This should get some good discussion / denial!
652 thoughts on “Evolving Wind Turbine Blades”
Leave a Reply
You must be logged in to post a comment.
The Skeptical Zone
"I beseech you, in the bowels of Christ, think it possible that you may be mistaken."
By definition, it is what the word is defined to mean.
Definitions aren’t circular, they’re just definitions. Circularity pertains to logical arguments, or reasoning from premises to concusions in logic. Fitness is not a conclusion of an argument. For something to be circular, there would have to be at least one premise and at least one conclusion, and the conclusion would have to be in the premise.
Fitness is defined as reproductive success. It merely means “how many offspring does carrier of this allele have on average”. No premise is stated and no conclusion is derived from it.
If it was circular reasoning(which is also a form of question-begging logical fallacy, which means you are somewhere assuming what you are trying to prove), then the definition of fitness would have to take the form of an argument, or a line of logical reasoning with multiple steps leading to a conclusion, instead of merely an explanation for a concept.
But “fitness” is not an argument, it is a concept. It is not supposed to constitute an argument. When you describe the fitness of something, you are not arguing, you are explaining or describing a property.
The property of reproductive success. Properties aren’t circular, they are either observed or definitional.
When somebody says “this organism has a relative fitness of 1.02 compared to it’s ancestor”, they are not arguing. They are not reasoning in order tro try to convince you of something. They are just stating one of the observed properties of the organism.
You’ll have to elaborate.
No idea what you think I changed my mind about. What happened between me posting about your lack of clarity and my comment about “environmental design” is your comment with quoted text that you had picked out phrases including the word “design”.
Your subsequent comment suggests you may be thinking that GAs and EAs smuggle information in via the fitness filter.
Alan Fox: So you’re still arguing that there is design going on with a GA model. I agree. I’ve been making this point for years now.
This is a meaningless statement. “science” is just as amenable to logical propositions and reasoning as any other intellectual human endeavour. You can, in fact, be guilty of circular reasoning when formulating scientific conclusions.
The problem is individual definitions of words (the conceptual referents of technical terms used within some scientific field, such as biology) just aren’t arguments in propositional logic, as such they cannot constitute circular arguments or circular reasoning, because the mere definition of fitness is not seeking to argue something or to reach a conclusion from a premise.
You people are so profoundly confused.
They really struggle with logic, it’s so embarrassing.
Going by their “logic”, Newton’s second law is also circular, because mass determines acceleration and acceleration determines mass. Sigh.
But the worse part is not that they make basic mistakes like those, it’s that even if you explain it a thousand times it still doesn’t sink
phoodoo,
Wrong. The specific properties that enable a particular type to produce more offspring depend upon the specific environmental circumstances in which it finds itself.
Your question and answer amounts to “The properties of the fittest organisms are those that survive and reproduce more”. That’s incorrect. Surviving and reproducing more on the average is not a property, but a consequence of a property – it’s not the thing that makes you survive and reproduce more, any more than winning races is the thing that makes you faster. As usual, the fact that it makes no sense leads you to think that all evolutionists talk nonsense, rather than that you may have made an error.
The properties of the fittest organisms are the things that lead them to survive/reproduce more. Which could be fur, disease resistance, speed, size, metabolic efficiency … ‘survival of the fittest’ is just a slogan, an attempt to briefly encapsulate the complexities, in a statement that is hardly wrong by being apparently tautologous. Do you think organisms that survive/reproduce more will not persist more? What a strange position that would be.
Tautologous does not mean ‘wrong’. The fastest runner was the one that won the race. Stop the presses, I have just tautologously nullified every race ever run!
It’s like they’re in this strange state of mind where anything, no matter how trivial or seemingly mundane, must be objected to, derided, protested and argued against. This desperate need to reject and downplay absolutely everything even remotely related to the subject of evolution makes them commit to these utterly inane ideas and statements, and there just cannot possibly be the slightes concessions on any subject.
In a discussion with William Murray here a few days ago about truth, how to define it and to go about knowing it, he eventually admitted that he doesn’t care about what’s true. At least he was being honest about it. Phoodoo, Frankie and Mung OTOH want to redefine truth to mean whatever they please and then claim victory.
They don’t understand that fitness is quantified differently in GA’s than in nature because god forbid that GA evolved a crocoduck instead of a wind mill!
Rumraket,
“It is difficult to get a man to understand something, when his salary depends on his not understanding it.”
— Upton Sinclair
The same is true when his religious beliefs depend on not understanding.
But wavelength does equal frequency.
I guess that’s some sort of joke? I’ve seen it before here but not sure what it’s all about
EDIT: OK found it. Just another one of Joey’s gems, LMFAO
Erik,
No. To do so would be to misunderstand it.
It would be rather easy to disprove the work of John Sanford using ‘phoodoo-fitness’, if that was all evolutionists did. If the fitter genotypes are merely those that survive a particular contest, there is no problem of genetic meltdown, since deleterious mutations cannot (by the definition of ‘phoodoo-fitness’) accumulate. Of course the argument would be spurious, and in any case this is not how fitness is defined, though phoodoo will never accept that however clearly or often one states it.
But Sanford relies upon the pop-genetic approach to fitness for his argument. He doesn’t attack it as merely empty, he questions the power of selection and recombination. And he uses GAs as models of evolution. I wonder if phoodoo et al think his work is spot-on too, which would generate a certain degree of dissonance, which may not be cognitive. Fitness is rubbish, GAs don’t do anything biologically relevant and Sanford is on to something!
Allan Miller,
That’s what makes me wary of calling this GA an example of evolution at work. It models darwinian evolution, but as already stated, adding neutral and (slightly) deleterious mutations would improve it’s capabilities to explore more of the landscape, and would better model evolution. But that would make it more resource intensive (would take longer to reach the same fitness but better in the long run)
Of course the idea was to find a good design for a wind mill, not to accurately model biological evolution, so it’s perfectly valid to make simplifications like fitness = reproductive success in this case, and the fact it works for this and many other applications is a great case for evolution: the closest it models it the better it works
Do the fittest organisms survive best. That is the claim made by evolution. but is it actually true?
Well, guess what, it is true. In fact it can’t not be true! Fitness is survival! It sure makes it a lot easier to have a theory whose definitions preclude it from being false.
I guess the only way it can’t be true would be if the fittest didn’t actually survive best…but oh, wait. they have to! Hoho.
Why can’t we pesky creationist just accept it..we are just so stubborn.
dazz,
Yes, the criticism seems to be that using evolution as a tool, making design trade-offs due to your specific goals and constraints – the very fact of having goals, and tailoring accordingly – invalidates any and all GAs as models of evolution, not just this one. But it doesn’t invalidate Sanford’s … !
phoodoo,
But that isn’t the case. Sometimes the fittest organisms get hit by an asteroid, eaten anyway, crushed by a log …Fitness is not survival. Fitness tends towards survival, true enough. That’s why Natural Selection has causal power to change genotype frequencies. But frequency change is stochastic, not deterministic. Your investment can go down as well as up, even for the fitter genotype.
See above. How can deleterious mutations fix in a population, if the fittest are defined as the ones that survive? Answer: they aren’t.
You tell me.
You think genotypes with greater average output (you can measure this) will become rarer? Stay the same? Fluctuate with no direction? Fluctuate with an upward tendency towards the higher? What? Very easy to investigate with bacteria in a chemostat. Guess what happens? A bit harder with giraffes on a savannah, but there is no obvious change in principle.
Is the concept of fitness a theory in itself? Is it the entirety of the theory of evolution? No and no.
By now I don’t think your posts can even get any dumber. Seriously, I am positively amazed by how inane your objections have become.
The “claim”, the definition must be tested. There are countless experiments where cultures of viruses or bacteria are put under selective pressure for infection, or motility, etc… if they evolve infectivity/motility, and the populations that do spread quicker (get a reproductive edge) then the definition is supported by experimental evidence. Guess what? tons of experimental results support the power of selection to give fitter organisms a reproductive edge. That’s not to say that every single fixation is adaptive, but it’s not all in the definitions in science, unlike your religious faith.
It’s ridiculous that you guys attack science on the basis that it takes it’s definitions on faith when it’s crearly false, but ironically that’s exactly what you do
The idea that the statement “the fittest are fittest” is even something one needs to test before accepting is idiotic in the extreme. It’s like saying you want evidence that “God is God”. What fucknuttery is it to even start an argument about something so trivial?
Not sure if that’s aimed at me, but what I meant is that the correlation between fitness and reproduction success (natural selection) can be tested
Except that correlation is only true if the genetic changes were happenstance occurrences. If the changes were directed to occur, as in GAs, it isn’t natural selection.
Natural selection is blind and mindless whereas GAs are neither
I’ll be clearer- by having one unchanging fitness function the program is directing all solutions towards a final solution. That isn’t how NS works
Are you guys admitting that you don’t understand computers or GAs? Thank you
Or is it that you don’t grasp theoretical concepts?
One last opportunity for you to make some sense:
What models those changes in the GA?
That doesn’t even make any sense. The program drives the changes and it then directs them towards a solution. It is evolution by design, not NS.
AGAIN- GAs are search heuristics actively searching for a solution they were designed to solve. NS is blind and mindless- it is passive. The two are not alike.
No indeed. GAs and EAs are models. NS works in reality. The environment is multi-dimensional and constantly changing.
No indeed? To what?
NS works? How do you know? How can we test the claim that NS can produce protein complexes?
Then why don’t all runs converge into the same solution?
And how do you know you’ve reached the solution?
You have no idea, because you’re wrong, that’s why
dazz,
LoL! All runs are CONTINGENT and we know we reached the solution when the output stops changing.
Doesn’t make sense? that alone proves you didn’t understand anything:
Watch there and see how in the GA, changes (new candidates) are produced by mating of two previous relatively fit candidates by combining their spines length, and by producing occasional random mutations or variations in the spines.
So there you have it, changes in the GA are driven by random variation, just like in evolution. Therefore, since GA’s work, they support evolution
phoodoo,
Jesus, phoodoo. We schooled you and William on this a year ago (in Barry Arrington digs up the ‘tautology’ argument). Do we have to do it all again?
I commented then, in exasperation:
The GA (as in the example in the OP) is acting like a filter or sieve, selecting the best of the randomly generated candidates. Obviously, the only parameter in this case is theoretical power output generated by the candidate designs. The objective is higher output. Nobody is designing the shape of the blades. This is the same principle as biological evolution but at a much simpler level.
I’m disagreeing with your use of “directing” and agreeing that this is not how biological evolution works.
They support DIRECTED evolution.
LoL! The GA eliminates the less fit, Alan. You didn’t watch it, did you?
And again it is a search actively searching for a solution to the problem it was designed to solve. NS is nothing like that.
dazz,
That doesn’t make any sense. What causes the changes is the program,
You do realize that wind turbine blades do not reproduce. And by programming them to do so you are starting with the very specified complexity that needs to be explained. GAs are design all the way down
Alan Fox,
GAs direct the solutions towards a final solution. That is what they do. Are you actually disagreeing with that? Really?
So I show you that GA’s implement random variation just like in evolution and all you can do is that? Some bald assertion?
OK, are you ready to put your money where your mouth is?
Answer this. If a GA is directed. What should we observe if we re-run the same simulation over and over again, starting with the same exact population? Should it follow the same exact path over and over again?
dazz,
A GA is a search heuristic that actively searches for a solution it was designed to solve. It directs all minor solutions to that end. Natural selection is nothing like that.
Deal with that, or not. I don’t care. The objective world can see it for what it is.
Or selects the best candidates for the job.
The OP vid? Yes.
The GA fitness function selects the candidate blade designs (by their performance in the simulated wind tunnel) It chooses the best randomly generated design on offer, it does not direct or produce the design.
Well, biological evolutionary processes are something like that, but real biology is much more multi-dimensional.
Yes, I am disagreeing with that. Really!
So you are disagreeing with reality. Wow.
I’m talking about a GA and you’re talking about evolution. Yet more evidence that GA’s don’t work that way.
Doesn’t Joe Felsenstein’s remark that it would actually be possible to model natural selection in a GA interest you in the slightest?
Well, I am a “specialized geek of genetics”. Fitness is an important concept in biological studies of natural selection.. I have been using it and thinking about it for over fifty years. To make use of it I don’t need to get into “philosophical”, “moral” or “vital” concerns (though it helps to be “logical” and to have a “commonsense” understanding of the concepts).
We can define it, in real cases you can measure it for different genotypes or different phenotype classes. We can use models to predict future changes in the genetic composition of the population. We can reconstruct past changes. We can investigate how different characters affect fitness.
That’s not nothing.
As for the phrase “survival of the fittest” I don’t use it. I have an online population genetics text (here) that is more than 500 pages long. Guess how many times the phrase “survival of the fittest” appears in it?
Zero. There’s plenty to do without using it.
Anyway it’s misleading. Because survival is not the only component of fitness, even in simple discrete-generations models. There’s fertility too. Almost everyone in this argument seems to have forgotten that. It is quite easy to show that a genotype that has lower viability (probability of survival) can have higher fitness.
Now I have to go back to my work, which today is trying to prepare a presentation for tomorrow. It is a talk to the Joint Mathematical Meeting of the American Mathematical Society (AMS) and the Mathematical Association of America (MAA) about a model of multiple traits evolving on a phylogeny. It models change in those traits as an Ornstein-Uhlenbeck process, using equations from quantitative genetics. The objective is to use a known phylogeny, inferred from molecular data, to infer the additive genetic covariances of the characters and also, simultaneously, the shape of the fitness surface, where we approximate that by a multivariate Gaussian curve. That’s right, estimating the fitnesses.
Alan Fox,
Of course the GA produces the design- that is what it is doing-> producing designs that can either be kept or eliminated, just like engineering trial and error. It takes those designs and refines them until some pre-specified optimum is met.
I’m giving you the opportunity to prove what you’re asserting. Why are you chickening out?
If a GA is directed, why doesn’t it always follow the same route starting at the same point like a hill climbing algo does?
And I’ll double down. I’m willing to bet real money with you here: if one removes random variation (mutation) from a GA, then I say it will also follow the same route every run, proving that random variation makes it unpredictable, undirected, and proving you wrong.
Are you willing to put your money where your mouth is or not Joe?
Joe Felsenstein,
A “geek of genetics” that cannot tell us what makes a human a human nor a cat a cat. Not to worry though, no one can do that.
dazz,
CONTIGENCY, just as I have already told you. Look don’t blame me because you don’t understand GAs
A GA is a search heuristic that actively searches for a solution it was designed to solve. It directs all minor solutions to that end. Natural selection is nothing like that.
Deal with that, or not. I don’t care. The objective world can see it for what it is.
I dare you to show that is wrong. Put your money where your mouth is
I’m also talking about GAs. Clearly they must carry the good forward more than the bad to be effective. What do you think GAs were inspired by?
Sure.
That’s so cool Prof. Felsenstein. I wish I knew more of this stuff already. It’s fascinating. Please keep us posted