When we consider information specified in a code of some kind, the specification can be objectively validated when the code is broken.
In some cases, but not in others. The One Time Pad, used correctly, is completely unbreakable. It is unbreakable because without prior knowledge of the key, all possible messages of that length may be present.
However, you example has a more general problem, that of a
prior specification. Dr Dembski, rightly, likened a
post-hoc specification to painting the target round the arrow after you have fired it. The specification has to come before the calculation, not after it. You cannot validate the specification after the event. Any validation must be done before the event.
The question remains whether the specified complexity in DNA could have arisen through chemical and physical regularity. The answer to that question is increasingly being found to be “highly unlikely” or even “no chance.” Douglas Axe has determined the probability of assembling a 150 amino acid chain into a minimally functional protein to be about the same as finding a single atom in the entire Milky Way.
Does Dr Axe’s calculation include the effects of natural selection? If it does not, then his calculation is useless. His result is also suspect, since it is relatively easy to find functional ribozymes in the lab from a soup of random RNA. See
Structural diversity of self-cleaving ribozymes for just one example. Agreed, these are usually shorter than 450 bp, but they are far closer to the probable origin of life than Dr Axe’s amino acid chain. Real life experiment confirms that Axe’s calculation is overly pessimistic.
You must realize this is duplicity on your part. The string containing what you claim to be zero specified information (Va gur ortvaavat Tbq perngrq gur urnira naq gur rnegu. …) contains the same specified information as the KJ quote but merely coded so that when the ROT13 decoder is applied you get the form of the KJ text.
You are changing the specification mid-calculation. Once you pick a specification, then you have to stick to it through t the end of the calculation. If I can change specifications arbitrarily in mid-stream then I can easily get any result I want out of the calculation.
My example shows that specified information can be created by a regular process. The ROT13 string does not contain the same
specified information, since it meets a different specification: “The ROT13’d text of the KJV”. If arbitrary changes of specification are allowed, then I can change the specification to: “The list of all even prime numbers greater than two.” which will instantly reduce all values of specified information to zero.
Once you have picked a specification, then that specification must remain the same throughout the entire calculation. Given that reasonable condition, then my ROT13 example shows clearly that a regular process, ROT13, can create arbitrarily large quantities of specified information.
It is also duplicitous in another way. You are intentionally attempting to conflate regular or chance processes with “natural” processes to implicitly claim that since your ROT13 as a regular process can create specified information, therefore natural processes can as well.
I took the trichotomy of regular, chance and design processes from Dembski’s explanatory filter. If you have a problem with that classification, then I suggest you take it up with him. There are also plenty of examples of genetic algorithms producing specified information, such as: “An efficient solution to the Travelling Salesman Problem.”
Meyer on Dembski:
Read Chapter 8 - Chance Elimination and Pattern Recognition in Meyer’s Signature In The Cell
The section on Genetic Algorithms in Chapter 13 is also revealing.
Also view for Dembski’s own words.:
Thanks for the references.
rossum