DennisJones at UncommonDescent has published an argument that Behe’s irreducible complexity has resisted all attempts at refutation.
…The only logic fallacy would be to draw a conclusion while resisting further examination. Such is not the case with irreducible complexity. The hypothesis has endured 17 years of laboratory research by molecular biologists, and the research continues to this very day.
An irreducibly complex system is one that
(a) the removal of a protein renders the molecular machine inoperable, and
(b) the biochemical structure has no stepwise evolutionary pathway.Here’s how one would set up examination by using gene knockout, reverse engineering, study of homology, and genome sequencing:
I. To CONFIRM Irreducible Complexity:
Show:
1. The molecular machine fails to operate upon the removal of a protein.
AND,
2. The biochemical structure has no evolutionary precursor.
II. To FALSIFY Irreducible Complexity:
Show:
1. The molecular machine still functions upon loss of a protein.
OR,
2. The biochemical structure DOES have an evolutionary pathway
My (limited) understanding is that (1) will nearly always be confirmed.
The problem is confirming (2), the assertion that there is no evolutionary pathway.
The first problem I see is in methodology. There is no reason to believe that gene knockout is equivalent to stepping backward in the history of the protein. The first clue would be that knocking out the protein, by definition, disables the function in a way that would require a specific point mutation to restore. I don’t know of anyone in biology — even Behe — who believes this is the way proteins evolve. In fact, my understanding is that Behe doesn’t argue that proteins can’t evolve. This is simply a pointless experiment.
Then there is the larger problem of demonstrating no pathway at all. Douglas Axe seems to have taken a whack at that, but I am unaware that he has tested any realistic evolutionary scenarios, as has Thornton.
To explain my point of view I need to digress a bit and discuss three dimensional chess. Fans of Star Trek might recall the three tiered chessboard that allows jumping pieces up or down as well as through the usual grid.
When discussing evolution you need to be able to envision many dimensions, not just one or two or three.
A typical discussion of evolutionary change invokes just one dimension: the string of bases along a strand of DNA. Modify one, and you change or disable a function.
Except that somewhere in that 17 years since Behe expounded irreducible complexity, the picture has become more complicated. We now know that the majority of positions on a protein coding string can be changed to any arbitrary value without drastically modifying the fold.
So even in the one dimensional case, there are many paths to any current configuration. But to represent functional space, you need to consider all the possible equivalent strings. You need to be able to jump from one to any other.
And having jumped to another string (neutral mutation) you have subtly altered the game.
In the Lensky experiment, at least two neutral mutations occurred prior to any change in function. These mutations had no noticable effect on function, but were necessary to enable the functional mutations. Evolution jumped to another board in another dimension.
The assertion that there is no evolutionary pathway is hollow if it does not test all possible pathways. I have not seen anyone at UD give any attention to the multidimensionality implied by neutral mutations. They seem completely enamored of isolated islands and never mention multidimensional connected space.
Is it reasonable to believe that evolution can negotiate multiple dimensions? Check out the Lensky experiment.
I hope some folks here can add some links to relevant research and flesh out this argument. Or correct anything stupid I may have said.
And perhaps someone would like to address the rest of Behe’s argument, that complex functions require multiple genes or proteins, and this cannot be reached incrementally.
The other problem with the knockout approach is that the system we see may have come from a system with more parts, not fewer. Evolution is not only additive.
This is the “arch analogy” that shows one of the fatal flaws in the concept of irreducible complexity. An arch is irreducibly complex, but can be constructed with the appropriate scaffolding (which may be as simple as a mound of earth).
There are many problems with IC, but I think Dennis is only addressing the construction of sequences. This seems to be gpuccio’s argument also, although he hasn’t mentioned knockouts.
The problem I see is that ID advocates can never seem to get away from the concept of target. They cannot grasp that there are many ways to get from Chicago the New York.
Even worse, they cannot grasp the idea that evolution cannot even see targets.
Because evolution involves ratcheting at several levels, it is not a trivial matter to rule out an evolutionary pathway to a current sequence. The main contributors are (1) the effective irreversibility of many mutations, (2) the ratchet of Natural Selection, and (3) co-evolution.
(1) Although one could repeatedly sample the point-mutational neighbourhood of a sequence reasonably easily, mechanisms involving multiple residues are much less easy to address, and contain vastly more possibilities. A simple frame shift, or a multibase deletion, or an insertion from elsewhere in the genome (or another genome) … trying to reverse that is a near-impossible task. And then there is likely another step, and another … you aren’t trying to reconstruct just any old path, you want the actual path, in reverse. Chaos soon intervenes.
2) Demonstrating that the pathway is plausible requires a factoring in of long-ago environments, both inside and outside the cell, before one can determine that the protein was not detrimental in that environment.
3) The rest of the genome is under evolution too. A change here enables a change there, and if the pairing becomes beneficial, it is locked in where originally it was not.
Which may all seem like special pleading to the ID die-hard. But it’s the landscape we have. IC is (IMO) impossible to prove, and I’m sure Darwin knew this too in penning his famous phrase (“If it could be demonstrated that any complex organ existed[,] which could not possibly have been formed by numerous, successive[,] slight modifications[,] my theory would absolutely break down.“).
+1.
A hilarious title! = ))
Though ‘evolution’ doesn’t hold a belt in my books, I’m willing to take a shot at “Knocking Out Evolution” in more than 2/3rds of the Academy, probably more like 4/5ths or 5/6ths. In a (vast) majority of scientific and scholarly fields, evolutionary theory makes too little sense for people to give it much attention or respect. It’s a B-league lower theory for biologists and naturalists, not fit for higher discourse about human existence and meaning.
For those persons here at TSZ who cannot or will not admit anything is ‘real’ other than what is ‘natural,’ you are already knocked out by clinging to naturalistic evolutionism.
Behe is not a major player in this discussion (nor is any IDist). And the main issue here is not about reducibility or irreducibility. It is about elevation or lack thereof.
Ratchet-ratchet…pray.
I’ll just say that the first book I read on evolution was Chardin’s Phenomenon of Man. Required reading for everyone at my college.
50 years ago, almost to the month. But I had been reading about evolution for about seven years prior.
Can you cite an example of something which is “real” without also being “natural”? I don’t think there is any such thing, myself, but I could be wrong. So if you’ve got a specific, identifiable example of something which possesses the real-but-not-natural property, I’d be interested in seeing it.
The problem with Behe’s IC-equals-unevolvable argument is that it ignores two-thirds of all evolutionary changes.
There are three general classes of steps in an evolutionary (“Darwinian”, Behe says) pathway: One, there’s add a new part. Two, there’s remove an existing part. Three, there’s modify an existing part.
If the final step in the evolutionary pathway to System X is add-a-new-part, the immediate evolutionary predecessor to System X must be System-X-minus-one-part. And by Behe’s definition, an IC system that lacks one part cannot function. So, okay, add-a-new-part cannot be the final step in any putative evolutionary pathway by which that IC system is supposed to have been created.
If the final step in the evolutionary pathway to System X is remove-an-existing-part, the immediate evolutionary predecessor to System X must be System-X-plus-one-part. Problem for Behe: Since when does adding a part to an IC system make that IC system stop working? In some cases, sure, adding an extra part could screw up the rest of the system. Alas for Behe, his argument requires that an IC system must stop working in all cases where an extra part is added to an IC system. Good luck supporting that, Behe.
If the final step in the evolutionary pathway to System X is modify-an-existing-part, the immediate evolutionary predecessor to System X must be System-X-with-one-modified-part. Again, we’re talking about something which could cause an IC system to stop working in some cases—but Behe’s argument demands that it must cause an IC system to stop working in absolutely all cases.
Behe used to be a real scientist. What happened to him, that he can’t see a gaping hole of this magnitude in his reasoning?
By actual research, it appears that duplication is the most likely route for adding a part, and in most cases, duplication does no damage.
Joe seems to be following this thread on UD. He doesn’t seem to have noticed that the argument here is not about the source of variation — whether it is blind or directed — but whether there is a pathway.
With so many ID balls in the air it must be hard to keep track of them, but DennisJones is not talking about whether a designer exists, but whether the history of genomes could be incremental. Dennis’ argument would apply even if a designer reached into the genome and made each change.
From the UD thread:
I believe the Lensky experiment meets that criterion for irreducible complexity. Two necessary but unseleted mutations prior to the selectable mutation.
So the problem is that evolutionary theory is neither physics nor metaphysics? Therefore wrong?
This Academy of which you speak – is it like the Stonecutters in the Simpsons? I was never offered induction when I was at university.
Behe (in his CCC threshold argument) omits the contribution of recombination, which is highly significant in its effect on ‘exploring’ a fitness landscape and generating the apparently irreducible without the restriction of having to do it serially. Lenski’s setup also excludes recombination, but still demonstrates the ability of serial neutral changes to explore the landscape much more effectively than the strawman ‘Darwinists think every change was beneficial’ viewpoint. Put recombination on top of that and the dimensionality increases greatly, both in terms of the space accessible to a single genetic unit, and the separation of genetic units into quasi-independent elements that can each ‘explore’ a different space.
In the organisms generally wished to be divorced from an evolutionary scenario (US, DAMMIT!) recombination has been a major and continuous factor from eukaryogenesis onwards.
The failure to grasp the implications of multiple dimensions in evolution is a real stumbling block, I think.
It’s one of the reason IDists seem to think that it is so “obvious” that in silico models of evolution have “smuggled” something in – because we usually have a single criterion for fitness, and it stands out like a sore thumb. There are so many fitness functions in nature – so many ways in which things can vary that can affect their chances of breeding, IDists don’t seem to notice that they are there.
It’s a good example of how oversimplifying it can hinder understanding rather than help.
One of the reasons the AVIDA simulation was able to demonstrate that IC features could evolve (including IC pathways that went via quite steep reductions in fitness, is that AVIDA has several different criteria that contribute to fitness – there are a range of functions that a critter can perform, so even if one gets knocked out, there are still other ways it can survive
The irony is that IDists knock AVIDA because it is “too simple”. The point is that AVIDA is at a disadvantage compared to biology because it is so simple, but at least it is complicated enough to evolve some seriously IC functions by some seriously IC pathways.
Thanks for this post, Gregory.
Very illuminating.
Still, it’s very kind of you to summon up the condescension to communicate with hoi polloi every so often.
Me:
recombination has been a major and continuous factor from eukaryogenesis onwards.
And of course outside the eukaryotes as well, but with a different dynamic. One important feature of a recombinant sexual population is the networking of ‘solutions’. Separate beneficial alleles can integrate into future genomes, where, without such recombination, entire indivisible genomes are restricted by the weight of the net effect of all the different components of these composite evolutionary ‘alleles’. It’s one reason I don’t consider sex a mystery, though I’m told I should (must write that OP …).
Allan Miller,
Do you think a better term to use instead of “explore the landscape” might be to “link previously disconnected landscapes”.
I mean that each of the potentiating mutations identified in Lenski’s experiments were, in themselves, functionally unconnected (climbing their own local hills, or climbing no hills at all because neither conferred a selective advantage in the experimental environment).
But when they appeared together in the same genome, a new landscape was created, with a new hill (successful citrate metabolism).
If this is true, then the IDiots are barking up the wrong dimensions. New fitness landscapes will appear (with new dimensions) whenever any novel combination of genetic states arises. All that is required is that the new combination of gene states be expressed in the organism’s phenotype and that the new phenotype be more or less fit than the population average. Then natural selection will get to work to sort out the fit from the unfit.
As a by-blow, can anyone tell me whether high-dimensional “landscapes” are almost always mathematically “smooth” (so that, if dimensionality is high, there will always be a way to get to there from here). Intuitively, I imagine they are, but I don’t know.
Well, I wouldn’t want to be too pedantic (it is, after all, only a metaphor), but I think of the whole space as a ‘landscape’. The combinatorial permutations are (imaginarily) plotted as a 2 dimensional grid and some comparative metric takes the dimension of height. It then depends what your combinatorial strings are – proteins, or whole genomes. The vertical metric would then be selective advantage (relative fitness) or ‘expected’ offspring output (absolute fitness).
The citrate-metabolising E Coli genome already ‘existed’ in the space, it just hadn’t been visited by any actual organisms yet. Drift allows a genome to wander about the permutation space, either to be ‘captured’ by a peak, or for a peak to suddenly rise up nearby. Recombination (also involved in the potentiating mutations) allows longer-distance probing without having to go the long way round (if such a route is even permissible).
Gregory,
I must say, I really don’t get you’re behavior here. What are the coyness and taunting suppose to accomplish? If you have a point or perspective that you think is valid, what does hinting at it, but not stating it get you? Take this for instance:
A couple of things to note. First, if we “cannot” or “will not” admit anything is real other than what is natural, yet you are aware of things that are real and not natural, how does not providing an example and not explaining our misunderstanding help your claim? Won’t those of us who hold onto naturalistic reality continue to propagate such an inaccurate view throughout science? This can’t be what you want, yet I don’t see any effort on your part to correct this erroneous perspective. How can you expect any of us to adopt this rather radical point of view if folks like you won’t share it with us? That just doesn’t make any sense to me.
I think of “the landscape” as a list of genome strings. The Library of Mendel, as someone at AtBC says. The list of all possible genomes is a big list, as KF correctly argues.
Somewhat smaller, but still huge, is the list of viable genomes. I think of this as like a loosely woven fabric, mostly empty. Actually I think of it as a multidimensional sponge, mostly empty space, but still a single connected entity.
The trick for ID advocates is twofold. They need to demonstrate in the history of life that there are illigitimate children somewhere. Children whose genomes are not the plausible descendents of their parents. Children whose parentage includes gods or designers or the like.
And assuming the can do that, they need to show, at least in a thought experiment, how a designer would navigate the Library of Mendel.
Robin,
Gregory studies scientists like lab rats, and lab rats don’t ask questions.
Face it, none of us are smart enough to engage in dialogue with him. He comes here occasionally to open the cage and poke and prod. Just take it like a lab rat and squirm. Although he seems to take too much pleasure in his experiments, he has other labs to attend to, so it’ll be over soon.
I’m OK with probing questions. But I get the feeling after it goes on for several weeks that the prober is just searching for a gotcha. Perry Mason tactics. If anyone remembers Perry mason.
I think in the long run the best strategy is simply to make one’s case. If one’s case is that science (and biology in particular) don’t offer much light on moral and social issues, then so be it. Chocolate isn’t beefsteak either.
Joe still can’t read:
the Behe quote is:
Joe accepts that two necessary but unselected mutations occurred.
Behe was wrong about HIV also:
http://pandasthumb.org/archives/2007/10/an-open-letter-3.html
http://endogenousretrovirus.blogspot.com/2007/08/michael-behe-please-allow-me-to.html
http://scienceblogs.com/erv/2010/05/18/hiv-1-and-vpu-behe-is-still-re/
etc.
Pardon me for editing and adding stuff to this post. I just think it is clearer than adding more posts.
Perhaps he’ll change it.
KF has re-published an Eric Anderson essay and sanctioned it as an officially approved UD definition of ID.
In conclusion it reads:
Without resorting to jargon, and trying to avoid pejorative language, it appears to me that ID rests on asserting that there are genomic sequences that could not have been reached via evolutionary pathways.
I reach this conclusion by assuming that reducible complexity — that which has an evolutionary pathway — is not applicable to the calculation of probability.
If this is in fact the core of the ID argument, then ID rests on the attempt to prove a negative.
Regarding the Lensky experiment, Joe responds:
We know that they did not enable any increase in colony growth. that’s the definition of selection. As for whether they were directed, we know that they were among many neutral mutations. Enough to sample all possible point mutations.
What kind of director buys all the lottery tickets?
What kind of director wastes his time making sure E Coli can grow on citrate? Not just making sure the mutations happen, but making sure those organisms get sucked into the new-culture pipette! 😯
This exposes the desperation of the ‘directed’ argument. G-d intereferes in every nook and cranny unless you can show he doesn’t – even the things that are supposedly completely non-controversial: ‘microevolution’. Oh, and we humans have free will. The incompatibility of these two positions will not register, but G-d can’t have a genetic plan and let his organisms make their own marital choices (or any antecedent ones that deflect from the Path).
Since this site seems to be experiencing technical difficulties, I’ll just post a link to the UD thread.
Nightlight seems to claim that mutation is nonrandom. He is certainly welcome to come here and make his case.
I would like to see a separate discussion on the randomness of mutation.
More and more ID advocates seem to be acknowledging the historical fact of common descent, and the fallback position seems to be that mutations occur as a result of some deep underlying intelligence. It seems almost mystical.
Who can address this? I know that a lot of research has been done, but I really don’t have the knowledge needed to summarize it.
petrushka,
Having the mutations be guided by a “deep underlying intelligence” would seem to be contradicted by the fact that a great many more of them are deleterious than are advantageous. Why would this intelligence go around breaking things?
On the other hand that is exactly what we would expect if the mutations were random. For any reasonably well-adapted system, there are a lot more ways to damage the system than there are to improve it. So random changes would be expected to be mostly deleterious.
(Incidentally, the fact that many more mutations are deleterious than are advantageous does not mean that the system will deteriorate over time. Creationists like to say that, particularly YECs who have religious reasons for believing that everything is falling apart since Adam and Eve. They just have not done the population-genetic calculations to see what the probability of fixation of advantageous and deleterious mutations are.)
I understand that argument, but are there any overviews of the statistics accessible online? Ones that might be understandable to a layman.
My own take is that if it were possible to anticipate need and function, evolution would find a way to bias mutation. Maybe it has, as with the various high level mutations.
But my question to the intelligent mutation crowd is how do you anticipate emergent properties, and how do you respond to competition, predation, environmental change.. And how do you balance competing needs?
The thankfully banned JoeG claims to have evidence that at least some mutations are directed; and he further claims to have found it in Spetner’s book. I think that there’s nothing in there that could be dignified with the name of “evidence”, but perhaps some fan of Spetner could drop in and explain? Perhaps anyone holding the view that mutations are directed could come and defend it?
(Not holding my breath)
I notice there is a somewhat amusing and refreshing argument going on at UD. Nightlight is asserting that the God of the Gaps argument is a loser because the gaps keep getting smaller. He asserts some sort of mystical cosmic intelligence that can direct mutations. I would like to see him come here and explain this to us.
It’s not fine tuning; it’s more like continuous creation. To me it seems not unlike Newtons angels that continuously tweak the orbits of planets.
KF doesn’t like this because his argument depends on unbridgeable gaps. He does not seem to buy into the notion that even directed mutations can bridge between the isolated islands.
If Joe is watching, perhaps he’ll relay this question back to UD.
It might be significant that Nightlight complained about having posts in moderation.
Petrushka
As in ‘aimlessness’, not equiprobability, presumably?
The problems start right at bottom level. The mutation rate observed is in the ballpark of the error rate of DNA polymerase in vitro, which seems to suggest that its stochastic basis is simply the slight uncertainties and spontaneous corruptions at the molecular level. Unless someone is guiding those test tube errors as well … ?
DNA polymerase has no access to phenotype. It has no way of telling whether a base is part of a promoter, a coding region, intron, etc. It just copies. So even without the problem of access to the environment which will judge the mutation’s phenotypic effects, the process can’t even access the expression level through which those effects are implemented.
The rate of mutation does not positively correlate to regions whose sequence matters, which one might expect if they were ‘directed’. Of course, ID-ers think every bit of sequence probably matters, but there is still a curious pattern to explain – why does ‘direction’ concentrate mutations in regions of apparent neutrality? If it’s not really neutral, how come it is so mutatable? Do we have two mechanisms, the ‘natural’ one explaining pattern and the indistinguishable ‘directed’ one explaining adaptation?
Additionally, transitions (like-for like substitution) outnumber transversions (switch from puirine to pyrimidine or vice versa) by about 2 to 1. There is an obvious mechanistic reason for this – it’s easier to make a mistake when the units resemble each other than when they differ more. There is no reason why ‘directed’ mutations should show this bias.
So all-in-all, directed mutation seems to explain nothing of the mutation pattern. And let’s not get started on cancer …
I’d like to see NightLight come here and defend his thesis.
NightLight is, in my opinion, cleverly pushing the interventionist argument that lies hidden under the ID conceit to the extreme of absurdity. Who knows the limits to the Master Manufacturer’s wiles? Why assume that there are any limits?
petrushka,
Apparently this is a position taken from Thomas Nagel’s Mind and Cosmos, where he builds a (non-religious) case that there is some deep teleology implicit in reality and especially biology.
My reading is, Nagel (a philospher, not a scientists) is simply rewording the argument from incredulity. He simply can’t believe the evolutionary process as he understands it could produce such marvels as we see, so evolution must be running down channelized streams.
Nagel, on page 16, writes: “My guiding conviction is that mind is not just an afterthought or an accident or an add-on, but a basic aspect of nature.”
So, no, Nagel is not making a case for teleology. He is assuming it up front as an initial assumption, that the purpose of the universe is to produce minds similar to those of humans.
If one begins with that assumption, then of course evolution will seem implausible unless mutations are guided toward meeting that purpose.
Presumably it is an ego thing. Nagel doesn’t like the idea that he might be an accidental result of contingencies.
One of those ideas that occupied late, dateless nights in the dorm.
Neil Rickert,
I know a number of people of whom I would much rather think that they were the accidental result of contingencies.
The idea that they have been designed to be as they are involves there being a pretty nasty Designer, who/which should be ejected from the universe, before more damage is done.
“DennisJones at UncommonDescent has published an argument that Behe’s irreducible complexity has resisted all attempts at refutation.”
He must be unaware of the numerous refutations posted here at TSZ.
Yes, it would seem so.