Randomness in Evolution: Valid and Invalid Usage

[Belz...]

Not really, what is so random about reproduction?

Well, according to Mijo's definition, since your girlfriend may or may not have a headache that night, it becomes a probability distribution which makes the act of lovemaking completely random!!

 

[Belz...]

It is a dogression, but chaotic systems are hypersentistive tio initial conditions so that a classical treatment requires accuracy that doesnt actually exist.

EVERYTHING is sensitive to initial conditions. That has nothing to do with randomness.

I seems that the most common response to your questions about chaos and randomness would be "your description makes everything random".

No, YOUR definition makes everything random as everybody in this thread but you has noted.

 

[Belz...]

First of all, you are correct that quantum fluctuations can have large effects on chaotic systems.

I'm aware they could have an effect, but a large one ? How ?

 

[Dancing David]

That would be more plate tectonics/vulcanology than gravity. Analogy fails.

Sure, it was a metaphor, I should have said , erode. History of water , wear and gravity.

 

[sol invictus]

I'm aware they could have an effect, but a large one ? How ?

The definition of a chaotic system is that if you change the initial conditions slightly, the system's state diverges exponentially with time from where it would have been had you not changed the initial conditions. The rate of that divergence depends on the size of the initial deviation and a system-dependent parameter. But after enough time the divergence is always large regardless of the size of the initial perturbation, and the time required grows only very slowly (logarithmically) as you make the initial perturbation smaller.

jimbob's example of the billiard balls was a good one - the claim was that after 12 billiard ball collisions, the Heisenberg uncertainty in the initial positions of the balls translates into total uncertainty in their positions. I haven't checked the number 12, but even if it's wrong the correct answer won't be very much larger.

 

[Dancing David]

It is a dogression, but chaotic systems are hypersentistive tio initial conditions so that a classical treatment requires accuracy that doesnt actually exist.

Here is a discussion about a very simple system (from the Israel physical society)



Twenty-four collisions ahead, and there are twelve sets of collisons where the accuracy required would be beyind the uncertainty principle.

That is different than saying QM has an impact on the out come.

You haven't demonstrated that the outcome is dependant on the uneven distribution of quantum events. The gross scale interactions are the significant ones. The even distribution of QM events will strill occur.

i want your demonstration that a partcile being in postion P of a probable distribution is electron shell S is going to have significant impact on m and v momentum of a snooker ball at some particular point.

You haven't shown that at all.

Why is it wrong to describe the behaviour of the snooker ball that far ahead as random? Not just unpredictible, but random.

Why don't you demonstrate your theory, that unpredictable equals random. Your usage , your defense.

You would be wrong to describe the behaviour three collisions ahead as random, but twenty four, or thirteen collisions ahead?

This is a simple system but it does involve classical mechanics in a chaotic system, and thus for far-enough ahead is random.

Unpredictable is not random.

Chaotic systems are determined and causal.

Weather systems, which affect individual organisms is similarly a chaotic system, so I cant see why this behaviour far enough ahead isn't truely random.

Indeed I can't see why any choatic system, far enough ahead, isn't random.

Okay, so that is how you want to use the word. What meaningful purpose comes from saying that unpredictable means the same as random.

Does it provide you comfort? Does it make communication clearer or is it just some bizzare point of semantics.

I can also say that all biological organisms are dog, does that mean anything?

ETA:

Sol agrees with you Jimbob, but I won't edit out what i just said. I will continue to think about why I feel it doesn't matter and see if I can restate my thought.

 

[sol invictus]

ETA:

Sol agrees with you Jimbob, but I won't edit out what i just said. I will continue to think about why I feel it doesn't matter and see if I can restate my thought.

It's OK to disagree with me! Sometimes I'm even wrong .

I think QM events can strongly affect chaotic systems after relatively short amounts of time. I'm pretty sure 99% of physicists would agree with me.

I don't think it's useful to distinguish between "random" and "unpredictable" when we're discussing physical processes. I'm not sure how many physicists would agree with me on that (although I think I could convince them).

But none of that prevents us from predicting with extremely high confidence that July in Saskatoon will be warmer than January in Saskatoon. As they say - "weather is chaotic, but climate is predictable" (or something along those lines). As for evolution, it has both weather-like and climate-like aspects.

 

[Walter Wayne]

I wanted to reiterate what articulett said: Those who write the most scientifically accepted books about evolution, tend not to describe evolution as random, at all. And, sometimes, they take great pains to show how it is not random. Dawkins' book The Blind Watchmaker is one, for example. Here is an excerpt from its preface:

It is almost as if the human brain were specifically designed to misunderstand Darwinism, and to find it hard to believe. Take, for instance, the issue of 'chance', often dramatized as blind chance. The great majority of people that attack Darwinism leap with almost unseemly eagerness to the mistaken idea that there is nothing other than random chance in it... if you think that Darwinism is tantamount to chance, you'll obviously find it easy to refute Darwinism! One of my tasks will be to destroy this eagerly believed myth that Darwinism is a theory of 'chance'.​

It may still be valid to call Evolution "random" in some contexts of the word. But, I feel that it is best the leave the word out, because of all this confusion it causes. It is perfectly possible to describe every aspect of Evolution with other words besides "random".

I agree the word random can be avoided in explaining evolution. But it is because of the confusion the word creates that I think the term must be used in evolution's teaching. Students will inevitably come across the term, most likely in a debate with a creationist. With descriptions of natural selection being laden with terms that don't imply certainty ("more likely", "tends to", ...) and mutation being "blind", how will our hypothetical student respond to a challenge based around the word random. I am guessing he would be ill equiped to deal with it.

I also take issue with describing evolution like the "experts" do. The newspapers articles cited claiming evolution is "not-random" actually link to papers that don't say that. This isn't an area that has no debate in it, but one can get the impression that there is only one expert opinion from the discussions.

"Understanding" is being painted as agreeing with a few specific experts (the most prominent lay writers). One may hold an opinion on what "level" natural selection works on, but if one teaches only "gene centered evolution" one is not giving students an understanding of the field. Thus understanding becomes identified with parrotting back one particular view, rather than knowledge of the subject. In another thread, when I disagreed with someone's characterization of natural selection being what "builds" the complexity, the person replied that what they meant was that selection was a sort of rachet for complexity. Now it so happens, that "rachet" is the precise metaphor that Dawkins uses. Should I then believe that the person actually understands Dawkins's metaphor, or that they interpreted in wrongly and are thus unable to express it in their own words?

Walt

 

[jimbob]

It's OK to disagree with me! Sometimes I'm even wrong .

I think QM events can strongly affect chaotic systems after relatively short amounts of time. I'm pretty sure 99% of physicists would agree with me.

I don't think it's useful to distinguish between "random" and "unpredictable" when we're discussing physical processes. I'm not sure how many physicists would agree with me on that (although I think I could convince them).

Surprisingly, I'd agree with that, but several people have been proposing "Laplacian Determinism", i.e.

An intellect which at any given moment knew all the forces that animate Nature and the mutual positions of the beings that comprise it, if this intellect were vast enough to submit its data to analysis, could condense into a single formula the movement of the greatest bodies of the universe and that of the lightest atom: for such an intellect nothing could be uncertain; and the future just like the past would be present before our eyes.

In such a I have been reading their argument an being that evolution is completely deterministic, so that if we were able to measure the initial conditions accurately enough, the evolutionary outcomes would be always the same.

I started out thinking there was simply a semantic difference, then came to the conclusion that the difference was more fundamental, hence my above reference to Laplacian Determinism.

The following post for example, with most of it hidden for brevity:

Sigh. I am going to have to point out the counter-intuitive nature of infinities again?

Jimbob - yet again I must ask: do you get the difference between using probabilities to model a system where the probabilities stand in effectively for a set of complexly interacting variables?

Consider the evolutionary models below:

Model A will be much like Ev - we will have a pool of creatures with a target genome. Mutations to the target pool occur at some rate and distributed with probabilities. Survival is dictated by how well the organisms score against the target genome.

Model B will extend Ev - we now introduce the concept of a target functionality. Instead of being scored against a genome we score against the ability for a given organism to act like the functional target. We will now have a notion of how the genotype effects phenotypes.

Model C extends model B to introduce the notion of competition. Instead of selecting based on a simple score we now provide a physical environment for our phenotypes to interact in. We have a more 'natural' sense of survival in this instance - now having to exist in a physical space organisms will have to contend with the thing you had a problem with earlier. Essentially being, "in the wrong place at the wrong time."

Now we could model some of the new things we will see occur in model C in models A and B using probabilities to represent the idea of there being "bad luck" in one's positioning.

Now, as we extend model C with new aspects that make the simulation more akin to what we see in the real world clearly we are increasing the complexity of the model. Now do you see why this increase in complexity would lead to the appearance of it being random? It's because the more variables you have - the richer your model - the more difficult it is to predict the behaviour of the model even if it is deterministic.

If you think it is always easy to predict how a deterministic model will behave then I would have to ask why you would think that would be so. If you get that the difficulty as to how to predict its behaviour increases with the number of variables in the model and how aperiodic it is then it is only a short step to see that the more one increases the number of variables the more akin it is to seeming to be random in nature. (And in the infinite case it is mathematically equivalent).

So, is the randomness in the model a reflection of how things are actually occurring or just the sum total of our ignorance about the variables involved in giving rise to it?

Can you see why for a sufficiently rich model probabilities would not have to be involved?

But none of that prevents us from predicting with extremely high confidence that July in Saskatoon will be warmer than January in Saskatoon. As they say - "weather is chaotic, but climate is predictable" (or something along those lines). As for evolution, it has both weather-like and climate-like aspects.

I have already agreed with your description:

I would argue that selection is probabilistic ("random" gust of wind etc affecting survival). However, we can still see how different traits affect the odds of producing reproducing offspring. I think a valid analogy might be between weather and climate. The individual slection event might be "random" but the efffect over a large enough population means that some beneficial traits will propagate.

Doing the sums, I would conclude that most "beneficial" traits that arise probably don't survive more than one generation.

This is because the odds are against any individual organism reproducing, for virtually any species (possibly except our own currently). For example, the Barn Owl population is roughly stable, but it tends to have a clutch sizes of about 3-7 and sometimes breed twice a year, and live for 1-5 years in the wild (25 years in captivity). Of the total brood size over the lifetime of the pair, on average only two offspring will breed if the population is stable.

Say this equates to 5 clutches, of 4 birds. Then there is 90% chance of any individual not breeding, and a 10% chance of it breeding. To get an evens chance of a particular trait making it past the first individual, it would need to confer a 500% advantage compared to its peers...

However we are dealing with big numbers, and some (enough) advantageous traits will survive and get passed on...

Disadvantageous traits are almost certain to vanish very quickly. In the barn owl example, a neutral trait already has a 90% chance of not getting passed on.

 

[cyborg]

In such a I have been reading their argument an being that evolution is completely deterministic, so that if we were able to measure the initial conditions accurately enough, the evolutionary outcomes would be always the same.

If everything is the same everything is the same - even you have to acknowledge this.

QM is a Red Herring - it's an appeal to the argument, "yeah, but in reality if you wind back the clock because QM is random not everything can be the same, hence you can't ignore the impact of that on evolution."

Which makes a mockery of the concept of modelling phenomena as a class of events.

 

[jimbob]

So you still stand by your view that:

Can you see why for a sufficiently rich model probabilities would not have to be involved?

 

[cyborg]

Yes. If I replace the die roll with a list of numbers I have a deterministic set of data - irrespective of whether or not the list of numbers has maximum entropy.

 

[Wowbagger]

I agree the word random can be avoided in explaining evolution. But it is because of the confusion the word creates that I think the term must be used in evolution's teaching. Students will inevitably come across the term, most likely in a debate with a creationist. With descriptions of natural selection being laden with terms that don't imply certainty ("more likely", "tends to", ...) and mutation being "blind", how will our hypothetical student respond to a challenge based around the word random. I am guessing he would be ill equiped to deal with it.

We can teach them that there are valid and invalid usage of the word "random", similar to my OP. Of course, if my OP is wrong about something, I hope the education professionals would be able to put together something more accurate.

We don't need the R-word to explain the process to them, but I agree that, in order to be equipped with challenges (in this case, semantic ones), they should also be taught some of the proper rebuttals.

I also take issue with describing evolution like the "experts" do. The newspapers articles cited claiming evolution is "not-random" actually link to papers that don't say that. This isn't an area that has no debate in it, but one can get the impression that there is only one expert opinion from the discussions.

I disagree. I think the newspapers should learn to describe science the way experts do it. Perhaps Dawkins-style analogies would not be a bad comprimise, though, if they are carefully constructed.

"Understanding" is being painted as agreeing with a few specific experts (the most prominent lay writers). One may hold an opinion on what "level" natural selection works on, but if one teaches only "gene centered evolution" one is not giving students an understanding of the field. Thus understanding becomes identified with parrotting back one particular view, rather than knowledge of the subject.

This has nothing to do with the topic (which is more steeped in semantics), but I do basically agree with it. Comprehensive understanding of the field, is better than focusing solely on some narrow aspect of it.

This thread has now accumulated over 7,000 views. Is it really that popular?!!

 

[mijopaalmc]

If everything is the same everything is the same - even you have to acknowledge this.

QM is a Red Herring - it's an appeal to the argument, "yeah, but in reality if you wind back the clock because QM is random not everything can be the same, hence you can't ignore the impact of that on evolution."

Which makes a mockery of the concept of modelling phenomena as a class of events.

You are completely missing the point. Quantum randomness has an orderly classical (or semiclassical) limit. Quantum randomness therefore stands as an example of a random process that has orderly long-term and large-scale behavior and the example does not rest upon everything's ultimate origin in quantum mechanics.

 

[cyborg]

You are completely missing the point.

No, it would seem you are.

Quantum randomness has an orderly classical (or semiclassical) limit. Quantum randomness therefore stands as an example of a random process that has orderly long-term and large-scale behavior and the example does not rest upon everything's ultimate origin in quantum mechanics.

In other words: the whole QM discussion is largely irrelevant to how we talk about evolution. It is a large red fish.

 

[mijopaalmc]

In other words: the whole QM discussion is largely irrelevant to how we talk about evolution. It is a large red fish.

Except that it is an example of a random process that has orderly long-term and large-scale behavior, demonstrating that there need be no underlying deterministic framework for evolution by natural selection to display the phenomenology it does.

 

[Reality Check]

Except that it is an example of a random process that has orderly long-term and large-scale behavior, demonstrating that there need be no underlying deterministic framework for evolution by natural selection to display the phenomenology it does.

This is the inverse of quantum mechanics: QM is a deterministic process (see the Schrödinger equation) where a measurement has an expectation value. It is the measurement that has a probability and I think that you could call that the "random" part of QM.
If you want to use QM as an analogy for evolution (or demonstration of evolution) then there must be the equivalent of the Schrödinger equation in evolution, i.e. by your argument evolution has a "underlying deterministic framework".

 

[mijopaalmc]

This is the inverse of quantum mechanics: QM is a deterministic process (see the Schrödinger equation) where a measurement has an expectation value. It is the measurement that has a probability and I think that you could call that the "random" part of QM.
If you want to use QM as an analogy for evolution (or demonstration of evolution) then there must be the equivalent of the Schrödinger equation in evolution, i.e. by your argument evolution has a "underlying deterministic framework".

The quantum mechanical wave function is a random variable as it describes the probability amplitudes at different points in space. Since the Schrödinger equation takes the wave function its argument, it is a probabilistic equation.

 

[sol invictus]

Surprisingly, I'd agree with that, but several people have been proposing "Laplacian Determinism", i.e.

An intellect which at any given moment knew all the forces that animate Nature and the mutual positions of the beings that comprise it, if this intellect were vast enough to submit its data to analysis, could condense into a single formula the movement of the greatest bodies of the universe and that of the lightest atom: for such an intellect nothing could be uncertain; and the future just like the past would be present before our eyes.

Interesting. I'm pretty sure I can prove on the basis of physical law that such an "intellect" is literally impossible.

 

[sol invictus]

The quantum mechanical wave function is a random variable as it describes the probability amplitudes at different points in space.

Wrong. The wavefunction is a probability distribution (or really, the square root of one), not a random variable. All these posts and you can't even get the basic mathematical terminology right?

The time evolution of the wavefunction is totally deterministic, as it obeys a differential equation.