What evidence is there for evolution being non-random?

[articulett]

You are confusing random marks on a page from random marks that look like punctuation and change the meaning of the words entirely, Mijo. A scratch on a CD disk might be random. But that random effect could be huge if the disk becomes unreadable because of it, or unimportant if all the info. is the same--or result in a subtle change in whatever it is the disc is "designed" to do in some way with consequences of some sort...

When you call evolution random you ignore that some randomness has an incremental or altering effect on the info. while most does not or destroys it completely. If you have people doing all sorts of tweaks on programs some tweaks are bound to improve the "program" and be incorporated...hence the evolution of the program. The info. evolves-- Windows 95 is still windows 95--Vista is the current evolutionary specimen in that "species"--

You truly are confusing the info. itself with the the vehicle carrying the info.

 

[articulett]

Cyborg, I have been chastized on this thread and another for accusing mijo of being an "intelligent design proponent", because he argues exactly like Behe, and can't seem to comprehend the difference between the blueprint and that which it "designs". But it could be, that he really is not able to understand... I think, that like Von Neumann, he's smart--but he's using his intelligence to purposely obfuscate or not understand a relatively simple point. Do you think this is the case? Or am I a bad girl for making unfair accusations?

He has started two threads with known creationist arguments and failed to absorb pages of detailed answers to his questions.

To me...mijo sounds like Egnor too...though I think Von Neumann is more like Egnor...
http://killtheafterlife.blogspot.com/2007/06/michael-egnor-is-retarded.html

 

[cyborg]

Cyborg, I have been chastized on this thread and another for accusing mijo of being an "intelligent design proponent", because he argues exactly like Behe, and can't seem to comprehend the difference between the blueprint and that which it "designs". But it could be, that he really is not able to understand...

Quite possibly. There is a strong emotional attachment to the idea that we are not in some way 'mere' machines. The very notion that we are really more like computers than gods upsets people and can lead to insane things like QM explanations for consciousness that verge on woo. (Oh and some computational arguments that annoy me more as a Comp Sci).

I think, that like Von Neumann, he's smart--but he's using his intelligence to purposely obfuscate or not understand a relatively simple point. Do you think this is the case? Or am I a bad girl for making unfair accusations?

Since I am fully aware of how an intelligent person, like myself, can screw up so badly but a lot of them aren't I wouldn't be surprised. It is an easy mistake to make to think you are correct about everything if you are correct about things most all of the time. I do it to but at least I am aware of my failings. That is the only thing that makes me better; otherwise I am only quite average.

Science should have taught us by now the value of being wrong and knowing when that is.

To me...mijo sounds like Egnor too...though I think Von Neumann is more like Egnor...
http://killtheafterlife.blogspot.com/2007/06/michael-egnor-is-retarded.html

Not familiar with him but interesting. Von Neumann probably would at least understand some of the computing stuff.

 

[cyborg]

When you call evolution random you ignore that some randomness has an incremental or altering effect on the info. while most does not or destroys it completely.

Actually this is another mistake - the total amount of information is irrelevant, it is the fact that it is changing that is important.

To use the computer analogy it is much like assuming that if you write more lines of program then that program must be better than one with fewer. It turns out that this is not the case at all - there are optimal ways of solving a computation and a lot of the time they're much smaller than the alternatives that also perform worse. There are lots of ways of writing a program that is equivalent to another one. This holds for DNA as much as it does for computing.

As such it is not how long your program is, it is what it does. When information in DNA changes - whether or not you are increasing or decreasing information - the important bit is that you are trying another experiment. Evolution can only work from random processes because of the large numbers involved. Another thing people have difficulty comprehending.

 

[mijopaalmc]

Because you are hung up on the fact that biological machines have their blueprints encoded using a base 4 chemical digital sequence and manufactured machines have their blueprints stored in base 2 electronic digital sequences and also on dead trees with symbols formed using chemical dyes. As such you are not comprehending that it is the change in the blueprint, not the expression of the blueprint, which is important here.

Well, I see you obviously taken lessons from articulett in argument mischaraterization. You seem not to want to acknowledge that technological objects are planned from their very inception as an idea by an entity that is capable of anticipating the consequences of its action before it takes those actions. Biological organisms are not assembled in this way. There is no idea behind their inception, no purpose to be fulfilled. It is only after the mutation has happened that it can be select for or against by the environment, and even then it is not because the environment "wants" the organism to have the mutation but because the organism wants to survive in the environment and can do so better with the mutation. Badly designed technologies, however, are thrown out because they do not serve the purposes of their designers or end users. Furthermore, no matter how diverse biological organism have become through evolution they are still based on the same basic information, DNA. Time pieces, though, are based on fundamentally different types of information. Their parts maybe analogous as in the pendulum in a pendulum clock and the spring or quartz crystal in a pocket or wrist watch, but as these part actually represent the information of the time piece, they are not analogous role to DNA in living organisms because these parts of the respective time pieces are fundamentally different structures*. (Yes, I too know how to analogize; I am calling your analogy false because the point of correspondence on which you are basing it namely information is in fact not a point of correspondence between the alleged analogs, as explained in the immediately preceding sentence). I do realize the expression of such information with reference to time pieces is identical because they all that the time, but that to is the exact opposite of how organisms function because the information in the DNA of organisms is expressed in extremely diverse ways.

*If I have misapprehended what constitutes the information in time piece, I welcome alternatives be proposed.

 

[cyborg]

Well, I see you obviously taken lessons from articulett in argument mischaraterization. You seem not to want to acknowledge that technological objects are planned from their very inception as an idea by an entity that is capable of anticipating the consequences of its action before it takes those actions.

HOW they are designed is IRRELEVANT. The same design is the same design.

There is no idea behind their inception, no purpose to be fulfilled.

A car does not need ideas in order to move. It moves because of its construct - not who constructed it.

It is only after the mutation has happened that it can be select for or against by the environment, and even then it is not because the environment "wants" the organism to have the mutation but because the organism wants to survive in the environment and can do so better with the mutation.

You cannot say your plane will fly until you build it.

Furthermore, no matter how diverse biological organism have become through evolution they are still based on the same basic information, DNA.

Irrelevant. I am also limited if I don't allow myself to use a certain subset of chemistry in building a machine.

Time pieces, though, are based on fundamentally different types of information. Their parts maybe analogous as in the pendulum in a pendulum clock and the spring or quartz crystal in a pocket or wrist watch, but as these part actually represent the information of the time piece, they are not analogous role to DNA in living organisms because these parts of the respective time pieces are fundamentally different structures*.

UGH. Why should I not be surprised that you are comparing the wrong information? The design, on paper or in machine, is the blueprint. When you copy an existing idea the original does not become a blueprint - it becomes a master. What we are dealing with are the abstractions for the thing, not the thing itself.

A strand of DNA is just a strand of DNA. It is only an abstraction, an encoding, for a lifeform. Until you 'run' the DNA program you don't know what you're going to get because then the program is subject to being interfered with by the environment - nature vs nurture. Phenotype vs genotype.

We are talking about the abstractions, which change, not the expressions, which don't.

A monkey cannot become a man - a sufficiently rewritten monkey program can.

Yes, I too know how to analogize;

But you did poorly at it.

 

[Schneibster]

They win because the pay off is not 1 to 1 -- the results are still random and disordered and constrained within the system that they are in. Each result is (relatively) equally likely--but we aren't using that definition of random, are we. What's the definition you're using again?

You sure you have a good handle on the subject under discussion here? I think you just said that king high is just as likely as a royal flush.

 

[Ivor the Engineer]

You sure you have a good handle on the subject under discussion here? I think you just said that king high is just as likely as a royal flush.

Articulett was correct: her reply was in relation to my examples of a roulette wheel and a lottery ball picker. Poker was _not_ mentioned.

Play fair.

 

[Ivor the Engineer]

<snip>
As I said previously that is a priori selection. You still need to build the thing and one still needs to deal with the realities of the limits of design with respect to the abstractions used. Designs are wrong all the time. Anybody who thinks any successful engineering project can be neatly divided into planning, implementation and testing is simply wrong: doing it that way has been shown time and time again to fail to achieve the stated goals. This is iterative design; it is evolutionary in nature.

Stated like someone who's actually done some real engineering

 

[articulett]

You sure you have a good handle on the subject under discussion here? I think you just said that king high is just as likely as a royal flush.

He mentioned the lottery and roulette--If the odds are even, if the chances of winning are 10 to 1 then you would win ten dollars for every one dollar played...but in Vegas, you would win $9.00, and the house would keep a dollar--you are paying for the chance of winning.

In blackjack the you win twice what you paid, but because a tie goes to the house...the house has the favor.

 

[mijopaalmc]

Articulett was correct: her reply was in relation to my examples of a roulette wheel and a lottery ball picker. Poker was _not_ mentioned.

Play fair.

Uh....this from Ivor "if you change the limits of integration the moments become finite" the Engineer?

 

[mijopaalmc]

Could someone please define "information"?

If were are discussing "information" as it pertains to information theory, information theoretic quantities are all defined by random variables and probability distributions. In other words, discussing evolution in terms of a decrease in mean information entropy, which is, as I recall, the original topic of Annoying Creationists or at least the the authors of ev define evolution of binding sites, because you can't divest yourself of the fact that information theory is grounded in probability theory.

 

[articulett]

Actually this is another mistake - the total amount of information is irrelevant, it is the fact that it is changing that is important.

To use the computer analogy it is much like assuming that if you write more lines of program then that program must be better than one with fewer. It turns out that this is not the case at all - there are optimal ways of solving a computation and a lot of the time they're much smaller than the alternatives that also perform worse. There are lots of ways of writing a program that is equivalent to another one. This holds for DNA as much as it does for computing.

As such it is not how long your program is, it is what it does. When information in DNA changes - whether or not you are increasing or decreasing information - the important bit is that you are trying another experiment. Evolution can only work from random processes because of the large numbers involved. Another thing people have difficulty comprehending.

Yes. Just because a recipe has a lot of ingredients doesn't mean it's more delicious. And oak trees have more DNA and chromosomes than people. But complicated genomes have evolved and the the length of genomes has changed via many processes--non-disjunction, insertions, translocations, methylation, etc. I just pointed out the additive part because in some ways chromosomes are like communities of genes and cells are like communities of organisms that allow different cells to differentiate... a bundle pack of software might be akin to a community of humans or ants-- information that attracts or incorporates other information can form symbiotic relationships if it increases the replication. Eukaryots (multicellular organisms) began via the fusion of prokaryots. Generally extra stuff in a genome isn't good... (an extra chromosome), but on occasion it can confer a benefit. And though genomes don't really evolve one base pair at a time, they have clearly grown over the eons. With molecular genetics we can SEE what changed between two organisms and their last common ancestor. We even have genes that are "installer packages" that are used in fetal life, but never again.

 

[Schneibster]

INFORMATION is what is evolving--"design instructions"--not the vehicles that carry them. You die the species you are. The airplane doesn't become another airplane--the information about building prior airplanes is built upon based on what worked.

Absolutely. See, this is why I don't understand why you don't get "random." Random implies that everything will eventually develop- if you apply a sorting algorithm, you'll get something interesting. Just about any sorting algorithm will do- for example, a bias in the probabilities favoring certain interactions over others yields classical physics from the randomness of quantum mechanics. For another and much more germane example, natural selection.

BOTH the randomness AND the probabilistic biases are required- if there is no randomness, there's nothing to select from, and if there is no selection, you get chaos at the high level as well as the low.

I've seen arguments that DNA evolved the way it did because it provides just the right amount of random input. Any more, and you get too many products that are non-viable in ANY environment. Any less, and you don't get enough new characteristics to be likely to be able to deal with environmental contingencies that modify the fitness landscape. Kauffman makes substantially that argument, but extends it further down into chemistry than my presentation here does.

INFORMATION. You guys just don't seem able to distinguish information from the vehicle it is contained in. DNA is information. Blueprints are information. This thread is evolving because the information is being added to, built upon. I guess this is just not comprehensible to some people while being obvious to others. "Hone as you go" is a lot different than random combinations resulting in sudden "winners".

The fact that it's obvious- and equally obvious, at least to me, that without randomness no information exists to be selected from- is a core point in my argument that there cannot be evidence that evolution is non-random: without the randomness, there'd be nothing to evolve.

The fact that BOTH are obvious to anyone who's taken the time to think about it is also a core point in my argument that it is negative information to state that evolution is not random to people who know better.

Cyborg, you may save us yet.

 

[articulett]

Could someone please define "information"?

If were are discussing "information" as it pertains to information theory, information theoretic quantities are all defined by random variables and probability distributions. In other words, discussing evolution in terms of a decrease in mean information entropy, which is, as I recall, the original topic of Annoying Creationists or at least the the authors of ev define evolution of binding sites, because you can't divest yourself of the fact that information theory is grounded in probability theory.

a recipe is information; a book or conveyance of some sort is the vector, and the result is what can be made from the recipe.

The word "stop" is information. A stop sign is the vector. Cars stopping is the result.

Marking the rotation of the earth is the blueprint behind all time keeping devices. Sundials were a result and watches evolved to do the same purpose--and then the blueprint evolved further and was honed to keep time more precisely and easier so that all people could be on the same "time"-- Those making sun dials weren't designing today's atomic clock--but the first "atomic clock" designer didn't bring forth the notion of timepieces from scratch.

Yes...humans are designing...but it's design as you go. Nature does the same, but the design just is what can get replicated--what reproduces best...what survives environmental impacts...if you've got a good way for the information to be passed on--and an imperfect copying system in place, you've got a good chance of "being selected" by natural selection.

Information changes "randomly", but it acts in vectors (organisms) and the environment interacts with these vectors, and some survive (are selected)...and some of those that survive make lots of copies of their info.(further selection) and some of this progeny is very successful at making progeny (and yet more selection).

 

[Schneibster]

UGH. Why should I not be surprised that you are comparing the wrong information? The design, on paper or in machine, is the blueprint. When you copy an existing idea the original does not become a blueprint - it becomes a master. What we are dealing with are the abstractions for the thing, not the thing itself.

A strand of DNA is just a strand of DNA. It is only an abstraction, an encoding, for a lifeform. Until you 'run' the DNA program you don't know what you're going to get because then the program is subject to being interfered with by the environment - nature vs nurture. Phenotype vs genotype.

Your CS training and experience tells you the same things mine tells me. "The map is not the territory," or more properly, "the source code is not the running program." To put this another way, you can design and code a wonderful program- which then fails to work when you attempt to compile and run it.

I think, though, that where the randomness comes in is, if you think of DNA as a program, then it's kind of like a program in memory that's subject to random alterations of the machine code. Most such alterations result in uselessness, core dumps or race conditions or whatnot, but a few would be improvements, and if a constraint is imposed, then the useless ones would be winnowed out over time and the improvements would accumulate. That's how I see evolution. The point of the thread being, without the random alterations, there'd be no improvements. And the argument of the cretinists being, how can there be improvements unless someone designs them- which is ridiculous, if you actually understand what "random" means.

 

[articulett]

The fact that it's obvious- and equally obvious, at least to me, that without randomness no information exists to be selected from- is a core point in my argument that there cannot be evidence that evolution is non-random: without the randomness, there'd be nothing to evolve.

The fact that BOTH are obvious to anyone who's taken the time to think about it is also a core point in my argument that it is negative information to state that evolution is not random to people who know better.

Cyborg, you may save us yet.

Nobody is doubting there is randomness. The randomness is obvious. But evolution is "random mutation" with natural selection. Natural selection is is far different in regard to randomness. It is determined by the quality of the programs exhibited in the pool of randomness. If the program (DNA) in the vector confers advantages in whatever environment it finds itself in, then the organism just may escape the brutal conditions of life long enough to make some copies of the program that runs it and some new vectors that will carry the program into the future whereby randomness can have a chance to act again on the information. It really IS different than other random processes scheibster...different than the spiral-ness of galaxies--because the information replicates...nothing in physics does.

I don't think anyone is saying evolution is non-random...just that it's not explanatory to call it random, because selection is just not a memoryless process it is non-random in comparison with the relative randomness of mutation (which is not really even completely random)...it's determined entirely by what there is to choose from...and built on a chain of such successful selections. That is all Dawkins is saying.

The accumulations you speak of, schneibster...how would they accumulate--they would be selected...you'd run the programs in vectors and the changes would accumulate via human selection...or computer selections (not randomly!)--nature does such selecting too--on a much larger scale...and the stuff that "works" can often make tons of copies of itself any of which can contain a mutation or useful recombination.

Without selection you only have your original pool of randomness to pick from.

 

[Schneibster]

Could someone please define "information"?

In what context? What is "information" in one context is "noise" in another.

If were are discussing "information" as it pertains to information theory, information theoretic quantities are all defined by random variables and probability distributions. In other words, discussing evolution in terms of a decrease in mean information entropy, which is, as I recall, the original topic of Annoying Creationists or at least the the authors of ev define evolution of binding sites, because you can't divest yourself of the fact that information theory is grounded in probability theory.

I don't think this argument's got legs. You're ignoring the environment, which is necessary to provide a context in which noise becomes information. Raw DNA floating around in a chemical soup is just another chemical. Inside of a cell, it becomes a template for the production of proteins. So, which is it, just another chemical, or a template- i.e., information? Only the context can tell you that.

 

[mijopaalmc]

In what context? What is "information" in one context is "noise" in another.

I don't think this argument's got legs. You're ignoring the environment, which is necessary to provide a context in which noise becomes information. Raw DNA floating around in a chemical soup is just another chemical. Inside of a cell, it becomes a template for the production of proteins. So, which is it, just another chemical, or a template- i.e., information? Only the context can tell you that.

Maybe I jumped the gun, but I do think that it is interesting that when one tries to come up with a quantitative measure of information rather than a qualitative description, which is what articulett has done, uncertainty and probability become essential components of the concept of information, regardless of the context.

 

[articulett]

Maybe I jumped the gun, but I do think that it is interesting that when one tries to come up with a quantitative measure of information rather than a qualitative description, which is what articulett has done, uncertainty and probability become essential components of the concept of information, regardless of the context.

I'm not sure what you are talking about...as mentioned before, more isn't better... DNA is coded in 4 base pairs which are arranged in various ways to make all the living things we know of...just like 8 basic notes are the basis of all music...and the 26 letters of the alphabet make all English words-- But more words or more notes doesn't make books or music "better". The 4 base pairs are "read" in triplets which code for amino acids--some are redundant, though they can change the resulting folding of a protein without changing the protein itself. The proteins are coded in genes which are turned on and off at various times due to the instructions also in the DNA and spliced per instructions also--This creates the living organism and it's evolution towards maturity--maybe one day to make copies of it's own DNA--maybe even many copies.

You are getting confused like Kleinman now. The genome wasn't built up one single base pair at a time. And a single base pair can have a huge effect (it can cause dwarfism, for example)--or make a complete frameshift with a new protein. But genomes only get to change if their vector gets to reproduce--if their vectors are selected by being able to survive and reproduce successfully. You don't get to be a part of the "randomness" until you've been selected from a long line of successful replications.

If you were making copies of a poem on a copy machine, random dots or ink specks that didn't affect your ability to read the poem, would be noise. But if there was a dot that looked like a period, and somehow changed the meaning of the poem--that would be randomness that changed the information. If the poem was made better by that bit of randomness and the poet kept the error because it read better--that would be selection. Nature selects too. If the change confers any survival and replication advantage in the vector carrying it...it can escape the many trials nature has in store.