness + selection can come to a target faster than randomness alone. So evolution requires a target? How would that target be determined?tribune7_{March 18, 2009
March
03
Mar
18
18
2009
12:53 PM
12
12
53
PM
PDT}

Patrick and Clive, Zachriel is about the most polite commenter in the blogosphere that I have ever come across. Why is he banned here?Arthur Smith_{March 18, 2009
March
03
Mar
18
18
2009
12:35 PM
12
12
35
PM
PDT}

Hoki,
The purpose of the program is simply to show that iterative randomness + selection can come to a target faster than randomness alone.
Of course. The problem is natural selection knows nothing of targets; therefore, the whole exercise is invalid. See my question @65—which choice would you make?SteveB_{March 18, 2009
March
03
Mar
18
18
2009
10:01 AM
10
10
01
AM
PDT}

Zachriel's program has been discussed to a limited extent before: https://uncommondescent.com/intelligent-design/behe-refuted-yet-again/comment-page-1/#comment-90063 https://uncommondescent.com/intelligent-design/dodgen-daily/ https://uncommondescent.com/intelligent-design/low-probability-is-only-half-of-specified-complexity/comment-page-1/#comment-145598 And since we're discussing English-based examples I'll copy over my previous thoughts on this subject: the Explanatory Filter can take multiple types of inputs (which also makes it susceptible to GIGO and thus falsification). Two are (a) the encoded digital object and (b) hypothetical indirect pathways that lead to said objects. My name “Patrick” is 56 informational bits as an object[each letter is represented by 8 bits]. My name can be generated via an indirect pathway in a GA. An indirect pathway in a word-generating GA is likely composed of steps ranging from 8 to 24 informational bits. Let’s say you take this same GA and have it tackle a word like “Pseudopseudohypoparathyroidism” which is 30 letters or 240 informational bits. It can be broken down into functional components like “pseudo” (48 informational bits) and “hypo” (32 informational bits). Start with “thyroid” (56 informational bits). For this example I’m not going to check if these are actual words, but add “ism”, then “para”, and then “hypo”. “hypoparathyroidism” is a functional intermediate in the pathway. The next step is “pseudohypoparathyroidism”, which adds 48 informational bits. Then one more duplication of “pseudo” for the target. That may be doable for this GA but what about “Pneumonoultramicroscopicsilicovolcanoconiosis” (360 informational bits) or, better yet since it’s more relevant Dembski’s work (UPB), the word “Lopadotemakhoselakhogaleokranioleipsanodrimhypotrimmatosilphiokarabomelitokatakekhymenokikhlepikossyphophattoperisteralektryonoptokephalliokigklopeleiolag?iosiraiobaph?traganopterýg?n” (1464 informational bits). I’m not going to even try and look for functional intermediates. And I’d add that none exist, although the entire word consists of functional components. So someone could argue that an indirect pathway could duplicate all of them from other words and somehow assemble them into a coherent whole.Patrick_{March 18, 2009
March
03
Mar
18
18
2009
08:42 AM
8
08
42
AM
PDT}

tribune7: IOW, it was the design of the program to show how things can come about without design. The purpose of the program is simply to show that iterative randomness + selection can come to a target faster than randomness alone.Hoki_{March 18, 2009
March
03
Mar
18
18
2009
08:21 AM
8
08
21
AM
PDT}

It is exactly like a movie in which a computer can supposedly crack a password by cumulatively selecting one character at a time — something that would be impossible to do without some kind of access to the final answer. There is no reward for guessing 99 percent of a safe's combination.tribune7_{March 18, 2009
March
03
Mar
18
18
2009
07:19 AM
7
07
19
AM
PDT}

Because that was the goal of the program. IOW, it was the design of the program to show how things can come about without design. IOW, evo--taken to the anti-ID level -- is irrefutably irrational.tribune7_{March 18, 2009
March
03
Mar
18
18
2009
07:12 AM
7
07
12
AM
PDT}

Here is a pair of house building illustrations of the War Games fallacy that evolutionary sentence building programs repeatedly fall into. Suppose one writes a program to simulate the construction of houses or other buildings, perhaps out of discovered/mutated building blocks, rather than the construction of English sentences. War Games Fallacy, Version #1 In this version, the program is designed to eventually produce a specific target house. The program cheats by comparing the work in progress against the complete target house. So if it discovers that a roof piece matches at such and such distance from the ground, it can preferentially preserve it there, floating in midair, even if the supporting structure is not yet in place. The Dawkins WEASEL program commits this version of the fallacy. War Games Fallacy, Version #2 In this version, there is no specific target house. Rather, the program has access to a generalized understanding of complete houses. (This is comparable to a program operating not according to a specific English sentence, but rather according to a dictionary of English words or phrases and an understanding of what is required to form a complete valid English sentence.) Nevertheless, the program is still cheating by taking advantage of this knowledge of the complete future picture. If it discovers it is appropriate in general that a roof piece should exist at such and such distance from the ground, it can still preferentially preserve it there, floating in midair, even if the supporting structure is not yet in place. If a program, such as Zachriel (cf. JT's comments above) or others like it, preferentially preserves words such as "of" because they are in the dictionary and will eventually become useful in meaningful sentences (even though they are currently useless for meaningful sentence construction), then that program is committing the second version of the same fallacy. Notice that the distinction between pursuing a specific predefined target or a generalized predefined target makes no difference to the error. The problem is that selection is allowed to take advantage of knowing future utility. It is exactly like a movie in which a computer can supposedly crack a password by cumulatively selecting one character at a time -- something that would be impossible to do without some kind of access to the final answer.ericB_{March 18, 2009
March
03
Mar
18
18
2009
06:36 AM
6
06
36
AM
PDT}

Because that was the goal of the program. Goal reached program stops.Joseph_{March 18, 2009
March
03
Mar
18
18
2009
06:25 AM
6
06
25
AM
PDT}

Something else to ponder: Why does Dawkins program stop with METHINKS IT IS LIKE A WEASELtribune7_{March 18, 2009
March
03
Mar
18
18
2009
06:17 AM
6
06
17
AM
PDT}

David Kellogg --But all languages, including English, do evolve, one step at a time. That is a very interesting point, and you are correct. But seriously, how could language evolve without intelligence i.e. understanding and agreement that particular sounds depict particular objects? And of course all alphabets are created.tribune7_{March 18, 2009
March
03
Mar
18
18
2009
06:15 AM
6
06
15
AM
PDT}

If it represents a “completed” evolution of an organism, wouldn’t each successive generation have to have to have some natural advantage to the previous one? You would think so, yes.tribune7_{March 18, 2009
March
03
Mar
18
18
2009
06:08 AM
6
06
08
AM
PDT}

Joseph, re 62: To get implicit latching without reversions, the population as well as mutation rate actually need to be mutually tuned, so that you get steady advances without letter substitutions [one flicks back while another moves ahead]. In short, for a given mutation rate, there is an "optimal range" of population size in which the number of mutations per member of a generation is sufficient that advance will happen, but not so large that it will converge over-rapidly relative to the sort of run lengths we see as probably good runs: 40+ and 60+. Also, the population has to be small enough and the rate for mutation small enough that we are unlikely to get the sort of double-mutations that would revert one while advancing another. (Observe, in 200 possible places for that, we never see it happen in the 1986 outputs.) Explicit latching is probably significantly easier to achieve than such tuning that achieves implicit latching or quasi-latching. But, if Mr Dawkins said that he did not latch explicitly, then I conclude that he most likely did some tweaking to get what he then thought was "good" results. And, that boils down to implicit letter latching or quasi-latching. GEM of TKI PS: One apparent feature of the 1987 output is the "winking" on the reverting correct letters in what I take to be the "generational champions" string. That is, they come back on target fairly quickly once they wander off. On average ADDITIONAL letters are going correct once every 2 - 3 or even 4 generations, to get runs of 40+ and 60+, which means that often -- 1/2 or more ot the time, it is the 0-mutations case that wins the contest. At 5% rate, there will be 24% of the population in that state. (That would be suggestive of such flicked back letters being specially treated.)kairosfocus_{March 18, 2009
March
03
Mar
18
18
2009
03:52 AM
3
03
52
AM
PDT}

Mr Kellogg: Re :your silly accusation that the program fixed letters in the first place Mr Kellogg, all that this cited statement reveals is that you are still unwilling to look at the evidence that is staring you in the face, not least at the head of this thread. (I will ignore the ad hominemistic tone.) 1 --> Please, first, LOOK. And, COUNT.
(You don't have to take my counts, do your own. Having counted, could you let us see your results: (i) how many letters appear after the first line in each run as published? (ii) of these how many are "correct"? and (iii) how many of the correct letters revert to incorrect status? And, (iv) are you able to comprehend the just stated point and questions -- if not, why not?)
2 --> I believe, you will see that in the 1986 program outputs, there are some 300 places in which letters might change. 3 --> As pointed out, of these 200+ go correct and once they are correct, NONE reverts in the sample space. A sample space of 300 instances in which we see 200 of a phenomenon, and none of the phenomenon that would point away, is not "no evidence" or a "silly accusation" or the like. It is in fact, well within the law of large numbers, and we have no reason to believe that sampling every tenth line will correlate with the program's behaviour such that thew results will be strongly atypical. [Do you understand what the adverted to "law odf large numbers" speaks about? Do you understand what "typical" in a sampling context is about? Do you understand that scientists will routinely publish their "best" results?) 4 --> Thus, as I noted from December on, we can see that the output plainly has "latched" letters, and indeed, it does so in a context that Mr Dawkins admits, rewards non-functional increments towards the target. [NB: As I have pointed out since Dec last, the fact of marching towards a target without reference to functionality is enough to show that Weasel is not an example of a BLIND watchmaker -- the thesis it allegedly supports -- but of in fact intelligent, foresighted, goal-oriented DESIGN.] 5 --> The only issue of consequence on is what mechanism gives rise to the observed latching of the output. On that question being raised as an issue in the previous thread, at 346 - 7 I put forth three alternative candidate reverse engineered algorithms:
T1 -- monkeys typing at random with hit/miss; T2 -- explicitly letter-latched search with warmer/colder; T3 -- implicitly letter-latched search with warmer/colder.
6 --> Of these, T1 is by Mr Dawkins own words exactly what he is trying to provide an alternative to. this, because as he states, it will not find the target on a PC within the available search resources, most notoriously. [But, observe: the increment in informational functionality to make the Weasel sentence is far smaller than that for origin of life or of body plan level biodiversity, which are plainly beyond the search capacity of the observed cosmos. This is of course being studiously evaded and ignored.] 7 --> As it turns out, BOTH T2 and T3 can supply the observed results, the difference being simply if one is explicit or implicit. Indeed, it is also the case that for the different 1987 behaviour, BOTH T2 and T3 can again fulfill the observed behaviour, as Apollos showed by writing a code that explicitly latches but occasionally reverts. 8 --> Thus, strictly, only inspection of credible code would be demonstrative. But, on the recent report above that Mr Dawkins has stated that he did not explicitly latch, on the charitability principle of accepting testimony unless there is reason to specifically reject it, I am willing to agree to T3 as the current best explanation. Never mind, that I suspect that such an algorithm is probably going to struggle to get the sort of "typical" outputs as are published in the 1986 case, if the 1987 program output is also typical of full runs for the 1986 cases. 9 --> To illustrate how implicit latching could appear, with 28 letters and a 5% per letter "mutation" rate [the parameters are unreaalistically favourable relative to the bio-world, of course . . . ], typically 0, 1 or 2 letters will change from one generation to the next: e.g. there is about a 24% chance that no letters will change for any given member in a given genration. So for a smallish to modest populations, the program will normally preserve currently correct letters or will advance a step or two. (With a bigger population and that sort of per letter mutation rate, with the closest to the target selected for, the program would probably find a case with two or even more further advancing letters in each generation, and would jump to the solution rather quickly, maybe even better than the deterministic, explicitly letter-latched search that is guaranteed to hit the target in 27 generations. [The 1987 version does not seem to do that sort of rapid convergence, just the opposite.]) 10 --> By shifting parameters a bit to slow down approach to the target, the program will be more likely to show advances but with occasional flickbacks, e.g. if double-letter changes pull one back and push one forward. But the problem here is that the 1987 flick backs then typically rather rapidly flick forward to correct state. (Hence the flickering, letter winking effect.) 11 --> But, strictly on the evidence of the 1986 output, the simplest -- most natural -- explanation would be explicit latching, as we can reasonably be assured that Mr Dawkins published samples are TYPICAL of "good" runs of the program at that time. In short, latching without reversion was the typical "good" output behaviour of Weasel at that time. (All the emphasis on how Weasel does not latch is based on subsequent developments. indeed, the 1987 video of outputs that fairly frequently flick back on correct letters, as already noted, was on the order of months to a year after the 1986 published data. There are no good grounds to equate the 1986 and the 1987 outputs.) 12 --> Algorithms implemented for subsequent versions of Weasel are irrelevant to the behaviour of the program as at 1986, as published. _____________ Onlookers, we continue to see the damaging impact of selective hyperskepticism. GEM of TKIkairosfocus_{March 18, 2009
March
03
Mar
18
18
2009
03:32 AM
3
03
32
AM
PDT}

JT, when you next have even a spare minute, SteveB has posed a fine question at 65. I would also be interested in your response. As a speaker of English, you are falling into the trap of unjustifiably assuming too much, based on your existing fluency in English as a whole. Without realizing the unfair edge this gives you, you are backtracking from the meaningfulness of the whole to a subset. For example, you wrote (with my emphasis added):
63: It hardly seems “arbitrary” to identify words and groups of words in a sentence as carrying the functional information of that sentence.
and
66: Consider the phrase “Tomorrow, I think I will wash the car.” The sentence fragment “Tomorrow, I think I will” is certainly generally functional, and millions upon millions of sentences can start that way.
and
66: So, we can imagine an eye being developed incrementally and some component that was being developed could certainly be associated with innumerable other potential organs - not just an eye.
and
66: As far as individual words, they do exist in isolation as well, in my mind (and I would say my brain - not some platonic dimension).
But of course, mindless Darwinian evolution cannot use a mind and it cannot select based on potential organs, just as it would be illegitimate to defend the preservation of a word or phrase in isolation based on the meaning it would have in a sentence (that doesn't yet exist). You cannot legitimately have selection take out a loan against future meaning or future functionality. Yet, without realizing it or thinking about it, that is exactly what you as an English speaker with a mind are doing when you consider fragments to have meaning/function based on potential completed sentences.
But you implied such a sentence would be a complete function.
I am allowing, uncritically and for the sake of argument, that a complete and correct English sentence is considered to have function, so that we can examine whether Darwinian sentence building programs show us anything worthwhile -- even if we grant that assumption. I am supposing, hypothetically, that a sentence might be considered to be something like a working bacterial flagellum. The question then is clearly this: How does one justify preserving all of the needed words (proteins) while we have not yet reached a working sentence (flagellum)? Notice that we cannot assume we have a working flagellum (sentence) and then backtrack to the idea that all the parts have function. That is illegitimate. Going in a forward direction, one must contend with finding a basis for selection when the isolated sentence parts have never yet made any functioning sentence at all. If such programs were made realistically comparable to real world requirements for undirected change, they would become just as feeble as Darwinian mechanisms are in practice (cf. Behe's The Edge of Evolution).ericB_{March 17, 2009
March
03
Mar
17
17
2009
05:30 PM
5
05
30
PM
PDT}

JT:
So, we can imagine an eye being developed incrementally and some component that was being developed could certainly be associated with innumerable other potential organs - not just an eye.
Only if it was designed to do so. That is an eye being developed incrementally.Joseph_{March 17, 2009
March
03
Mar
17
17
2009
10:08 AM
10
10
08
AM
PDT}

ericB [60]: ericB: I think I see what you're getting at now. Consider the phrase "Tomorrow, I think I will wash the car." The sentence fragment "Tomorrow, I think I will" is certainly generally functional, and millions upon millions of sentences can start that way. In fact, while writing "Tomorrow I think I will..." I could have a change of heart and instead of writing "Tomorrow I think I will wash the car" I could write, "Tomorrow I think I will go to the beach". So, we can imagine an eye being developed incrementally and some component that was being developed could certainly be associated with innumerable other potential organs - not just an eye. But your point would be that "Tomorrow I think I will" could not exist on its own - would not be a valid functional target on its own - there would have to be something with it, it could not be a target in isolation. But in what sense is the sentence "Methinks it is a weasel" functional in isolation? Could you write a post that only contained the phrase "Methinks it is a weasel?" Everyone would certainly consider that an unacceptable post and incomplete. But you implied such a sentence would be a complete function. And furthermore, even though the "Tomorrow, I think I will" ostensibly cannot exist in isolation, the fact is it does. Until I finish the sentence the phrase exists in isolation for a discrete interval of time. Its not as if the phrase "Tomorrow, I think I will wash the car." materializes instantaneously. And nor am I locked into one target merely by beginning a sentence "Tomorrow, I think I will" As far as individual words, they do exist in isolation as well, in my mind (and I would say my brain - not some platonic dimension). Anyway, I have more intuition as to what you were implying now, and have a sense how I want answer it, but just don't have any more time at the moment.JT_{March 17, 2009
March
03
Mar
17
17
2009
09:09 AM
9
09
09
AM
PDT}

JT et al: If we’re going to do simulations, let’s at least make them conform to the theory. Consider the following string of letters (which happens to be a scrambled English clause—underscores represent spaces to give you a head start if you like puzzles). This string represents some kind of biological proto-structure. wsoldin_je_d_swodx Now, consider the following “mutations” to the original, A and B. A wsocdin_je_d_swodx B wsoldin_je_d_srodx You, representing natural selection, do not know the target—in fact, you don’t even know what a target is, as you have no goals. Additionally, you cannot predict, cannot reason, and certainly cannot write computer code; you don’t have any lifelines and you can’t buy a vowel. All you can do is evaluate the “fitness” of each variation in the here and now. Your options are: select A, B, both, or neither. Which do you choose and why.SteveB_{March 17, 2009
March
03
Mar
17
17
2009
08:49 AM
8
08
49
AM
PDT}

Oops- the locking is a consequence of the programming. IOW it is an inevitable outcome- relatively speaking input to output- given the proper parameters. There isn't any need for a specific coding sequence to lock in any letters.Joseph_{March 17, 2009
March
03
Mar
17
17
2009
08:11 AM
8
08
11
AM
PDT}

ericB [60]: Obviously, if one claims that it is functional by mere definition, one can define anything at all to be functional. ... The use of arbitrary definitions does not salvage an inappropriate and non-functioning analogy to biology. As best I can tell, any analogy would be inappropriate to you. Its obvious you don't have any definition in mind yourself as to what would constitute functionality within the context of an English Sentence. A computer program is a series of words too. Those words are generally speaking actual functions. If the target were a C statement as opposed to a sentence, what would be the functions in that. It hardly seems "arbitrary" to identify words and groups of words in a sentence as carrying the functional information of that sentence. If you have no concept in mind yourself as to what constitutes functionality in a sentence, then any objection to the weasel program for preserving non-functional states is meaningless, as to you, any intermediate state is non functional. JT, explain to me the justification for calling the lone word “of” functional, to use one simple illustration. What is its function, apart from other words in a meaningful sentence? What about the word "apple"? Any problem with that term being defined to carry functional information? If you personally want to derive a exacting set of exceptions, for example ruling out single connector words like "of" as legitimate targets by themselves then fine. Why don't you just imagine the model with all necessary revisions that would render it sensible in your mind. Or is your potential list of objections actually endless.JT_{March 17, 2009
March
03
Mar
17
17
2009
07:33 AM
7
07
33
AM
PDT}

Put another way: Given 1- A Target 2- A small enough mutation rate 3- A large enough sample size The output will never be less than the input if the output is "closest to the target". IOW there doesn't have to be coded statement that locks the matching letters. The locking is a byproduct of the program.Joseph_{March 17, 2009
March
03
Mar
17
17
2009
07:25 AM
7
07
25
AM
PDT}

ericB:
Can you salvage Dawkins’ WEASEL words (or similar attempts) from the War Games fallacy?
I believe I addressed this in #32 above. AFAICR, the Weasel assumes the existence of a gradual path from source to target; it's not meant to prove the existence of one.anonym_{March 17, 2009
March
03
Mar
17
17
2009
07:07 AM
7
07
07
AM
PDT}

JT, explain to me the justification for calling the lone word "of" functional, to use one simple illustration. What is its function, apart from other words in a meaningful sentence? Obviously, if one claims that it is functional by mere definition, one can define anything at all to be functional. One might also "define" a lone protein as functional, despite the fact that its function will be future in the context of other proteins. To be selectable, it needs current function, not future function. The use of arbitrary definitions does not salvage an inappropriate and non-functioning analogy to biology.ericB_{March 17, 2009
March
03
Mar
17
17
2009
06:52 AM
6
06
52
AM
PDT}

JT:
Someone made statements previously to the effect that the reason that English is amenable to this type of incremental search (i.e. an efficient imcremental cumulative search that hits intermediary islands of functionality ) is because English is intelligently designed. Well then if you say biology is intelligently designed, then you are forced to to admit that it also is amenable to the same type of search.
Wrong. Not all that is intelligently designed has to amendable in the same way. However with ID biology does have a target and the resources to help reach that target.Joseph_{March 17, 2009
March
03
Mar
17
17
2009
06:47 AM
6
06
47
AM
PDT}

David Kellogg, If you want to get technical the program does lock in letters because of the probability thingy and the target. Those two factors, working together, will most certainly produce "survivors" that have at least the match that the parent had and you will never see any reverses- that is given a small enough mutation rate and a large enough sampling size. Take away the target and see what happens. Monkeys pounding away on a keyboard do not have a target- not that they know of anyway.Joseph_{March 17, 2009
March
03
Mar
17
17
2009
06:44 AM
6
06
44
AM
PDT}

EricB [48]: The same type of question would apply to “the Zachriel algorithm ... What is the biologically equivalent justification for preferentially preserving partial progress toward function even before function is reached (e.g. before we reach a functional sentence)? The Weasel algorithm has not definition of functionality other than one particular sentence. The Zachriel scheme is completely different. Here is what it defines as functional: 1)Any English word. 2)Any group of English words that can legally appear contiguously. (And the longer a legal word or legal group of words is, the more adaptive it is.) At no state of the search is anything preserved that does not meet the above criteria of functionality. ----------------------- A legal sentence is an example of a legal group of words. Any contiguous subset of words in a legal sentence is also a legal group of words. If you accept a sentence as being functional (which evidently you do judging from your above statements ), then you have no rational basis for denying that a smaller valid set of contiguous words is also functional (albeit to a lesser degree). If you accept a word as functional, you have no rational basis for denying a set of words that can legally appear contiguously together in English is also functional (and functional to a greater degree than a single word). As far as a "biologically equivalent justification" specifically - I personally would say that the parallel to English is valid to the extent that there are "grammatical" rules governing how biological components can be configured together, and to the extent there is a hierarchical and modular organization of biological components, being built up from reusable subcomponents and the like, and to the extent that there is semantic meaning arising from the arrangement of biological components which determine what larger biological components can be associated together. And to reiterate, it would be reality itself that determines if a biological organism is functional. It is reality itself that parses the chemical and physical syntax and semantics of some biological configuration, and determines whether that configuration confers some selective advantage. I don't really feel inclined to debate this a lot longer, though. Here are the websites: http://www.zachriel.com/mutagenation http://www.zachriel.com/phrasenation --------------------------- Someone made statements previously to the effect that the reason that English is amenable to this type of incremental search (i.e. an efficient imcremental cumulative search that hits intermediary islands of functionality ) is because English is intelligently designed. Well then if you say biology is intelligently designed, then you are forced to to admit that it also is amenable to the same type of search.JT_{March 17, 2009
March
03
Mar
17
17
2009
06:43 AM
6
06
43
AM
PDT}

Joseph, you certainly understand better than kairosfocus! :-) Kairosfocus, I find your writing opaque as always, but I think you may be notpologizing for your silly accusation that the program fixed letters in the first place -- which people have been telling you for weeks! If that is out of the way, it might be possible to correct your other misunderstandings, but the return on investment is very low, so I'm not going to make that effort.David Kellogg_{March 17, 2009
March
03
Mar
17
17
2009
06:30 AM
6
06
30
AM
PDT}

Prev 1 … 8 9 10 11 12 Next