The Edge of Evolution: The Search for the Limits of Darwinism, by Michael J. Behe (Free Press 2007) [B&T Books] QH367.B44 2007.

In his newest offering since Darwin's Black Box, Behe contends that random mutation and natural selection plays only a minor role in the evolution or transmutation of species. The phrasing of that statement is important because Behe does not deny either common descent of species (e.g., yes humans and apes do—according to Behe—have a common ancestor) or that these changes took place over many millions of years or that the universe is incredibly old. No, Behe's contention would seem to be that even given a near infinite amount of time, natural selection just cannot fit the bill. The alert reader might also note that Behe is no longer arguing that certain discrete organism display what he had previously called irreducible complexity, though this might simply be a result of a more modest program. Behe seems perfectly willing to see mutation and natural selection at work in the mid-range of biological classification but once one gets more specialize than species level or more general than orders, other forces are at work.

Behe makes this case by examining the relationship between malaria and sickle-cell anemia and the presumed paradox that while the HIV virus mutates at a very high rake, it is still relatively unchanged. I must admit, I found the relationship between malaria and sickle-cell fascinating. Apparently the reason that sickle cell is found in people of African decent is that having one copy of the gene mutation offers protection against malaria (though, of course, having two copies is deleterious) and so that gene was selected for. There have been other mutations found that confer a similar benefit—though without the price—but these have not occurred in areas where there is malaria and so the mutation is very rare.

This would seem to be a prefect illustration of how natural selection works. But, says Behe, it also shows its limits. There have been drugs to counter malaria and they worked for a few decades and then failed. Malaria mutated to compensate for these drugs, but, notes Behe, it did so at the expense of being a weaker strain overall. That is to say, the drug resistant strain does not function as well overall, save in an environment with certain drugs. Natural selection, says Behe, is not so much an arms race (where each species comes up with new better tricks to survive) than a war a attrition (learning what to live without in a hostile environment). Behe underscores the point by noting while malaria has quickly over taken the drugs to keep it in check, it has not yet found a way around the sickle cell gene—even after having to deal with it for thousands of years. Behe's explanation is that the number of simultaneous beneficial mutations for malaria to counter sickle cell is just too great. A single mutation would confer no benefit and would likely be deleterious and so selected out before it had a chance to be paired with some other mutation, which combined would confer a benefit and be selected.

From what I can tell of the less vociferous reviews from evolutionary biologists, Behe ignores the phenomenon of drift (mutations that are not selected for because they confer no benefit, but are not even expressed) where multiple mutation could occur and that Behe is too narrow in his dealing with types of mutations. Jerry Coyne, writing for The New Republic notes that Behe fudges with the notion of randomness. When geneticists speak of randomness, they mean something more like indifference to benefit rather simple stochastic predictability. One gene is not as likely to mutate as any other. It is for this reason that malaria has so much trouble with the sickle cell gene and why HIV frantically mutates in a narrow range.

This last point, however, is something that Behe implicitly employs later. He argues that there is fitness landscape where traits are optimized for fitness. Of course a mutation the confers fitness to a species in one area commits it to a certain tract from which it cannot retrace itself should it hit its inherit maximum fitness. Behe's picture, if true, would present a problem. However, it assumes there is a single inherit maximal fitness or that fitness can be modeled on single axis (ironically enough, this is denied by fellow ID proponent William Dembski in No Free Lunch, whose theory of conservation of information demands an infinite number of axises), that fitness does no alter with the environment and that a species cannot retreat in fitness once a maximum is reached. In short, Behe being too restrictive in his conditions and (say the critics) artificially makes natural selection more improbable than is warranted.

I am neither a mathematician nor a biologist, so much of the above I am taking if not from authority, certainly from a disadvantage when it comes to a critical understanding of the topic. I do, however, know something about William Paley's so-called teleological¹ argument which Behe manages to smuggle in and then folds in the fine-tuning argument to boot.

Paley's argument runs with an analogy. Imagine that you are on the beach and discover a watch. Upon examining it, even if you don't know what a watch is, you conclude that unlike the shells, rocks, and other such items, this an artifact of human contrivance. Behe follows in the same tradition by arguing since certain organisms display a high level of complexity and internal coordination that they cannot be the result of either chance of the working out of natural laws. These things require a designer. The problems with such arguments are severe. One must first note that for all intents and purposes, the universe is a one-off affair. It is not as if we can look about and distinguish between designed universes and happenstance ones. Moreover, all the instances of artifice with we are undeniably acquainted are those of human invention. We can tell something of the maker of these artifacts because we are generally acquainted with the artificers. The analogy breaks down with the designer of the universe because we have no such general acquaintance. We have no way of knowing whether the designer is either intelligent and purposeful or blind and mechanistic. It would be one thing to have an argument for God's existence and to have such an argument so that God is intelligent and purposeful and from there argue from providence how nature displays God's wisdom and purpose, but Paley's analogy (and Behe's appropriation) puts the cart before the horse. Finally, for Behe's argument to work, some things cannot be part of the intelligent artificer's design. If everything is intelligently designed then no contrast exists and the analogy fails. If some things are outside of the design, either the doesn't care about those things or the designer is incompetent to design those things. Neither of these options would seem to fit in with the program of intelligent design, although Behe seems to embrace the former.

Behe's use of fine tuning argument seems somewhat confused and his replies to objections (such and the many-universes conjecture made by string theorists) displays a certain lack of understanding of these objections and of modal logic. Part of the confusion exists in his insistence upon treating certain happy states of affairs (the earth being just so far from the sun, the moon being just the size and position it is, etc.) with the values of certain constants of the universe (the strength and extent of the fundamental forces and the like). The latter defines this universe and any universe like this one, the former contingent events within this universe. What they have in common is that if one grants the existence of God or some intelligent designer, one can argue that these things display a certain amount of providential care. Where they differ is that happy events are events and (if one grants God's existence) may be said to be consistent with an interaction between God and the cosmos. In the case of physical constants, we simply have nothing certain about whether these too arose from happy events before the universe came into being or are in fact strictly governed. It is a composition fallacy of treating what is within a class as if it were the same thing as the class itself.

This is not to say that The Edge of Evolution is without value. Behe's most vocal critics fulminate over his conclusions and his methods of getting there in part because they don't like his conclusions, not that the questions he raises don't have merit. It may be that Behe has jumped to intelligent design, but that hardly warrants the conclusion that we understand how evolution works at all levels.

Jimm Wetherbee

If The Edge of Evolution looks good, here are some other interesting Baker and Taylor Books. . .

• Avoid Boring People by James D. Watson.
  Call Number:QH3.W4 A3 2007
• The Language of God, by Francis S. Collins
  Call Number: BL240.3.C66 2006
• Scientists Confront Intelligent Design and Creationism, edited by Andrew J. Petto and Laurie R. Godfrey.
  Call Number: QH367.B44 2007

1. I should note that I do not object to teleological arguments in general, but the analogical versions have always struck me a inherently flawed. [return to review]

Updated August 14, 2008