Michael J. Behe is a professor of biochemistry at Lehigh University in Bethlehem, Pennsylvania. With the publication twenty-five years ago of Darwin’s Black Box, Behe began to advance an argument against Darwin’s theory that evolutionary traits said to alter a species by natural selection typically consist of multiple component parts, resulting from separate genetic mutations occurring on different occasions. The completed traits are “irreducibly complex,” because they cannot function until all their component parts have first reported for duty. So, contra Darwin, the survival advantage eventually conferred by the functioning trait cannot possibly have “naturally selected” all the trait’s necessary component parts, which had to have made their appearance before the trait could function at all.

Behe concludes that the only possible explanation for the appearance of irreducibly complex traits that aid survival and transform species is that they were intelligently designed in advance.

Now comes Behe’s latest book, A Mousetrap for Darwin.  It is a 556-page compendium of more than 100 brief essays by Behe responding to the many criticisms of his argument over the years.

But however much has already been said about Behe’s idea (most recently, for example, in Nathan Lents’s Skeptical Inquirer article “Behe, Bias, and Bears (Oh My!),” March/April 2021), there is at least one thing about it that cannot be said enough: What he lumps together as “intelligent design” is not a single hypothesis but an amalgam of two very different hypotheses. The first qualifies as a scientific claim, while the second does not.

Behe’s first hypothesis (“irreducible complexity”) is that Darwin’s theory of random mutation/natural selection is inadequate to account for available data on how evolution actually works. Whatever you think of it, that is a claim made within the boundaries of biological science and amenable to scientific inquiry. It may be proven right, wrong, or something in between. But whatever it proves to be, the proof will be evidentiary, and Behe’s fellow biologists will be the judges of what the evidence shows. Most of the essays in Mousetrap—over 75 percent by my count—address only this specific claim of irreducible complexity.

Behe’s second hypothesis (“intelligent design”) goes like this: because random mutation/natural selection cannot account for evolution, only an intelligent actor can. The premise of this second hypothesis, Darwinian inadequacy, remains in question. But even if that premise were proven, Behe does not explain why the conclusion of supernatural design would be compelled by it or why that should be the sole alternative to random mutation/natural selection.

As for the source of intelligent design, Behe isn’t talking. “I argue simply for the conclusion of design itself,” he says. “My argument … is directed merely toward the conclusion of design. … The core claim of intelligent design theory is quite limited. It says nothing directly about how biological design was produced, who the designer was … or other such questions,” Behe wrote in a letter to Science in 2000.

Nevertheless, despite Behe’s best efforts to write a play about Hamlet without the Prince of Denmark, we can hardly miss his point: The unnamed Generator of Design (“GoD” for short) is intelligent like we are. Because no merely natural process such as random mutation/natural selection can produce the outcomes the GoD intends, the GoD has obviously intervened at one or more points over time to achieve its designs. (For an irreverent look at some of these needed “interventions,” see the recent Creationist Funhouse series in SI.) Science cannot observe how the GoD has done this, but science has accumulated sufficient evidence of irreducibly complex results that no natural process can explain to conclude that the GoD, whatever the GoD may be, must have done whatever it is the GoD does to make species evolve.

Unpacking Behe’s second hypothesis provides insight into why science operates by the rule of thumb known as “methodological naturalism.” This is the working assumption that answers to questions about how the natural world works should be sought exclusively within the same natural world from which the questions arose. Methodological naturalism preemptively forecloses supernatural explanations such as intelligent design for natural phenomena such as evolution.

A supernatural GoD, by definition, cannot be located in space/time, those ultimate coordinates of the only arena of scientific inquiry there is. Scientific inquiry can therefore learn nothing about the properties and characteristics of a GoD. The mechanism by which a GoD might efficiently interact with the natural world to control the course of biological evolution is categorically beyond the capacity of science to even guess at. And if nothing can be either known or ruled out about the location, the nature, and the instrumentalities of Behe’s GoD, then no further data will be forthcoming from Behe’s second hypothesis—either about evolution or about the GoD.

The intractable problem with intelligent design is not that it is a scientific hypothesis that is unproven or false. The problem is that it is not a scientific hypothesis at all. You might say its terms are irreducibly unfathomable, an inquiry-ending feature that invariably accompanies a resort to the supernatural. Methodological naturalism is not an irrational bias that needlessly restrains scientists from following the evidence to wherever it leads. It is a wisdom arising from the collective experience of scientists across the centuries that irreducibly unfathomable explanations for how the natural world works don’t lead anywhere.