Do you deny that the organization and complexity of the surface of the Earth is driven by radiation from the Sun?

No. Neither do I deny that automobiles are driven by fuel. Shall we reason that the energy stored in petroleum caused the organization of metal and other materials into automobiles?

I maintain that unintelligent self-organization is an optimistic fantasy. But this hypothesis – “we know that a local increase in organization can be traded for a larger increase in entropy throughout the larger system” – why not test it six ways from Sunday and see what sort of organization results?

Take an example of the Sun/Earth where massive increases in entropy in the Sun drive modest amounts of compleity and organization here at the surface of the Earth.

Modest complexity being DNA, human intellect, millions of life forms, Shakespeare, etc. What would be an example of immodest complexity?

re the other 7 planets, whatever it is it is. Do you deny that the organization and complexity of the surface of the Earth is driven by radiation from the Sun?

Think about Miller-Urey type experiments. A small yield of amino acids for large energy inputs. Some fraction of the energy helped build the amino acids, most got radiated away as waste heat.

The sun shines on eight other planets in this solar system. Where is their increase in organization?

If the hypothesis is that increases in entropy drive complexity and organization, how would we test that? It sounds like something we should be able to repeat.

Without intelligence all things go unavoidably towards disorder, that is exactly in the opposite direction of organization.

I fear you have returned to an assertion that we know is not true. As a general point, we know that a local increase in organization can be traded for a larger increase in entropy throughout the larger system. Take an example of the Sun/Earth where massive increases in entropy in the Sun drive modest amounts of compleity and organization here at the surface of the Earth. Or take individual chemical reactions that create complex results but also heat, or water or some other high entropy product. The total entropy of the reaction products may have increased, but the increase is unevenly distributed.

may be I must clear a misunderstanding. I agree perfectly with you that in science probability matrices do exist and are used to model many phenomena. For example in mathematics there is the theory of Markov chains, which are stochastic processes based on probabilistic transition matrices (the passage from a state to the next one is not deterministic because depends on a probability). Therefore I don’t deny at all probability matrices (or arrays). Simply I deny that the genetic code is a probability vector. In the translation process the passage from an RNA codon to the codified amino acid doesn’t depend on a probability but on a fixed rule.

I think if you accept that physical processes can change aspects of the code, you are open to evolution of the code from a simpler state to a more complex state. This strikes me as a frontloading/TE kind of position, as you say design time vs run time. Abiogenesis is design time and evolution is run time.

If with “more complex” state you mean “more organized” then the passage to more organization always implies the intervention of intelligence (because organization implies CSI). To try returning to the issue of my post, when you write software you are organizing data, processes and events. There is no other way to organize things than to apply intelligence. Without intelligence all things go unavoidably towards disorder, that is exactly in the opposite direction of organization. You say “abiogenesis is design time and evolution is run time”. Here I could again agree with you, if you concede that evolution has only a passive role, insofar it develops only the possibilities that at design time were inserted in the systems. No doubt ultra complex systems (as the biological ones) have large potentialities of variation. In the informatics terms we could say that the biological software is highly configurable and parametrical. Moreover many configuration changes are triggerable by environmental events. All these aspects can be grouped under the name “evolution”. No one denies that organisms change. What ID denies is that changes implying increase in organization might arise thank to randomness and laws only.

Thank you very much for your kind words.

Here is an example of a code where you might use probabilities as entries. In cryptography, it is important to disguise the letter (and bigram and trigram)frequencies of a substitution cipher as much as possible, since these frequencies are the easist path to solving the cipher. Let us say that E is 10 times more frequent than Q. I construct a cipher where E is probabilistically mapped to 10 byte codes, but Q is mapped to only one. In this way I disquise the identities of E and Q.

I bring this up only to point out that a matrix with probabilities may have some uses in different computer applications. Related to the subject of evolution, there are also Estimation of Distribution Algorithms, in which an array or martrix is populated with probabilities which are updated as the algorithm is run. EDAs can be a very compact way of representing an entire population. For example, an EDA WEASEL (everyone’s favorite example) would start with an array 28 entries long and each entry would be a set of 27 probabilities, one for each possible character. In the first generation, each character in each space has the same probability, 1/27. When the algorithm has worked for a while, the probabilities have changed such that in slot 1 the probability of M is now higher than all the other letters, etc. (Yes, I am skipping how you change the probabilities based on fitness!)

I think if accept that physical processes can change aspects of the code, you are open to evolution of the code from a simpler state to a more complex state. This strikes me as a frontloading/TE kind of position, as you say design time vs run time. Abiogenesis is design time and evolution is run time. I’m not going to push you too hard on that now.

Thank you for your always challenging and interesting remarks. You are a person that the ID movement would be glad to have on board (and by the way I don’t consider impossible that you in the future will be entirely on our side).

I think you are wrong about whether a rule matrix can be populated with probabilities instead of certainties.

Here I am not sure to understand your point. I said that rule matrices (which specify codes) must be populated by constant values, not probabilities. The issue of my post was focused on the usefulness of computer programming to have an idea of what happens inside the biological systems. When a programmer must specify a code, usually he fills an array or hash with constants or eventually inserts them into a file or library module. Here probabilities don’t matter, only certainties. May be you mean the probability of the random arise of a code. You know that for me this probability is null.

We can also observe that in practice the implementation of the genetic code by the machinery of the cell is not perfect. Protein assembly errors do happen.

Abstract models are near perfect per se but when they are implemented in matter entropy intervenes and errors happen. No wonder about that, it’s everyone everyday experience.

On the subject of mapping only 15 amino acids, I’m asking you to imagine a form of life that only needs 15 amino acids. So a map that only contains 15 is fine. For example, suppose that there used to be only one ancestral map, which mapped 22 different amino acids, but our cells are a reduction of that ancestral map, while mitochondria are another reduction.

I don’t know what would be the technical consequences of a 15 or 22 amino acids life. I suspect the repercussions would be many and deep. However it seems clear to me that these considerations are frameworked into an ID perspective.

I’m not sure what mechanism you think exists that prevents a physical, material system which can be related to one abstraction from changing such that it is now related to another abstract system. The system doesn’t “know” it is related to an abstraction.

The system doesn’t “know” it is related to an abstraction but the designer of the system must know, otherwise he couldn’t design it!

My laptop has a CPU that is related to the abstraction of a Turing machine. It can be destroyed by a cosmic ray, nonetheless. [...] Being related to an abstraction, even a powerful abstraction, doesn’t offer any protection from the slings and arrows of outrageous fortune.

Of course. The relation to an abstract model is a must at the design level but at run time it doesn’t protect against entropy and all its harmful effects.

Therefore, I don’t see how you can maintain that all the genetic codes we see today in the world _must_ have been independently designed to be exactly what they are today. And if these maps _can_ change, there is the opening for evolution.

Any engineering variation in an ultra complex information processing system has to be designed, except the changes that the system is able to do by itself . . . according to possibilities that were frontloaded into the system just from the beginning by the designer!

Keep in mind also that complexity must be paired with specificity to indicate design.

Are you suggesting that Microsoft Windows could be simplified and still maintain all of its features, including compatibility with previous versions?

Yes. Same with the EU and the Tax Code.

And while I agree that the US Tax Code could be simpler, I think you’ll have trouble defending your argument that is came about by design.

Okay, I hope you’re joking. I’d admit that there’d be humor in it if you were.

Therefore, parsimony does not indicate design better than complexity.