Tom Schnieder’s ‘the And-multiplication Error’ Article Refuted

[Fred Williams 0]

From the 'Does Dna Contain A Code?' topic:
 

 

What "usual probability multiplication error" are you referring to?
Fred

Its the error of assuming the probability of subsequent events happening is independant of the previous events. I assume this is what he is referring to. Its the old 'Boeing in Junkyard" argument. There are some good articles about it around and about. Try this one

Multiplication Error

...As a brief summary- Schneider sets up his ev program so that it cuts selection out of the picture. The result is no information increase (the actual result is 0.00e+00 +/- 4.66e-10). As soon as selection is introduced...information content increases. This would seem to indicate that selection is extraordinarily important.

 

The Claim

Some time ago Dr Tom Schneider wrote the paper 'Evolution of Biological Information', only to have it thoroughly refuted by Dr Royal Truman (link). In Schnieder's response, he offered very little in defense other than an article he wrote for the internet 'The AND-Multiplication Error' (Dr. Schnieder's excuse for his scant rebuttal was "I have other things to work on"! (see link).

Several evolutionists have latched on to this defense, essentially parroting Schneider without realizing the speciousness of the article.

The illusion

It turns out the only defense Schneider offered is built on illusion. It is a strawman argument, followed by invalid assumptions on population genetics that have no support from even evolutionist journals!

Schneider writes "“The multiplication rule does not apply to biological evolution".

For starters, almost all of the protein chain calculations used by creationists are used to refute abiogenesis. Evolutionists love to point out time and time again how "“evolution" is not abiogenesis! If we are to accept the evolutionists complaint and keep these two separate (IMO this an equivocation, but that's another debate), then the evolutionist is forced to admit that Schnieder has erected a strawman!

He then writes: "“We then find the card that has the most coins with heads up and we throw away all the other cards. So if even one card has an extra head, it will be found"¦ That is what happens in nature."

This is not true and easily disproved. This is called "“truncation selection", which doesn't happen in nature! Aside from common sense, we can establish this using evolutionist's own words from their journals [1] and from it's absence from college textbooks. Truncation selection would be a powerful mechanism, but it only happens with man-made intervention like artificial selection (that's the only time you'll see it mentioned in the college biology books), or in programs like Dr. Schneider's (thus one of many reasons his program is bogus). He tried to sneak it past us with the just-so statement "“that is what happens in nature", but there is not even the smallest shred of evidence this is true. It is an illusion masquerading as "“science".

To further enhance the illusion, he mentions the dandelion. The fact that a dandelion can maintain or increase it's population rapidly due to huge numbers of progeny, has absolutely nothing to do with truncation selection! This was a remarkably short-sighted analogy by Schnieder, and I suspect evolutionists would have a hard time finding even one population geneticist to agree that this somehow supports truncation selection!

Conclusion

In summary, while it is true that the Multiplication Rule requires the events to be independent, Schnieder uses the argument as a strawman (if you accept the evolutionist terminology that differentiates between evolution and abiogenesis). He ends with assumptions about population genetics that are not even remotely true, assumptions that have no support even from evolutionist literature.

Fred Williams

---

[1] Schnieder applies extreme truncation selection. A less severe version of truncation selection is essentially called "synergistic epistasis", and even evidence for it is scant: "“Although there is some theoretical support for synergistic epistasis (Szathmary 1993; Peck and Waxman 2000), there is little experimental support for this type of gene interaction (Willis 1993; Elena and Lenski 1997)." - Agrawal and Chasnov 2001. Also, "“Current evidence is equivocal as to whether the required levels of epistasis exist." - Siller 2001

[Modulous 0]

I think the program is meant to show information increase when a selection method is used. Truncation selection (and his example), I'm sure, is a grossly simplified selection method. The program isn't bogus though since it clearly shows exactly what it porports to:-

 

With random mutations and a selection method, information content can increase.

 

Remove the selection, and information content decreases.

 

Now - I agree that implying that natural selection and truncation selection are synonymous may be stretching things a little, and I am not in a position of sufficient knowledge at this time to fully discuss this. However, I think it is a mistake to think that Schneider has attempted to fully model natural evolution with his program. I believe (though I haven't studied it fully yet so I could be wrong), it also shows (perhaps as side effect) how IC structures can evolve.

 

Anyway, back to reading.

[Guest 92g]

I need a little help with this. He seems to be trying to demonstrate the potential of taking a "randon" genome with a uniform disutribution of bases, and try to figure out the chances of getting a sequence that binds to another genome.

 

Its not hard to imagine that if I have a goal of getting to the binding sites, and arrange my program to do it, that eventually it will happen.

 

How can this be an example of how the gentic code evolved? He's essentially starting with the letters, simply rearanging them until he finds the word he's looking for that he can attatch to a presumably existing sentence that makes sense to begin with.

 

He seems to have already assumed a genetic code, and he's just trying to imagine how the genome grew with more information.

 

The ev model quantitatively addresses the question of how life gains information, a valid issue recently raised by creationists

This assumes a living being with a functioning code, and decoding mechanism. I don't see how it could be any evidence for how the code came into exisitance.

 

the ev program also clearly demonstrates that biological information, measured in the strict Shannon sense, can rapidly appear in genetic control systems subjected to replication, mutation and selection

http://www.lecb.ncifcrf.gov/~toms/paper/ev/latex/node5.html

 

I think its more of an attempt to answer Spetner's charges than Gitt's.

 

How does selection pressure work in a simulation?

 

Terry

[Modulous 0]

Its not hard to imagine that if I have a goal of getting to the binding sites, and arrange my program to do it, that eventually it will happen.

 

I think the 'goal' is to make least 'mistakes'. So in the end the goal is to survive.

 

 

He seems to have already assumed a genetic code, and he's just trying to imagine how the genome grew with more information.

This assumes a living being with a functioning code, and decoding mechanism.  I don't see how it could be any evidence for how the code came into exisitance.

http://www.lecb.ncifcrf.gov/~toms/paper/ev/latex/node5.html

 

 

He's not trying to demonstrate how the code came about in the first place, but how random mutations combined with selection can result in an increase in information.

[chance 1]

Its not hard to imagine that if I have a goal of getting to the binding sites, and arrange my program to do it, that eventually it will happen.

 

How can this be an example of how the gentic code evolved? 

Terry

1343[/snapback]

 

I am just researching into this, so I don’t have much to add, but I will make one observation.

The ‘ev’ program is somewhat falsifiable, because it could have (but did not) demonstrate that you could not get from A to B, and there was no way of knowing that in advance of running the program. The arguments now focus on who is using the ‘correct’ assumptions in the ev program.

[Guest 92g]

I think the 'goal' is to make least 'mistakes'. So in the end the goal is to survive.

He's not trying to demonstrate how the code came about in the first place, but how random mutations combined with selection can result in an increase in information.

 

Maybe you need to have a discussion with Mr. Anagnostopoulos.....

 

Tom Schneider's Ev program creates a code from random DNA.

Terry

[Guest Paul C. Anagnostopoulos]

Ev starts with a population of creatures with random chromosomes. Part of the chromosome contains a "gene" with a weighting matrix and a threshold. The other portion of the chromosome has certain positions chosen to be the binding sites.

 

On each cycle, *every* position on the chromosome is evaluated with the gene to see if it binds there. Mistakes are counted when the selected positions are not bound, and when other positions are bound by mistake. The half of the creatures with the most mistakes are killed off and replaced by clones of the better half. Mutations are performed each cycle.

 

The goal of Ev is to show that the Shannon information content of the chromosome can increase from zero to a (turns out) predictable number of bits over time. It does so. Ev has nothing to do with abiogenesis or the evolution of the modern protein synthesis mechanism.

 

Ev does demonstrate the evolution of a code, however.

 

~~ Paul

[Guest 92g]

Ev does demonstrate the evolution of a code, however.

It attempts to show the change in the information contained in the genetic code. The "code" is not evolving. It still uses the same symbols and syntax.

 

Terry

[Guest Paul C. Anagnostopoulos]

It attempts to show the change in the information contained in the genetic code. The "code" is not evolving. It still uses the same symbols and syntax.

The chromosome is a sequence of 2-bit numbers that just happen to be displayed as A, C, G, and T. These "bases" are taken in groups to form weights, thresholds, and binding sites.

 

The weights form a code (not the "genetic code") as the creatures evolve. In particular, the code could be used to correct binding site sequence mutations.

 

Hey look, we're famous:

 

http://www.lecb.ncifcrf.gov/~toms/paper/ev...rytale2005Apr21

 

~~ Paul

[Guest 92g]

One can only believe that if he has time to pay attention to this wonderful web site, then he has time to answer Royal Truman's critique.

 

Terry

[Modulous 0]

One can only believe that if he has time to pay attention to this wonderful web site, then he has time to answer Royal Truman's critique.

 

Terry

1355[/snapback]

Somebody might have sent him an email. I doubt he spends all day monitoring the entire internet for mentions of his program.

[Guest Paul C. Anagnostopoulos]

I sent him an email.

 

~~ Paul

[Fred Williams 0]

I am just researching into this, so I don’t have much to add, but I will make one observation.

The ‘ev’ program is somewhat falsifiable, because it could have (but did not) demonstrate that you could not get from A to B, and there was no way of knowing that in advance of running the program.  The arguments now focus on who is using the ‘correct’ assumptions in the ev program.

1345[/snapback]

It’s been some time since I looked at Tom's program, but I recall that it is not falsifiable at all because it will essentially work every time. I was not surprised at all his program did what he predicted it would, given the input parameters. I would have been surprised it did *not* work! It's rigged for success, and does not emulate naturalistic accumulation. Too many unrealistic assumptions (i.e. truncation selection, high favorable mutation rate, etc).

 

Fred

[Fred Williams 0]

Ev does demonstrate the evolution of a code, however.

 

~~ Paul

1347[/snapback]

No, it doesn't. I would be surprised if he stated as much in his email to you. As I believe Terry mentioned, a code is properly defined as symbols, syntax, and semantics. IMO this is sufficient to define a code, though Gitt attaches "prgamatics" (action) and "apobetics" (purpose or intent). I think the last two are obvious once you get to the semantics level.

 

So how in the world did his program evolve a new "code"?

 

Hmm, I see in the link you provided that Tom makes it clear his program does not evolve a genetic code. I guess this particular claim of yours is now defunct since the very author of it doesn’t think it evolves a code.

 

Fred

[Guest Paul C. Anagnostopoulos]

I did not say it evolved a/the genetic code. No one is claiming that code is synonymous with genetic code, are they?

 

Here is Terry's list of requirements. Are we still agreed on these?

Necessary Conditions:

1) A uniquely defined set of symbols is used

2) The sequence of the individual symbols must be irregular

3) The symbols appear in regular structures

4) At least some symbols must occure repeatedly.

So let's see how this pertains to Ev:

1) The set of symbols are the four values ("bases") 0--3.

 

2) The sequence of values is irregular.

 

3) The values appear in the regular structures making up the weighting matrix, the threshold value, and the binding site sequences.

 

4) The values occur repeatedly.

 

 

~~ Paul

[Fred Williams 0]

I did not say it evolved a/the genetic code. No one is claiming that code is synonymous with genetic code, are they?

 

Here is Terry's list of requirements. Are we still agreed on these?

 

So let's see how this pertains to Ev:

1) The set of symbols are the four values ("bases") 0--3.

 

2) The sequence of values is irregular.

 

3) The values appear in the regular structures making up the weighting matrix, the threshold value, and the binding site sequences.

 

4) The values occur repeatedly.

~~ Paul

1430[/snapback]

What is the syntax? What is the meaning attached to the syntax? (example: codon = syntax, codon encodes amino acid = meaning).

 

Here is another example:

 

Symbols: English alphabet

Syntax: words comprised of symbols, ie "evolution"

Meaning attached to syntax (words): ie "evolution" = fairytale.

 

Why don't you ask Tom if he thinks his program evolved a new "code"?

 

Fred

[Guest Paul C. Anagnostopoulos]

A codon is not syntax. It's just a "word" composed of three "letters." I suppose a sequence of codons followed by a stop codon might be considered a "sentence," but the sum total of the syntax is the period at the end of the sentence. There are no other characteristics that we associate with syntax.

 

As far as meaning is concerned, if any sequence of letters or numbers or molecules performs a function, then it has meaning. That is, if what we mean by meaning is "An interpretation of the message over and above the sequence itself."

 

~~ Paul

[Guest Paul C. Anagnostopoulos]

Schneider thinks a code evolves, albeit a very simple one. The set of all possible site-length sequences are coded into two subsets: those that are recognized by the weight matrix and those that are not.

 

~~ Paul

[Fred Williams 0]

A codon is not syntax. It's just a "word" composed of three "letters." I suppose a sequence of codons followed by a stop codon might be considered a "sentence," but the sum total of the syntax is the period at the end of the sentence. There are no other characteristics that we associate with syntax.

1447[/snapback]

Yes, there is. In programming, while syntax encompasses the set of instructions and their required order (similar to the traditional definition you gave above), it can also refer to individual opcodes.

 

Fred

[Fred Williams 0]

Schneider thinks a code evolves, albeit a very simple one. The set of all possible site-length sequences are coded into two subsets: those that are recognized by the weight matrix and those that are not.

 

~~ Paul

1449[/snapback]

I'd love to see Tom put that in writing somewhere on his site. So don't get me wrong, I'd have to see him say this in writing before I believe he would make such a blatantly erroneous statement.

 

BTW, let's clarify something. I asked you if Tom said his program can evolve a new code. You responded that Tom "thinks a code evolves". Of course he would say this, every atheist is required to say this or else he is admitting he is not an atheist. But that is not what I asked. It sounds like you are dodging my question. Does Tom think his program evolved a new code, yes or no.

 

Fred

[Guest Calipithecus]

In programming, while syntax encompasses the set of instructions and their required order (similar to the traditional definition you gave above), it can also refer to individual opcodes.

 

Fred

1535[/snapback]

I don't see how you figure that. Can you provide an example?

[Modulous 0]

I think there is a little bit of confusing terms here. But its not a big issue. The code isn't evolving anymore than the DNA code is evolving. DNA uses the same 'symbols' to represent the same protiens universally. The point of the program is to show information increase due to random mutations and selection. Which it does.

[Fred Williams 0]

The point of the program is to show information increase due to random mutations and selection. Which it does.

1566[/snapback]

No, it doesn't, it's an illusion. The illusion is built around two things: 1) low-level statistical informaton (Shannon) masquarading as real, tangible information (ie a code) and, IMO more importanty, 2) the word "selection". The selection in Tom's program is not real-world, not even close. New information was not created naturalistically. Artificial Shannon information, perhaps, random with natural selection, no.

 

So in short Tom's program did not even create new Shannon information naturalistically. At best it did so artificially. It's a bogus attempt to deal with the information problem.

 

Fred

[Modulous 0]

No, it doesn't, it's an illusion. The illusion is built around two things: 1) low-level statistical informaton (Shannon) masquarading as real, tangible information (ie a code)

 

Its not masquerading. If you look at the paper its quite clear that its Shannon information it discusses.

 

and, IMO more importanty, 2) the word "selection". The selection in Tom's program is not real-world, not even close. New information was not created naturalistically. Artificial Shannon information, perhaps, random with natural selection, no.

 

Its not Artificial Shannon information, it IS Shannon Information. The selection method used wasn't 'natural selection', but that doesn't mean it isn't Shannon Information. I think Tom dealt with this in his single paragraph about this thread. Don't like the selection method? Change it and see what happens. Shannon Information doesn't rely on a selection method, its just a measure of information. I don't even know what a definition of Artificial Shannon information would be.

 

Its not meant to be 'real world'. Its meant to show that random mutation coupled with selection increases the Shannon Information content. It does this.

 

It's a bogus attempt to deal with the information problem.

 

The information problem being what? That information content can't increase through mutation and selection? Well, I think the program certainly shows that Shannon Information does increase with mutation and selection, and that is all it sets out to do. If you want to develop new definitions of Information and explain how that cannot increase with mutation and selection and how your model is closer to the DNA issue than Shannon Information then go right ahead. Write a program, submit it for peer review and win a Nobel prize (or at least a phd :-D)

[Guest Paul C. Anagnostopoulos]

More comments from Tom Schneider:

 

http://www.lecb.ncifcrf.gov/~toms/paper/ev/blog-ev.html

 

~~ Paul