More (Un)Intelligent Design
This article at Panda’s Thumb is quite good at showing the problems with intelligent design (ID) research. Reed Cartwright, Ian Musgrave, and Steve Reuland takes a look at an article by Michael Behe and David Snoke, Simulating evolution by gene duplication of protein features that require multiple amino acid residues. In this paper Behe and Snoke simulate the amount of time necessary to arrive at a specific type of mutation and conclude that the amount of time necessary to arrive at this specific mutation is too long, hence Darwinian evolution can’t be the only answer. The following is from Behe and Snokes conclusion,
Although large uncertainties remain, it nonetheless seems reasonable to conclude that, although gene duplication and point mutation may be an effective mechanism for exploring closely neighboring genetic space for novel functions, where single mutations produce selectable effects, this conceptually simple pathway for developing new functions is problematic when multiple mutations are required. Thus, as a rule, we should look to more complicated pathways, perhaps involving insertion, deletion, recombination, selection of intermediate states, or other mechanisms, to account for most [multi-residue] protein features.
It looks pretty bad. The real kicker here is the nebulous comment “…or other mechanism, to account for most [multi-residue] protein features.” Basically this is opening the door for things like Intelligent Design, and once you have the thin edge of that wedge in the crack you keep driving it in until you get the whole edifice to crumble. The end result being the re-introduction of some sort of Creation myth into the science classroom.
But is it really that bad? Note that the authors are using a very specific mutation. Basically they are setting the result. This paper violates one of the basic tenets of another intelligent design “theorist” William Dembski. Dembski has criticized evolutionary algorithms for “smuggling in design”. Dembski pointed to the fitness functions in these algorithms and relied on the No Free Lunch Theorems by Wolpert and MacReady. While Dembski’s complaint ultimately turned out to be false, it is ironic he doesn’t raise a similar complaint here with Behe & Snoke’s work.
In Dembski’s work he came up with the notion of complex specified information (CSI). The complex part means that the event in question has a really small probability. Specified means that the event has a recognizable pattern. And example of this are the letters in a person’s name. For example the sequence ‘Steve Verdon’ (tossing in the space as a possible “letter”) would constitute specified information to everyone reading this post. The problem with Behe & Snoke’s work is that they are sneaking in specification. By selecting a specific mutation they are specifying the result they want and hence are rigging the “game” to produce a highly unlikely result. Highly unlikely results are expected to take a large number of trials to be observed (on average). This in turn means a large amount of time. Or as Cartwright, Musgrave, and Reuland put it,
However, the model has many restrictive assumptions that prevent it from supporting the conclusion that Behe and Snoke make. In reality, the paper says that if you have a protein function that requires two or more specific mutations in specific locations in a specific gene in a specific population, and if the function is not able to be acted on by natural selection until all mutations are in place and if the only form of mutation is point mutation, and if the population of organisms is asexual, then it will take a very large population and very long time to evolve that function. This is not unexpected.
No kidding it is not unexpected. In fact, one has to wonder if perhaps it was not set up this way purposefully to get precisely the result Behe & Snoke wanted. Further, this assumption turns the whole Darwinian evolutionary process on its head. A crude way of thinking about is that there are random mutations going on quite regularly. Then if there is say a change in the environment if one of the mutations simply provides a selective advantage then those organisms with that mutation would be more successful than those without. Ultimately, the mutation would become dominant and the non-mutations could very well go extinct. There is no “right” or “correct” mutation. It is more of a, “is it good enough to survive” rule vs. it has to be “just so”. This is kind of funny considering how many anti-evolutionists complain of the “just so stories” from the pro-evolutionists.
It is funny how ID “theorists” end up violating their own rules, objections, and doing the things they themselves complain about. In this case, it turns out that Behe & Snokes have done exactly the opposite of what they have set out to do. They have calculated how long it would take for specific mutations to occur and have provided a sizeable over-estimate. In other words, because of their restrictive assumptions the time span they come up with is quite high. So for less restrictive and more reasonable assumptions the time spans fall to much more reasonable levels and the paper actually does not, as William Dembski notes, undermine Darwinian evolution.