More Articles on Evolution

Where's the Evidence?

Robert DiSilvestro

In Darwin's Black Box, Michael Behe emphasizes the idea of irreducible complexity: he claims that complex biochemical systems need many components present simultaneously to work, and argues that such systems could not evolve without design. Allen Orr's review makes the very strong statement that the book is "just plain wrong." I am always reluctant to use such strong language unless I have an airtight defense of my position. I do not think that Orr provides such a defense of his claims.

Before explaining my opinion, though, I need to comment on why I am qualified to respond. At the beginning of his review, Orr introduced the issue of qualifications: he noted that a biologist's ideas about evolution currently draw more attention than a physicist's ideas, and that everyone seems to think themselves qualified to join the debate. The observation and concern seemed reasonable. At the end of the review, however, Orr went a step further and suggested that the evolution debate should be limited mainly to those who specialize in evolutionary biology.

This limitation seems too severe. For comparison, a physical biochemist working with electron spin resonance techniques in vitro is apt to discuss the implications his or her work holds for living organisms. Even if this person has never worked with living organisms, the discussions would generally be judged on their merits. I think the same principle should apply to my response to the review of Behe's book. Though I don't directly work on evolutionary theory, my research in protein characterization and nutritional biochemistry has been interpreted by others in light of evolutionary theories. Thus, I have a stake in this area. I may be hesitant to critique certain technical aspects of such theories, but that shouldn't keep me from raising more general issues.

One such issue is Orr's claim that irreducible complexity can be explained by natural selection. In supporting this claim, Orr, to his credit, rejected two simplistic explanations. But his own alternative proposal seemed similar to one of those he rejected. According to that argument, the components of complex systems evolve to perform one function, but then fortuitously come together into an irreducibly complex system that performs some other function. Orr says that this is "unlikely" and therefore not plausible as a "general solution to irreducible complexity." According to Orr's alternative, a complex system's components were originally advantageous for a particular function, but later became essential for that very function: unlike the rejected alternative, there is no shift of function. Still, the argument also involves a functional transition of sorts, and seems to me to require, as did the other model, a series of unlikely events. Molecules still had to fall into the right arrangement.

Orr never explained why this was likely to happen. He did give some possible scenarios (for example, the transformation of air bladders into lungs, which start out advantageous and then become essential), but without data to say why those scenarios were feasible. He also stated that his idea was not just hypothetical. His principal illustration, however, was not the evolution of a biological system, but an analogy to computer program writing. This is not an appropriate analogy because the program writer is not operating by totally random processes. Orr did also raise the issue of gene duplication and structural similarities. I would agree that this issue could be given more attention in subsequent writings on irreducible complexity. But gene structure similarities can't account for all the complexities of a given system. For instance, I sometimes ask students to list seven nutritional deficiencies that cause anemia by different mechanisms. These mechanisms involve many diverse structures and processes.

One final comment about the review: Orr, like so many biologists, mentions that there is lots of time for biological systems to make lots of changes. But--again like so many biologists--he doesn't relate time available to time needed to make a particular series of events likely. Unless that is done, statements about long time periods and biological changes don't mean much.

In conclusion, Orr's review continues a practice that I find bothersome: it equates a speculation about how a system might have evolved without design with evidence that this system did evolve without design.

---