If evolution is not a search, why is the term “evolutionary search” not an oxymoron?
Over at Uncommon Descent Elizabeth posted the following:
“…any “search” algorithm worthy of the name of “evolutionary search” comes with its own moderately smooth fitness landscape built in.”
So evolution is a search if it comes pre-built with its own moderately smooth fitness landscape built in?
Neil Rickert:
So?
What did I say that needs to be retracted?
Elizabeth:
So you agree with DEM?
So “evolutionary search” is not an oxymoron? But evolution is not a search, right?
Elizabeth:
So in the papers by DEM “search” is a metaphor? A metaphor for what?
A bit of background.
Mung:
Joe Felsenstein:
I explained in the thread that you started here that, yes, you were wrong. You will see my answer here.
How was I wrong?
Elizabeth Liddle:
Why?
Neil Rickert:
That’s your definition of evolution?
Question posed in the OP:
This question was posed in response to the following assertion by Elizabeth Liddle:
Still waiting for an answer.
Hey Joe, where you goin’ with that gun in your hand
Mung,
One can search for treasure on foot. This does not make all pedestrian journeys searches. Evolutionary searches utilise the basic processes of evolution (generating heritable variation and applying selection) in order to search for something – ie, as a tool. But evolution itself if no more a search than a stick is a spear.
(It’s Allan-with-an-a BTW. You’re by no means the only one that misses that)
Because evolutionary theory posits a population of self-replicators that replicate with variance.
If the offspring of a self-replication was no more likely to be similar to its parent than totally unlike, then we wouldn’t call it “self-replication”. We’d just call it “something-else-producing”. For instance if you had a population of geometrical shapes, and each produced a baby geometrical shape, and the baby of a triangle was no more likely to be a triangle than a square or a trapezium, we wouldn’t call it a population of self replicators. We’d just call it a population of things that produce other things, at random.
If, on the other hand, every triangle produced an identical triangle, and every square an identical square, the things would be self-replicators alright, but they wouldn’t reproduce with variance.
So any candidate for the descriptor “evolutionary”, has to be a population in which the offspring of things are similar to their parents, but not identical. Which means that the only circumstance in which such a system would not evolve across a smooth fitness landscape would be if most properties of similar things were not also similar.
Which is theoretically possible, but unusual in the real world we live in. A big rock tends to behave like a slightly bigger, and a slightly smaller rock. A spherical pebble tends to behave like an ovoid pebble. A sharp stone tends to behave like a slightly sharper stone.
In other words, in the real world, similar genotypes have similar phenotypes which have similar abilities to survive, and you don’t need (or at any rate, Ewert, Dembski and Marks don’t, as far as I can see, argue that you need) to invoke a designer to account for the tendency of similar things to have similar properties. In fact, you’d have to try quite hard to design a system in which they didn’t.
The “Active Information” you need for an evolutionary system (i.e. one in which self-replicators replicate with variance) to result in adaptation (i.e. find “targets” consisting of populations with specialised survival properties) is intrinsic to the tendency of similar things to have similar properties.
Which is really no biggie.
Provided above. You may have missed the memo, Mung, but I am pretty short of time right now, and that precludes systematic thread searching. But I’ve done it now.
It’s a bit OT but I was struck by this remark by mung at UD directed at Aurelio Smith
So Smith should stop criticizing ID to earn a lifting on his IP block?
I’m shocked (shocked, I say) that mung is happy that UD management “silently” block IP addresses of unwelcome commenters. Where were Barry Arrington and mung when integrity was handed out? 🙂
Alan Fox,
Queuing for hypocrisy?
Talking of hypocrites, did anyone else notice Winston Ewert disowning Kairosfocus and his FIASCO?
It seems KF has to publish or perish! 🙂
Kairosfocus has just posted another OP on active information, needles in haystacks, FIASCO, and Darwinist rhetorical tactics, illustrated with the usual “borrowed” graphics and an exploded view of the Holy Fishing Reel of Montserrat. Seeing the same stuff again and again makes me want to howl with despair.
Are we allowed to hope for the latter?
Piotr:
If it helps, remind yourself that KF is just being KF. Given his genes and his environment, he had to turn out this way.
I wonder if KF will start accusing Winston of being an enabler to cyber stalking.
I remind myself that about people other than KF every day. I guess it’s no hardship to add KF to the list of “he can’t help the way he turned out”. Poor baby.
Acartia,
People have been banned for smaller offences than dissing FIASCO.
I know from personal experience. Numerous times. But nothing gets you banned faster than pointing out that Barry only bans ID opponents even though he supports abusive comments by people like Joe, KF and Mapou.
Hi Allan,
No doubt! My brother’s name is Alan, so while not trying to place an ‘a’ or an ‘l’ where one does not belong I do indeed sometimes overcompensate. I feel blessed that you and Allen MacNeill do not often post in the same thread!
While some folks claim that search is the metaphor, you seem to be claiming that it is evolution that is the metaphor.
Now that Elizabeth is back moderation seems to have gone to hell here at TSZ.
Mung,
Nope. When someone writes a search program using evolutionary methods, it’s an evolutionary search. Biological evolution itself isn’t really a search, though some metaphorically say it is. But it is evolution, non-metaphorically.
Elizabeth:
Did the memo mention the time you are spending at UD?
I don’t think I demanded that you be the one to respond. I could be wrong.
I’m hoping Joe Felsenstein is out enjoying the beautiful weather we’re experiencing here in Seattle.
But I would still like an explanation of why Joe thinks I am wrong.
Joe F: DEM call every process a “search”, even crazy ones.
Mung: DEM do not call every process a search.
Allan Miller:
But evolution isn’t a search.
That’s true, naturally occurring evolution is not a search. But the evolutionary processes of variation and selection can be used in a search by manipulating the environment to direct the processes towards a predefined goal.
Why the IDiots keep equivocating over this is anyone’s guess.
Mung,
Correct. I think I have explained it as well as I know how. I take full responsibility if you are still having trouble with it.
Allan, given that evolution is not a search, it [evolution] cannot be modeled as a search.
Do you agree or disagree?
Adapa:
explain. please. non natural evolution is a search?
@ mung
While, technically, you have the ability to delete comments in your own threads, it is against the rules. The current rule is that comments deemed outside the rules will be moved by admins on request.
And there is no 24 hour mod service here. Usually, it is not required.
Mung,
What one does with any evolutionary computation is to put competing genotypes in a finite arena with replication. One is not modelling evolution as a search by doing that, one is modelling evolution.
The purpose of a model of evolution is not generally to find a target, so it is not generally a search. One of the simplest models would be one which starts with a population of labelled genotypes and allows them to breed with no selection. There’s one here. In such a model, one genotype always emerges as the sole ancestor of all descendant genotypes – variation is always eliminated. I wouldn’t call that a search, I’d call it a model.
That is precisely why design is the better explanation. Design prefers the direction leading to survival.
Evolution does not prefer a particular direction because it neither designs nor manufactures. It maintains.
Thats why bacteria produce billions to keep millions, insects produce thousands to keep hundreds, snakes produce hundreds to keep several, rabbits produce several to keep a few.
Reproduction is how design solves the problem of an ever changing environment and entropy. Each organism produces the quantity of progeny that will ensure a minimum percentage survives, while the ones that dont survive contribute to the reproductive success of other organisms.. It allows life to perpetuate in any environmental condition.
It also circumvents the 2nd law by reproducing just before entropy erodes the individual organism. Though any particular organism eventually dies due to entropy, Life continues on indefinitely.
It is pretty logical and obvious that evolution is a subset of design, not an alternative to it.
Since most taxa coexisting at any particulat time eventually die out leaving no descendants, this “design” looks recklessy wasteful to me. Were all the non-avian dinosaurs, not to mention pterosaurs, ichthyosaurs, plesiosaurs and miscellanous whatchamacallasaurs, designed to go extinct? So much effort invested for almost 200 million years, and not a single species left? If the designer was bored with the big dinos and wanted to make more living space for birds and mammals, why did nearly all Mesozoic birds and mammals, save for a few lineages, have to die out as well?
So does survival 🙂
An interesting argument, if only it were true.
Speaking of birds and mammals, there’s a reason why IDists with IQs above room temperature go back to the Cambrian for solace. Once you start having fossils, it’s pretty east to trace the gradual steps of evolution.
Evolution works as a feedback system that follows changes in the environment and acts to keep populations at local reproductive success maxima.
How does “Design” know ahead of time what changes to the environment are going to occur so it knows in advance what to “design”? For example, how did “Design” know what to create to survive the Chucxulub asteroid impact of 66 MYA?
But it is incorrect to imply that evolution only maximizes adaptation.
Evolution continues at pretty much a constant rate even when the apparent outward form is static. New an unnecessary features appear in static environments. Populations are not uniform.
We have a word for populations that lack diversity.
Endangered.
Moved a post to Guano. Remember the game rules: Assume the other poster is posting in good faith.
Ta.
Elizabeth,
@ Adapa
I moved the other comment that mung had deleted. Hence my comment upthread.
Here’s omething I’ve known for a long tim, but it’s good to see it from a major IDist.
Winston Ewert
The whole inverted pyramid of ID rests on IC.
This is the most bizarre excuse for preferring “design” as an explanation that I have ever heard. Reproductive life is truly extravagantly wasteful, therefore, what … design!
But it’s somehow apt. It’s the closest an IDist has come to admitting their Calvinist roots. You were designed to die. You were designed to be food for vultures and worms. You were created just to “contribute to the success of other organisms”.
Your mother was created to bear “several” children in sorrow and “keep a few”. Aren’t you the lucky man that you are among the “few”!
Get on your knees and thank the Lord Designer for not creating you to contribute to the reproductive success of a female Anopheles gambiae. Be extra thankful you weren’t designed to contribute to the reproductive success of a Carcharodon carcharias. That would really hurt. But who are you, lower than a worm, to harbor doubts about the Designer’s bloody methods? Don’t!
Just be grateful that you are among the chosen few, and humbly beg that by the Designer’s grace you will continue to be among the chosen.
This seems rather cruel. The Yellow Perch is known to produce a very high number of offspring and then eat most of them because the smaller fish can utilize a food source (small plankton) that the adults can’t. Wouldn’t a better design be to design the adults to be able to utilize a smaller food source?
There’s a science fiction story that sounds a bit like this.
http://en.wikipedia.org/wiki/The_Legacy_of_Heorot
But Soylent Green tastes soooo good.
Mung:
Thanks for the good wishes. I am currently up at Friday Harbor Labs — we certainly are having beautiful weather here.
You were. Very wrong. Let me explain. What DEM call a “search” is not the same thing as what most of us would call a search algorithm. Here I will discuss DEM’s “search”. I hope in a comment soon to discuss whether evolution is a search or not, using search in the more conventional sense.
DEM have a finite search space Ω of points. One example would be DNA sequences 1000 bases long. For DEM a “search” is a probability measure on the space. Any probability measure.
One “search” would be “choose a point in Ω at random”. That (or rather the resulting uniform probability distribution of points) is one of their “searches”, in fact the “default search”.
Another would be “Assign fitnesses to the points x1, x2, … , using the particular function f(x). Pick 50 points at random, return as your final answer the one with the lowest fitness. If two or more tie for lowest, pick one of the tied ones at random and return that as the answer.”
Now that of course, from the point of view of evolution, results in a crazy “search”, one which gets fitnesses that are worse than random, not better. But it creates a probability distribution on the results which is within DEM’s definition of “search”.
I am not saying that DEM were themselves being crazy, just that readers (apparently including you) might misunderstand what their space of “searches” included.
Tom English and I were making the point in our Panda’s Thumb post that once one has organisms that reproduce, and fitnesses that tell us how many offspring each is expected to have, any “evolutionary search” involving them will not be nearly as crazy as many of the “searches” in DEM’s space of “searches”. So just from having organisms that have fitnesses you already have a fair amount of Active Information, in their terms.
I do have to apologize to you for missing your response to my query, and for saying that all I heard were “crickets”. You did respond.
But your response was not an effective reply. Yes, DEM’s “searches” include probability measures resulting from all sorts of random and crazy “searches”. You argued that DEM’s “searches” did not include these. You are wrong about that. I was right about that.
I trust that you will now openly acknowledge that error.
Alan Fox:
Piotr Gasiorowski:
OP or OPerish.
Piotr,
Appealing to a slew of extinct taxa does not invalidate a design paradigm. If the concentration of life on Earth was diminishing at a rapid rate due the extinctions, your observation may have a point.
As to the your assertion of wastefulness, Life wastes nothing. When rabbits produce several to keep a couple, the others become food for snakes, fox, big cats, etc. etc. Absolutely nothing wasted.
again, the rabbits’ overproduction of rabbitos ensures the survival of at least enough rabbitos to ensure rabbits will continue to exist. In turn, the excess rabbitos will help ensure the survival of snakes, fox, big cats.
What is fascinating is the reproductive rates of organisms seems finely tuned so that each organism contributes a proportionate amount of offspring to ensure the stability of the food chain.
Truly a designed win/win cooperative effort in the name of Life.
Life wants to live and makes it happen.
In contrast, evolution is not interested in survival per se. How can it be? It takes no sides on the issue. It doesnt do goals, purpose, direction.
So it settles for second fiddle at best. First duster is more likely, though..