Evolution Not Random

[sol invictus]

From what I read, though, it's not clear to me whether physicists think the wave function is a convenient mathematical expression of the way particles behave, or if it's the actual way that they behave.

My guess is that most physicists believe that if there's such a thing as a fundamental theory of everything, it's quantum mechanical. The question of whether nature "actually" behaves that way doesn't have a functional meaning, so far as I can see - the only question we can answer is whether the model adequately describes that behavior, and whether it's logically consistent and complete. QM describes the world very well, but it is not complete, both because we don't have a complete quantum theory of gravity, and because of the collapse of the wavefunction.

Does it ? (really. does it ?) Or does it say we simply can't have it all ?

What it really says is that particles are described by probability distributions, and that any theory in which they are not necessarily gives results which are inconsistent with experiment.

 

[sol invictus]

BTW, isn't the 'many worlds' interpretation of QM Deterministic?

That's a good question. The Schrodinger equation is obviously a deterministic equation (because it's a differential equation in time), but the wavefunction it governs is a probability density for the results of measurements. In the standard interpretation of QM one simply says that when a measurement is made, the wavefunction collapses (with probability given by the wavefunction before the measurement) to that value. This is non-deterministic, but it is also incomplete because it does not define "measurement" rigorously.

To define "measurement", one could in principle solve the Schrodinger equation for a situation where a real measurement is made: a small quantum system interacts with a macroscopic measuring device, for example. The evolution of the wavefunction in such a case is smooth and deterministic. Many worlds posits (probably correctly, in my opinion) that the wavefunction splits into several very narrow and almost totally disconnected pieces, one where the small system is in state A and measuring device reads A, plus one where the small system is in state B and measuring device reads B, plus etc., with the coefficients weighted by probability appropriately.

That's great, but it doesn't explain why, when our brain's wavefunction splits in half like that, we experience only one option with the correct probability. So many worlds is incomplete as well, and without knowing what the completion is I don't know what to think about that. In the end, the results are identical to the standard interpretation as far as anyone can tell, so we might as well call it non-deterministic.

Sometimes I wonder whether there is any sense in which this debate (over QM determinism) is meaningful at all, given the existence of classical chaos.

 

[hammegk]

....very few people grant free will to microbes.

Who would argue a microbe does not have and utilize at minimum compatibilistic free will. As to free will per se, who knows. Certainly they exhibit will-to-survive.

 

[Belz...]

My guess is that most physicists believe that if there's such a thing as a fundamental theory of everything, it's quantum mechanical. The question of whether nature "actually" behaves that way doesn't have a functional meaning, so far as I can see - the only question we can answer is whether the model adequately describes that behavior, and whether it's logically consistent and complete. QM describes the world very well, but it is not complete, both because we don't have a complete quantum theory of gravity, and because of the collapse of the wavefunction.

What it really says is that particles are described by probability distributions, and that any theory in which they are not necessarily gives results which are inconsistent with experiment.

Which is all that science deals with. Got it. And thanks again.

 

[Belz...]

Who would argue a microbe does not have and utilize at minimum compatibilistic free will. As to free will per se, who knows. Certainly they exhibit will-to-survive.

Mechanism-to-survive, more likely.

 

[Fredrik]

That always strikes me as rather a bold assertion whenever someone comes out with it.

QM is a [very good] model of reality, but it is still not reality.

For example, what if our 'view' on reality is limited such that we cannot access the necessary knowledge?

Then Bell's inequality would hold, but experiments show that it doesn't. (You quoted #352, but I had already posted the answer to your question before you asked, in #354).

If you're thinking that we can't be exactly 100% certain of this, then you're right. However, a deterministic theory that is to replace quantum mechanics must explain how empty space can appear to be rotationally invariant and still not be rotationally invariant. This is a lot like trying to prove that we actually live on the inside of the Earth's surface even though we appear to be living on the outside.

Does it ? (really. does it ?) Or does it say we simply can't have it all ?

You're quoting #354, which is the post where I answered this question, or at least I thought I did.

 

[Fredrik]

If he moves the Queen to that square by Bishop will be under attack. There is a probability associated with that move.

Therefore Chess is a random game.

I actually think this is a good argument against the claim that evolution is random. Unfortunatly there are just as good arguments against the claim that evolution isn't random.

BTW, isn't the 'many worlds' interpretation of QM Deterministic?

Not any more than the other interpretations. (I feel that I can keep this answer short, since Sol has addressed the issue of what's deterministic in QM and what isn't in several of his posts).

How does your knowledge of the event change the event in any way ? Whether you have complete knowledge of it doesn't change the outcome.

I'm not sure what you're asking here or why. My knowledge won't change anything.

Sometimes I wonder whether there is any sense in which this debate (over QM determinism) is meaningful at all, given the existence of classical chaos.

Classical chaos is definitely enough to guarantee that evolution can't be described as "deterministic", so the QM discussion doesn't really belong in this thread. (It should have a thread of its own).

 

[cyborg]

I actually think this is a good argument against the claim that evolution is random. Unfortunatly there are just as good arguments against the claim that evolution isn't random.

So why can't we all conclude that the statements:

"X is random"

"X is non-random"

Don't really tell us anything about X?

 

[sol invictus]

I'm still waiting for someone to answer my polka dot question...

 

[jimbob]

Given that

1) We have (mostly) cleared up why several of us were contending that QM is random, and
2) that chaotic systems are affected on a macroscopic level by differences of the size of quantum effects,
and
3) that it is possible for e.g. asteroids to be in chaotic orbits.

Dosen't this mean that events like the KT impact could have been affected significantly by random events?

This is why I am saying that over geological timescales, the environment in which the organisms evolve is affected by random events that are large enough to completely alter the selection pressures.

If the selection pressures can be affected by random events over geological timescales, then over these geological timescales, the whole "direction" of evolution will be affected by random events.

 

[Fredrik]

So why can't we all conclude that the statements:

"X is random"

"X is non-random"

Don't really tell us anything about X?

They would tell us something about X if they were true, but in this case we have an X with the property that both claims are false, or rather nonsensical.

I'm still waiting for someone to answer my polka dot question...

The shirt is green goddammit, and if you say it isn't, I question your motives. You must be a creationist in disguise. If you tell people it's not green, they will think it's a 5-dimensional lizard that's going to eat us.

 

[Meadmaker]

A truck tire sheds its retread and the pieces fly all over the road. Every car driving past randomly hits the pieces and they randomly fly about. But the end result is not random, the pieces end up off the driving surface. It may be that the pieces fall in random places and the time it takes to get off the road is random, but the pieces end up predictably off the road every time. The cars hitting the tire shreds were not random.

Those tire bits can be big enough to cause traffic accidents in following cars. Engineers at tire making companies have to consider the possibility that one of their exploding tires might cause death. There's a tradeoff involved between safety, cost, and convenience. Somewhere along the line, someone has to make a decision about what to do. When they make that decision, they use random numbers in their calculations.

Yes, all those pieces end up off the road. If that's what you are wondering about, then you can say that the result is not random. However, some people care, with good reason, about how and why they end up off the road. In order to model the process, you have to use randomness.

That is not a random process.

You are using a phrase, "random process", that has a very specific meaning, and your example is indeed a random process. When engineers study random processes, the most common applications are control systems and communication lines. Everyone who designed your cable TV system had to study random processes. That doesn't mean that your TV is showing random images, although sometimes it seems that way.

Engineers care, not just about the final destination, but about the path. They aren't wrong to do so. For them, that tire shredding phenomenon is a random process. Similarly, evolutionary biologists may very well care about the path, not just the destination. To model evolution at a numerical level, you must use random processes.

On the other hand, to describe what happened at a coarse level, or to describe trends (the "whole"), you don't need to use random processes. It all depends on what aspect you are studying.

Not a thing. There is one small part of the process of evolution that is random. I don't believe a single person in this thread disagrees with that statement.

There's at least one, and I suspect more. One small part is the mutations. Those are random. Without it evolution wouldn't happen. Another small part is the final outcome. That's random. Another small part is the actual events which, when aggregated, become known as "natural selection". Those are random (in the mathematical sense). The only thing that isn't random is the pattern. So which is more important? I don't think either, myself.

So I repeat my question, reproduction is a part of evolution. Is it accurate to describe evolution as a birthing process?

No.

Is it even accurate to describe evolution as a process of reproduction?

No.

Would anyone think you were describing evolution, or would they recognize you were describing some part of the process, not the process?

No.

Of course, "birthing process" and "random process" are very different sorts of phrases, so I'm not sure how the analogy really applies.

How random is that yersinia pestis' acquisition of specific antibiotic resistance genes against antibiotics that other organisms have been subjected to but the yersinia organism has not?

Are you talking about an individual organism, or the species as a whole. On an individual level, very random. However, the large numbers involve mean that somewhere, some individual will likely get the resistant genes, and so it is highly likely that the species as a whole will get the resistance.

What aspect are you interested in? Are you interested in whether or not it develops the resistance, or the mechanism by which it develops the resistance. If you are interested in the latter, the mathematics of probability will be very useful to you in your studies.

How is the yersinia developing resistance to antibiotics which have never actually exerted selection pressures on the organisms? Certainly not through random processes.

Again, the phrase "random process" has a very specific meaning, and the development of antibiotic resistance can be modelled as such.

 

[Belz...]

I'm not sure what you're asking here or why. My knowledge won't change anything.

Yes, that is exactly what I meant. Whether we CAN know everything is irrelevant to the functioning of the thing itself.

 

[articulett]

I'm still waiting for someone to answer my polka dot question...

Your shirt is considered to have enough green to wear on St. Patrick's day to avoid getting pinched, and if someone tells to tell you it's a red shirt, they may be trying to keep you from this knowledge. If you just wanted to know the color of the shirt, you'd say it was red and green polka dotted... if you wanted to know how the shirt is useful on St. Patrick's day, you emphasize the fact that it has green... the red is irrelevant and confusing. But some people have a need to call that shirt red no matter what the context.

 

[Belz...]

That hurt my head, articulett.

 

[jimbob]

Skeptgirl,
Do you see why I think evoluiton is significantly affected by random events over geological timescales?

Some adaptations (e.g. flight) are common because there is such a large selective sdvantage, others are less so, and might not occur in a particular ecosystem.

What form the adaptation takes is also random and depends on the "available" mutations, and whther they survived or not.

Mijo is not arguing for "random" as "unbiased" or "haphazard". This is a difference between him and Behe that Articulett refuses to accept. (That and mijos repeted assertions that humanity evolved from ape-like ancestors with no guiding supernatural deity, indeed almost a "drunkards walk" and influenced by many chance events. (What would our genome look like if humanity had not got almost wiped out 70k-yrs ago by Tambora's eruption?

Unfortunalty "random" is sometimes associated with the word "haphazard" so I don;'t like it. That is why, when talking about natural selection I preferthe term probabilistic. If it conjours up images of casinos, and the fact that over time the house will win, so much the better.

 

[jimbob]

Originally Posted by jimbob
The best evidence is that quantum events are random, the particle doesn't "know" where it is either. Identical systems with identical inputs would not always give the same output.

Total knowledge would have to include knowledge of what the output is otherwise it is less than total.

The concensus is that the universe is not like that. The random nature of a QM event means that it is not determined until it has happened.

You can't know a decision if it hasn't been made.

If the system was just very complex, "the decision" will have been taken, even if nobody could analyse it.

Quote:
If you ignore the inputs, outputs and mechanism of a process, then you don't have much left to understand.

Thanks for the strawman.

What are you saying then? You seemed to imply that inputs were not important.

Yoy also seemed to imply that the process wasn't important, that only left the outputs...


[/QUOTE]

Quote:
Selection would be different if it weren't probabilistic. Evolution would be faster as beneficial traits would always prosper and deletereous traits would always vanish.

Grossly wrong as I've tried to explain to you before but you just won't listen because you don't seem to get the difference between having a trait and being able to exploit it for gain - i.e. it being used.
[/QUOTE]
Is this your argument that if there are two fledglings in a nest, and one is eaten it was because that fledgling was by definition less well adapted to its environment?


What else dop you mean, and why is it a better treatment than a probabilistic treatment of natural selection. I have even given you the distribution and typical parameters.

Quote:
Artificial selection is faster than natural selection for that reason. The selection pressure is far closer to 100% than in natural selection, where 1% is a more realistic number.

Yes. Very realistic numbers pulled straight from the proverbial bottom.

Artificial selection is near 100% when selecting for a trait, as that is what breeders do.

I chose the 1% because of it being larger than the selective pressure that Richard Dawkings gave when he talked about a selective advantage as low as 1:1000.

Even a 50% selective advantage for a single trait, which would be massive in an ecosystem, is only about half as efficient as artificail selection.

Quote:

Its population is cyclic. This means that if a trait occured in an individual during a population boom it is very likely that it would spread (even if it were slightly deletereous). If the same trait occured during a population crash it is highly likely that it would vanish (even if beneficial). Of course, if the trait was established in the population and was beneficial, and survived to the next population boom, it would be very likleyu to survive.

To understand the process and its implications, you need a probabilistic treatment.

Speaking probabilities and talking games again I see.

Are you going to explain that comment?

How do you get any useful numbers without a probabilistic treatment?

 

[Beth]

Skeptgirl,
Mijo is not arguing for "random" as "unbiased" or "haphazard". This is a difference between him and Behe that Articulett refuses to accept. (That and mijos repeted assertions that humanity evolved from ape-like ancestors with no guiding supernatural deity, indeed almost a "drunkards walk" and influenced by many chance events. (What would our genome look like if humanity had not got almost wiped out 70k-yrs ago by Tambora's eruption?

Let me second this. Jimbob is correct. Mijo's insisting that by the proper definitions of 'random' and 'deterministic', evolution is random. It is. I'm not interesting in debating the matter. I'm a professional statistician; I know what the word 'random' means! But 'random' is also quite commonly taken to mean things that are properly termed "haphazard" "unbiased" or "uniformly distributed". By that usage, you are correct and evolution is anything but.

 

[articulett]

It's the fact that he claims to believe it, but can't say how... goes out of his way to obfuscate such understanding... he does not seem to understand or "convey" natural selection because of his need to use the word random. And, THAT, is identical to Behe-- who also concedes common descent. In fact, I dare you to find a difference in his explanations of evolution.

People who cannot convey understanding of evolution for whatever reasons are often hung up on the random and miss the far more important part of natural selection--what is ... in fact, both Mijo and jimbob do something funny in regard to tenses to make natural selection seem random because unpredictable event can affect it... but an unpredictable environment is part of the selection process... organisms that are better suited to deal with unpredictability--preferentially survive.

But the apologists cannot hear the experts on the topic--they are so sure they already know everything there is to know. They think they are clearer than Dawkins and the Science article and peer reviewed scientists though no one but fellow apologists who also cannot explain evolution seem to have this opinion of them.

Tsk. You can fool yourself in your head all you want--the smart people understand who is conveying information and who is using words to say nothing of value at all. Whatever your goals may be, it appears they are working more in your head then in reality.

 

[kjkent1]

Nosing in again for a moment...

It seems to me that the issue with evolutionary change is not so much deterministic vs. stochastic as it is goal vs. non-goal oriented.

A primarily deterministic system has a certain predictable goal. For example, in a casino, while every spin of the roulette wheel is guided by random chance, in the long run, the amount that the average player will lose is ascertainable in advance with reasonable certainty (-5.26% of money wagered, given a wheel with 1-36, half red -- half black, 0 and 00). The "goal" in roulette is for the house to win 5% of all the money -- and it's known in advance.

But, evolution is not like the roulette wheel, because with the roulette wheel, the environment for the ivory ball is invariant throughout the set of trials -- while with evolution, the environment is just as unpredictable as are the mutations. It's as if the evolutionary roulette wheel were constantly changing the status of the slots into which the ball may fall. A number can change, as can a color, and so can the number of slots. And, while these changes are generally slow to occur, there is no certainty of that, either.

This is why evolution is so difficult to nail down with a mathematical model: the odds are in constant motion.

The outcome of any evolving organism is not reasonably ascertainable in advance, despite the existence of deterministic selective pressures, because those pressures may, and usually do, change over time.

Evolution has no goal. The "house" has no acertainable expected value, other than the survival of the fittest -- fittest being relative to the instant environment.