Evolution Not Random

[sol invictus]

It's really simple:

deterministic mutation + deterministic selection = detereministic evolution

deterministic mutation + random selection = random evolution

random mutation + deterministic selection = random evolution

random mutation + random selection = random evolution

Every single statement in your post is meaningless unless you define "random".

If you define random as "[o]f or relating to a type of circumstance or event that is described by a probability distribution", as you did above, every single statement in your post is meaningless, because that (silly) definition implies that everything is random.

Did you have a point?

 

[articulett]

Every single statement in your post is meaningless unless you define "random".

If you define random as "[o]f or relating to a type of circumstance or event that is described by a probability distribution", as you did above, every single statement in your post is meaningless, because that (silly) definition implies that everything is random.

Did you have a point?

That IS his point. It always has been. Sometimes he'll cede a wee bit and say "evolution is not non-random" which is about as clear as saying "evolution is random per Mijo's definiton of random".

 

[mijopaalmc]

Every single statement in your post is meaningless unless you define "random".

If you define random as "[o]f or relating to a type of circumstance or event that is described by a probability distribution", as you did above, every single statement in your post is meaningless, because that (silly) definition implies that everything is random.

This is exactly why I say that this discussion just highlights the ignorance of those who argue that evolution is non-random, especially on the topic of probability theory. The definition of "random" as "[o]f or relating to a type of circumstance or event that is described by a probability distribution" does not make everything "random" because it refers to a very specific class of mathematical function (i.e., probability distributions), which are themselves defined in a very specific way. In particular, differential equations that do not take probability distributions as their argument, such as those used to describe weather (and, in a larger sense, fluid dynamics), cannot be described by a probability distribution and are therefore not random, making weather (and, in a larger sense, fluid dynamics) a deterministic (although chaotic) system.

 

[Walter Wayne]

Even this wouldn't be random if you could model the exact trajectory that you threw the dice, you know the exact function of the fluid the dice if moving in, the landscape surface on the table, density of dice, it's elasticity, loss of energy due to bouncing/friction, ....

This isn't entirely true. Starting with an over-simplyfied model, you can see that in some cases the die roll will be random, and others it won't be. Starting with a very unrealistic model, I drop the die, from a dead stop, with no rotation on it from about 1 mm in height. Its obvious the die will just fall and roll to which-ever side is closest to straight up. So it looks very deterministic. However, there are some unstable solutions as well. If I dropped with the the center of gravity directly over a vertex or edge we find that we are at a bondary point. In a perfect newtonian system, the die would come to rest on the vertex or edge, but in practice what we have is a boundary condition where a small perturbation in initial state will change the outcome. Around these points a physical system can be affected by apparently insignificant variations and quantum affects.

Now it would be a bit silly in insisting on calling the dropping of a die a short distance with no rotation random, when the initial conditions that produce uncertainty in the results is such a tiny portion of the whole.

But what happens if you know roll the die from higher up, with spin onto a somewhat elastic surface. The interaction of the die with particles, like air molecules on the way down will cause a small uncertainity in the velocity, angular momentum and angle it has when it hits the surface. This uncertainty may be very small, but it will grow each time it bounces. For one, angular momentum it has when it rebounds depends on the angular momentum it had when it touch down, the angle and the velocity it hit at. In the same way, the velocity it has when it rebounds will also depend on the state of the three variables at impact. The affect is that the uncertainity it has on rebound will be much larger than that on impact. Since the die will bounce multiple times, the uncertainty grows.

As your uncertainty grows the set of initial conditions that produce random results grows, and the set of initial conditions which yield predictable results shrinks. Now a die and a single surface is a fairly simple system, so I don't know how high you would have to drop one from to get it so that the majority of initial conditions would lead to random results.

In relation to evolution, many physical system are sets of coupled, non-linear sub-processes which can lead to Chaotic behaviour. When random elements are adding to chaotic systems, the uncertainty will be amplified, so instead of getting a determistic type result with some small amout of noise, one can get results that are genuinely random at the macroscopic level.

Walt

P.S. Seems this thread has started to grow like the others. Some other things I like to respond to, but don't know if I can get to it before this thread is pages on.

 

[sol invictus]

The definition of "random" as "[o]f or relating to a type of circumstance or event that is described by a probability distribution" does not make everything "random" because it refers to a very specific class of mathematical function (i.e., probability distributions), which are themselves defined in a very specific way. In particular, differential equations that do not take probability distributions as their argument, such as those used to describe weather (and, in a larger sense, fluid dynamics), cannot be described by a probability distribution and are therefore not random, making weather (and, in a larger sense, fluid dynamics) a deterministic (although chaotic) system.

That's false. Any evolution - even an ideal and absolutely deterministic one - can be described by a probability distribution (perhaps a very specific one - a Dirac delta function, for example). Not only that, but in the real world everything is uncertain to some degree, and hence everything (especially the weather) is described by a non-trivial probability distribution.

When the definition of a word includes language as incredibly vague as yours does, it's totally useless. The OED definition I gave above is much more precise, and according to it evolution is not random.

 

[Meadmaker]

But the research was addressing a specific aspect of Darwin's theory. (From artic's 3 links) Verdict in: Darwin's evolution theory confirmed

Not really. The researchers were comparing two different views of evolution. Both of those views are perfectly compatible with Darwinian evolution. When they cited papers in their references that explained the two theories, they weren't citing things from the Discovery Institute, they were citing things from reputable journals.

I think they did a disservice because the journalists covering the paper somehow seem to think that there was something about Darwinism that was at issue here, as if somehow there was some need to confirm Darwin's theories. That's not the case. They were just filling in some details.

 

[joobz]

This isn't entirely true. Starting with an over-simplyfied model, you can see that in some cases the die roll will be random, and others it won't be. Starting with a very unrealistic model, I drop the die, from a dead stop, with no rotation on it from about 1 mm in height. Its obvious the die will just fall and roll to which-ever side is closest to straight up. So it looks very deterministic. However, there are some unstable solutions as well. If I dropped with the the center of gravity directly over a vertex or edge we find that we are at a bondary point. In a perfect newtonian system, the die would come to rest on the vertex or edge, but in practice what we have is a boundary condition where a small perturbation in initial state will change the outcome. Around these points a physical system can be affected by apparently insignificant variations and quantum affects.

Now it would be a bit silly in insisting on calling the dropping of a die a short distance with no rotation random, when the initial conditions that produce uncertainty in the results is such a tiny portion of the whole.

But what happens if you know roll the die from higher up, with spin onto a somewhat elastic surface. The interaction of the die with particles, like air molecules on the way down will cause a small uncertainity in the velocity, angular momentum and angle it has when it hits the surface. This uncertainty may be very small, but it will grow each time it bounces. For one, angular momentum it has when it rebounds depends on the angular momentum it had when it touch down, the angle and the velocity it hit at. In the same way, the velocity it has when it rebounds will also depend on the state of the three variables at impact. The affect is that the uncertainity it has on rebound will be much larger than that on impact. Since the die will bounce multiple times, the uncertainty grows.

Of course I was refering to the rediculous case where you were able to calculate all of those perturbations. your point is that it's highly unlike to be able to know all of those things, and I agree. but my point is that even in the example of a random system, it's not really "random".

In relation to evolution, many physical system are sets of coupled, non-linear sub-processes which can lead to Chaotic behaviour. When random elements are adding to chaotic systems, the uncertainty will be amplified, so instead of getting a determistic type result with some small amout of noise, one can get results that are genuinely random at the macroscopic level.

which doesn't describe evolution at all. evolutionary mechanism at a global aspect is deterministic.
We know that the living system will come to some evolutionarily adapted form with it's surroundings, we just can't exactly predict what that form will take.

It's sort of like the constructal theory in thermodynamics. we know the gross forms things will take, but not the percise path.

 

[Walter Wayne]

Of course I was refering to the rediculous case where you were able to calculate all of those perturbations. your point is that it's highly unlike to be able to know all of those things, and I agree. but my point is that even in the example of a random system, it's not really "random".

No, it is random. The pertubation that causes uncertainties are subject to variation at the quantum level, like the air molecules I mentioned as an example. Knowing the wave function of every molecule of air in the test will not allow you to determine how they will affect the die. It may be a very small perturbation, but in complex system the affect of the random pertubation can grow. It has nothing to do with uncertainity.

which doesn't describe evolution at all. evolutionary mechanism at a global aspect is deterministic.
We know that the living system will come to some evolutionarily adapted form with it's surroundings, we just can't exactly predict what that form will take.

That's like saying a die roll isn't random because it will come up as an integer between 1 and 6. No impossible result will happen. In evolution, each species is a culmination of mutations. Each one does not exist if the chain of inherentance back to the origins of life is broken. I'm not going to call a die roll deterministic just because a 7 won't come up.

It's sort of like the constructal theory in thermodynamics. we know the gross forms things will take, but not the percise path.

We don't know the gross form things will take in evolution unless you abstract out to a ridiculously high level. Take humans, we are an insignificant twig of the evolutionary tree by any means, but as little as 70000 years ago it is believed the population was 2000 people. How many brushes with extinction has our line had in the last million years, going back? Break one link at we would not be here. But as insignificant as we are, we have had a drastic affect on a large number of species in a short period of time. Many have learned to live with us, others are going extinct. And each and every species is a product of a similar unbroken chain back to the dawn of life on earth, each a product of a culmination of mutations and selection, and influenced by the species that arose around it.

It is no where near gross form in the same sense of thermodynamics.

Walt

 

[mijopaalmc]

delete me

 

[Skeptic Ginger]

I'm sorry, but how the **** do you get someone to give you money to study nematodes' vulvas?

The Human Frontier Science Program

supports novel, innovative and interdisciplinary basic research focused on the complex mechanisms of living organisms; topics range from molecular and cellular approaches to systems and cognitive neuroscience. A clear emphasis is placed on novel collaborations that bring biologists together with scientists from fields such as physics, mathematics, chemistry, computer science and engineering to focus on problems at the frontier of the life sciences.

Many studies of the evolution of developmental processes have focused on either the conservation of mechanisms (such as signaling pathways and anterior-posterior patterning mechanisms) or on how differences have evolved among organisms. In contrast, the present study traces the remarkably dynamic evolution of developmental mechanisms underlying a highly conserved organ in a large number of closely related species. It makes use of the developmental processes that underlie formation of the nematode vulva, which is exceptionally well studied in the model system Caenorhabditis elegans.

Maybe it's a guy thing.

 

[Skeptic Ginger]

The problem is that the single piece of evidence that would settle this debate (i.e., whether all individuals with a given collection of phenotypes all reproduce while all other individuals with other phenotypes don't) doesn't seem to exist. What my stridence (for which I apologize profusely), has concealed was that my original argument was that we simply don't have the evidence to rule out the possibility that evolution is random, but simultaneously that possibility that evolution is random doesn't contradict anything that is predicted by the Theory of Evolution and observed in the lab or in the field.

Mijo, it is you who has defined the supposed criteria which would "settle this debate" (in your mind, anyway). But that fixation keeps you from looking at anything else being discussed such as the points I have made about different aspects of the system being non-random and only selectively pulling from the random pool. As well you ignore the discussion about the organism's mechanisms which control the random processes to speed them, slow them or negate them altogether such as in highly conserved portions of the genomes. Nor have you addressed the possibility that your image of how evolution occurs might be oversimplified and out of date given advances in genetic science.

Instead of actually discussing the merits of these arguments your knee jerks out and you repeat yourself like a xerox machine. You don't even engage anyone in why the arguments do or don't have merit. Well dude, sorry to burst your bubble but it isn't just the members on this forum using terms like NOT stochastic process. It is the top genetic researchers in the field of genetics. And from what I gather, you believe they are wrong because you know something about math theory you don't believe they know. From where I sit it looks more likely to me that they know something about genetic science you don't know.

 

[Wowbagger]

It's really simple:

deterministic mutation + deterministic selection = detereministic evolution

deterministic mutation + random selection = random evolution

random mutation + deterministic selection = random evolution

random mutation + random selection = random evolution

random anything + deterministic convergence factors (based on properties of physics) = practically inevitable evolutionary pathways

 

[Skeptic Ginger]

All deterministic systems are predictable if you are able reproduce the initial conditions with infinite accuracy. The "unpredictability" of chaotic deterministic dynamical systems comes from the fact that it is impossible to reproduce the initial conditions with infinite accuracy.

In this scenario, the same mutations would occur if you reproduced the initial conditions with infinite accuracy.

 

[articulett]

But he wants to call selection "random" because you can't predict what will and won't survive and thus selection could be "described by a probability distribution" (his definition of random) which leads him to conclude option 4. That's all he can hear.

 

[Skeptic Ginger]

When I ask for evidence that evolution is "random" by one definition and receive evidence that evolution is "random" by another definition, that is not a straight answer. In fact, equivocation is one form of fallacious argumentation that people here like to slam creationists for.

You claim only one answer satisfies your question. But you haven't supported your premise that the answer you determined is required is actually required.

You claim only X will satisfy as evidence a system is not random. I claim your underlying premise is false.

 

[Skeptic Ginger]

You might be right about his motives but that doesn't invalidate his position.

As I said in my initial post in this thread, I typically characterize evolution as a deterministic function with non-deterministic inputs. Unfortunately, technically, that necessitates that I admit evolution is random if someone holds my feet to the fire in a technical discussion. The very definition of a stochastic process is one where a given state contains insufficient information to predict the future state. As far as I understand it anyways.

So, yes, once you introduce randomness you've buggered the whole thing up. That randomness might be curtailed and shaped by a very non-random process but no matter what you may wish for, you can't predict the future state of a gene based on its current state. It's a stochastic process.

So, please (please, please, please), explain where I am mistaken without calling me a creationist. Please.

But there are things about the future state you can predict and things about the future state you cannot predict. So then what do you have? Mijo would prefer if there are predictable and unpredictable future states that the unpredictable states just negate anything predictable. In addition, mijo is ignoring the biological mechanisms which determine the random nucleic acid mutation processes. He is convinced the single random input in the system overrides everything else in the evolution process. Some aspects in the process of evolution are not random at all. My problem with how those processes are being described is the description is inaccurate, to say the least.

No one has addressed my example yet. Suppose you had a random number generator, but you determined which numbers you were going to use when they were generated. Now the only thing that is random is how much time it will take you to accumulate the numbers you want. But you determine completely which numbers those will be. The time is random but the outcome is determined.

In mijo's tunnel vision, he would continue calling that a random process. Most people would view the fact the time involved was random as unimportant to the end result and would call that a deterministic process.

 

[Skeptic Ginger]

You are not mistaken. Most people including biologists would say something to the effect that mutation is relatively random while selection is not-- it selects the "fittest" from the pool of randomness. So although randomness plays a role, what we observe and the appearance of design comes from the accumulated selection and exponential replication of the "fittest" random mutants. ...

Dang woman, and I didn't even read your post before composing mine.

 

[Skeptic Ginger]

It's really simple:

....random mutation + deterministic selection = random evolution...

According to you. But you have not made your case that this mathematical model applies to the processes of evolution. It might apply in an oversimplified version of evolution theory. But when that deterministic selection takes on certain properties, the quantity and quality of randomness in the system is overwhelmed by the quantity and quality of the deterministic processes.

If time is random but gene selection is not, are those two components distinct and separate enough to be considered unrelated components? And if only time is random but gene selection is not, then who cares? The random component for all intents and purposes is irrelevant.

 

[Walter Wayne]

In this scenario, the same mutations would occur if you reproduced the initial conditions with infinite accuracy.

Not true. Mutation isn't an affect of a chaotic process, but a random one. Mutations are governed by chemical bonds of sub-atomic particles, so are described by quantum physics. If all initial conditions are reproduced with infinite accuracy, then different mutations can happen.

 

[Skeptic Ginger]

Not really. The researchers were comparing two different views of evolution. Both of those views are perfectly compatible with Darwinian evolution. When they cited papers in their references that explained the two theories, they weren't citing things from the Discovery Institute, they were citing things from reputable journals.

I think they did a disservice because the journalists covering the paper somehow seem to think that there was something about Darwinism that was at issue here, as if somehow there was some need to confirm Darwin's theories. That's not the case. They were just filling in some details.

I see.