What exactly is ID?

quote:

---

While Dembski;s CSI MAY be quantifiable in simple cases, that isn't the way to do it. You've made two big assumptions (that the actual sequence is the only way to get the function and that the sequence is assembled entirely randomly). Neither is likely to be true in a real case. You need to calculate the probability of the specification being met in the absence of design - and you haven't done that.

---

1.) We are trying to get the probability of this sequence, not some other. There may very well be other ones that work well, but we are working with this one right now. What you actually have to show is that there is a possibility of a considerably large amount of sequences than can code for a flagellum.If you read Dembski's NFL, in his calcualtion he defined the E. Coli as consisting of flagellum as consisting of 4289 proteins, which is 4,639,221 base pairs. Out of which 50 proteins are used for the flagellum. And gave the possibility of 10 interchangeable parts. For which there is no evidnece than the flagellum can be modified by that much.2.) The sequence is calcualted that way because the NFL tehorem itself says that averaged over all sequence sapces, no algorith outperforms any other. And is thus no better than random chance if it does not take into account any prior problem-specific information. Since evolution is an algorithm, that means that if there is no prior input from an intelligence, it's as good as random chance.

quote:

---

Also, you've hit the big flaw in Dembski's CSI in that a specification based on the observed pattern is NOT the same as a specification produced without that knowledge. Dembski has recognised the problem by won't to the best of my knowledge has not yet found a way of dealing with it (and probably won't bother because the practical problems of using CSI in biology were already insurmountable rendering the whole thing utterly useless to ID).

---

You have to have some knowledge in order to descibe the observed pattern. It's called a descriptive language, aand can be any language. Mostly English because it's the easiest to use.

quote:

---

1.) We are trying to get the probability of this sequence, not some other. There may very well be other ones that work well, but we are working with this one right now. What you actually have to show is that there is a possibility of a considerably large amount of sequences than can code for a flagellum.

---

If you are just trying to get the probability of a specific sequence then you are NOT using Dembski's CSI. If you are trying to use Dembski's CSI then YOU have the burden of dealing with all the sequences that meet the specification. I don't have to show anything other than that you have failed to correctly follow Dembski's method.

quote:

---

If you read Dembski's NFL, in his calcualtion he defined the E. Coli as consisting of flagellum as consisting of 4289 proteins, which is 4,639,221 base pairs. Out of which 50 proteins are used for the flagellum. And gave the possibility of 10 interchangeable parts. For which there is no evidnece than the flagellum can be modified by that much.

---

I've done better than that - I've read The Design Inference. And to measure the variations possible even within the limits of the E Coli flagellum (which itself is taking too narrow a view) you need to consider what variations witin the protein sequences are possible without disrupting function.

quote:

---

2.) The sequence is calcualted that way because the NFL tehorem itself says that averaged over all sequence sapces, no algorith outperforms any other. And is thus no better than random chance if it does not take into account any prior problem-specific information. Since evolution is an algorithm, that means that if there is no prior input from an intelligence, it's as good as random chance.

---

That would be averaged over all fitness spaces. Unfortunately that doesn't tell you how well evolution will do at finding a working (NOT necessarily optimum) solution given the actual situation. The NFL theorems aren't much use to you.

quote:

---

You have to have some knowledge in order to descibe the observed pattern. It's called a descriptive language, aand can be any language. Mostly English because it's the easiest to use.

---

Using an observed pattern IS the problem. The probability of getting 500 heads in 500 tosses of a coin is 2^-500. The probability thatt 500 tosses of a coin can be completely specified is far higher. (500 tails is specified, alternating heads and tails is specified - in fact you can probably specify any sequence if you work hard enough).Thus specification derived from observation is not good enough.

quote:

---

If you are just trying to get the probability of a specific sequence then you are NOT using Dembski's CSI. If you are trying to use Dembski's CSI then YOU have the burden of dealing with all the sequences that meet the specification. I don't have to show anything other than that you have failed to correctly follow Dembski's method.

---

And that's why I said that out of 50 protein parts, 10 parts are assumed to be interchangeable.

quote:

---

I've done better than that - I've read The Design Inference. And to measure the variations possible even within the limits of the E Coli flagellum (which itself is taking too narrow a view) you need to consider what variations witin the protein sequences are possible without disrupting function.

---

How is that better, since basicly NFL came out after the Design Inference?Anyway, as I said above, Dembski assumed 10 interchangeable parts. If we are going to be more specific we can look at Doug Axe's work.

quote:

---

Using locally-randomized sequence libraries of a portion of the antibiotic resistance enzyme Beta lactamase, Axe calculated that somewhere between 1 in 10exp50 and 1 in 10exp77 150 amino acid-long protein folds form configurations with a Beta lactamase function (7). Of these one in 10exp50 to 1 in 10exp74 form folded structures that might perform any number of alternative functions (7).

---

http://www.arn.org/...nimal_complexity_relegates_life_originYou see, by modifying the already existing protein with mutations, untill it loses all function we can know which sequences would corespond to the original working specification. The number is somewhere between 10^50 and 10^70. So we take this number of 10^20 working combinations and this increases the chance of protein evolving. So if a certain molecular machine had a 1:10^200 chance in evolving. We can now calculate it at 10^180. These are all relevant sequences that fit the specified pattern.

quote:

---

That would be averaged over all fitness spaces. Unfortunately that doesn't tell you how well evolution will do at finding a working (NOT necessarily optimum) solution given the actual situation. The NFL theorems aren't much use to you.

---

It tells me that if evolution does not use any prior knowledge it's probably as good as random chance. Do you have any evidence that the laws of nature are set up in such a way that evolution does perform better than random chance in inputing novel information into the genome? Or do you accept the evidence from genetic entropy that clearly shows it's not working so well?

quote:

---

Using an observed pattern IS the problem. The probability of getting 500 heads in 500 tosses of a coin is 2^-500. The probability thatt 500 tosses of a coin can be completely specified is far higher. (500 tails is specified, alternating heads and tails is specified - in fact you can probably specify any sequence if you work hard enough).

​

Thus specification derived from observation is not good enough.

---

But it's not an independently given pattern, therefore, it's not a specification. It has to describe something else.For an example. Any hill side is complex and has a pattern. But only one hill side has 4 US presidents on it. And therefore it has a specifiaction. You see, that hill side, Mount Rushmore, has an independently given pattern to it. And that's why it's differnet from other naturally occuring hill sides. And that's why we know it's designed.Edited by Smooth Operator, : No reason given.

Thanks for your response! I have been wishing to communicate with proponents of ID. Yes and adding to the description, the images on Mount Rushmore have a "specified" relationship to the people it depicts. We recognise the images as a designed sculpture with a high amount of details that refers to something else that is real.It now sounds to me that Dr. Dembski has checkmated the materialistic evolutionists. It is certainly worthy of more investigation.

quote:

---

And that's why I said that out of 50 protein parts, 10 parts are assumed to be interchangeable.

---

Probably ALL of them could be changed for similar proteins. At least one could bne left out altogether without loss of function.

quote:

---

How is that better, since basicly NFL came out after the Design Inference?

---

Because The Design Inference is (supposedly) an academic level book aimed at describing Dembski's method. NFL is a popular level book where Dembski botches his own method.

quote:

---

Anyway, as I said above, Dembski assumed 10 interchangeable parts. If we are going to be more specific we can look at Doug Axe's work.

---

Which demonstrates that quite serious mutation is needed to remove function from a protein. Therefore indicating that if you were allowing for sequence variations all of the proteins could be replaced.

quote:

---

It tells me that if evolution does not use any prior knowledge it's probably as good as random chance.

---

Then you've been fooled by Dembski. The NFL theorems don't tell you any such thing.

quote:

---

Do you have any evidence that the laws of nature are set up in such a way that evolution does perform better than random chance in inputing novel information into the genome?

---

Yes. The change of fitness related to changes ing traits is not purely random. You will rarely find a small variation in a trait having a huge effect in fitness nor another small step in the same direction having a huge effect in the opposite direction. Thus the fitness landscape will not be the random mess which the NFL theorems focus on. Any reasonably well-behaved landscape will be more conducive to search algorithms than a random one. And that is just one problem with applying the NFL theorems.

quote:

---

Or do you accept the evidence from genetic entropy that clearly shows it's not working so well?

---

No, I don't accept that speculation (which is not evidence).

quote:

---

But it's not an independently given pattern, therefore, it's not a specification. It has to describe something else.

---

If you don't accept a specification derived from observation as valid at all you are completely rejecting Dembski's method.

quote:

---

For an example. Any hill side is complex and has a pattern. But only one hill side has 4 US presidents on it. And therefore it has a specifiaction. You see, that hill side, Mount Rushmore, has an independently given pattern to it. And that's why it's differnet from other naturally occuring hill sides. And that's why we know it's designed.

---

Unfortunately we know it is deisgned because of our background knowledge of human beings and human activities. That's why we don't need to do Dembski's probability calculations. The same, alas for ID, cannot be said for anything in biology.

Two of the flagellum's proteins are not interchangable. I'm just correcting you because you stated you don't accept speculation (which is not evidence). The question I have, does this loss of function mean loss of specific function? Also I have read that the active site for a specific function in relation to the size of the protein as a whole can be quite small. This suggests to me there can be quite a large amount of changes to a protein before that specific function is compromized. This also suggests to me certain proteins are predesigned to adapt to new functions.As pointed out in "Signature in the Cell", the precise shape and charge distribution enables DNA strands to coil efficiently around the nucleosome spools and store an immense amount of information in a very small space. Thanks in part to nucleosome spooling, the information storage density of DNA is many times that of our most advanced silicon chips.

quote:

---

Two of the flagellum's proteins are not interchangable. I'm just correcting you because you stated you don't accept speculation (which is not evidence).

---

Either a)you are asserting that there are two proteins in the E Coli flagellum which will tolerate no substitutions AT ALL or b)you didn't understand the point. The first is far more speculative than anything I wrote.

quote:

---

The question I have, does this loss of function mean loss of specific function?

---

It means losing the function that it had. Axe didn't look for other functions.

quote:

---

Also I have read that the active site for a specific function in relation to the size of the protein as a whole can be quite small. This suggests to me there can be quite a large amount of changes to a protein before that specific function is compromized.

---

That's correct.It also suggesst that counting the entire length of the protein as "specified information" will give you a serious overestimate.

quote:

---

This also suggests to me certain proteins are predesigned to adapt to new functions.

---

That is something of a leap. Perhaps you would like to explain your reasoning ?

Nope. Averaged over all possible landscapes. Sequence space is an infinitesimal subset of all possible landscapes.

quote:

---

Thanks for your response! I have been wishing to communicate with proponents of ID.

---

No problem, glad to be of help.

quote:

---

Yes and adding to the description, the images on Mount Rushmore have a "specified" relationship to the people it depicts. We recognise the images as a designed sculpture with a high amount of details that refers to something else that is real.

​

It now sounds to me that Dr. Dembski has checkmated the materialistic evolutionists. It is certainly worthy of more investigation.

---

Well not yet actually, but very soon. The next peer reviewed paper by Dembski is supposed to be published soon. It talks about vertical NFL theorems. And as Dembski said it, with the horizontal NFL theorem in place, this one is going to be the final nail in the Darwinian coffin. Watch out for my short explanation in the next few posts...

quote:

---

Probably ALL of them could be changed for similar proteins. At least one could bne left out altogether without loss of function.

---

Could be, but we don't know that for sure.

quote:

---

Because The Design Inference is (supposedly) an academic level book aimed at describing Dembski's method. NFL is a popular level book where Dembski botches his own method.

---

No, he actually improves on it and deals with the NFL theorem to boot. The paper he wrote in 2005 went even further. Basicly the CSI defined in The Design Inference is the oldest possible model you could use.

quote:

---

Which demonstrates that quite serious mutation is needed to remove function from a protein. Therefore indicating that if you were allowing for sequence variations all of the proteins could be replaced.

---

By how much different proteins? The chance of the flagellum forming by chance according to Dembski in NFL is 1:10^2954. If we add the possibility of modification and the flagellum still working fine, according do Axe's work I showed you, we would be justified to cut off 10^20 from this number. And would still be way over 10^120. Which is the limit of computation of the whole universe.

quote:

---

Then you've been fooled by Dembski. The NFL theorems don't tell you any such thing.

---

Oh, well than please tell me, why is any algorithm better than random chance averaged over all fitness functions?

quote:

---

Yes. The change of fitness related to changes ing traits is not purely random. You will rarely find a small variation in a trait having a huge effect in fitness nor another small step in the same direction having a huge effect in the opposite direction. Thus the fitness landscape will not be the random mess which the NFL theorems focus on. Any reasonably well-behaved landscape will be more conducive to search algorithms than a random one. And that is just one problem with applying the NFL theorems.

---

That is not what I asked for. I asked you to show me why do you think evolution is going to input CSI into living organisms. The way you are going to do that is to show me novel biological functions that evolved in nature or in a lab.

quote:

---

No, I don't accept that speculation (which is not evidence).

---

No, it's not a speculation, it's a FACT. And I already presented this link few posts ago. Did you miss it? It talks about the evidence of the accumulation of slightly deleterious mutations in 110 mammalian species. It would seem that evolution is not cut out to do the job you thought it was.Just a moment...

quote:

---

If you don't accept a specification derived from observation as valid at all you are completely rejecting Dembski's method.

---

That is not a specification, that's a fabrication. You obviously didn't read Teh Design Inference very well, or you read it a long time ago. Please note the difference between a specification and a fabrication.

quote:

---

Specified information is always patterned information, but patterned information is not always specified information. For specified information not just any pattern will do. We therefore distinguish between the "good" patterns and the "bad" patterns. The "good" patterns will henceforth be called specifications. Specifications are the independently given patterns that are not simply read off information. By contrast, the "bad" patterns will be called fabrications. Fabrications are the post hoc patterns that are simply read off already existing information.

---

http://www.arn.org/docs/dembski/wd_idtheory.htm

quote:

---

Unfortunately we know it is deisgned because of our background knowledge of human beings and human activities. That's why we don't need to do Dembski's probability calculations. The same, alas for ID, cannot be said for anything in biology.

---

But if we didn't know it was designed, we would still infer design. With or without Dembski's method. Simply because our intuition would tell us it's designed.A clear evidence for what I have just said is the Rosetta Stone. Nobody ever saw it get designed. But it is a clear case of design. And when it was found, design was infered.

I don't believe I'm understanding this statement the way you intend me to.The Rosetta Stone is a human artifact, made by humans, with human writing on it, in two languages that we were familiar with, from other human artifacts, but did not understand, and one language that we did understand, at the time that it was discovered. There is no doubt that it was designed, by humans.But the purpose that it is famous for, deciphering Demotic and hieroglyphic script and understanding that the latter was phonetic rather than pictographic, as had been previously believed, was not the use for which its makers intended it in any way. Its actual reason for existence is to glorify one of the Ptolemies in any language his literate subjects might understand.This is a good example of irreducible complexity, in that if any of the three languages was missing -- Greek for the meaning, Demotic for the phonetic relationship to Coptic, hieroglyphic for the meaning of the pictorial alphabet -- then it would not have been able to perform this unintended and serendipitous function. And this irreducible complexity appeared entirely by accident, through a slow reduction in other features, ie every other potentially decipherable polylingual Egyptian text that had ever existed.Unless you mean God had the libraries of Alexandria burned and then sent Napoleon to deface the Sphinx in order to be able to later accomplish this miracle?

quote:

---

Could be, but we don't know that for sure.

---

It's more certain than anything you've offered. Unless you want to argue that Axe's work is inapplicable in this case.

quote:

---

No, he actually improves on it and deals with the NFL theorem to boot. The paper he wrote in 2005 went even further. Basicly the CSI defined in The Design Inference is the oldest possible model you could use.

---

That's odd because he hasn't owned up to any real improvements. The NFL theorems aren't even relevant to the method.

quote:

---

By how much different proteins?

---

You mean you want me to provide data that Dembski should have used in his calculations ? Isn't it in his book ?

quote:

---

The chance of the flagellum forming by chance according to Dembski in NFL is 1:10^2954.

---

Which is based on his idea of an individual of an individual flagellum forming by chance. The idea that bacteria GROW flagella seems to have escaped him. It would have been muuch better had he calculated the probability that a bacterium would evolve a means to grow some sort of motility aid...

quote:

---

If we add the possibility of modification and the flagellum still working fine, according do Axe's work I showed you, we would be justified to cut off 10^20 from this number. And would still be way over 10^120. Which is the limit of computation of the whole universe.

---

Why only 10^20 ? That isn't allow for manhy other options at all. If there were a mere 10 possible sequences for each of the 50 proteins that would be a factor of 10^50. And that is going to be a huge underestimate - before even looking at other factors.

quote:

---

Oh, well than please tell me, why is any algorithm better than random chance averaged over all fitness functions?

---

Obviously you've been fooled so badly that you don't even see the distinction between Dembski's claims and the NFL theorems. Sorry, but your question is an irrelevant strawman.

quote:

---

That is not what I asked for. I asked you to show me why do you think evolution is going to input CSI into living organisms. The way you are going to do that is to show me novel biological functions that evolved in nature or in a lab.

---

Then your question is nonsense. Any information that dis not the produce of design is not CSI by Dembski's definition. I answered the real point, which is that there is reason to believe that evolution will do better than chance.

quote:

---

No, it's not a speculation, it's a FACT. And I already presented this link few posts ago. Did you miss it? It talks about the evidence of the accumulation of slightly deleterious mutations in 110 mammalian species. It would seem that evolution is not cut out to do the job you thought it was.

---

Your misinterpretations are not evidence either. Genetic entropy is only a problem under certain circumstances.

quote:

---

That is not a specification, that's a fabrication. You obviously didn't read Teh Design Inference very well, or you read it a long time ago. Please note the difference between a specification and a fabrication.

---

No, obviously YOU failed to understand it. Dembski's own explanation starts with observation of a pattern - whih is used to produce the specification.

quote:

---

But if we didn't know it was designed, we would still infer design. With or without Dembski's method. Simply because our intuition would tell us it's designed.

---

Background knowledge (e.g. knowledge of statues) helps rather a lot.

quote:

---

A clear evidence for what I have just said is the Rosetta Stone. Nobody ever saw it get designed. But it is a clear case of design. And when it was found, design was infered.

---

And again it's an example of background knowledge leading us to prefer a positive design hypothesis over non-design. How unlike Dembski's method which avoids positive design hypotheses like the plague.

quote:

---

It's more certain than anything you've offered. Unless you want to argue that Axe's work is inapplicable in this case.

---

What's certain is that you can't on average change the protein beyond 20%. Anything else is a complete loss of function.

quote:

---

That's odd because he hasn't owned up to any real improvements. The NFL theorems aren't even relevant to the method.

---

Of course they are relevant. They show that you can't get new CSI by the use of algorithms. Algorithms are only used to transmit CSI.

quote:

---

You mean you want me to provide data that Dembski should have used in his calculations ? Isn't it in his book ?

---

Nope, I told you, it's about 20%. You can even look at the critique on Panda's Thumb of Axes work. They don't agree that this work support's ID, but what do they know anyway. The point is, that the number of 20% is the limit of change.Axe (2004) and the evolution of enzyme function

quote:

---

Which is based on his idea of an individual of an individual flagellum forming by chance. The idea that bacteria GROW flagella seems to have escaped him. It would have been muuch better had he calculated the probability that a bacterium would evolve a means to grow some sort of motility aid...

---

You missed the point of the entire book. First of all, it doesn't matter that bacteria grow flagellums. The information they have to grow one still has to be acounted for. Teh reason Dembski got that number is that because without prior knowledge any algorithm is as good as any other averaged over all fitness functions, including random chance. Therefore, the number is correct. Evolution is not going to help you because it's an algorithm without prior knowledge. Unless you want to calim it actually does have prior knowledge. But in that case you have to explain where it got it.

quote:

---

Why only 10^20 ? That isn't allow for manhy other options at all. If there were a mere 10 possible sequences for each of the 50 proteins that would be a factor of 10^50. And that is going to be a huge underestimate - before even looking at other factors.

---

The best possible estimate is 20% change. That would amount to chnace of 1:10^14770. Therefore, that's still way over 10^120.

quote:

---

Then your question is nonsense. Any information that dis not the produce of design is not CSI by Dembski's definition.

---

That's not the point. We are talking about transmition of CSI here, for which algorithms are very well suited. The question is, is evolution well suited to transmit the CSI from nature into living organisms.

quote:

---

I answered the real point, which is that there is reason to believe that evolution will do better than chance.

---

Yes, you answered it, by simply asserting it. Where is the evidnece evolution can actually do it?

quote:

---

Your misinterpretations are not evidence either. Genetic entropy is only a problem under certain circumstances.

---

What misinterpretations are you talking about? And what "certain circumstances" are you talking about?

quote:

---

No, obviously YOU failed to understand it. Dembski's own explanation starts with observation of a pattern - whih is used to produce the specification.

---

But the point is that not every observed pattern is a specification.

quote:

---

Background knowledge (e.g. knowledge of statues) helps rather a lot.

---

Of course it hels, that is why it's here for.

quote:

---

And again it's an example of background knowledge leading us to prefer a positive design hypothesis over non-design. How unlike Dembski's method which avoids positive design hypotheses like the plague.

---

What are you talking about? What positive design hypothesis is Dembski avoiding?Edited by Smooth Operator, : No reason given.

quote:

---

The Rosetta Stone is a human artifact, made by humans, with human writing on it, in two languages that we were familiar with, from other human artifacts, but did not understand, and one language that we did understand, at the time that it was discovered. There is no doubt that it was designed, by humans.

---

If you were walking down the street one day, and you came across a piece of paper that had written on it "Mark wrote this". Tell me, how would you know htat the person who wrote that was not named John? How would you know, that it wasn't actually a woman? How would you know, that it was not a trained monkey that wrote that? Or when, why and how it was written?If you want to argue that you could know anything what I meantioned above, than you are completely wrong.The same goes for the Rosetta Stone. For all we know aliens could have designed it. How would you know the difference? You wouldn't obviously. We only assume that people did it. And it's a good assumption that is probably about 99.9% correct. But it's still an assumption. And we DO NOT, and I repeat, we DO NOT, know who, when, why and how designed that stone. Yet we still infer design, without knowing the identity of the designer.If you disagree, please tell me, how would you tell apart this Rosetta stone, and an identical one that was made by aliens.

I wouldn't know, nor would I claim to know any of that, nor would any scinetist, if all he had as information was the writing on the paper. It could have been pooped out by a hippo with three horns on his butt as well.... Yes, that's how science works. We work with the data we have. Since there is no evidence of any aliens ever having visitred this planet, we assume it was created by the only intelligence we know capable of producing it: man. Yes, this is true for all science. We infer that because it has written language on it, we know people used in the past. And we know people wrote on stones. No, nor does anyone claim they would. That's science for ya, tentative. We go with the best assumption we have untill more evidence is available that shows otherwise.So, if you have evidence for this intelligent designer, show it, and we can then discuss whether or not he designed life on this planet. Untill then we go with the explanation that is simplest: Life arose by natural means and evolved by natural means.

---

I hunt for the truthI am the one Orgasmatron, the outstretched grasping hand
My image is of agony, my servants rape the land
Obsequious and arrogant, clandestine and vain
Two thousand years of misery, of torture in my name
Hypocrisy made paramount, paranoia the law
My name is called religion, sadistic, sacred whore.
-Lyrics by Lemmy Kilmister of Motorhead