What evidence is there for evolution being non-random?

[Meadmaker]

I'm not saying the argument is correct. I'm saying that selection doesn't get you out of the random problem. If a guy comes to you saying that you can't be assembled by chance, you aren't going to win by telling him that you can be assembled by chance, but only if you add an additional constraint. That's what you are doing by trying to convince him that selection makes the process non-random. You are avoiding his problem, which is a mental block on the idea of random assembly, and you are substituting an even harder problem, which is random assembly with constraints.

If you want to counter the argument that we are too complex to have come about by chance, there's no way around dealing with the argument. There's no point in denying it. We can, and did, come about by chance. Selection doesn't negate the chance aspect, because the creation process is still pure, 100% random mutation. Selection just adds an additional layer of difficulty on top of a problem that the creationist already thinks is unsolvable.

 

[ImaginalDisc]

I'm not saying the argument is correct. I'm saying that selection doesn't get you out of the random problem. If a guy comes to you saying that you can't be assembled by chance, you aren't going to win by telling him that you can be assembled by chance, but only if you add an additional constraint.

If he understands that the variability in the traits is randomly determined, but selection on those traits is nonrandom, usually directional, disruptive, or stabilizing, (though disruptive selection is rare) then he would understand current scientific theory. The inputs (traits) are random, but the outputs (fitness) are not.

 

[articulett]

No.



Your assertion that selection plays into the hands of creationists is based on a faulty assumption: that incremenental changes before the "final" product are useless. Nothing could be further from the truth. I'm mildly nearsighted in one eye, but the poor vision it provides me is substantially better than having an extremely myopic eye, or no eye at all. Even a 20/20 human eye only detects three colors (except fr tetrachromats) and is plainly pitiful when combared to the resolution of a falcon's eye. A stubby wing is better than no wings at all, for flying, and weak claws are better than no claws at all, for predators that claw at their prey. The vast majority of traits which are slected for exist in a broad specturm, with lower utility at one end, and higher utility at the other.

You also cannot forget the cost of developing and maintaining traits. A fleet footed prey organism that is introduced into a habitat without any predators is essentially wasting energy in developing that much speed. Selection would tend to drive the population in the direction of lower speeds. Evolution is not goal oriented.

I agree. I just can't make sense out of what Meadmaker, Walter, Whitey, or Mijo are saying...and I don't think they make sense to each other--though they all think they are being more explanatory than Dawkins and "more honest" and more "intellectual rigorous"...they all seem to think they know more about teaching evolution without even being able to describe it in any useful way--or understandable comprehensive way. I don't think they are being more honest. I think they are sort of just talking to themselves like Kleinman. I think they haven't a clue as to how people come to understand natural selection...I can't even tell if they understand it --especially with the way they define "fitness" (as physically fit--stronger, faster, etc.) Fitness of information passing (which is what a genome is) describes the how good the information is at getting itself copied. The same with a virus and computer virus or a catchy meme. A chain letter has a built-in "trick" for getting itself copied, but that has nothing to do with physical fittness. Darwin didn't coin the term "survival of the fittest", but in the world of DNA (which Darwin never even saw)--fitness only refers with that which confers preferential survival and subsequent reproduction and/or keeping competitors from doing so.

I don't even think they have a basic understanding of the impression people have of evolution in general and what it is that starts to make them "get it".
I don't even think Whitey or Mijo understand natural selection. And Meadmaker doesn't understand the "wedge strategy" and common creationist conundrums (not that I think anyone can change the mind of a male creationist over 40)-- He actually denied that creationists bring up this thinking snafu despite multiple references of such--moreover, he thinks he knows how to unhinge the argument without even showing an interest in those who HAVE unhinged that argument and edified many. I know that, for myself, once I understood natural selection, the rest fell into place. I marvel at the random aspects, but it's the process itself that I find miraculous...and it is no more "random" that food processing or the birth process or word processing.



I found this better definition with the use of the word"stochastic":

Evolution is partly a stochastic opportunistic process and partly an optimizing selective response to bifurcations in the eco-landscape

That is certainly clearer and simpler than any of them have said, isn't it? And it answers mijo's question. And I would start with the simpler principle--of "random mutation coupled with natural selection" because that is what worked for me--it works in Dawkins' books and it works on my students. I can't imagine anyone learning anything about evolution from the sources above--and I don't think they understand each other. Don't they notice that they are not really getting any feedback--no one is "getting them" or understanding evolution better. I think the criticisms they levee at Dawkins apply more to them. I mean it's interesting to watch--it is so Kleinman like and Behe like--it's like they are each in their own conversation from my perspective. And they just can't hear anyone say that the words they are using, don't explain anything--they aren't more honest or intellectually rigorous--though like religious folks, they sure do think they are. They don't clear up the made hurdle to understanding evolution. They aren't saying anything of value on the topic or answering the question or clarifying anything--just criticizing those who don't understand why they'd want to explain something so important in such an unclear, uninformative way.

I wish they'd just tell me if they understand each other--or could sum up each others argument. Or define a non-random process versus a random process via analogy and show how they perceive evolution being more like the latter.

I don't understand their problem with Dawkin's definition or Talk Origins ways of describing thing nor is anything they are saying clearer. I think Whitey is way more dishonest than Dawkins would ever be. I don't think they are being more honest or clearing up conundrums with their oddly twisted imprecise definitions of "random" and "stochastic". I mean meadmaker actually says "Mate selection is a random process". What does that mean? I mean, sure there a serendipitous events involved--but why would you describe it so vaguely if you wanted to convey any information.

I mean aren't they all just saying--"evolution is a process that contains non-uniform probabilities". So is seat belt wearing. It just doesn't convey any useful information.

I should just let them sit back in engage in dialogue...and see what definition they find better than Dawkins and more intellectual rigorous and honest while answering the question at the top of this post and clearing up creationist conundrums to boot. Now that would be amazing.

 

[mijopaalmc]

articulett-

From what I understand about numerical weather prediction, the difficulty in obtaining accurate long-term forecasts stems from the fact that the equations of fluid dynamics and physics used to describe meteorological systems cannot be solved exactly and that it is nearly impossible to obtain precise and detailed enough data. Thus, not only do we have to use approximate numerical models, but the problem of weather prediction is underdetermined in so far we do not have enough detailed data to make the detailed predictions that we want.

The same is true of celestial mechanics, another deterministic problem. The equations for an N-body system will always yield the same posterior positions a velocities given the same prior positions and velocities. However, as with the equations of fluid dynamics used in weather prediction, the N-body equations have not been solved exactly and the data used to predict posterior states of the N-body system are often not detailed enough. Thus, we have the same problems of approximation and underdetermination that we had with weather prediction.

Evolution, on the other hand, is fundamentally a statistical process, because it is possible to get different outcomes for the input. In other words, not all fit individuals survive and not all unfit individuals perish. Obviously, this is a different situation from that of numerical weather prediction and celestial mechanics, because, given exact solutions and detailed data, both weather prediction and celestial mechanics would yield the same results with the same initial conditions.

 

[ImaginalDisc]

Evolution, on the other hand, is fundamentally a statistical process

Emphasis added.

Ding!

Randomly distributed values tend to conform to a Gaussian curve, Poisson distributions, and other statistical measures. We can state with extremely high degrees of certainty that, for example, the fitness of an impala is correlated with its peak ground speed. That some fast impalas die without reproducing and some slow imapalas get lucky, reproduce often, and die old, is a given. We are extremely confident when saying that their fitness is not random because randomly distributed fitness values would not have any relation to any traits.

 

[articulett]

Emphasis added.

Ding!

Randomly distributed values tend to conform to a Gaussian curve, Poisson distributions, and other statistical measures. We can state with extremely high degrees of certainty that, for example, the fitness of an impala is correlated with its peak ground speed. That some fast impalas die without reproducing and some slow imapalas get lucky, reproduce often, and die old, is a given. We are extremely confident when saying that their fitness is not random because randomly distributed fitness values would not have any relation to any traits.

Yes. And moreover--fitness can be used to describe physical qualities--but lots of things can give an organism a physical/reproductive edge--lots of stuff that is coded in the genes. In looking back at evolution, we can only know that the genes that are prevalent in populations are key to those organisms survival and reproductive success--they were selected. They confer or did confer some sort of fitness at one time or sorted with genes that did so. We can also say that humans that beget the most humans are the humans whose genetic traits will appear in the future. If the people who produce the most humans tend to have poor impulse control or greater tendencies towards religiosity--then we'll see more of such--even though it has nothing to do with "fitness"-- The sex drive evolved not because it means one is more fit--but because with the strongest most urgent sex drives preferentially passed on their genes--including their strong urgent sex drive.

I do dislike how the creationists and the like get mired in the "survival of the fittest"--the fittest at what? The fittest at surviving, reproducing, and eliminating competition or forming symbiotic relationships with other life forms.

Walter says this definition is circular--it's just that "survival of the fittest" is so easily misconstrued and used in creationist conundrums that I prefer preferential survival. The only fitness that "matters" to any "selfish gene" is the ability to replicate itself on into the future and accumulate refining changes along the way.

Ugh...just as random is such an unclear word when talking to the clueless--or those who are listening to hear what they want to hear--fitness is as well. Do all the randomites conceive of "fitness" as some sort of physical strength?
Females drive selection in many species--they aren't the strongest or the prettiest--but they drive, both sexual dimorphism and salient physical traits in the males. A peacock's tail did not evolve randomly or stochastically--but incrementally, because each increment in the genome that caught the peahen's discriminating eye, got passed on more often by virtue of the fact that the male possessing the "best plumage" got the ladies. How the hell does stochastic processes and anti Dawkins "intellectually rigorous" truth convey that? Sure, some males that could have had prettier tails could have been eaten by a predator while still in the egg, and some less beautiful peacocks might sneak in a mating on the sly when the good looking guys aren't around-- but that doesn't make the mating process, evolution, or anything about the issue--"random", and I can't imagine how the use of the word random or synonyms thereof can describe this process any better than that. It's just so vague.

If you someone cannot understand why natural selection is the opposite of random or why anyone would refer to it with those terms, then I suspect they are a lost cause. For some reason they NEED to see evolution as "random" with whatever nebulous definition they've given that term. (Non-uniform probabilities?--that's the same as the seat belt example. It doesn't convey any useful information to call seat belt wearing a random process in regards to saving lives.)

Once again...Give us an analogy of a not random process that evolution is not like... and another random process that evolution is like. I understand that mijo thinks weather is deterministic in a way that evolution is not. But I want him to tell me another random process so I can see what the hell he is talking about. I don't think that weather is more predictable than Mendelian genetics and amniocentesis. I think mijo is bogged down in tangential details to be able to describe evolution as a "stochastic process"--and can't see the forest for the trees. He misses the principle completely while thinking the exceptions are meaningful in some way for defining the principle.

 

[articulett]

If he understands that the variability in the traits is randomly determined, but selection on those traits is nonrandom, usually directional, disruptive, or stabilizing, (though disruptive selection is rare) then he would understand current scientific theory. The inputs (traits) are random, but the outputs (fitness) are not.

Exactly. Cities aren't built by chance...nor is the internet. Technology doesn't evolve by chance though serendipity and randomness come into play. These are incrementally designed from the bottom up by what came before. Random components do not a random process make.

A "random" sperm made each of us...but we would not describe ourselves as "random"...because random implies a lack of connection--a one time event. The birth process isn't random and the genetic information that the spermatazoa carry certainly isn't random if you consider it's journey through time. Evolution is the compounded interest on the best the randomness came up with over the eons. I wouldn't talk about the random influences of natural selection until someone got the basics. And mijo sure doesn't. He asked a question in this OP. It was answered well by many, but rejected because some people still think it's clear or "truer" to sum up the entirety of evolution as "random" or a "stochastic process"-- even though lots of people have told them they aren't conveying anything, and in many ways they sound as muddled as a creationist in the explanations.

I can't tell if Meadmaker doesn't understand natural selection though he thinks he does...or if he just doesn't get how little is conveyed by his choices of words. Does anyone other than meadmaker think that meadmaker sums up evolution better than Dawkins. Or even other people on this forum? I mean his better than mijo...and more coherent than walt and whitey--but I want to know if his words convey information about the nature of evolution to anyone. I already know that I can do so with my words, because I have. I just don't understand where meadmaker's arrogance and opinion of his explanation being better than everyone else's comes from. But then again, he thinks Paul and all others are wrong about the creationist obfuscation with the word chance. And he thinks mijo isn't an "intelligent design proponent". I think it's obvious that he is.

 

[mijopaalmc]

Evolution, on the other hand, is fundamentally a statistical process

Emphasis added.

Ding!

Randomly distributed values tend to conform to a Gaussian curve, Poisson distributions, and other statistical measures. We can state with extremely high degrees of certainty that, for example, the fitness of an impala is correlated with its peak ground speed. That some fast impalas die without reproducing and some slow imapalas get lucky, reproduce often, and die old, is a given. We are extremely confident when saying that their fitness is not random because randomly distributed fitness values would not have any relation to any traits.

I find it interesting that you lopped of the end of the quoted sentence. Here is the entire sentence:

Evolution, on the other hand, is fundamentally a statistical process, because it is possible to get different outcomes for the input.

The point is that selection is fundamentally statistical (or stochastic). One can only determine that that the fitter individual will survive on average more often than the less fit individual. On the individual level, however, there is a fundamental uncertainty about the fitter individual's survival relative the less fit individual's survival.

 

[ImaginalDisc]

The point is that selection is fundamentally statistical (or stochastic). One can only determine that that the fitter individual will survive on average more often than the less fit individual. On the individual level, however, there is a fundamental uncertainty about the fitter individual's survival relative the less fit individual's survival.

You have not described a random process. You have described the very essence of a non-random system. If you insist on discussing statistics you must accept the statistical implications of the words "random" and "nonrandom."

 

[ImaginalDisc]

Addendum to the above post: You are succumbing to the "no middle ground" fallacy. Systems that exhibit uncertainty are not automatically random.

 

[jimbob]

Because evilition isnt random, the farce was able to make humanity inevitible

(that was a praraphrase)

The above quoted post is why I say it is important to say that evolution is random, but with underlying laws.

Jim

 

[Meadmaker]

I thought of an analogy.

A:”I know of a guaranteed way to win the lottery”
B: “You can’t win the lottery. It’s random”
A: “My method eliminates the randomness.”
B: “That’s impossible. It’s random.”
A: “But I throw away the losing tickets!”


That’s what you get if you just say that evolution isn’t random because of the selection process. Of course, it’s not a great analogy. Let’s try the one below.

A: “There’s a new lottery game. It’s called the ‘Methinks it is a weasel’ Lottery”.
B: “How do you play?”
A: “They draw one set of winning numbers. There are 100 balls each numbered 1 to 100. You have to match all the numbers to win. (In order) They never redraw the numbers. Every week there is a drawing. The first player to get all the numbers wins. I’m going to win.”
B: “But you can’t win. It’s random”
A: “It isn’t random. There’s a catch. If you get a partial match, you get to play the same numbers again.”
B: “So? If it lost before, it will lose again. To win, you have to draw a random number, and that might as well be impossible. The chance of that is 100 to the 100th power. You can’t draw a random number like that”
A: “I’m not drawing random numbers. I’m copying my numbers from the old bet slips.”
B: “But they lost. You can’t change losers to winners.”
A: “Yes, I can. You get one extra ticket with the same numbers for every number you match. .”
B: “So how do losers change to winners? You’re wasting your money.”
A: “ But, to replay, you hand your ticket to a clerk, and he types in the numbers”
B: “And?”
A: “Sometimes he makes mistakes!”
B: “Are you crazy? He’s going to randomly type in the wrong numbers and BAM, you’re going to match 100 numbers?”
A: “There’s nothing random about it. Sooner or later, I’ll win.”
B: “But even if you were amazingly lucky and got, say, 5 numbers, when he made a mistake on one of the numbers at random, it’s a lot more likely that you will lose a hit than gain one. You’ll never win.”
A: “There’s nothing random about it. Losing tickets won’t get played. Winning tickets will win more tickets. When they get lucky and change to even better winning tickets, you’re that much closer.”
B: “So you’re counting on getting lucky? And random mistakes will get you a winning ticket? Yeah, right. You don’t have time!”
A: “Why do you keep saying random. There’s nothing random about it. Winners will get more tickets. Losers will get less. The better the ticket, the more chances you will get to play.”
B: “But the only way you get better is with a random mistake”
A: “Why do you keep saying ‘random’” It’s the exact opposite of random. It’s a guaranteed win.”
B: “It is not. You have to assemble a specific string of 100 numbers, and you have to do it based on random mistakes in the copying process.”
A: “But it isn’t random. Random means every number is just as likely as any other number. In this game, winning numbers get more chances to play than losers. There’s nothing random about it.”
B: “But in order to change the numbers, you depend on a random event”
A: “There’s nothing random about it. The outcome is guaranteed.”
B: “Yeah, right. The outcome is guaranteed. If you draw the right numbers. And you get lucky with copying errors. Sounds random to me.”
A: “It isn’t random. Sure there’s random parts in the middle, but having random parts does not make for a random process. When you say ‘random’, you’re trying to say that you have to pick all the parts at once. This process isn’t random.”
B: “I’m not saying that. I’m saying that every part of this process is random. Even replaying the tickets is random. What if you lose a ticket now and then?”
A: “But it’s the exact opposite of random.”
B: “Sure. The problem is that at every step, you have a random error, but somehow, it’s supposed to get you closer to the winning ticket. Most of the time you lose winning numbers with a copying error, and when you do, you lose tickets. You don’t get any reward until all of them pay off.”
A: “Why do you keep saying random? The rules that determine the winner are deterministic”


It would seem that A above has the correct argument, but he’s hung up trying to insist that it isn’t random. It is random. In fact, the success of the algorithm depends on the randomness. If there were no copying errors, there would be no winning ticket. B recognizes the randomness involved, but is skeptical.

And maybe he should be…
B: “OK. I’ll buy a ticket. How long does the game run?”
A: “A year”
B: “A year! That’s not enough time! You have to get just the right number of random errors and you only have a year? There’s no way!”
A: “That’s plenty of time! And it’s not random!”
B “It is random. Are the copying errors guaranteed?”
A: “Theres’ got to be some.”
B: “But we don’t know when they will happen. That means random. Well, how often do they happen?”
A “I don’t know.”
B: “Huh?”
A: “I don’t know.”
B: “Then how do you know you’ll win? You have to get a certain number of random errors, and in a year, it has to exactly match the winning numbers. How do you know there’ll be enough errors that just happen to be made so that the ticket you get back is even better than the one you put in.”
A: “Because the guy in the big house on the hill won last year.”
B: “I saw him praying. Maybe God told him the numbers…”


And finally:

B: “So, where can I buy tickets?”
A: “I don’t know, but the guy with the big house bought one. There must be someplace around here to buy a ticket.

 

[jimbob]

Again, we seem to be stuck on the very definition of random. I do find it slightly interesting that there is a split between the engineers/physicists on one hand and biologists on the other.

The problem seems to be that, from the POV of the physicists, if identical inputs produce different outputs, then the process is random, and that is axiomatically so. A nonrandom process, by definifon means that identical inputs always produce identical outputs.

Chaotic systems are nonrandom, because although imperceptible differences in the starting conditions will eventually produce perceptible differences in the outcomes; identical starting conditions would produce the identical outcomes.

Randomly distributed values tend to conform to a Gaussian curve, Poisson distributions, and other statistical measures. We can state with extremely high degrees of certainty that, for example, the fitness of an impala is correlated with its peak ground speed. That some fast impalas die without reproducing and some slow imapalas get lucky, reproduce often, and die old, is a given. We are extremely confident when saying that their fitness is not random because randomly distributed fitness values would not have any relation to any traits.

But I would imagine that impalas *would* show a distribution according to their speed, quite possibly a poisson distribution, with very few abnormally fast and extremely few slow ones.

As Ivor said, just because a process is random doesn't mean that you can't model it, and make predictions.

It is not the tornado in a junkyard producing a 747, but a blind mechanic trying to assemble *any* form of transport, with many attempts, and an imperfect memory of the previous designs. After a billion attempts, a car, or an aeroplane might appear. Not a 747 but an aeroplane.

What are the chances of the mechanic producing an identical design from the same start? Nil.


What are the chances of the mechanic producing an aeroplane from the same start? Pretty high.​

 

[ImaginalDisc]

Again, we seem to be stuck on the very definition of random. I do find it slightly interesting that there is a split between the engineers/physicists on one hand and biologists on the other.

The problem seems to be that, from the POV of the physicists, if identical inputs produce different outputs, then the process is random, and that is axiomatically so. A nonrandom process, by definifon means that identical inputs always produce identical outputs.

Chaotic systems are nonrandom, because although imperceptible differences in the starting conditions will eventually produce perceptible differences in the outcomes; identical starting conditions would produce the identical outcomes.



But I would imagine that impalas *would* show a distribution according to their speed, quite possibly a poisson distribution, with very few abnormally fast and extremely few slow ones.

As Ivor said, just because a process is random doesn't mean that you can't model it, and make predictions.

It is not the tornado in a junkyard producing a 747, but a blind mechanic trying to assemble *any* form of transport, with many attempts, and an imperfect memory of the previous designs. After a billion attempts, a car, or an aeroplane might appear. Not a 747 but an aeroplane.

What are the chances of the mechanic producing an identical design from the same start? Nil.


What are the chances of the mechanic producing an aeroplane from the same start? Pretty high.​

I have no idea what you are talking about. Any system that deviates from a normal distribution is non-random.

P.S. Whats this "POV of a physicist?" Who cares? We're not discussing physics, we're discussing evolution. Evolution is studied by ecologists, geneticists, botanists, and biologists, not physicists. How is a physicist's POV relevant?

ETA: Ok, maybe "who cares" is asinine. However, the usage of "random" in physics is entirely irrelevant to the usage of "random" with regards to evolution. You don't mix jargon across scientific fields.

 

[CapelDodger]

I have no idea what you are talking about.

For what it's worth, it seems clear to me.

Any system that deviates from a normal distribution is non-random.

P.S. Whats this "POV of a physicist?"

Pretty much what you've just articulated, if I read jimbob accurately. Evolution is a process, not a system; normal distribution might be expected in a snapshot - of a species, or an entire ecology - but not across successive snapshots.

We're not discussing physics, we're discussing evolution. Evolution is studied by ecologists, geneticists, botanists, and biologists, not physicists. How is a physicist's POV relevant?

ETA: Ok, maybe "who cares" is asinine. However, the usage of "random" in physics is entirely irrelevant to the usage of "random" with regards to evolution. You don't mix jargon across scientific fields.

I understand jimbob as referring to mindsets rather than specific disciplines.

If I'm wrong (and it has been known) jimbob will put me straight.

 

[Dancing David]

The emotional appeal and the rational need not conflict. Surely there is truth in what you say, but it's an incomplete explanation, at best. Also, there's plenty of emotional appeal on both sides, and there's nothing wrong with that. Indeed, I think that the refusal to use the word "random" is primarily an emotional, not logical, appeal, but more on that later.

Quite so!

I mean that "A human is so complex it can not have evolved" and other creationist arguments that just appear intuitive, give people a privileged place in the universe, have an emotional appeal. And that is why they are hard to argue with.

Just as it is hard to argue with people that intelligence is not the pinnacle of creation, or language.

 

[articulett]

Addendum to the above post: You are succumbing to the "no middle ground" fallacy. Systems that exhibit uncertainty are not automatically random.

It's worse than that. He's changing definitions mid stream. It's like taking the theory of gravity asking how it's "non-random" upon learning that air resistance and the like can affect gravity. He's confusing the principle with incidentals and saying the incidentals makes the process random (or rather stochastic).

Evolution is "modification with descent" or "change through time" or "random mutations coupled with natural selection"--that's the principle--the formula. That's the process. Then there are the details...but all the details of what came before are already in the puzzle--the genomes that were the best at getting passed on and added to--did so in regards to the past. We can only predict some pieces of the future, but we can say for sure that every life form is still in the process of evolving or dying out (or both). It doesn't mean anything to talk about fitness unless it's fitness in regards to something. And the only fitness that matters in evolution has nothing to do with what humans find "fit". But mijo and walter and meadmaker seem to miss this. Fitness only means the bits of information that are best at getting themselves copied and added to and built upon. Evolution is about culling the random--refining and honing and building upon it and pruning away the obsolete--that process. The evolution of computer viruses "released in the wild' and the programs that evolve to stop them really are a pretty good model for understanding the basics. If you understand a little bit about how the internet works--you can grasp both it's complexity and how it came to be. I aim to do that when discussing evolution. And the more I participate on these forums and teach this subject, the more convinced I am that "natural selection" is the key to understanding.

 

[I less than three logic]

articulett-

From what I understand about numerical weather prediction, the difficulty in obtaining accurate long-term forecasts stems from the fact that the equations of fluid dynamics and physics used to describe meteorological systems cannot be solved exactly and that it is nearly impossible to obtain precise and detailed enough data. Thus, not only do we have to use approximate numerical models, but the problem of weather prediction is underdetermined in so far we do not have enough detailed data to make the detailed predictions that we want.

The same is true of celestial mechanics, another deterministic problem. The equations for an N-body system will always yield the same posterior positions a velocities given the same prior positions and velocities. However, as with the equations of fluid dynamics used in weather prediction, the N-body equations have not been solved exactly and the data used to predict posterior states of the N-body system are often not detailed enough. Thus, we have the same problems of approximation and underdetermination that we had with weather prediction.

Evolution, on the other hand, is fundamentally a statistical process, because it is possible to get different outcomes for the input. In other words, not all fit individuals survive and not all unfit individuals perish. Obviously, this is a different situation from that of numerical weather prediction and celestial mechanics, because, given exact solutions and detailed data, both weather prediction and celestial mechanics would yield the same results with the same initial conditions.

You seem to just be making unsupported assertions in place of arguments. What evidence do you have that given exactly the same input evolution wouldn't yield the same output like the weather. How do you know, giving the exact same input, that exactly the same creatures wouldn't live and mate, or get eaten, run over, etc? What definitively removes evolution from the realm of chaotic systems and places it in the truly random category?

You seem to be saying that if we went back in time and put every particle of the universe exactly where it was so long ago. That every thunderstorm or other weather that ever happened will happen exactly the same way, in the same place, at the same time as it happened the first time, but evolution wouldn't follow the same course this time around. Why? What evidence supports this? As far I know, the debate over whether or not the universe is determined has yet to be definitively resolved. Perhaps you can clear that up quick before moving along.

 

[I less than three logic]

I thought of an analogy.

Looks like you thought of a straw man to me. The analogy has nothing to do with evolution. There is no predetermined outcome that evolution is aiming for, there are no winning numbers.

 

[articulett]

Again, we seem to be stuck on the very definition of random. I do find it slightly interesting that there is a split between the engineers/physicists on one hand and biologists on the other.

The problem seems to be that, from the POV of the physicists, if identical inputs produce different outputs, then the process is random, and that is axiomatically so. A nonrandom process, by definifon means that identical inputs always produce identical outputs.

Chaotic systems are nonrandom, because although imperceptible differences in the starting conditions will eventually produce perceptible differences in the outcomes; identical starting conditions would produce the identical outcomes


But I would imagine that impalas *would* show a distribution according to their speed, quite possibly a poisson distribution, with very few abnormally fast and extremely few slow ones.

As Ivor said, just because a process is random doesn't mean that you can't model it, and make predictions.

It is not the tornado in a junkyard producing a 747, but a blind mechanic trying to assemble *any* form of transport, with many attempts, and an imperfect memory of the previous designs. After a billion attempts, a car, or an aeroplane might appear. Not a 747 but an aeroplane.

What are the chances of the mechanic producing an identical design from the same start? Nil.


What are the chances of the mechanic producing an aeroplane from the same start? Pretty high.​

The word random was part of the OP. Many have said that it's a bad word...at least by itself. You are very clear about the assembly process. Whatever words a person uses to convey evolution you need to emphasize that part--the assemblage through time. The question in the OP was about the ways evolution is not random. And if you are using the lay person definition of random or the random of the creationist conundrum--then you've got a word that implies "willy nilly"--without direction--a one time occurance--and thus the improbability of it all. But, as Dawkins writes so eloquently, you can climb Mount Improbable...incrementally--step by step, with each step preceding from there. I suppose in physics terminology, evolution might be a chaotic process--but as you said, that makes it "non-random"-- as the fly article I linked earlier notes.

Yes, it wasn't purposeful--but it was built on a succession of successful reproductions--not whipped up from nothing like a tornado whipping up a 747.
I understand the position of most "randomites"--I understand their meaning when they talk about the randomness involved. I don't think that mijo understands natural selection. And I'm not sure the physicists and engineers are understanding what "fitness" means in regards to evolution. It is only in reference to a particulary strand of DNA's ability to make more of itself. If it sounds circular--it's because that is how evolution works. Each internet connection and hub and router makes more connections possible...the synapses of our brain branch out like the branches on a tree-- To a biologist when you say "evolution is random" it's as vague as saying tree growth is random. The dying process is random. It just doesn't convey anything useful or essential to understanding and it misses the "assembly" through time--the blind watchmaker...the endless experiments...the few chosen to go to the "next step".

The definition used for random in random mutation is different than the definition used to declare "natural selection" a "random process"---and to mix the two definitions and then sum it all up with the uninformative phrase "evolution is random" or "evolution is a stochastic process" isn't likely to get anyone to understand evolution nor how complexity arises from "random" mutations.