Plants background with biochemistry structure.
Image licensed via Adobe Stock
Michael J. Behe A (R)evolutionary Biologist

“Reducible complexity” in PNAS


Dear Readers, Recently a paper appeared online in the journal Proceedings of the National Academy of Sciences, entitled “The reducible complexity of a mitochondrial molecular machine” ( As you might expect, I was very interested in reading what the authors had to say. Unfortunately, as is all too common on this topic, the claims made in the paper far surpassed the data, and distinctions between such basic ideas as “reducible” versus “irreducible” and “Darwinian” versus “non-Darwinian” were pretty much ignored. Since PNASpublishes letters to the editor on its website, I wrote in. Alas, it seems that polite comments by a person whose work is the clear target of the paper are not as welcome as one might suppose from reading the journal’s letters-policy announcement (“We wish to provide readers with an opportunity to constructively address a difference of opinion with authors of recent papers. Readers are encouraged to point out potential flaws or discrepancies or to comment on exceptional studies published in the journal. Replication and refutation are cornerstones of scientific progress, and we welcome your comments.”) ( My letter received a brusque rejection. Below I reproduce the letter for anyone interested in my reaction to the paper. (By the way, it’s not just me. Other scientists whose work is targeted sometimes get the run around on letters to the editor, too. For an amusing / astounding example, see Call me paranoid, but it seems to me that some top-notch journals are real anxious to be rid of the idea of irreducible complexity. Recall that last year Genetics published a paper purportedly refuting the difficulty of getting multiple required mutations by showing it’s quick and easy in a computer—if one of the mutations is neutral (rather than harmful) and first spreads in the population ( Not long before that, PNAS published a paper supposedly refuting irreducible complexity by postulating that the entire flagellum could evolve from a single remarkable prodigy-gene ( Not long before that, Science published a paper allegedly refuting irreducible complexity by showing that if an investigator altered a couple amino acid residues in a steroid hormone receptor, the receptor would bind steroids more weakly than the unmutated form ( and (That one also made the New York Times!) ( For my responses, see here (, here (, here (, and here ( So, arguably picayune, question-begging, and just plain wrong results disputing IC find their way into front-line journals with surprising frequency. Meanwhile, in actual laboratory evolution experiments, genes are broken right and left as bacteria try to outgrow each other ( Well, at least it’s nice to know that my work gives some authors a hook on which to hang results that otherwise would be publishable only in journals with impact factors of -3 or less. But if these are the best “refutations” that leading journals such as PNAS and Science can produce in more than a decade, then the concept of irreducible complexity is in very fine shape indeed. ************* To the editor: Reducible versus irreducible systems and Darwinian versus non-Darwinian processes The recent paper by Clements et al (1) illustrates the need for more care to avoid non sequiturs in evolutionary narratives. The authors intend to show that Darwinian processes can account for a reducibly complex molecular machine. Yet, even if successful, that would not show that such processes could account for irreducibly complex machines, which Clements et al (1) cite as the chief difficulty for Darwinism raised by intelligent design proponents like myself. Irreducibly complex molecular systems, such as the bacterial flagellum or intracellular transport system, plainly cannot sustain their primary function if a critical mechanical part is removed. (2-4) Like a mousetrap without a spring, they would be broken. Here the authors first postulate (they do not demonstrate) an amino acid transporter that fortuitously also transports proteins inefficiently. (1) They subsequently attempt to show how the efficiency might be improved. A scenario for increasing the efficiency of a pre-existing, reducible function, however, says little about developing a novel, irreducible function. Even as evidence for the applicability of Darwinian processes just to reducibly complex molecular machines, the data are greatly overinterpreted. A Darwinian pathway is not merely one that proceeds by “numerous, successive, slight modifications” (1) but, crucially, one where mutations are random with respect to any goal, including the future development of the organism. If some mutations arise non-randomly, the process is simply not Darwinian. Yet the authors say nothing about random mutation. Their chief data are sequence similarities between bacterial and mitochondrial proteins. However, the presumably homologous proteins have different functions, and bind non-homologous proteins. What is the likelihood that, say, a Tim44-like precursor would forsake its complex of bacterial proteins to join a complex of other proteins? Is such an event reasonably likely or prohibitively improbable? Clements et al (1) do not provide even crude estimates, let alone rigorous calculations or experiments, and thus provide no support for a formally Darwinian process. Their only relevant data in this regard is their demonstration that a singly-mutated bacterial TimB can substitute for Tim14 in mitochondrial transport. While that is certainly an interesting result, rescuing a pre-existing, functioning system in the laboratory is not at all the same thing as building a novel system step-by-random-step in nature. Biologists have long been wary of attempts to fill in our lack of knowledge of the history of life with imaginative reconstructions that go far beyond the evidence. As I have discussed (5), extensive laboratory evolution studies over decades offer little support for the plausibility of such felicitous scenarios as Clements et al (1) propose. The authors may well be overlooking formidable difficulties that nature itself would encounter. References

  1. Clements A, et al. (2009) The reducible complexity of a mitochondrial molecular machine. Proc Natl Acad Sci USA doi/10.1073/pnas.0908264106.
  2. Behe, MJ (1996) Darwin’s Black Box :The Biochemical Challenge to Evolution (Free Press, New York).
  3. Behe MJ (2000) Self-organization and irreducibly complex systems: A reply to Shanks and Joplin. Phil Sci 67:155-162.
  4. Behe MJ (2001) Reply to my critics: A response to reviews of Darwin’s Black Box: the biochemical challenge to evolutionBiol Phil 16:685-709.
  5. Behe, MJ (2007) The Edge of Evolution: the Search for the Limits of Darwinism (Free Press, New York).


Michael J. Behe

Senior Fellow, Center for Science and Culture
Michael J. Behe is Professor of Biological Sciences at Lehigh University in Pennsylvania and a Senior Fellow at Discovery Institute’s Center for Science and Culture. He received his Ph.D. in Biochemistry from the University of Pennsylvania in 1978. Behe's current research involves delineation of design and natural selection in protein structures. In his career he has authored over 40 technical papers and three books, Darwin Devolves: The New Science About DNA that Challenges Evolution, Darwin’s Black Box: The Biochemical Challenge to Evolution, and The Edge of Evolution: The Search for the Limits of Darwinism, which argue that living system at the molecular level are best explained as being the result of deliberate intelligent design.