Cornelius Hunter seems very confused.
…This brings us back to the UC Berkeley “Understanding Evolution” website. It abuses science in its utterly unfounded claim that “natural selection can produce amazing adaptations.”
In fact natural selection, even at its best, does not “produce” anything. Natural selection does not and cannot influence the construction of any adaptations, amazing or not. If a mutation occurs which improves differential reproduction, then it propagates into future generations. Natural selection is simply the name given to that process. It selects for survival that which already exists. Natural selection has no role in the mutation event. It does not induce mutations, helpful or otherwise, to occur. According to evolutionary theory every single mutation, leading to every single species, is a random event with respect to need.
He has forgotten what “adaptation” means. Of course he is correct that “Natural selection is simply the name given to [differential reproduction]”. And that (as far as we know), “every single mutation …is a random event with respect to need”.
And “adaptation” is the name we give to variants that are preferentially reproduced. So while he would be correct to say that “natural selection” is NOT the name we give to “mutation” (duh); it IS the name we give to the very process that SELECTS those mutations that promote reproduction. i.e. the process that produces adaptation.
Cornelius should spend more time at the Understanding Evolution website.
ETA: CharlieM points out below that…
When CH says that natural selection does not produce adaptations he is talking about individual organisms. He is discussing mutations in individuals and adaptations in individuals. Natural selection has nothing to do with the first appearance of an adaptation in an individual.
And that of course is the confusion – I hadn’t seen just where Cornelius’s confusion lay. Because, of course, the term “adaptation” is a population-level concept. At the level of the individual, the equivalent would be “advantageous mutation”. And that makes the Understanding Evolution website absolutely correct.
keiths,
If a population can wander unrestrained through genetic space, the vast bulk of its productions will not be of the complex variety. The fact that, in such a population, a genetic series can be telescoped (it would have to be asexual for that to happen) does not mean that the ‘complex’ features have any special significance. They only have significance as adaptations – because we look at complex features and regard them as special and worth mentioning as distinct from the run of genetic wanderings. In ‘infinite world’, we would only know that complex features were worth mentioning because we had been to ‘finite world’.
Adaptations don’t need to be complex. Nonetheless, there is a theoretical possibility that complexity can be produced neutrally
With exponential population growth in bacteria (infinite resources), you just get a “J” curve. Vary the reproductive rate and you get a different “J” curve. Mix populations with different reproductive rates and you get a “J” curve for each. What’s the point?
From Allan’s link:
However, much of contemporary evolutionary genetics departs from the conception of evolution underlying neo-Darwinism, resulting in a widening gap between what formal models allow, and what the prevailing view of the causes of evolution suggests. In particular, a mutationist conception of evolution as a 2-step origin-fixation process has been a source of theoretical innovation for 40 years, appearing not only in the Neutral Theory, but also in recent breakthroughs in modeling adaptation (the “mutational landscape” model), and in practical software for sequence analysis. In this conception, mutation is not a source of raw materials, but an agent that introduces novelty, while selection is not an agent that shapes features, but a stochastic sieve.
Seems to me without some non-random element we are waiting for monkeys with typewriters to produce the complete works of Shakespeare. Given infinite resources and infinite time, it must happen.
Still not seeing how drift contributes to innovation. Still reading up on it.
Alan Fox,
… !
Allows non-adaptive genes to reach appreciable frequency in a population. From these, new parts of genetic space not reachable in 1 step can be attained. Further, with greater standing variation at significant frequencies, the population does not have to wait for mutation to occur before it can respond to an environmental change – drifting alleles can become selected.
No, I see that: I really do. But whilst I see the mission statement element, what I am missing is the methodology. It’s probably my fault. 🙁
Alan Fox,
🙂
It’s because both the drift and selection parts of evolution are really sampling, not exploration or progress or anything. Offspring sample the variation available in the parents, and variable copy numbers are made. Drift is sample error – the ‘random’ departure of each new population from frequencies in its predecessor – and selection is the biased element of that departure.
OM’s toy shows what happens when you run the sampling process with selection and mutation both turned off. One colour fixes. But if you were to add mutation, you would see tiny flashes of new colour. It might disappear immediately, or start to progress a little and then go backwards, but the new mutant has exactly the same chance as everyone else to dominate the future population through drift, so every N mutations will become fixed, not lost, in a population of N.
The reality is dynamic. Mutations are added all the time, and variation is being eliminated all the time. So if you were to start the simulation with a single colour, and allow mutation, new colour flashes would start to appear, and not all would disappear. After a while there would be a dynamic equilibrium – a mosaic of colour, constantly shifting as new ones arrive and old ones die out, but with a limit to the number of colours in existence and their ranges of frequency at any time. This pattern is standing variation, generated by drift.
Allan,
Sure, but complex features will be produced. William is right about that.
Why do you say that asexuality would be required?
\
I think William’s point was that complex features — features that would be definitely be regarded as adaptations if we encountered them in “real life” — could be produced even in the absence of natural selection in the no-death scenario.
He’s right about that, but the conclusion he derives from it is faulty. The fact that variation alone can produce complex features in one highly unrealistic scenario does not prove that it can do so in realistic ones.
Alan,
If the monkey population increases exponentially, it can happen a lot faster than you might expect.
keiths,
Well, it rather depends on the ‘eliminating-selection’ scenario you allow for. If it is necessary (as it is in a ‘real’ sexual population) that zygotically and meiotically compatible gametes be produced, then you can’t just produce something that required N mutations in ‘finite world’ in N generations. A lineage can’t depart rapidly from type and remain reproductively compatible.
But of course this is a form of selective constraint. The only way you can ‘really’ get rid of selection in a sexual population is essentially to get rid of phenotype – to regard every organism as a genotype, and every possible genotype as viable, and every genotype as compatible with any other. Some of your organisms may build a phenotype, but it can’t take part in sex, otherwise you have introduced compatibility issues – ie selection.
(This bug in WordPress, that it fails to include nested quotes in the Quote in Reply, is really annoying).
Yes, sure, among the genotypes produced, sooner or later, will be one that combines all the mutants you need.
It will present at a nearly infinitesimal frequency. Getting it to take over the population, and the probability that it will not be lost first, is a whole other topic.
For example, with a 500-base stretch of DNA, with symmetrical mutation, there are 4^500 or about 10^150 possible haploid genotypes. People will be familiar with this number from William Dembski’s calculation of “probabilistic resources” — it is approximately the number of individual events that happened since the Big Bang to all the particles in the universe. An individual event might be a change of orbital of an electron in a helium atom on Alpha Centauri ten million years ago.
If each such occurrence were instead a DNA sequence of 150 bases, one of them, somewhere in the universe, would be the 150-base sequence you seek.
With unlimited resources, one could play all.possible chess games in a few minutes. Someone tell me why this claim is wrong.
Joe,
William wasn’t arguing that any particular complex feature would take over the population. He was simply pointing out that complex features would be produced despite the absence of natural selection, which is correct.
The thought experiment is fine. It’s William’s post-experiment logic that is faulty. He is arguing thus:
1. Variation by itself, with no natural selection (in the Darwin/Mayr/keiths sense), will produce/build complex features.
2. Natural selection by itself, with no variation, will not produce/build complex features.
3. In these two scenarios it is variation alone that produces/builds complex features.
4. Therefore variation alone produces/builds complex features in the real world.
Steps 1-3 are fine, but step 4 is problematic. In the real world, we won’t get complex features in evolutionary time without both variation and selection. Both are involved in producing/building complex features.
William tries to salvage his argument by relying on a constricted and inaccurate interpretation of the words “produce” and “build”, but he ends up shooting himself in the foot. His idiosyncratic definitions are so narrow that he inadvertently rules out variation itself as a producer/builder of complex features. By his definition variation only produces/builds the sequences, not the features themselves.
So neither variation nor natural selection produces complex features in the real world, by William’s silly definitions.
Allan:
keiths:
Allan:
You’re forgetting that nothing dies in William’s scenario. You’ll end up with reproductively compatible individuals by chance if nothing else.
There would be no need to play. You’d simply have a tree of moves that could be traversed move by move.
It reminds me a little of this: http://boingboing.net/2009/11/02/mechanical-computer.html
Alan Fox, to Joe:
keiths:
Alan:
keiths:
I hope you’ve learned your lesson, Alan. The pop gen definition is not the only scientifically valid definition of natural selection. Joe was mistaken, and it wasn’t “chutzpah” for me to challenge him on this point.
OMagain:
Or to put it slightly differently, complete traversal amounts to playing exhaustively.
William’s scenario can be viewed in a similar fashion.
In real-life chess playing, only a tiny portion of the tree is covered. Likewise with real-life evolution.
keiths,
If any individual has become reproductively incompatible with its parent lineage, and it remains obligately sexual, then there is selection – its pool of potential mates is reduced which limit its offspring compared to the ‘less-evolved’ parental group. And if it has to wait (say) a billion years for such a compatible mate to emerge and for them to locate each other, this is not really happening ‘quite quickly’.
Consider the possibility that my failure to learn your lessons may be a reflection on your teaching skills.
Allan,
Nothing dies in William’s scenario, so the parents of the incompatible individual are still around, along with their close relatives. Why do you think it would necessarily take a billion years for a reproductively compatible partner to be produced?
In any case, time is not essential to William’s point. He is simply arguing that variation by itself is capable of producing complex features, and he is right about that. It’s his follow-up reasoning that’s faulty.
Alan,
And yet again you try to blame your failures on others. It was your mistake, Alan. Take responsibility for it.
I’ve lost track. Which mistake are you referring to?
@ Joe
I find highlighting the text I want to copy then clicking “quote in reply” seems to work for me.
keiths,
So you get partners by inbreeding? That would slow things down, not speed them up. It’s also importing new assumptions. Why do offspring stay close to their parents? Why isn’t everything kept nicely stirred? Neither has to be the case, but they make a difference.
If there is no selection, all possible genotypes are being produced [eta – if there is no selection, there is no such thing as an impossible genotype]. This makes finding your compatible partner something of an uphill struggle, in all that mess. A billion was just a proxy for ‘a long time’
It’s your ‘quite quickly’ point that I’m arguing, specifically in relation to sexuals. ‘Infinite’ and ‘fast’ are two different ball games. Making things infinite slows them down – unless you aren’t limited by partner compatibility. If you want to get N mutations in N generations, but you’ve made a generation X times longer, you haven’t necessarily sped anything up.
keiths,
I did make a mistake and a big one! I didn’t realise you were drawing on an example of WJM for irony (have to admit I don’t tend to read much WJM now).
Yes, but I don’t think that was Joe’s issue.
If I highlight text containing a quote, then click “quote in reply”, the quoted text shows up as flat (no indentation for the enclosed quote, because no blockquote tags were included). When I run into that on my own replies, I usually add appropriate tags.
No, what’s faulty is your understanding of my point. I didn’t say that in the real world RM could produce complex features by itself; in fact, I stated exactly the opposite – that NS was essential in order for RM to generate adaptations in the real world. I reiterated several times that just because something is essential to the building of X, doesn’t mean that thing actually builds X.
All we disagree on, keiths, is whether or not it is appropriate to refer to NS as “producing” or “building” features without even referring to RM (at the site page in question and the glossary), when RM is doing all of the actual building.
That’s it. That’s all we disagree on – whether or not the terms “build” or “produce” are appropriate characterizations of the NS contribution to the production of adaptations.
Neil Rickert,
Sure, I do that too. What with the page bug as well…
Frustrating!
William:
William,
You’ve thrown the baby out with the bath water. By your unreasonably narrow definitions “build” and “produce”, natural selection doesn’t produce complex features — but neither does variation!
Variation merely produces new sequences. It doesn’t produce (in the William sense) the features that derive from those sequences.
If you stick to your definitions, then consistency requires you to say that neither NS nor variation “produces” or “builds” complex features. That’s goofy.
But if you drop your restrictive definitions and use English as she is spoke, then NS and variation together produce complex features, which is what we’ve been telling you all along.
Allan,
We could argue about what “quite quickly” means in William’s scenario, but what’s the point? He was simply arguing (correctly) that variation without NS can produce complex features — in a highly unrealistic scenario where there is no death.
Under a realistic scenario, though, you don’t get complex features without selection. NS is just as essential as variation, as William seems to acknowledge, but if he insists on using his idiosyncratic definitions of “build” and “produce”, then neither NS nor variation “builds” or “produces” complex features. It’s an absurd position.
Thanks. I knew something useful would come out of this thread.
Joe, to Alan:
Joe,
You learned at least two other things, as well:
1. There isn’t one single valid scientific definition of natural selection.
2. Some evolutionary biologists do use the term in the Darwin/Mayr sense.
Why don’t you check why your posts which violate the “good faith” rule don’t get sent to Guano?
Two mistakes, actually.
The one where WJM made a statement and keiths misunderstood and it was WJM’s fault for not being more clear.
And then the one where keiths made a statement and you misunderstood and it’s your fault for his not being more clear.
I’m convinced! I was that last “that’s goofy” bit that swung me over to the keiths side of this entire debate.
OK, I’ll concede. Your definition is perfectly scientific. After all, Ernst Mayr (praised be he) agrees with you.
Useless, but scientific.
Not Useless. Useful for winning an internet debate.
But Scientific. Because useful for winning an internet debate.
Ain’t science grand!
Meanwhile, an operational definition of natural selection is required lest it join the category of god woo (not objective, not empirical, plus, I just don’t plain like it).
keiths,
Exploring the behaviour of impossible scenarios, of course! 😉
Allan,
Good point. We agree entirely too much on evolution in real life. We need more impossible scenarios to argue over.
Joe,
Such bitterness, Joe!
It isn’t just Mayr, it has nothing to do with Mayr worship, and it isn’t useless, whether you like it or not.
I haven’t seen keiths using his definition to win any debates. Unless you count the debate with me over whether the definition “is scientific”.
You claim to have read university level textbooks regarding biological evolution yet ask for a definition for natural selection to elevate it out from “woo” status.
Inconsistent, to say the least.
I always get confused by infinite populations. I know their mention in population genetics owes something to sampling theory and the LLN – a theoretical infinite population from which our finite population is drawn – but when talking of the effect on selection and drift, I still give the population members ‘real’ characteristics in my mental model – particularly geometric ones. They both occupy and move around in space, and this takes variable amounts of time depending on how far they are going. But in some models, the cows aren’t just spherical, they are singularities which are connected at no distance to every other organism! Random mating isn’t scale invariant. Which is fine of course, depending on what is being modelled – it’s not always necessary (or even possible) to model everything.
Allan Miller,
Infinite populations remove the need to work out all possible outcomes, to mess about with Markov chains and diffusion equations. They permit the theory to be much simpler and deterministic. You have to start with that to develop the theory, and that’s what happened in the 1900s and 1910s.
The first theoretical treatments of genetic drift did not come until the 1920s.
It’s similar to the situation in physics where classical physics started out with homogeneous spheres moving in space. Worrying about what the effect of mountain ranges is just gets in the way at that stage.
I used the term “woo” to refer to natural selection?
You really ought to work on reading in context. There was an ongoing debate between keiths and Joe. I never said the definition keiths was using was not scientific or otherwise referred to it as “woo.”
And yes I’ve read both Mayr and Fisher.