More Recent Comments

Tuesday, December 11, 2012

On the Evolution of Complexity

Can you go from some simple character to a more complex feature without invoking natural selection? Yes, you can. Complex features can evolve by nonadaptive means. Just think of our complex genome and read The Origins of Genome Architecture by Michael Lynch.

Want a more simple example? Read the latest post by PZ Myers: [αEP: Complexity is not usually the product of selection]1.

This is an important point. You can't just assume, without question, that a complex trait must be an adaptation and must have arisen by natural selection. That applies to molecular complexes and also to complex behavior.


1. See The Evolution of Enzymes from Promiscuous Precursors for supplementary information.

68 comments :

Unknown said...

I understand this, but nor can you assume that it happened by chance. Isn't this a direct contradiction to what Dawkins argues in The Blind Watchmaker?

Anonymous said...

No. What Dawkins argues in TBW is that biological complexity can arise from natural processes.

Unknown said...

But Dawkins is, in fact, and adaptationist. So he does take the view that selection is the more important process. He's also a proponent of evolutionary psychology, which is where this discussion began.

Unknown said...

"It is precisely the inability of selection to cull all variants that deviate from the successful, well-tested norm that allows novel complexity to emerge." -- Myers.

Doesn't this actually make selection the more relevant process when considering outcome?

Also, this:

"One of the most important and controversial issues in population genetics is concerned with the relative importance of genetic drift and natural selection in determining evolutionary change. The key question at stake is whether the immense genetic variety which is observable in populations of all species is inconsequential to survival and reproduction (ie. is neutral), in which case drift will be the main determinant, or whether most gene substitutions do affect fitness, in which case natural selection is the main driving force. The arguments over this issue have been intense during the past half- century and are little nearer resolution though some would say that the drift case has become progressively stronger. Drift by its very nature cannot be positively demonstrated. To do this it would be necessary to show that selection has definitely NOT operated, which is impossible. Much indirect evidence has been obtained, however, which purports to favour the drift position. Firstly, and in many ways most persuasively is the molecular and biochemical evidence..." (Harrison, G.A., Tanner, J.M., Pilbeam, D.R. and Baker, P.T. in Human Biology 3rd ed. Oxford University Press 1988 pp 214-215)

So the more relevant thing to study seems to be selection, which is what EP does. There's no denial of genetic drift going on. Moreover, I have not seen proof of the assertion that genetic drift is the more important force. In fact, it's selection (including whether it occurs or not) that appears to be the driver.

So if we want to study causality, we must look to selection. The rest is random (and interesting), though not directly relevant to the issue of whether selection occurred, or whether a particular behavior is an adaptation. Chance is always a given. But when you're dismissing natural selection as unimportant, you're basically dismissing the only thing that can provide an explanation (regardless of whether it does or not).

Saying that something is random is not an explanation, and it's usually my default assumption. But twin studies, for example, clearly demonstrate that much of our behavior is not learned, and thus it seems appropriate to examine whether a particular behavior is the product of selection -- or not.

I understand the point that selection is not necessary for complexity, so that's a given.

steve oberski said...

I think PZ Myers addresses this in his summary:

evolution relies on neutral or nearly neutral precursor events to produce greater functional complexity

If I understood the article correctly, neutral events serve to increase the pool of genetic diversity for selection to act on.

He also seems to say that selection acts to decrease genetic diversity.

So the two processes have to act in tandem in order to generate complexity.

Unknown said...

There are three processes at play, actually -- random mutation, natural selection (on those mutations) and genetic drift (then selection or mutation). EP is focuses on the study of whether selection occurred in respect to certain behaviors. Genetic drift is simply not the focus of EP.

Also, I think it's a bit disingenuous to confuse the type of complexity that PZ is referring to (basically meaning the existence of gene families) with the complexity of the human brain. The quantitative difference is tremendous, and by PZ's own gambling analogy in his prior post, it would be *more probable* to assume that selection occurred at some point.

David Pinsof said...

This all depends on how you define "complexity." Dawkins takes great pains to define this word for biologists in the first chapter of The Blind Watchmaker. It is not enough for a feature of an organism to merely be heterogeneous or have a lot interacting parts -- drift can easily produce that. What makes a feature adaptively complex is that these different parts are improbably well-organized for carrying out a specific function that aids the organism in survival or reproduction. In other words, there is a coordination between the phenotypic properties of the organism and the structure of the environment such that when they interacted they reliably meshed to promote fitness. I concur with Dawkins that this kind of complexity is best explained, and perhaps only explainable in terms of, natural selection.

steve oberski said...

Saying that complex traits can arise by non adaptive means is not the same as saying that they never arise by adaptive means.

I wasn't event taking EP into consideration in my comment but I don't think I would single out the human brain for preferential treatment in the discussion of complex systems.

Consider the fact that much of our brain and associated subsystems such as audio, visual, olfactory, motor control etc. can be traced back to less complex ancestors, and our current human brain is a cobbled together rube golbergesque, onion like construct with a recent primate neocortex on top of the older mammalian limbic brain with an ancient reptilian core.

At what point in the evolution of our brain would your tremendous quantitative difference become a factor ?

And I'm puzzled as to why the importance of the various evolutionary processes is of such concern, they all seem to play an important role, the system would not work if any of them were missing.

Unknown said...

I concur with you, Wilkins, and Dawkins. The beginning of a proper argument begins with the definition of terms. The fact that the parties to this debate are using the same terms to mean different things makes it hard to follow. Myers, especially, seems to rely on that as a diversion tactic. It is highly improbable, almost impossible, that the human brain evolved solely by genetic drift, which makes Myers' whole post irrelevant. And the prior one, too.

Unknown said...

Because some people are implying that to focus on the study of selection is wrong, I suppose. I agree with you completely, otherwise. And I'm certainly not saying that the study of genetic drift is useless.

Sergio A. Muñoz-Gómez said...

Complex functional features that are adaptive have complex histories. Usually both drift and positive selection (and of course purifying selection) are involved.

PZ Myers said...

Weird. So your take on the story is that I'm arguing that the brain evolved entirely by drift?

Huh. OK. If that's your interpretation, I know what your opinion is worth.

David Pinsof said...

It's actually a reasonable interpretation of your post, because the only way evolutionary psychology could be useless is if the brain evolved solely by genetic drift. If there are any adaptations at all housed within the human brain, then adaptationist (i.e. evolutionary psychological) investigations would be the best way to understand them.

Anonymous said...

simpleminded

Yet, there is one well-known contemporary philosopher(in his own words, not a creationist) for whom even the genetic drift and selection aren't sufficient to account for biological complexity:

"no viable account, even a purely speculative one, seems to be available of how a system as staggeringly functionally complex and information-rich as a self-reproducing cell, controlled by DNA, RNA, or some predecessor" (T.Nagel. Mind and Coscmos, 2012)

So he calls for... uh... teleology to justify the emergence of this complex systems:

'The teleological hypothesis is that these things may be determined not merely by value-free chemistry and physics but also by something else, namely a cosmic predisposition to the formation of life, consciousness, and the value that is inseparable from them.' (ibid.)

He modestly describes his vision as 'a heroic triumph of ideological theory over common sense'
But I think he would agree with your criticism of EP:

'...in order to understand our questions and judgments about values and reasons realistically, we must reject the idea that they result from the operation of faculties that have been formed from scratch by chance plus natural selection, or that are incidental side effects of natural selection, or are products of genetic drift' (ibid.)

Unknown said...

I'd also like to point out that "complex trait" is not the same thing as "complex organ." So the attack on Wilkins is a strawperson.

Anonymous said...

simpleminded

Sorry for the cropped first quote. The full quote is:
'no viable account, even a purely speculative one, seems to be available of how a system as staggeringly functionally complex and information-rich as a self-reproducing cell, controlled by DNA, RNA, or some predecessor, could have arisen by chemical evolution alone from a dead environment'

Shawn said...

'no viable account, even a purely speculative one, seems to be available of how a system as staggeringly functionally complex and information-rich as a self-reproducing cell, controlled by DNA, RNA, or some predecessor, could have arisen by chemical evolution alone from a dead environment'

...its not that difficult to imagine. But it is the jump to complex cells and even humans etc (and leapfrogging over the principle and concept of chemical evolution) that causes people to become confounded,and to default to teleology and then to gods.

Robert Byers said...

Surprising to hear this.
Creationists point out the unlikelyness of complexity from selection and BANG here its said.
HMMM.
Is it really possible to make very complicated things anew from selection on existing things or mutations from things??
It seems impossible.
Their is a barrier of complicatedness after all.
Its only a special case that selection, might, rearrange butterfly populations in the Amazon.
Butterfly into butterfly but not into armadillo.
Who thought this up??

Shawn said...

Butterfly into butterfly but not into armadillo.

no god could be so powerful

Piotr Gąsiorowski said...

Butterfly into butterfly but not into armadillo.

That's correct. Armadillos did not evolve from butterflies.

Mikkel Rumraket Rasmussen said...

Therefore he's postulating his own, purely speculative and fantastical "cosmically predispositioned" one.

God is given the weirdest names by these crackpots.

By the way, the blanket statement that there are no viable accounts, EVEN PURELY SPECULATIVE ones, about the origin of life, is an outright lie. There are plenty of speculative accounts, the problem is, of course, that they are speculative.

The Other Jim said...

Maria Maltseva wrote;
"Because some people are implying that to focus on the study of selection is wrong,"

I think what is being said is that "One cannot assume adaptation is the only mechanism leading to this complexity". Not what you wrote.

Have you seen this video linked to in this post? The point here is about the question. Asking "How did trait X arise?" is different than asking "What adaptive process lead to trait X".


The Other Jim said...

@ David Pinsof,

You say, "It's actually a reasonable interpretation of your post,"

I have to disagree with your interpretation. His last paragraph is this;

Even in something as specific as the physiological function of a biochemical pathway, adaptation isn’t the complete answer, and evolution relies on neutral or nearly neutral precursor events to produce greater functional complexity."

(Emphasis mine).

@ Maria Matlseva,

"I'd also like to point out that "complex trait" is not the same thing as "complex organ." "

You are walking into the old "How did a bird wing evolve, then" trap. In a purely adaptive scenario, how can you explain the transition of a theropod hunter's arm to fully functional wing? Nearly-neutral intermediates, again...

Piotr Gąsiorowski said...

Nearly-neutral intermediates, again...

And a lot of random mosaic combinations of variable traits in different theropod lineages, with one lucky combination (small body size, light skeleton, long forelimb feathers etc.) becoming the "preadaptation" for flight and enabling natural selection to take over as the driving force.

steve oberski said...

Even a stopped clock, monkeys pounding away at keyboards, and Robert Byers.

Anonymous said...

Thomas Nagel also admits in his book that he's getting his knowledge of science from popular science writings - he doesn't have a scientist's understanding of biology.

Anonymous said...

simpleminded

even worse, he likes to read creationists' books, uses ID like 'arguments' and sincerely believes that
'Even if one is not drawn to the alternative of an explanation by the actions of a designer, the problems that these iconoclasts pose for the orthodox scientific consensus should be taken seriously. They do not deserve the scorn with which they are commonly met. It is manifestly unfair' (p.10)

therefore it is not strange that, for him, complexity is not explainable in terms of the chance and selection.

Unknown said...

I agree that the first question should be, "Is the behavior in question adaptive (or is the lack of it adaptive)?" But wouldn't EP be the right place to ask the question? Then we need to establish whether it's heritable, etc. I'm not sure that EP is the field that needs to tie it to particular genes or neural processes, however.

We know, in a very general sense, that some things about the brain are, indeed, adaptive. How can you figure out which ones those are without asking the questions you're curious about?

Unknown said...

I don't think that PZ is arguing that. What he's trying to do is diss EP because not all "complex" traits are adaptive. But modern EP has a reasonable standard for determining if they are. Also, in an earlier post, PZ (perhaps jokingly?) proposes a bright-line test for telling good EP from bad EP -- 90% of EP is bad. If the hypothesis has to do with gender or race, it's bad -- so I don't understand the purpose of his post. His criticism of Wilkins is unwarranted; I'm sure that Wilkins was referring to complexity in the same sense Dawkins was in TBW, and saying that the EP/SB is the right place to ask related questions.

Unknown said...

Sorry for the typos. I tend to send w/out proofing...

David Pinsof said...

@ The Other Jim,

Again, we're back in semantics. This depends on what you (or PZ) mean by "the complete answer." Evolutionary psychologists are interested in: 1) Is this aspect of our psychology an adaptation, a byproduct, or noise? 2) If it is an adaptation, what is it's function? It seems to me that you can infer via experimentation that the function of fear is danger avoidance without doing an analysis of all the neutral precursors that had to be present in order for the fear system to evolve. Though the neutral precursors are interesting and worth studying, they are not the primary objects of study for evolutionary psychologists or adaptationists in general.

Unknown said...

Right, I've seen no reasonable argument to suggest why the field should focus on the null hypothesis, or why it's wrong to explore a known alternative possibility. People should never be afraid to wonder, or to ask questions of those with more relevant experience. (Which is, BTW, all I'm doing here. I'm not a scientist, and I'm not giving a talk on the subject of EP or planning on it. Ever.)

Chas Peterson said...

"You can't just assume, without question, that a complex trait must be an adaptation and must have arisen by natural selection. That applies to molecular complexes and also to complex behavior."

Because they're the same thing, and you say so. FFS.

Q: Is there a single behavior, complex or not, in any animal, that has ever been shown, with positive evidence, to have evolved by random genetic drift? Even one?
A: no.

Pedro said...

"EP is focuses on the study of whether selection occurred in respect to certain behaviors. Genetic drift is simply not the focus of EP."

I disagree. You seem to be implying that evolution amounts to selection. EP studies the evolutionary basis of behaviour. That evolutionatry basis, if it is present and is not the product of factors like culture, etc, can either be adaptive or not (genetic drift). But even if it not adaptive it still can have a biological basis. So EP is both about selection and drift. The problem is that most EP papers seem to ignore drift and everything is assumed to be selectice even when that cannot be clearly established (let alone when it can be explained by non-evolutionary factors).

Pedro said...

"Though the neutral precursors are interesting and worth studying, they are not the primary objects of study for evolutionary psychologists or adaptationists in general."

Adaptationists are people who think everything is a result of adaptation. *Evolutionary* Psychology should be about the *evolutionary* basis of behaviour, regardless of said evolutionary basis (if established) being adaptive or not.

If that's not what EP scientists are doing, then they should.

Pedro said...

Thomas Nagel doesn't understand science in any depth, but apparently it doesn't stop him from stating things like this:

"Even if one is not drawn to the alternative of an explanation by the actions of a designer, the problems that these iconoclasts pose for the orthodox scientific consensus should be taken seriously. They do not deserve the scorn with which they are commonly met. It is manifestly unfair."


Unknown said...

I guess I have to disagree on that point. EP studies focus on adaptation; other studies focus on drift. But drift is the null hypothesis. It is the fundamental assumption. EP dares to posit that there's another explanation, and tries to weed out adaptations from byproducts and noise(drift). That's the goald of EP, which again, makes Myers' recent post irrelevant and misleading. We do know that at least some parts of the brain and some behaviors are adaptive (as PZ admits), and that's all we need to know to attempt to figure out which ones they are. How well we do so is another story.

Unknown said...

goald = goal. Ugh.

Unknown said...

No, adaptationists (as Dawkins was characterized by Moran in prior writings), believe that natural selection is the driving force causing "complexity," AS DEFINED BY DAWKINS. They are by no means denying drift, as has been said repeatedly throughout this thread. Nor are they denying the necessity of mutation/drift to provide novelty.

Unknown said...

This. A million times, this.

Pedro said...

The point is not that adaptationists accept genetic drift, the problem is the general idea that virtually every phenotypical characteristic is the result of adaptation.

Pedro said...

Then "Evolutionary Psychology" should change itd name to "Adaptive Psychology".

Pedro said...

Has anyone read "The Origins of Evolutionary Innovations: A Theory of Transformative Change in Living Systems", by Andreas Wagner? Any opinions? How does it compare to Lynch's "The Origins of Genome Architecture"?

Thanks!

David Pinsof said...

No one, I repeat no one, believes that every phenotypic characteristic is an adaptation. That is a blatant, preposterous, unproductive straw man. If you disagree, I challenge you to find me one unambiguous quote from an evolutionary psychologist that says that everything is adaptive. Evolutionary psychologists seek to TEST, using the scientific method, whether traits show evidence of adaptation. They also test, using the scientific method, predictions derived from byproduct or spandrel theories of the trait in question. Different models are pitted against one another, and, as in the rest of science, the one that receives the most empirical support survives. A priori arguments about the importance of genetic drift do not tell you whether fear, humor, jealousy, anger, and morality are adaptations or spandrels or noise. Neither can evidence of their genetic basis. Only careful empirical investigation informed by evolutionary theory can tell you what function, if any, these psychological traits serve.

Larry Moran said...

paraphrasing David Pinsof ...

A priori arguments about the importance of natural selection do not tell you whether fear, humor, jealousy, anger, and morality are adaptations or spandrels or noise.

Shawn said...

Are you rejecting the assertion that evolution by random genetic drift is the null hypothesis?

Anonymous said...


@ Chas Peterson,

Oh goody! Someone has finally explained the adaptive advantage of gulls trying to flip green, oversized eggs back into their nest, instead of their smaller, brown spotty ones!(Baerends and Kruijt, 1973) Please share...

Larry Moran said...

Maria Maltseva writes,

I've seen no reasonable argument to suggest why the field should focus on the null hypothesis

Some evolutionary psychologists have told me that everyone in the field has read and understood the spandrels paper [see: A Critique of the Adaptationist Programme]. If that's true then either they all understand and adhere to the concept of the null hypothesis, or they have really good arguments for why it should be rejected. I'm not familiar with the arguments against the null hypothesis but perhaps I've just missed those papers in the evolutionary psychology literature. (The other possibility is that many evolutionary psychologists have NOT read and understood the spandrels paper but I'm told that this is not true.)

Here's some more discussion about the importance of the null hypothesis.

The Importance of the Null Hypothesis

Michael Lynch on Adaptationism

What's the Difference Between a Human and Chimpanzee?

The "Null Hypothesis" in Evolution

How to Frame a Null Hypothesis

SLC said...

But Booby Byers is descended from an ape ancestor that lived 6 mullion years ago. Judging by his moronic comments, he doesn't seem to have progressed much beyond that ancestor.

Anonymous said...

@ David Pinsof

Again, we're back in semantics.

I have to admit that I give up on people who scream "semantics" whenever a specific point, well described in other aspects of evolutionary study, are pointed out to be merely "semantics". Drift happens, and s << 1/Ne matters.

The Other Jim said...

Oops... Post above was me.

The Other Jim said...

@ Maria Maltseva,
I don't think that PZ is arguing that.

I'm just responding to that post alone, and not any other comments. That is the point, and it is a very important one.

Pépé said...

Larry is the product of the four fundamental forces of physic (the strong and weak nuclear, the electromagnetic and the gravity forces) and we all can see the result!

PS: I do hope the $1000 wager about junk DNA still holds!

David Pinsof said...

I'm actually curious: how, specifically, do you believe evo psych should be conducted differently? Say I'm interested in whether or not fear is an adaptation for danger avoidance. How would a gouldian pluralist proceed in answering this question? Would she test to see if fear increases heart rate, elevates blood pressure, tenses muscles, widens eyes, dilates pupils, focuses attention, prompts fighting/fleeing, etc.? Would she test to see if the inputs to the fear system correspond to the types of things that posed survival threat in the EEA? This is the adaptationist approach, and it seems to me these types of experiments would go a long way in answering whether fear is an adaptation. If it turned out that all of these experiments yielded null results, we would have to conclude that fear is not an adaptation for danger avoidance. In what way(s) would the "pluralist" approach differ from this type of work?

Unknown said...

Since I've lost track of all the threaded responses, I just wanted to say thanks to Larry for the links on the null hypothesis, for getting me to read the spandrels paper, and for hosting some of the best discussions on this topic. So -- whether I ultimately agree with his conclusion or not -- Larry, thank you so much!

Unknown said...

Oh, were those available in their EEA? Didn't know.

Unknown said...

No, what I think Chas is saying is that you can't show positive proof that selection did not occur.

Unknown said...

Previous comment is a response to Shawn. The one before, to Anon.

Larry Moran said...

Yes, I'm more than willing to take your money. We have to agree on how what criteria you will accept in order to concede that at lest 51% of our genome has no function.

Larry Moran said...

David Pinsof asks,

Say I'm interested in whether or not fear is an adaptation for danger avoidance. How would a gouldian pluralist proceed in answering this question?

I would want to know which fears were instinctive and which ones were learned behaviors. I would want to know which fears were universal—is it true, for example, that everyone is afraid of spiders, heights, and snakes?

I would want to learn as much as possible about irrational fears and what they could teach us about the genetic basis of fears. Is there a specific set of alleles that control fear of danger? Or is this an epiphenomenon related to intelligence? Was there ever a time in hominid history when most individuals didn't have such fear and then a mutation occurred that rapidly became fixed? If so, is there any evidence of such alleles?

I would be interested in finding out about the fears of our closest relatives, chimps and bonobos.

Unknown said...

Would a rose by any other name smell as sweet? I don't think a name change could hurt, to be honest. But it is a growing & improving field that overlaps with many other recognized (noncontroversial) sciences.

That said, I've never been a huge fan of the social sciences; I find neurology, biology, chemistry, and physics far more interesting.

Shawn said...

Pépé, you are being conned by a grifter...demand that it is your genome that is examined.

Anonymous said...

@Maria Maltseva

"Oh, were those available in their EEA? Didn't know."

That is an assumption with no evidence. There are no green eggs within the gulls, so probably not. Next attempt?

Pedro said...

I have to admit I kind of like Byers. His views are baseless, sure, and he can't understand why his views non-evolution and a young earth age are wrong, but at least he is *honest*. It's just that he doesn't get it and puts all his eggs on the faith basket.

Now contrast that with the local idiot Pépé.

Pedro said...

That's the whole point of these forums, I guess. We don't have to agree on everything, we just discuss things and try to learn something in the process from each other.

Unknown said...

Since no one said there's an adaptive advantage to gulls trying to flip green eggs, who exactly are you arguing with? You can't positively prove that selection did not occur; therefore, you can't conclusively assume that something occurred by random processes -- even if that's what happens most of the time.

I'm not sure how your gull example is even relevant to what the rest of us are discussing. It's interesting, but it proves nothing.

Unknown said...

If there were only some way to completely shed one's own confirmation bias...

Anonymous said...

"Since no one said there's an adaptive advantage to gulls trying to flip green eggs, who exactly are you arguing with?"

I am asking Chas to defend his position that no behaviors are the result of drift. If they are not the result of adaptive selection, what are they the result of?

I thought that you were invoking "must have been selected for in their EEA". I may have mistaken the point of you comment above.

"You can't positively prove that selection did not occur;"

Yes you can. If you know the allele(s) involved, based on the molecular evolution of the surrounding sequence. You can also rule it out is you can measure the coefficient of selection for the trait, and if you know the effective population size of the population in question. If both of these values are small, the allele is indistinguishable from a neutral allele, so would have fixed via a stochastic process.