What evidence is there for evolution being non-random?

[articulett]

articulett, what is so uninformative about describing something as "[o]f or relating to a type of circumstance or event that is described by a probability distribution" other than that you think that it describes anything an is therefore meaningless?

Um...I was a genetic counselor for several years, and we don't tell people that your chances of having a kid with Down Syndrome is related to a probability distribution.

In the article about the nozzles, it would have been silly to sum up the article as, "as you can see, the nozzle evolved due to being related to a probability distribution...therefore, it evolved randomly".

It just doesn't convey any information. It doesn't distinguish the probability distribution in mutation from the very different probability distributions you see upon selection (elimination rounds). It is the same as saying "scientists saying that all we see just came about by chance." It's uninformative, and leaves out the most important part of what we KNOW about evolution. No scientist thinks this all came about JUST by chance. Natural selection played a vital role and creationists go out of their way to garble understanding very similarly to the way you do.

How informative is the word or term if it applies equally to a junkyard 747 and the nozzle example?

And why do you ask questions and then never show any interest in the answer while, at the same time, ignoring all questions directed at you?

 

[cyborg]

You're not honestly saying that because I define the environment of two adjacent unhatched birds as being the same that I must define their environment as the same as a fishes are you?

Either I'm allowed to make location an important factor or I'm not.

 

[jimbob]

Even the article didn't say natural selection can be described by a probability curve--why? Because such a statement is uninformative. They showed the actual probabilities.

I would say it is because it is a trivially obvious statement, given the rest of the article. Not because it is wrong. I thought people were arguing that the premise (of probabilistic selection) was wrong.

My point is that if one accepts that selection is probabilistic, which is synonomous with random, one can then discuss/analyse the odds and begin to quantify how large a selective (dis)advantage is, in terms of probabilities. When genetic counseling, I imagine that one discusses the probabilities of the various scenerios, and their implications.

Because the odds of any individual small fry reproducing are well under 0.001%, there were selctive pressures to have broods in the millions. Whether any individual initially healthy cod spawn lives to reproduce is almost entirely "luck".

It just doesn't convey any information. It doesn't distinguish the probability distribution in mutation from the very different probability distributions you see upon selection (elimination rounds). It is the same as saying "scientists saying that all we see just came about by chance."

How about "mutation is random, and selection is probabilistic"?

The fact that one is using two seperate words there should imply that there is a difference between the two.

With the 747 in a junkyard, one could make the point that there has been no selection.

 

[articulett]

It's vague Jim Bob...but I give up.
To be selected in DNA is just to be copied...that's it. The first hurdle you must pass to be copied is to produce an organism that can make copies of DNA--If you can't or don't do that--you are the millions of sperm that never got close an egg. Then the organism you make must also make copies of DNA that will survive after the organism dies-- Being copied is being selected-- it doesn't matter who or how or how much luck or randomness plays a role--it only matters that DNA is selected. If anything in that selected DNA conferred traits that aided in it's selection process, those traits are passed on.

You just don't seem to really "get" it. And it's so simple. Your example doesn't distinguish the evolution of nozzles from the evolution of animals--

The randomness you speak of in an environment is "noise"--nothing in the environment matters except that some DNA somewhere escapes into a future organism before the current vector dies.

DNA that is selected is merely DNA that is copied. All living organisms are the results of passing at least trillions of rounds of successful copying (including successful mutatations) in DNA. The DNA has been through the selection sieve again and again. As in the nozzle example...what could work is not relevant--only the modifications that actually gets made get to enter the selection process. And only those that outperform the parent generation are further selected in the example. (in DNA only the DNA that makes organisms enter the selection process and only those that make organisms that reproduce are further selected)

I think the nozzle example was a great analogy to evolution and the words used were pretty clear. I personally don't see the point of using random or "it can be described by a probability distribution" because it's unclear. Either tell us what the probability distribution is or find a better way to describe natural selection if you want someone other than yourself to understand it-- find a way so it doesn't sound like the creationist canard "scientists think all this complexity arose just by chance"-- we don't think that. It's much more than chance. It's natural selection which hones organisms in the pool of randomness.

 

[articulett]

mijo, walt, and jim-bob are doing something weird. You are taking an artifact of the selection process and using it to define selection in an ambiguous way. It's as if you are pointing out that there are lots of soap nozzle modifications that could have been made but weren't...and that some of the modifications made could have been better if kept...and the engineers may have inadvertently thrown out some promising designs--therefore the nozzles were selected randomly!

The modifications were made to the nozzles randomly, they were selected based on what works. That's simple and all you need to know...the tidbits above are not relevant to understanding the process. Neither are all your examples and your views about what is and isn't "fit". Fitness just means a DNA strands ability to get copied. The more it does so, the more fit it is, and the more potential mutations there are likely to be--some of which could end up furthering future vectors replication success.

 

[articulett]

If you really want to see why it's not useful to describe natural selection with the terms you are using... plug in your term or modifier every time you see the word selection in this article. Nobody is saying that selection is non-random, but they go out their way to distinguish it from the common understanding of random. The terms and ways you want to describe evolution (mijo) are just so poor and uninformative in response to the ways others describe it. I can't imagine where or how it would be useful and not misleading.

http://www.nytimes.com/2007/06/26/science/26human.html?pagewanted=1

Are these people misinformed? Are you saying they should say, lactose tolerance evolved "randomly"? or "Skin color evolved through a process that can be described by a probability distribution"? You seem to have a need to sum up the entire process of evolution as random when one part of evolution is far closer to true "randomness" (thought still not completely random) and the other part is much closer to determined because it is entirely determined by what is replicated.

Another article...and so much easier to understand than those who want to call natural selection "random": http://www.nytimes.com/2007/06/26/science/26lab.html

 

[jimbob]

Articulett, I have probably not conveyed what I meant quite well enough.

With natural selection, because nature is "red in tooth and claw", only a small fraction of those "otherwise fit" survive to reproduce. I would argue that, as (I think) you do, that "fit" is a slightly poor concept. But the workround that I prefer would define "fitter" as having slightly better odds of reproducing .

I am not defining the probability distribution because it will be different for every trait/environment, but a hypothetical one could look like this

[img="http://www.internationalskeptics.com/forums/vbimghost.php?do=displayimg&imgid=6577"]http://www.internationalskeptics.com/forums/vbimghost.php?do=displayimg&imgid=6577[/img]

If the y-axis is odds of reproducing, and the x an arbitary, quantified trait, then in the middle of the trait the chances are best (actually, 100%, due to my poor choice of functions) at the extremes, the probability is zero.

This modulates the selection of the random mutatons, as the seed population for the second generation would look like that multiplied by a normal distribution (assuming the trait varies normally).

The next seed population would look even more so, as it its starting population would begin to follow the curve...

With artificial selection, this probability plot would look more rectangular, a step function.

 

[articulett]

Articulett, I have probably not conveyed what I meant quite well enough.

With natural selection, because nature is "red in tooth and claw", only a small fraction of those "otherwise fit" survive to reproduce. I would argue that, as (I think) you do, that "fit" is a slightly poor concept. But the workround that I prefer would define "fitter" as having slightly better odds of reproducing .

I am not defining the probability distribution because it will be different for every trait/environment, but a hypothetical one could look like this

[img="http://www.internationalskeptics.com/forums/vbimghost.php?do=displayimg&imgid=6577"]http://www.internationalskeptics.com/forums/vbimghost.php?do=displayimg&imgid=6577[/qimg]

If the y-axis is odds of reproducing, and the x an arbitary, quantified trait, then in the middle of the trait the chances are best (actually, 100%, due to my poor choice of functions) at the extremes, the probability is zero.

This modulates the selection of the random mutatons, as the seed population for the second generation would look like that multiplied by a normal distribution (assuming the trait varies normally).

The next seed population would look even more so, as it its starting population would begin to follow the curve...

With artificial selection, this probability plot would look more rectangular, a step function.

Maybe so. I'm guessing artificial selection like the nozzle example would follow a sort of S curve...slow to start and then "exponential" improvements and then a leveling off. But I think you understand why biologists go to great lengths to distinguish the relative randomness of mutation from random components that affect selection. Mutations are relatively random, the ones that end up copied widely is not. If some DNA is copied widely its because it, either conferred an advantage to it's vectors or it or was passed along in genomes that had some other very "reproductively fit" genes.

 

[steenkh]

With natural selection, because nature is "red in tooth and claw", only a small fraction of those "otherwise fit" survive to reproduce. I would argue that, as (I think) you do, that "fit" is a slightly poor concept. But the workround that I prefer would define "fitter" as having slightly better odds of reproducing .

I think that it is wrong to think that the selection is based on what is "fit" or "fitter". A more accurate description is that selection is weeding out those that are less fit. Everything that is neutral or "fitter" is left in place. Decidedly non-random in my mind. But I am also of the old-fashioned kind who believe that loaded dice are non-random.

 

[articulett]

I think that it is wrong to think that the selection is based on what is "fit" or "fitter". A more accurate description is that selection is weeding out those that are less fit. Everything that is neutral or "fitter" is left in place. Decidedly non-random in my mind. But I am also of the old-fashioned kind who believe that loaded dice are non-random.

Same here! And I'm sure any Vegas judge would agree.

 

[jimbob]

I think that it is wrong to think that the selection is based on what is "fit" or "fitter". A more accurate description is that selection is weeding out those that are less fit. Everything that is neutral or "fitter" is left in place. Decidedly non-random in my mind. But I am also of the old-fashioned kind who believe that loaded dice are non-random.

I would reverse what you say: Everything that is worse is almost certainly weeded out, but so is a lot of otherwis fit (hence the 4million cod-spawn, to have sufficient odds of two surviving.

It is a perfectly effective strategy, the genes do get passed on, but more than 99.999% fail

ETA: here is the sort of plot I wanted to show

[/IMG]

 

[steenkh]

I would reverse what you say: Everything that is worse is almost certainly weeded out, but so is a lot of otherwis fit (hence the 4million cod-spawn, to have sufficient odds of two surviving.

It is a perfectly effective strategy, the genes do get passed on, but more than 99.999% fail

I do not agree with you, because we are talking about separate things. I talk about the weeding out of genes, and you talk about the weeding out of individuals. Statistically, only unfit genes are weeded out. Neutral mutations are the most numerous, and they invariably survive.

With individuals there is no telling who survives and who do not. If any of the individuals have worse genes than the others, selection will ensure over the eons such individuals will not survive.

 

[mijopaalmc]

I do not agree with you, because we are talking about separate things. I talk about the weeding out of genes, and you talk about the weeding out of individuals. Statistically, only unfit genes are weeded out. Neutral mutations are the most numerous, and they invariably survive.

With individuals there is no telling who survives and who do not. If any of the individuals have worse genes than the others, selection will ensure over the eons such individuals will not survive.

It may be true that only genes are weeded out, but this is not done in a deterministic fashion. In other words, an unfavorable allele does completely disappear all at once. In fact stochastic modeling of evolution by natural selection (Rouzine and Coffin 1999, Rouzine et al 2001, Rouzine et al 2002, Rouzine and Coffin 2005) shows that if the reversible mutation rate is greater than zero and the effective population size is greater that then inverse of the mutation rate, the frequency of that allele will "settle down" into a "steady state" where it is equal to the ratio of the mutation rate to the selection coefficient. Thus, the unfavorable allele will never be fully weeded out of the population. When the effective population size is smaller that the inverse of the selection coefficient, however, the frequency of the unfavorable allele will fluctuate unpredictably between complete fixation and complete absence.

The research I am citing here was done on haploid asexually reproducing populations so it may not be directly extensible to vertebrate evolution, but it none the less provides an example o how evolution is governed by probability distributions and random variables and how it is useful to describe it as such.

 

[jimbob]

Maybe so. I'm guessing artificial selection like the nozzle example would follow a sort of S curve...slow to start and then "exponential" improvements and then a leveling off. But I think you understand why biologists go to great lengths to distinguish the relative randomness of mutation from random components that affect selection. Mutations are relatively random, the ones that end up copied widely is not. If some DNA is copied widely its because it, either conferred an advantage to it's vectors or it or was passed along in genomes that had some other very "reproductively fit" genes.

Yes, the sigmoid curve is one of the very common curves that crops up all over the place in biology.

The point is that if this is the probability of reproducing, in this enviromnent for this trait, it would soon migrate to the optimum via a statistical process.

I think that is very similar to what you are saying articulett?

With artificial selection, which is still modulated by nartural selection, these probabilities would be further modulated by a (near) step function) so the relative odds of survival for the desired traits would be almost infinitely greater than for the undesired traits. Farmers are very good at ensuring that their cattle are served by prize bulls and that "runts" don't breed.

The mutation part is fine for evolutionary algorithms and sleective breeding, but the selection is far more rigorous in terms of the desired traits, which had been decided upon before breeding

Artificial selection is a good starting point, but the only goal is reproduction, and as many strategies as species...

I just think that one has to be careful how far to take the analogy.

I suspect that my understanding of the process isn't actually too far from yours, articulett. My emphasis differs slightly.

Jim

ETA:

I want to keep away from any (possibly attractive to IDiots) idea of a "designer/farmer" creator "overseing" evolution and selecting according to a set of criteria, with a (divinely chosen) goal in mind

 

[articulett]

I do not agree with you, because we are talking about separate things. I talk about the weeding out of genes, and you talk about the weeding out of individuals. Statistically, only unfit genes are weeded out. Neutral mutations are the most numerous, and they invariably survive.

With individuals there is no telling who survives and who do not. If any of the individuals have worse genes than the others, selection will ensure over the eons such individuals will not survive.

You've hit the nail on the head. It's just the code that is or isn't selected. In order to even begin the elimination rounds the DNA has to copy itself into a vector that can copy the DNA.

But some people have a strong need to call natural selection "random" (or probabilistic) so they can sum up evolution as a "random process". I think some are doing this because they really don't understand that it's information that is mutating and must successfully form a vector (organisms) which the environment acts upon in a series of elimination rounds.

And others specifically do NOT want to understand and will use the most tortured logic possible to conclude that evolution IS random. (Note the continual use of the word "fitness" or "advantageous" or "favorable")--Mijo just can't understand that fitness only refers to being able to be copied. The reasons some information fails at getting copied is irrelevant as to evolution--only what IS selected counts. To me, it's like he wants to call the evolution of nozzles random because there were some designs the engineers never tried. Information that doesn't result in a vector (organism, nozzle) can't be passed on. It could be the best information in the world, but the physical environment cannot act upon it nor can it go through any elimination rounds until it produces something. The environment can't react to the information until it creates an organism.

To me it seems so simple, and all the links I've provided describe it so simply. According to mijo's tortured logic, the nozzle evolution is random as is the evolution of everything and all processes are random as well. Why, everything is random. And everything is a noun. But that doesn't mean that much information is being conveyed in saying so nor does it mean that random is a noun. (Truly, that's how he sounds to me; I wonder if anyone understands his point.)

 

[mijopaalmc]

articulett-

Instead of insulting me, you should try to explain why I am wrong. Organisms compete for limited resources and the ones that are more likely gain access to those resources are the ones that copy themselves more often. Because there is a range of phenotypes where organisms can compete for resources to varying degrees, it is not certain that an two organisms that have identical sets of genes that make it easier for them to compete for resource will survive. That, by definition, is why evolution is a random process.

 

[articulett]

articulett-

Instead of insulting me, you should try to explain why I am wrong. Organisms compete for limited resources and the ones that are more likely gain access to those resources are the ones that copy themselves more often. Because there is a range of phenotypes where organisms can compete for resources to varying degrees, it is not certain that an two organisms that have identical sets of genes that make it easier for them to compete for resource will survive. That, by definition, is why evolution is a random process.

There is no evidence that you are capable of understanding the answers to the questions you ask...not in any of your postings. Answering your questions or responding to you is identical to talking to a creationist. This is true in every post of yours. Your definition of a random process is so loose that everything is a "random process"--hence it's a useless definition. You are absolutely clueless when it comes to natural selection and purposely so. You just don't "get it" because you mistake the "noise" in the selection process with the process itself. You cannot and do not distinguish between the randomness of mutation and random components in the selection process--you just define it all with the same loose definition and then sum it all up as random. Your definition of random is as obfuscating as Behe's--as is your failure to understand all attempts at getting you to understand why biologists say that selection is not random.

Where is there any evidence that you have engaged in constructive dialogue anywhere on this forum? Where do you seem interested, even once, in the answer to your smarmy questions? What links provided have you gone to and commented on and how do they compare to your whining about nobody commenting on your links in a way that satisfies you? You have shown no ability whatsoever to understand anything in any thread except the point you are trying to prove. You are pedantic, garbled, and clueless and never really say anything. You don't make any coherent sense. You move the goalposts as you go.

I understand your example...it's identical to the nozzles were selected randomly, because not every possible change that could be done to a nozzle was even tried in a nozzle, and something could have been eliminated that worked with another mutation further on down the road. What nutcake other than you would define the nozzle selection process as random? It's certainly not of the same ilk as the "random" changes made to the nozzle.

Saying the nozzles evolved randomly is as misleading as saying natural selection is a random process. Saying evolution is random because natural selection can have random components is as misleading as saying Algebra is random because it contains random variables.

 

[articulett]

And identical DNA doesn't make entities identical--the DNA is only fitter if it can survive and replicate in whatever environment it finds itself in. Whenever you have identical DNA--then it's progenitor is fitter--it made multiple copies of itself some of which may make entities that survive into the future. Just like the millions of sperm per ejaculate--most of life is extremely wasteful--because abundance assures that something is likely to survive and get passed on. Fitness of all the other sperm that didn't make you is irrelevant tot he one that did. It's fitness is irrelevant because it never got to make an organism. Quit thinking that you know what fitness is, because that's what's making it impossible for you to understand natural selection. Fitness just refers how many replications are made (the more, the more likely that mutations will occur...the more mutations that occur, the more "beneficial" mutations exist in the genepool of successive generations. Fitness in regards to genomes only refers to the ability to get passed on. External randomness events are expected parts of selection--we EXPECT the environment to do all it can to cull and abort beginning life...it's expected that most life forms will never replicate...never copy their genomes...but just a small percentage needs to do so to keep the game going. Nobody expects all identical genomes to survive...it's a percentages game--if you have high survival rates, you can make low numbers of offspring, if you have poor survival rates, you need to make a lot more offspring for any to survive.

The "randomness" you keep trying to include in "natural selection" is already worked into the equation. Everything in the physical environment that acts on DNA vectors--including other DNA vectors--shapes what is selected and what isn't. What we think of as fit is only relevant if it is copied. It turns out, if you're a gene essential to organism building, you get copied a lot--because failure to do so make an organism unable to survive. So the info. for a spinal column only has to happen once...and from there all future surviving descendents will have the template. They won't all be selected, but spawn with a lack of such DNA won't even be able to make an organism--much less one that replicates.

 

[cyborg]

Put as simply as possible: you can't decide fitness first. Nature defines what is fit, not you.

 

[mijopaalmc]

So, articulett, if you think "randomness" has already been "worked into the equation" of natural selection, why do you resist calling natural selection random?