What evidence is there for evolution being non-random?

[mijopaalmc]

I know you.

That is also unhelpful.

But I'm glad you like it.

Uh....no, P(X) = 1 is not invalid for me, but I congratulate you on a yet another straw man.

The problem here is that P(X)=1 makes all other events in the sample space Ω impossible because, by the axioms of probability, P(Ω)=1. Therefore P(X^C)=P(Ω)-P(X)=1-1=0, or in other words any event that is not X cannot happen because its probability is zero.

While it is true that there are some mutations that make survival and reproduction impossible, there are no mutations that I know of that guarantee the survival of an individual. The P(X)=1 situation, while not impossible, is therefore unrealistic because it would require the all possessors of a certain mutation to survive and all others to perish.

 

[cyborg]

I'm glad you like your unhelpful definition of random. Truly happy. I'm also glad that you are able to miss the point with a high degree of reliability. It pleases me that you are satisfied with your terminology and find it descriptive. I am glad that your feel your back-peddling doesn't make you look silly.

All is good.

 

[Schneibster]

cyborg, I did warn you. Next time try actually answering the arguments, instead of playing semantic games with them. Good bye.

 

[mijopaalmc]

I'm glad you like your unhelpful definition of random. Truly happy. I'm also glad that you are able to miss the point with a high degree of reliability. It pleases me that you are satisfied with your terminology and find it descriptive. I am glad that your feel your back-peddling doesn't make you look silly.

All is good.

How is explaining to you that the axioms of probability ^{MW, WP} don't allow for there to be more than one possible event when the probability of a given event is 1 back-peddling?

I'm going to recommend to you, as I did to articulett, that you read Janos Galambos' Introductory Probability Theory or visit Virtual Laboratories in Probability and Statistics and try to actually understand how probability theory work before you insist that my explanations don't make sense.

 

[cyborg]

I see that neither of you get that this is about semantics and that both of you like to be sloppy.

No ****ing idea what games you think I'm supposed to have played but if you don't like the fact that a PTM is just a TM with analogical time for free or that randomness is about causality, not description then it is really not my problem that you will both flail about looking silly.

 

[Ivor the Engineer]

Ok, I've been out of this one for a while, but I have a question for the randomites. Do you believe what happens next is constrained completely by all of what happened before?

I'm guessing the answer will be "no". If so, how much wiggle room do you think there is for the outcome of the next event, given all of the events up to that point?

How would you tell the difference between x+1 = f(x) + err, where err is a truly random component and x+1 = f(x) + g(x-1, x-2, ..., x-N), where g() is a non-linear function from hell over all events till the present?

 

[mijopaalmc]

I see that neither of you get that this is about semantics and that both of you like to be sloppy.

So I see you yet again accusing us of playing semantic games because you have no substantive arguments.

 

[cyborg]

mjio, a test.

THTTHTTHTHTHTTHHTHTHTHT

Is that coin random? Can the trials be described by a probability distribution?

TTTHTTTTHTHTTTHTHTTTHTH

Is that coin random? Can the trials be described by a probability distribution?

THTHTHTHTHTHTHTHTHTHTHT

Is that coin random? Can the trials be described by a probability distribution?

TTTTTTTTTTTTTTTTTTTTTTT

Is that coin random? Can the trials be described by a probability distribution?

 

[jimbob]

If you have a mutation that makes some of your blood cells sickle shaped--you have a phenotypic change that gives you a survival/reproductive advantage in malarial regions. If you move out of malarial regions--the phenotypic change might hang out in your descendants through genetic drift. Or it might disappear. It becomes deleterious when an offspring is homozygous.

But it sure wasn't selected randomly! It was selected because it conferred an advantage...it was then passed on in these successful genomes even when it no longer conferred an advantage to its carriers because it was a neutral mutation in successful replicators--unless (or until) an offspring had two copies...in which case it became deleterious and usually those affected could not pass on their genes. And that is not "random" either. Having sickle cell anemia reduces reproductive fitness. Not randomly! Because having all your blood cells sickle shaped makes them get clogged up and produce infarctions and clots and priapism and other health problems. Yet you call this random! The first sickle cell mutations (it evolved numerous times and there are numerous mutations) came about randomly. But they were selected for or against based on pheotypic changes they code for in the vectors that copied them.

It was not selected "randomly", but the proportion of people with sickle-cell anaemia would depend on the percentage reproductive advantage this confers compared to individuals without this trait.

An example:

For simplicity, I am assuming that both full sickle-cell anaemia and maleria stop an individual from reproducing, and being a carrier of sickle-cell anaemia confers 100% protection.

If everyone in an area without any form of protection would get maleria, then the sickle-cell anameia carrierswould have an infinite advantage over non-carriers, as on average, 50% of their offstpring will be immune from maleria and not have sickle-cell anaemia.

In a different area there is no maleria, so these carriers would be at a 25% disadvantage, because on average that number would have full sickle-cell anaemia.

In between there would be different levels of infection, and different reproductive advantages/disadvantages for the sickle-cell carriers.

This can be assessed with a probabilistic treatment.
​

mjio, a test.

THTTHTTHTHTHTTHHTHTHTHT

Is that coin random? Can the trials be described by a probability distribution?

TTTHTTTTHTHTTTHTHTTTHTH

Is that coin random? Can the trials be described by a probability distribution?

THTHTHTHTHTHTHTHTHTHTHT

Is that coin random? Can the trials be described by a probability distribution?

TTTTTTTTTTTTTTTTTTTTTTT

Is that coin random? Can the trials be described by a probability distribution?

It depends, as there are (23?) throws, any of those set of outcomes are equally unlikely at 2^-23. If you are talking about the actual numbers of heads and tails, then it follows a binomial distribution.

Ok, I've been out of this one for a while, but I have a question for the randomites. Do you believe what happens next is constrained completely by all of what happened before?

I'm guessing the answer will be "no". If so, how much wiggle room do you think there is for the outcome of the next event, given all of the events up to that point?

How would you tell the difference between x+1 = f(x) + err, where err is a truly random component and x+1 = f(x) + g(x-1, x-2, ..., x-N), where g() is a non-linear function from hell over all events till the present?

I would say that for your last question that the only way of analysing that is statistically, and it is a moot point whether it is random, or pseudorandom. However, if the nonlinear function has positive feedback loops that magnify any random compomnents, then it is a moot point, because the randomness is magnified to a significant level.

Population bottlenecks are far more affected by such random events, and that is where a lot of evolution occurs.

After the KT impact, whether you think of that as random or pseudorandom, the population of surviving organisms was vastly reduced. That availiable ecological niches would be filled was a very near certainty. What would fill them, and how, was not.

Do you believe what happens next is constrained completely by all of what happened before?

Heavily influenced, plus magnification of true quantum randomness, as an other effect.

 

[jimbob]

Articulett, why dio you think it is imprecice to talk about selection according to probability functions.

To a first approximation the genetic mutations might be described as following a uniform distribution (maybe a 1/f ratio, but some similar form of distribution).

Selection, to a first approximation might be modelled by a poission distribution.

What is vague about saying the following sort of statement?

In general this species population is growing and the number of reproducing offspring per parent follows a poisson distribution with lambda=1.01.

There is an advantageous mutation, and of these, the number of reproducing offspring per parent follows a poisson distribution with lambda=1.1.​

You can then calculate how the chances of this beneficial mutation being fixed in the population, or the chances of it dying out, given an initial poplulation.

 

[cyborg]

It depends, as there are (23?) throws, any of those set of outcomes are equally unlikely at 2-23.

Wrong, but thanks for playing. You get partial credit for noting 'it depends'.

Articulett, why dio you think it is imprecice to talk about selection according to probability functions.

The real question is why do you think it is precise to call something describable by a probability function random even though it isn't.

 

[mijopaalmc]

mjio, a test.

THTTHTTHTHTHTTHHTHTHTHT

Is that coin random? Can the trials be described by a probability distribution?

TTTHTTTTHTHTTTHTHTTTHTH

Is that coin random? Can the trials be described by a probability distribution?

THTHTHTHTHTHTHTHTHTHTHT

Is that coin random? Can the trials be described by a probability distribution?

TTTTTTTTTTTTTTTTTTTTTTT

Is that coin random? Can the trials be described by a probability distribution?

All of these trials can be described by a single probability distribution, the binomial probability distribution^{MW, PM, WP}:

[latex]\begin{displaymath}f(k;n,p)=\left({n\atop k}\right)p^k(1-p)^{(n-k)}\end{displaymath}[/latex]

[latex]where[/latex] [latex]\begin{displaymath}{n\choose k} = \frac{n!}{(n-k)!k!} \end{displaymath}[/latex] [latex]and[/latex] [latex]\begin{displaymath}n!=n(n-1)(n-2)...\cdot 2 \cdot 1\end{displaymath}[/latex]

 

[cyborg]

So are the coins random or not? According to you they must be right?

 

[mijopaalmc]

The trials are random.

I have no idea what a "random coin" would be.

 

[jimbob]

If it was a fair set of throws, then any particular sequence of 23 throws would have the same odds of (1/2)^23.

I am assuming it was an oversight that removed my superscript in the quote and made it look as if I was saying 2-23, and not 2 raised to the power of minus 23.

The real question is why do you think it is precise to call something describable by a probability function random even though it isn't.

Do you consider quantum mechanical events to be random?

Is weather chaotic?

How long does it take for a quantum event to affect the weather?

Does weather affect which organisms reproduce?

Is this random due to chaotic magnification of quantum events?

I would say that how the odds of survival are modulated by mutations is the interesting part. TO understand this, one needs to use probability, as Dawkins does.

 

[mijopaalmc]

cyborg-

Why do you think that descriptions by probability distributions are inaccurate?

 

[cyborg]

The trials are random.

I say they are not.

Now - how do you know the difference?

If it was a fair set of throws

Yeah. 'If'.

Now tell me Jimbob - how do you know the difference? Does one not appeal to the properties of the coin rather than the outcome of the trials? What exactly has the probability distribution got to do with determining that? (It has bugger all of course - it's just one of the many statistical tools that are used to analyse the randomness or not of something; it does not impose randomness.)

Is anyone here getting the point yet - as I stated so many pages ago - that you are all shooting yourselves in the foot when you describe randomness as you do?

Do any of you even give a crap if you are using a stupid definition of random that leads to nonsense conclusions? Because it sure doesn't look like it.

Do you consider quantum mechanical events to be random?

They're the only thing in discussion here that have any claim to be truly random as arising from the physical nature of the event.

How long does it take for a quantum event to affect the weather?

How long does it take for a quantum event to cause a stray electron to cause a computational failure in your PC?

If you want to make everything random that is your choice. I'm guessing you would not find it helpful for me to describe your computer as 'random'.

I would say that how the odds of survival are modulated by mutations is the interesting part.

And I would point out that if it were random then it wouldn't be interesting. But then I'm using the word random with a meaning that actually allows one to actually categorise things as non-random too rather than being forced to conclude that everything is random.

 

[mijopaalmc]

If tosses of a coin are not random, what are they?

 

[cyborg]

I'd ask you to pretend that a coin does in fact have memory and that, for instance, the last coin cannot land heads up, one of the other's has to alternate etc... but I feel that would rather be wasted because you'd almost certainly miss the point that it is the coin, not the trials, that are either random or not. I mean I could point out that a coin that changed from T to H and H to T would have exactly the same probability distribution as a coin that is a little more fair. There are little things like autocorrelation and such to consider.

You people just don't have a clue what randomness is so it's no wonder you are using the word so poorly.

 

[jimbob]

Quote:
How long does it take for a quantum event to affect the weather?

How long does it take for a quantum event to cause a stray electron to cause a computational failure in your PC?

Significantly longer than for the weather. We aren't yet down to single electron devices in compters, so a very long time. (Alpha) radiation can cause soft errors, but the probabilities of this occuring can be calculated and the effects ameliorated by error checking.

Because weather is highly nonlinear, it takes less than a few months for quantum-sized differences to affect weather.