George Gilder and other ID advocates claim a truism as if it were somehow an argument against evolution: Proteins cannot put i

Physics, Information Theory, and Evolution

Feedback, discussion, comments, questions: Chris Cogan, ccogan@ou.edu

George Gilder, William Dembski, other ID advocates, and many conventional creationists claim a truism as if it were somehow an argument against evolution: Proteins cannot put information into the genes that was not there. In a narrow, but irrelevant, sense, this is true. For one thing, proteins would not know what information to put in the genes, because it's, well, just protein.

However, this is no argument against the theory evolution because the theory evolution does not claim that this happens, nor is it required for evolution to occur. Evolution claims that modifications in the genes occur for various physical reasons (the accidents of the physical processes involved) and that some of these modifications happen to produce fitness improvements in the organisms that have them, and so they tend to get passed on to future generations more than the relatively less-fit "normal" versions of the genes.

That's really all there is to it, not counting the interesting variations and additional mechanisms that have evolved for increasing randomness in some cases. Unless the creationists want to claim that "random" (i.e., not intelligently guided) changes occur (which seems unlikely, considering that they also claim that such "randomness" is a reason for rejecting it as a basis for supposing ("macro-") evolution to be a fact), then they cannot rationally argue that some such changes could, if only in principle, be beneficial to the prospects for the perpetuation of the genome that contains such modifications. If we take a million copies of a novel and randomly change one character in each copy, a few of them will probably be (by most standards) improvements. If we then discard all but those few and make millions more with similar modifications (but nearly all of them containing the changes from the previous cycle), there must, in this second generation, be a few that would be even further improvements.

Why would this work? Because while the modification process is "random," the selection process is not.

No, it's not that the selection process is intelligent, since, obviously, "improvements" could be achieved by clearly unintelligent selection processes as long as they are systematic. Unfortunately for ID/creationism, non-random selection is not synonymous with intelligent selection. Dirt dumped into a tank of water and stirred will settle out in much the same way every time the experiment is tried as long as the dirt and the water and the tank (etc.) are essentially the same every time. Large but compactly shaped and dense particles will settle to the bottom first, equally dense but very small particles will remain suspended longer, and so on. The order and systematic nature of this "sorting" depends on systematic factors such as the nature of the water and the particles, not on any intervention by an outside intelligence.

The laws of physics are, if nothing else, very systematic. That's why they give us the technological power they do. If they were incoherent, sometimes producing, for no good reason at all, very different results from essentially identical conditions, these "laws" would not even be laws of physics.

The physical environment of living organisms has, inherently, a certain informational structure. Evolution is a way for replicating things to adapt to the local conditions by taking on an internal informational structure (in the genes and the anatomy and physiology) that "maps" requirements of genetic survival to the systematic nature of the environment. This "taking on" is achieved by making unintelligent modifications to genes and then seeing if they work. If they do, so much the better. If they don't, the modifications go out of existence.

There is nothing any more "mysterious" about this process than there is about the fact that a lump of mud dropped over a rock shapes itself to the surface of the rock. Would even the most die-hard ID advocate claim that there must be intelligence guiding the shaping of the mud to the rock?

No? No?

Then why must there be intelligence guiding the "shaping" of the genes to the requirements of the survival of those very same genes when they are "dropped," organism and all, into an environment? How can the surviving genes, over many generations, not be forced to fit the environment, if all the genes that don't fit the environment are unceremoniously wiped out by that very same environment? What alternative is there, other than extinction (which, in fact, occurs very often in cases where the lack of guidance in producing modifications means that the "right" modifications do not occur in time to allow the genes to survive)?

But, ID advocates try to make the simple process by which "information" (i.e., nothing more than "fit" genetic modifications) gets into the genes and the "non-information" (i.e., "unfit" genes) gets "weeded" out into something mysterious that requires a "special" explanation (which, unfortunately, they don't really provide). This special explanation for this supposedly special fact is, of course, intelligent design.

But all of the information that needs to get into a genome for it to survive is information already present in the environment, by the very nature of the environment. For example, if the environment is the middle of the Pacific, then the information needed to survive in that environment is the environment, and any genetic modifications to an organism that don't happen to reflect one way to survive in this environment will be killed off.

The "information" in the genes is simply the sequence of nucleotides (with a little stored in the "machinery" of the reproductive process). But this information is not, despite the claims of Phillip Johnson and others, some special kind of "stuff" above and beyond the physical structure of the DNA. Unintelligent modifications to DNA can be information because they can cause the construction of one protein rather than another, and thus aid or harm the chances for survival of the whole genome in which they occur.

The mystical/magical view of information theory advocated by the ID folks is truly bizarre, and is not only rejected by biologists, but by all real information theorists as well.

Evolution works because it involves replication with modifications, inheritance of those modifications, and a systematic (non-random) "selection" process (i.e., some organisms do a better job of living and getting their genes passed on than others). None of these processes is intelligent, but the selection process is systematic. A water environment systematically "selects" organisms that are adapted to living in water. A desert environment systematically "selects" organisms that are adapted to living in desert conditions. All the rest die out.

Nature "pays" for the information in the genes by producing billions of "failures" in order to have a small number of "successes" (genes that work to enhance their own survival by enhancing the chances of their genome to be replicated into the future).

Notice three things about this explanation:

It does not require the introduction of any designers (not even alien designers).

It does not require the introduction of supernaturalism, not even of a minor life-designing spirit.

It does not involve either infinite regresses or arbitrary exemptions from the requirement of having an explanation, as does the intelligent designer theory.

It actually explains how life changes over time, whereas the ID theory cannot explain anything. It cannot explain, for example, why there were no humans sixty-five million years ago, even in environments where humans could have survived had God put them there. Evolution does explain why there were no humans there, or any other species of similar mental development: Evolution was forced to proceed slowly (precisely because it didn't have the advantages of intelligent guidance) to the development of human-level intelligence. Producing a brain like that of a human is not something that could occur overnight precisely because there was no intelligence guiding the process or forcing such rapid development.

It makes physical sense. It makes physical sense that genetic modifications that help the survival of the genome will tend to be the ones that get passed on to future generations. It makes physical sense that those organisms that have genes that are sufficiently unfit relative to their environment will get killed off, because they don't produce the right physical structures that in turn would enable the organism (and the genes themselves) to survive and reproduce. A gene for extremely high water requirements will not do well in camels living on their own in the Sahara, so the physiology that results from such genes will not be unable to survive physically in such an environment. When such an organism dies, it takes its genes with it, and it dies because it is unable to carry out the kinds of physical processes that would enable it to survive.

This applies even to brain function: An organism that has a brain that too frequently makes certain kinds of mistakes is an organism that will get itself physically destroyed, along with its genes.

This is why, information theory or not, we can describe the processes of evolution in purely physical terms. It is also why the re-writing of information theory (into the information theory nonsense) for the purpose of "refuting" evolution fails. Reality blithely ignores the lies we tell ourselves about it. Do the ID folks really think that the physical processes of nature really care about their information theory ideas? If so, they should be studying physics, not information theory.

Sound information theory cannot violate physical principles, because all information is stored in some physical form. Sound information theory cannot demand that the modifications of a genome always be ones that never result in a general improvement in biological fitness (which would violate both physics and probability theory as well). In the real world, physics trumps information theory. If there is a conflict between observational physics and information theory, it is information theory that is wrong, not physics. That is, a valid information theory for dealing with things in the real world must be compatible with physics, and ID-information theory amounts to a wholesale denial of the laws of physics to suit their religion.

Summary:

Evolution works by the "shaping" of life to the requirements of survival by making many modifications and only allowing those that "work" to be passed on to further living things. The ones that don't work are destroyed, leaving too few or no descendents.

Information is stored in physical form, so there is no need for any outside "source" of the information in genes. The "information" in the genes is nothing more than the physical sequencing of the DNA, and so is only information in a secondary sense. There is no special "stuff" called "information" that is somehow distinct and physically separable from the physical matter of DNA. It only "works" because of the specific form it is stored in. The individual nucleotides are physically distinct, and, grouped into codons, they cause distinct physical processes because of their physical structure and nature (not because, from our point of view, they contain "information").

The theory of evolution works and the "theory" of ID does not. It can't even explain a large majority of the facts that the theory of evolution explains. For example, it cannot explain the "tree of life" patterning of species, especially over space and time, without introducing additional ad hoc postulates about the designer's goals and constraints, with no explanation for them, resulting, eventually, in a theory so loaded down with ad hoc "epicycles" and special exemptions and exceptions that the result is a useless hodge-podge of drivel.

Finally, completely independently of any information theory ideas, evolution works for physical reasons. Genes are modified by physical processes, and whether a new modification works or not is determined by further physical processes. It is only from our perspective that genes are regarded as carrying and conveying information from one generation to another.

Note: Increasing randomness can have several uses in living organisms. One is in sensory sensitivity. Receptors that are pre-stimulated almost to the point of firing by a small random signal will respond to a smaller additional stimulus when receptors that have the same threshold but no "pre-stimulation" will not. This is because, if the receptor is almost at the point of firing, it takes relatively little additional energy to make it actually fire. Another use for "randomness" is in the processes that ensure (usually) that the heart beats in a reliable way, without "locking" up.

The human brain appears to use similar processes for sustaining awareness, for keeping vicious cycles from preventing proper functioning, and for cognitive processes. Some brains still get into vicious cycles, but this can be viewed as occurring when the randomization processes themselves fail. Further, there seems to be a natural "evolutionary" process of modification and testing that occurs to help prevent us from becoming too strongly "locked in" to responses that may not be optimal. Within a certain range, "randomization" can provide a means for progressive improvement, as successive successful modifications are kept and failed modifications get discarded. These processes seem to occur most at the "fringes," where our knowledge and understanding are not clear, and they help us to achieve understanding and clarity in such cases.

Finally, in humans, each offspring has small and apparently "random" modifications to its proteins that serve to make the body more nearly immune to micro-organism-caused diseases that are adapted to their parents' proteins but not to those of the new generation. This may be one reason why we have not been wiped out by such diseases; despite our generally slow evolutionary rate relative to those of micro-organisms, this mechanism enables us to be a "moving target" and thus require many generations of evolution of such micro-organisms to adapt. Further, because these modifications are different from one person to another, they help ensure that at least some people will be immune to almost any disease in any generation.

In short, do not scorn processes merely because they have no special uniformity or simple pattern, because randomness can be your friend.

Feedback, discussion, comments, questions: Chris Cogan, ccogan@ou.edu