Tuesday, April 02, 2013

Once More into the False Dichotomy Zone: "Naturalism vs. Design".

(Picture from http://www.faradayschools.com/re-topics/re-year-10-11/god-of-the-gaps/ )

Established evolutionary theory may be a good theory, but I wouldn't say I'm 100% convinced and I’ll continue to read with interest the views of the de-facto Intelligent Design community.  But in spite that I certainly don’t share what appears to the de-facto IDist’s well-motivated anti-evolutionary complex. On occasion this underlying complex manifests itself in a strenuous drive to find in principle refutations that short cut the work of debunking evolution. An example of this is Granville Sewell who is thoroughly beguiled by a belief that the 2nd law of thermodynamics provides in principle refutation of evolution.

Before I go any further with a critical look at a particular post on the IDist web site Uncommon Descent let me make it clear that I at least agree with the de-facto IDists on this: One can’t do natural history without assuming as a starting point a world with an initially high burden of information. That is, our world could not be as it is without being resourced by some very rare conditions, whether of actual up and running configurations or the appropriate algorithmic generators of those configurations. If from either of these starting points we use the principle of equal a-priori probabilities to convert rarity of case into high improbability, we infer that our cosmos contains an irreducible burden of information; in a nutshell this is William Dembski’s result. If one is so inclined this inevitable logical hiatus readily takes us into theology, but I won’t touch that subject here. Suffice to say that I agree with Dembski’s core thesis that the cosmos’s burden of information is irreducible. In fact it even applies to multiverse scenarios.

The recent Uncommon Descent post I am considering here, however, deals with less fundamental questions and serves to highlight where I would be depart from UD thinking.  Below I quote this post in its entirety and interleave it with my own commentary.

March 23, 2013           Posted by kairosfocus under Intelligent Design:-        
EA nails it in a response to an insightful remark by KN (and one by Box): “the ability of a medium to store information is inversely proportional to the self-ordering tendency of the medium”
Here at UD, comment exchanges can be very enlightening. In this case, in the recent Quote of the Day thread, two of the best commenters at UD — and yes, I count KN as one of the best, never mind that we often differ — have gone at it (and, Box, your own thoughts — e.g. here — were quite good too  ).

My Comment: The reason why Kairosfocus favours the two commenters he is about to quote is because he has very much backed the law and disorder vs. design dichotomy and he is quite sure that law and disorder have not generated life. He’s of the school of thought that once you have eliminated law and disorder from the enquiry that leaves intelligent design as the explanation for living structures. The trouble with this view is that it can give the false impression that law and disorder vs. design are mutually excluding categories. I have expressed doubts about this dichotomy on this blog several times. However, to be fair to Kairosfocus I’m sure he understands that if our cosmic law and disorder regime has generated living configurations then there still remains the irrefutable work of Dembski, work that, as Dembski admits, doesn’t in and of itself contradict evolution. 

But if Kairosfocus is right and the cosmic law and disorder regime is inadequate to generate life then this means that “design theory” becomes a very accessible and compelling argument; it is easy to picture some kind of homunculus molecular designer piecing together the configurations of life much like a human engineer. The OOL/evolutionary alternative requires one to grasp some rather difficult to understand notions employing information theory, fundamental physics and algorithmics.

Anyway, continuing with the UD post:

Let’s lead with Box:
Box, 49: [KN,] your deep and important question *how do parts become integrated wholes?* need to be answered. And when the parts are excluded from the answer, we are forced to except the reality of a ‘form’ that is not a part and that does account for the integration of the parts. And indeed, if DNA, proteins or any other part of the cell are excluded from the answer, than this phenomenon is non-material.

My Comment: This, I think, is an allusion to one of the de-facto ID community’s better ideas; namely irreducible complexity. In non-mathematical terms irreducible complexity can be expressed as follows: Organic components can only exist if they are part of an organic whole that maintains their existence. But conversely the survival of the organic whole is dependent on the individual components surviving. In other words we have mutual dependence between the parts and the whole. So, since organic wholes depend on parts and parts depend on organic wholes it appears that this mutual dependence prevents an evolutionary piecemeal assembly of an organism from its parts. The conclusion is that each organic form came into existence as a fait accompli.  However, this logic has a loop hole that evolutionists can exploit. The kind of incremental changes that can be conceived are not stuck at the discrete level of mutually dependent parts. It hardly needs to be said that organic components are composed of much more elementary components than organic parts, namely fundamental particles. Therefore the question naturally arises as to whether the organic parts themselves can be incrementally morphed at the particulate level and yet still leave us with a viable stable organic whole. This, of course, takes us into the fundamental question of whether configurations space with respect to these incremental changes is reducibly complex, a concept defined in the post here. But as I mention in that latter post there is an issue with reducible complexity: Given that the number of viable organisms is likely to be an all but negligible fraction of all the total possible configurations of atomic parts, it is certainly not obvious to me that a practical reducible complexity is a feature of our physical regime. But conversely I can’t prove that it isn’t a feature!

The point I am making here is that because the UD comments above remain at the discrete “part” level rather than the more fundamental particulate level they don’t scratch the surface of the deep theoretical vistas opened up by the reducible complexity question. But there is, I’ll concede, a prima facie case for the de-facto ID community’s skepticism of evolution, a case that particularly revolves round the idea of irreducible complexity; although this skepticism appears to be motivated by a narrowness of perspective, namely, the perspective that “Design” and “Naturalism” so called (i.e. OOL and evolution) are at odds with one another.

Now, it may well be that evolutionary theory as the scientific establishment conceives it is wrong, perhaps because irreducibility complexity blocks the incremental changes evolutionary theory demands. But one feels that if evidence came to light that unequivocally contradicted the defacto-ID community’s anti-evolutionism (if such is possible) it would mean a very drastic revision of their “design vs. nature” paradigm. The kind of argument above regarding the apparently all-or-nothing existence of organic structures, although in some ways compelling, is certainly not absolutely obliging. The UD argument I have quoted regarding the holistic nature of organisms does not classify as a killer “in principle” argument against evolution. The de-facto ID community is very enamored of the metaphor of the intelligent homunculus who works like a human engineer in contradistinction to the so-called “naturalistic” evolutionary mechanisms. But there is a great irony here: If physical regimes implying reducible complexity have a mathematical existence then the computational resources needed to find and implement such a regime could be put down to an intelligent agent. Ironically then, using the very principles the de-facto ID community espouse, a workable evolution can hardly be classified as “natural” but rather very “unnatural” and moreover evidence of a designer! If the de-facto IDists are prepared to espouse an all but divine designer, such a designer could be the very means of solving the problems of selecting a physical regime where OOL and evolution work!

KN, 52:  the right question to ask, in my estimation, is, “are there self-organizing processes in nature?” For if there aren’t, or if there are, but they can’t account for life, then design theory looks like the only game in town. But, if there are self-organizing processes that could (probably) account for life, then there’s a genuine tertium quid between the Epicurean conjunct of chance and necessity and the Platonic insistence on design-from-above.

My Comment: Self-organization, so-called, is not of necessity a tertium quid; it could yet be the outcome of a carefully selected Law and Disorder dynamic. In fact if evolution and the necessary OOL processes that must go with it are sufficient to generate at least an elementary form of life this would classify as “self-organization”. Richard Johns, who is an IDist, would agree on this point. In a published paper Johns probes the subject of self-organization using a cellular automata model. Cellular automata are based on a law and disorder paradigm and make use of no tertium quid. Of course, as a de-facto IDist Johns is somewhat committed to the notion that this form of self-organization cannot generate life, but his paper does not succeed in proving the case either way. In fact in order to support his prior commitment to the inadequacy of self-organization he hamstrings law and disorder as a means of self-organization with a habitual mode of thinking that has become fixed in people’s mind ever since Richard Dawkins coined the phrase “The Blind Watch Maker”. In Johns’ case he applies the general idea behind the Blind Watch Maker by taking it for granted that the law and disorder algorithms controlling his cellular system are selected blindly. Since it is a likely conjecture that life generating law and disorder systems are extremely rare cases amongst the class of all possible algorithmic systems (if indeed they have mathematical existence at all) then clearly blind selection of the cellular algorithms is unlikely to give us a system that generates living configurations! But if Johns believes in an omni-intelligent agent of open ended powers then that agent could just as well express itself through the selection of just the right life generating regime (assuming it has a mathematical existence) as contrive living configurations directly.  Given the ID culture Johns has identified with, he is likely to think of self-organization as a “naturalistic” method of generating life and so he hamstrings this notion by simply not allowing it to be set up via intelligent agency. Of course, if you disallow intelligence to express itself in this way and insist on the selection of physical regimes on a blind random basis then you are not likely end up with a life generator!

Notice that in the quote from KN he too is inclined to see self-organization and design theory as two competing scenarios whereby elimination of one leaves the other as the “only game in town”.   In fact self-organization is mysterious enough to KN that it classifies as neither law-disorder nor design, but a tertium quid. The naturalism vs. intelligence dichotomy is so fixed in his mind that it has never occurred to him that self-organization of the law and disorder variety leaves us with similar issues of logical haitus and computational complexity as does the idea that living configurations are simply a fait accompli.  He just doesn’t make a connection between the large measure of computational complexity implicit in the selection of the right physical algorithms and a design decision! I see this as yet another manifestation of the false dichotomy of God did it vs. Naturalism did it.

Self-organization is, in fact, a very bad term. The elementary parts of the cosmos could never self-organize but only do so because an imposed and carefully selected physical regime controls them. The term “self” is yet another subliminal signal of the “naturalistic” view that somehow the elementary parts of the cosmos possess some power of organization in and of themselves. But think about it: That’s not unlike claiming that the bits in say a Mandelbrot set have the innate power to organize themselves into intricate patterns!

EA, 61: . . .  the evidence clearly shows that there are not self-organizing processes in nature that can account for life.
This is particularly evident when we look at an information-rich medium like DNA. As to self-organization of something like DNA, it is critical to keep in mind that the ability of a medium to store information is inversely proportional to the self-ordering tendency of the medium. By definition, therefore, you simply cannot have a self-ordering molecule like DNA that also stores large amounts of information.
The only game left, as you say, is design.
Unless, of course, we want to appeal to blind chance . . .

My Comment:  EA is probably right about there being no evidence for self-organization; but only as an extra tertium quid factor. There is of course evidence for evolution as a form of self-organization arising from a cellular automata system, but just how obliging this evidence is and just how successfully the theory joins the dots of the data samples is what the debate is about!

 EA’s point about the conflict between information storage and self-organization is I think this: Self organization, at least as it is conceived by Richard Johns and myself, is a highly constrained process; though it may generate complex forms it nevertheless has low redundancy in as much as it is not possible to arbitrarily twiddle the bits of a self-organized configuration without the likelihood of violating the algorithmic rules of this process. In contrast arbitrary information storage allows, by definition, arbitrary bit twiddling and therefore one can’t use a self-organized system to store any old information. Self-organization only stores the information relevant to the form it expresses. For example I couldn’t arbitrarily twiddle the bits of a Mandelbrot set without violating the rules of the algorithm that generated it.

However, I believe EA has misapplied this lesson with some hand waving. If OOL and evolution have generated life using the algorithms of a cellular system it would classify as self-organization (albeit with “self” being a complete misnomer). OOL and evolution would work by virtue of the selection of what is likely to be a very rare algorithmic case and this rarity would imply a corresponding high level of information. Self-organized systems are algorithmic ways of storing the information found in the complex patterns they generate. Ergo, EA’s point about self-ordering systems and their lack of ability to store information is misleading; true they can’t store information about systems other than the forms they define, but they nevertheless do store information of a special kind.  What I think EA really means is that self-ordering systems can’t store arbitrary information.

The type of “think” that EA displays here is reminiscent of  an argument I once saw on UD (although I’ve lost the exact chapter and verse). It went along these lines: Self-organization requires “necessity”. Necessity implies a probability of 1 which in turn implies an information of zero. Therefore self-organization can’t store information. This argument is false and appears to  be based on the misleading connotations of the word “necessity”. What these IDists refer to as “necessity” is simply algorithmic constraint. Since the set of all algorithmic constraints is very large then the selection of a particular suite of constraining algorithms is highly contingent and is hardly a “necessity”. Conversely, a book of random numbers to the observer who first comes to it is very "contingent" and thus stores lots of "information". However, once the observer has used the book and committed it to memory, it's information value is lost. "Information" is observer dependent. In fact depending on the state of the observer's knowledge so-called "necessity" can be packed with information whereas so-called "contingency"  may have a zero information content.

EA, in thinking that he has chased self-organization out of the town, invokes the habit of mind which automatically separates out self-organization and design as two very distinct processes. He consequently concludes that design is the only game left in town. EA expresses no cognizance of the fact that, using William Dembski’s principles, he has also chased away what itself could classify as a form of design: For high improbability is also likely to found in the selection of the rare algorithmic cases needed to make self-organization work.

Kairosfocus finishes with this:

So — noting that self-ordering is a species of mechanical necessity and thus leads to low contingency — we see the significance of the trichotomy necessity, chance, design, and where it points in light of the evidence in hand regarding FSCO/I in DNA etc. END

My Comment: This statement identifies mechanical necessity with low contingency; I think that’s intended to suggest that mechanical necessity cannot be the information bearer required for life; a conclusion that as far as I’m concerned may or may not be true. 

Let me stress that I have no vested interest in evolution as a theory and will continue to follow the views of the de-facto IDists with great interest. But I certainly would not argue against evolution along the above lines. 

No comments: