Which Irreducible Complexity?

I was recently asked to explain again the shift that Behe made in his famous Irreducible Complexity argument. Most ID advocates claim it has never been falsified. That confident assertion is just false.

Below, I’ve excerpted a real and all too typical dialogue with an ID advocate. The conversation is pulled from this thread, with emphasis and a few typo corrections:

Evolution could not have taken place because these artifacts we’ve found are far too complex to have evolved

You really need to understand that when you embrace a definition of ID that can only come from an opponent of ID and something you would never hear from any competent proponent, you have deluded yourself - intentionally or unintentionally - into being content with the understanding of a straw man. Until you are willing to see ID for what its opponents say it is, your objections to it will remain completely off point.

He echoes the language of Behe himself and the Discovery Institute…

Behe, also a senior fellow at the Discovery Institute, argues for the idea of “irreducible complexity,” meaning certain aspects of living organisms, such as the mechanism for blood clotting or the development of the human eye, are too complex to evolve.

Now, there are ton more quotes just like this.

I would also ask you a question @deliberateresults. Which definition of IC do you subscribe to? There are two. Which one do you mean? Do you even know what the two definitions are?

I know you have been chomping at the bit for my answer to this question. In Darwin’s Black Box, Behe defines IC:

a single system composed of several well-matched, interacting parts that contribute to the basic function, wherein the removal of any one of the parts causes the system to effectively cease functioning.

Now I know you can’t wait to provide the “other” definition for me…

So this is one definition (I will call it IC1).

And I agree that the flagellum is IC1 by this definition. That is demonstrable directly with experiments. We can define the system, the function, and the parts. Then start removing parts one at a time, to see if the function is retained. For one example of flagella at least, removing some of the parts does not obliterate function, but removing many of them does obliterate function. Therefore (because removing some of the parts kills the function) it is an IC1 system. This [is an] objective [definition, directly observable and measureable, and that is its brilliant and strength. Everyone using the IC1 definition will agree what is and isn’t IC1.]

Now, Behe’s ID hypothesis is that IC1 systems cannot be evolved. We can test that hypothesis. We have, and we find that this hypothesis is false. There are several examples of IC1 systems evolving in laboratory systems. So that hypothesis, (that IC1 reliably identifies evolvable systems) has been clearly falsified. It was falsified even before Behe first made the argument in 1996.

One simple example is the labmbda phage that Venema discusses in his blog: http://biologos.org/blogs/dennis-venema-letters-to-the-duchess/the-evolutionary-origins-of-irreducible-complexity-part-4. The new species of virus is an irreducibly complex system that requires OmpF (on its prey), OmpF-binding proteins, genetic material and capsid proteins, etc. Remove anyone of these things, and it stops working, therefore the new phage is by definition an IC1 system. However, we observed directly that it has been produced by evolution. There are literally thousands of examples like this. IC1 does not reliably identify un-evolvable systems.

If the flagellum is not evolvable, this has nothing to do with it being an IC1 system. Evolution has no problem with IC1 systems, and this is demonstrable in experiments.

This is one reason that Behe himself has abandoned this argument. He does not make it anymore, because it fails.

So tell me now, what is the next definition of IC that was proposed (IC2 if you will)? You have read ID more than most people, you say. So tell me. What was the next definition? If you don’t know what it is, then why exactly do you think the IC argument is so strong? Maybe flagella are not evolvable, but without doubt the IC1 definition you use does not prove this so. Given overwhelming evidence, the argument fails basic logic.

I asked you to specify which version of IC you were referring to anyways, because this really changes how it needs to be responded to. After a side track covering my views on the origin of life (where I hope you see a great deal of common ground with me), you finally get around to answering my question. You apparently do not even know that there are more than one definition. You proposed the IC1 definition…

a single system composed of several well-matched, interacting parts that contribute to the basic function, wherein the removal of any one of the parts causes the system to effectively cease functioning.

And this is where we find ourselves now.

Before we move forward, let’s just remember the context of all this. Despite your doubts, I certainly do understand the ID case. The issue is not that I am clueless about it. Rather, I am not convinced by the argument.

So what is the next definition of IC, which I will call IC2? In 2000 Behe writes a different definition, that he carries forward to this day in the Edge of Evolution.

An irreducibly complex evolutionary pathway is one that contains one or more unselected steps (that is, one or more necessary-but-unselected mutations). The degree of irreducible complexity is the number of unselected steps in the pathway.

Now this definition suffers greatly compared to IC1, because it is not longer objective and directly measureable. Instead, we need to understand something of the evolutionary pathway, which is not directly observable.

Still, for the IC2, most knowledgeable scientists agree that IC2 systems exist in nature too. Similar to IC1 systems, we are convinced that evolution can evolve IC2 systems also. Embedded in this definition is an important straw man. He writes “one or more unselected steps.” Here, into the IC2 definition is being smuggled a strict enforcement of “Darwinian” evolution. Now, anyone following biology at a basic level of competence knows that strict-Darwinian evolution was falsified a long time ago, in the 1970s. We now know that a large proportion of changes are near neutral or even slightly deleterious in the short term. So this dominant pathway of change is intrinsically ruled out by Behe’s definition.

So, with that in mind, it is correct to say that modern biology agrees that IC2 systems appear throughout biology, because we thinking neutral mutations are often the evolutionary path to the systems we see, and they are not selected. Of course, this is all beside the point, because we are convinced that evolution can evolve IC2 systems using unselected steps (like neutral drift and draft, or even slightly deleterious steps).

Now, we could define a new version of IC (IC3 if you humor me here) that allows for all changes by neutral mechanisms. How might that fair? Unfortunately, this definition is even more unhinged from objective evaluation than IC2. We have moved entirely away from the concrete brilliance of IC1, to a fundamentally theoretical classification that is not directly testable in any way.

Well, no one has yet demonstrated that any system in biology is IC3. The claim that IC1 = IC3 is clearly false. So is the claim that IC2 = IC3. But how do we demonstrate that a biological system is IC3, and not just IC1 or IC2? No one knows. No one has proposed a way that seems remotely plausible to anyone outside the “already convinced” ID crowd. Instead we have a lot of sloppy arguments that somehow treat all classes as if they are the same thing.

Honestly, I do not know how to solve this problem. But I am not an ID advocate, so that isn’t my job. The fundamental problem for IC arguments is that biologists already agree that IC1 and IC2 systems exist, and we have even demonstrated their evolution directly in the laboratory.

Not part of the original thread, I’ll also point to this beautiful example of an IC1 system evolving.

An amoeba and a bacteria live fine apart. Then the bacteria infects the amoebae. Then the two lose the ability to live apart. Take away the bacteria, the amoeba dies. Take away the amoeba, the bacteria dies. This is, by definition, an IC1 system. It is very easy to evolve. The basic pattern for increasing complexity is.

  1. Start with something.
  2. Add something new.
  3. Make the new thing necessary.

“Add something, make it necessary.”


This easily builds up IC1 complexity by obvious mechanisms that most ID proponents would pejoratively call destructive evolution as if that ends the argument. Yet…

So that means that destructive evolution (what ever that is) can, apparently, evolve ever increasing amounts of IC1 complexity.

1 Like

One more Addendum.

This article has some really important historical information. The key point is that IC1 was shown evolvable well before Behe even proposed it. Quoting from this article below…


The Mullerian Two-Step

“Irreducible complexity” is a simple concept. According to Behe, a system is irreducibly complex if its function is lost when a part is removed. Behe believes that irreducibly complex systems cannot evolve by direct, gradual evolutionary mechanisms. However, standard genetic processes easily produce these structures. Nearly a century ago, these exact systems were predicted, described, and explained by the Nobel prize-winning geneticist H. J. Muller using evolutionary theory. Thus, as explained below, so-called “irreducibly complex” structures are in fact evolvable and reducible. Behe gave irreducible complexity the wrong name.

H. J. Muller predicted and discussed M. J. Behe’s “irreducibly complex” structures in two different papers, one in 1918 and one in 1939. This prediction was made long before the genetic material was known or anyone had seen the structure of a “molecular machine”.

“… thus a complicated machine was gradually built up whose effective working was dependent upon the interlocking action of very numerous different elementary parts or factors, and many of the characters and factors which, when new, were originally merely an asset finally became necessary because other necessary characters and factors had subsequently become changed so as to be dependent on the former. It must result, in consequence, that a dropping out of, or even a slight change in any one of these parts is very likely to disturb fatally the whole machinery; for this reason we should expect very many, if not most, mutations to result in lethal factors …”
Muller 1918 pp. 463-464. (emphasis in the original)

"V. The role of interlocking and diffusion of gene functions in hindering true reversal of evolution

“… an embryological or physiological process or structure newly arisen by gene mutation, after becoming once established (with or without the aid of selection), later takes more and more part in the whole complex interplay of vital processes. For still further mutations that arise are now allowed to stay if only they work in harmony with all genes that are already present, and, of these further mutations, some will naturally depend, for their proper working, on the new process or structure under consideration. Being thus finally woven, as it were, into the most intimate fabric of the organism, the once novel character can no longer be withdrawn with impunity, and may have become vitally necessary.”
Muller 1939 pp. 271-272.

H. Allen Orr has explained Muller’s explanation for “irreducible complexity” in several articles in the Boston Review criticizing Behe’s and William Dembski’s writings. Orr has emphasized the adaptive possibilities in the Mullerian two-step (i.e. improvement of function at each step). However, the mechanism is more general and does not even require selection, a point that Muller himself made originally, 50 years before neutral evolution was found to be important in molecular evolution.

“An irreducibly complex system can be built gradually by adding parts that, while initially just advantageous, become-because of later changes-essential. The logic is very simple. Some part (A) initially does some job (and not very well, perhaps). Another part (B) later gets added because it helps A. This new part isn’t essential, it merely improves things. But later on, A (or something else) may change in such a way that B now becomes indispensable. This process continues as further parts get folded into the system. And at the end of the day, many parts may all be required.”
Orr 1996

"… gradual Darwinian evolution can easily produce irreducible complexity: all that’s required is that parts that were once just favorable become, because of later changes, essential. "
Orr 1997

Note from @swamidass: “Darwinian” evolution was falsified a long time ago (neutral processes are important too), this is a quote of a scientist writing before that was known. Do not be confused by that anachronism. Modern evolutionary science is not Darwinian evolution. It turns out that Muller was wrong on this:

Turns out that most mutations are neutral, not lethal. See The Neutral Theory of Evolution. It seems that over 100 years, we’ve learned a thing or two of about evolution, because ware actively studying it. Remember, at this point in history, Muller didn’t even know what DNA was. No surprise he had a thing or two wrong.

Now, why exactly does anyone still think that IC1 is a good argument? How can we trust anyone advances yet again the IC argument and is ignorant of this history?


We can imagine several types of IC that should not be conflated…