Is evolution stochastic?

[jimbob]

I am still reading the paper, but later on Lenski supports what I said about even positive mutations being likely to vanish, not surprising, given that he is quoting a paper from 1927.

As Haldane (1927) showed, most beneficial mutations are lost by random drift before they become common enough for selection to drive them into fixation; the probability of substitution as a beneficial mutation is only about 2s, where s is the selective advantage. Thus, a mutation that xconfers a 10% advantage will be lost by drift about 80% of the time, whereas one with only a 0.5% advantage will disappear 99% of the time.

I have sai before in the discussion about natural selection, that evolutionary biologists discuss natural selection in probabilistic terms, and that the description of natural selection as nonrandom, was a rhetorical simplification or soundbite for getting a simplified and simplistic idea of evolution accross to those laypeople who *really* had toruble with the whole idea of evolution.

If you are *not* talking to that audience, but to numerate people, who understand that randomness does not mean "an equal chance of everything happening" then describing natural selection as "nonrandom" is misleading.


My position on natural is the scientific orthodoxy, despite what Articulett might claim.

My position on the randomness of evolution is one side in an active topic of discussion in evolutionary biology, but which Lenski's Long term Evolution experiment has provided strong support for: both in the case of Citrate+ metabolism, and in some of the evolutionary responses of cultures to cold-stress.

[sol invictus]

@sol invictus: Below is the response of susu and Marios to your questions.

I don't understand their answers. There are plenty of versions of evolution which are entirely deterministic - every computer simulation of it ever done, for example. Not to mention that nature itself (and therefore the evolution of life on earth) could be fundamentally non-random for all we know.... so the argument fails completely.

It's a semantic debate, and totally uninteresting.

<snip>

One of the truly bizarre aspects of this is that you bring that up (and I believe you did so also in the past) as evidence that evolution is random. On the contrary, it's evidence that it is non-random, because Lenski was able to engineer the appearance and survival of several very specific mutations. All the colonies evolved some of the same traits given enough time. The fact that one colony was first on any particular one is tautological - did you expect all of them to mutate the same way at precisely the same time?

It would be interesting to know whether one can define a mathematical limit of infinite time, and if so whether all colonies would asymptote to the same sequence. But regardless of that, it is completely obvious that the mutations that persisted were non-random - on the contrary, they were adaptive and advantageous in that environment.

[articulett]

Calling it "random" is far more misleading to the majority--which is why creationists insist that evolution is a theory of "random chance".

Ken Miller is understandable and understood far more than those who would classify things differently. He writes numerous text books and testified in the Dover trial. I think I'll stick with the actual experts. I understand them better than the self-appointed experts do.

[jimbob]

In answer to your question as to "why nobody calls weather random "

Will it rain within a square kilometer of lattitude 51°N, longitude 0°E anytime during the daylight hours of December 15th 2051?

Not only can't we predict the weather that far in advance; but, due to the highly nonlinear response of the weather, it hasn't even been determined by the universe yet. Weather is not predetermined, it is perturbed by random events. Over short timescales it can be forcast, because these perturbations take time to grow.

[articulett]

What the hell does that have to do with this thread?

Yes, we can find ways to prove that all systems have unpredictable or probabilistic elements-- that doesn't mean it's informative to call such systems "random". You sure do reach for some weird examples to justify the way you want to describe evolution.

[jimbob]

,snip.



One of the truly bizarre aspects of this is that you bring that up (and I believe you did so also in the past) as evidence that evolution is random. On the contrary, it's evidence that it is non-random, because Lenski was able to engineer the appearance and survival of several very specific mutations. All the colonies evolved some of the same traits given enough time. The fact that one colony was first on any particular one is tautological - did you expect all of them to mutate the same way at precisely the same time?

It would be interesting to know whether one can define a mathematical limit of infinite time, and if so whether all colonies would asymptote to the same sequence. But regardless of that, it is completely obvious that the mutations that persisted were non-random - on the contrary, they were adaptive and advantageous in that environment.

I disagree:

ETA: This is also Lenski's interpretation:

From the abstract to Lenski's paper (PDF (http://myxo.css.msu.edu/lenski/pdf/2008,%20PNAS,%20Blount%20et%20al.pdf)):

This potentiating change increased the mutation rate to Cit+ but did not cause generalized hypermutability. Thus, the evolution of this phenotype was contingent on the particular history of that population. More generally, we suggest that historical contingency is especially important when it facilitates the evolution of key innovations that are not easily evolved by gradual, cumulative selection.

More generally, we suggest that historical contingency is especially important when it facilitates the evolution of key innovations that are not easily evolved by gradual, cumulative selection.

Unless "historical contingency" is nonrandom...



Lenski was specifically looking to see if the same ancestor in the same environment would always evolve the same traits.

This did not happen.

The experimnetal design removed the effects of initial genetic differences, and also removed the effects of externally-applied environmental effects.

Of course, the traits that evolved were adaptive, and of course the more likely traits evolved numerous times. So what?

We know this from observing biology, where the eye has evolved many times, sometimes more than once in the same organism. We also know this from observing how many orginsims have evolved flight or gliding.

The initial potentiating mutation did make citrate evolution very probable, but this mutation was itself rare. Indeed this is the only observed variant of that strain of E.coli to have evolved citrate metabolism.


raising the spectre of infinite time is a red herring.

For the purposes of the experiment, the environment was kept uniform and simple (expecially compared to less artificial ecosystems). This made the outcome of any randomness less likely not more so, but also more obvious. Because there were fewer interactions between different types of organisms.

The fact that one colony was first on any particular one is tautological - did you expect all of them to mutate the same way at precisely the same time?

No but the others have had tens of thousands of generations and still not evolved this trait.

Timing is crucial when talking about interactions.

Suppose someone, 100 years ago in Botswana developed a mutation that woudl have protected that individual against HIV. This would have been evolutionary neutral, and not likely to spread.

If someone developed the same mutation today in Botswana, then that mutation would confer a large advantage.

What is the difference? The mutations of a different "player" in the ecosystem, the HI Virus.

This is a simple example. With more organisms, the nonlinear interactions become greater, and randomness becomes more important.

Sol, do you agree with my statement that chance events have played an important part in the evolutionary history of virtually any organism in that has ever lived.

[articulett]

All macro events are considered deterministic... it's how we know the way the world works.

Historical contingency may mean "random" to you-- it doesn't mean it to Lenski or anyone else. Historical contingency IS the selection process--and nobody of note would describe that as random.

As Miller said, to do so would be to confuse the sifted with the sifter. You cannot seem to differentiate the two. That's why you and your fellow friends and Behe are so hard to converse with. You need a certain answer to be "the truth"-- but that certain answer only makes sense if you are defining terms as broadly and uselessly as you are.

[jimbob]

Why do you think that articulett?

Lenski's stated aim was to see if the tape of evolution gives the same outcome when rerun.

It doesn't.

How is that nonrandom?

ETA: Different outputs for the same inputs, meets my definiton of "random".

EDIT2:

I am also arguing that "nonrandom" is wrong, *not* that "random" is correct.

It is more complex than that, but unless talking to innumerate laypeople, "nonrandom" is simply misleading...

[sol invictus]

ETA: Different outputs for the same inputs, meets my definiton of "random".

Ah, back to this... how refreshing to go around full-circle to a stupidity first expressed months and months ago.

According to that definition, evolution is (probably) random.

Now please name one single real-world process that isn't.

[articulett]

Why do you think that articulett?

Lenski's stated aim was to see if the tape of evolution gives the same outcome when rerun.

It doesn't.

How is that nonrandom?

ETA: Different outputs for the same inputs, meets my definiton of "random".

EDIT2:

I am aslo arguing that "nonrandom" is wrong, *not* that "random" is correct.

It is more complex than that, but unless talking to innumerate laypeople, "nonrandom" is simply misleading...

And this makes elections random, this thread random, smoke detectors random, and everything random. You've defined the term so that it is useless to anyone but... well... yourself and a few creationists. You are using the words of those who think they understand natural selection, but don't have a clue. And you seem completely unable to realize this.

Most people describe evolution as a 2 part process-- but because it is not completely random in any sense of the word, calling evolution "nonrandom" is far more accurate than calling it random to most people. In the same way that elections are nonrandom--random implies a lottery method of choosing. Non random allows you to explain how it's not like a lottery. And if you haven't understood this after years of posts on the subject, how can you ever understand it?

Whenever someone aims to clearly describe evolution, you claim that they are calling it nonrandom and that this is "wrong". The experts disagree. They would say that the way Ken Miller describes evolution is far more accurate than your silly semantics. Sorry. But that's just the way it is. You can convince yourself and others like you that you are saying something valid, but it appears to be valid only to the people on your level... even the experts you quote would not agree with you. What you are calling the same inputs are NOT the same inputs. All things that evolve have a common starting point... but once they diverge none of the inputs are truly identical. Lots of experts will say evolution is nonrandom and describe how. Of course they aren't using your definitions. None are calling evolution random. None. If we ran the tape from the beginning and EVERY input was identical... we'd expect the same output. But we cannot make every input identical. Just like every molecule in every storm would create the exacts same storm if events unfolded identically. When something different happens, it is no longer "identical". You just cannot seem to hear this.

The key to understanding evolution is understanding how it's nonrandom. Everyone (even creationists) can tell you how it's random. Few can tell you how natural selection (which is not random --according to Miller, Dawkins, Coyne, Ayala, et. al.) leads to the appearance of design. You are included in this latter category. You are clearly not using the term nonrandom nor random the way these experts are.

[jimbob]

ETA: Different outputs for the same inputs, meets my definiton of "random".
Ah, back to this... how refreshing to go around full-circle to a stupidity first expressed months and months ago.

According to that definition, evolution is (probably) random.

Now please name one single real-world process that isn't

I missed out the key word "significantly".

I would say that metabolising citrate or not is a significant difference. Especially as this has not been seen in any other variants of that strain of E.coli.

I can name many processes where similar or identical inputs produce significanlty similar outputs.

It is also a commonly accepted usage of the word "random":

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.


On planet smart people "random processes" are terms sometimes used to describe stochastic processes... the smart people don't consider the processes themselves random... they are noting that they have random components.
i would really like to understand all the confusion/argument here.

articulett, statisticians would generally not be "smart people" by your definition. random processes (aka stochastic processes) are indeed considered random, you seem to think this is a very restrictive condition, for the most part they do not.

not all random processes are IID (independent, identically distributed) or Gaussian distributed, "most" are subject to "deterministic" influences; they are random processes merely because there is a random term (in discrete or continuous time) onthe right hand side of the equations that define them.


do we all agree that a classical random walk is a random process?

and that so is a random walk on a tilted surface (that is a random process that contains a deterministic drift?)

if you agree the second is random, i do not see why you reject the idea that evolution is random, (given the typical alternatives: random and deterministic, would you want to say evoluiton is deterministic?)

articulett, you just seem to allergic to the use of the word "random" in connection with evoluiton, would you consider it OK to call a radiation induced mutation random?

cyborg could you tell me which post you explained things in, there are over a thousand on this thread!

thanks.

[articulett]

Yes... jimbob... normal people consider mutations (and the various recombinations) the random part of evolution, while they consider the selection process the part that accumulates and multiplies the randomness to achieve the appearance of design. We don't call the later random. You might think it's honest or informative to do so, but the experts do not. They do not define words as vaguely as you do because they don't have this bizarre semantic need to describe things in a way that is indistinguishable from the creationist straw man. Instead, they aim to clarify, and use more stringint definitions of words to do so.

Pretending that people are allergic to the word "random" because they find it vague and misleading in regards to evolution doesn't mean that anyone who is taken seriously in the field of science thinks that it's useful to describe evolution as random--that's why none do.

And that's exactly why creationists do refer to evolution as a "theory of chance"-- it works to befuddle folks so that they THINK they understand evolution, but they end up sounding like you. Such people ought to talk to each other, because they really aren't conveying information to anyone else--although they THINK they are... they think they are smarter than those who could give them a clue.

You don't want to describe evolution, you want people to agree that you are describing it coherently. So go find those people who agree and discuss whatever important things you imagine you are saying. The rest of us will describe evolution the way the actual experts do--because we have found that is the best way to convey understanding. Your descriptors are the best way to obfuscate understanding.

You don't need to put down me. I am only repeating what all the experts are saying. Not even Lenski, whom you imagine support you, would disagree with anything I've said. Nor would Dawkins or Miller. Just find one peer reviewed article that says evolution IS random or whatever it is you think you are saying... find one article where they are defining terms like you are. You guys have none. It's all semantics used to say nothing much at all.

If your arguments were valid or useful, it should be easy to find peer reveiwed scientists making them--without twisting their words or supplying your personal definitions to their terminology.

[articulett]

What is your point, and it doesn't help to mishear me as being allergic to the word random? I use it exactly as Dawkins and Miller do. But the fact is, RANDOMNESS is not essential to evolution-- VARIETY is... it doesn't matter how that variety is generated--randomly, through recombination, or through an alogrithm... there just needs to be a pool from which to select the most efficient teams of genes from. You are obsessed with randomness... as obsessed as any creationist. The word is completely unnecessary in describing evolution and can so readily lead to confusion.

That's it. You just can't get this. Whatever your point is, it is lost in your need to define things in terms of randomness--the least essential part of evolution. The part that confuses the most people. The creationist straw man.

How much more evidence do you need that you are not clear, and that no one really speaks of evolution like you do except for those who think they understand it, but don't? Those who are clear, publish and teach others. Those who are not, malign others with straw men and use lots of words to convey no information at all.

That's fine with me, but putting me down, reveals exactly what sort of person you are; the people here no me well enough to know that I am not as you'd characterize me. You are only fooling the fools-- you are miffed because other people have pointed out that you are not quite the expert you imagine yourself to be. You are miffed that I said the truth-- you aren't really particularly well regarded as an explanatory source on the topic of evolution. Like Mijo, your expertise seems to be entirely in your own mind. Entirely.

I mean really, Jimbob, what the hell is your point? Dont put me down--try to find someone of integrity who uses words in the cumbersoome non-explanatory ways you do to advance whatever point you are trying to make. Do you think Beth is someone others find "informative" on this topic? I teach this topic; I have a Masters in Genetics; I worked for years as a genetic counseler. What's your expertise again? I can't make much sense of Beth and neither can the people I respect--she's wishy washy sounding... like you... apologetic... misses the point... garbled. And lenny is okay, but he's payiing more attention to a straw man view of what I said, then what I actually said, and he's not exactly supporting your veiwpoint. I'll stick with the bigger experts like Dawkins et. al. I know exactly what such experts have to say on this subject.

You aren't exactly quoting highly respected people, you know... and when you do, they sure aren't saying what you seem to imagine them to be saying. You have no ability to improve your communication, because like the people referenced in my sig, you are so sure you know more than those who might clue you in.

Everyone just gets tired of it, and learns to ignore you. If your point was to converse, you need to be on the same page as everyone else. It doesn't help if you imagine yourself smarter than those who actually know more than you. You don't like it in other people--and people here, don't like it in you. Why wasn't Dr Adequate or Sol's responses good enough? Do you really imagine yourself smarter than them? I don't think the evidence supports this notion.

[six7s]

How much more evidence do you need that you are not clear, and that no one really speaks of evolution like you do except for those who think they understand it, but don't? I dunno about jimbob, but I have had more than enough, thanks :)

So, I'm unsubscribing

@arti and/or others who do understand:
If this or any other thread starts discussing randomness in a meaningful and coherent way, please do flick me a PM

Thanks

[jimbob]

And this makes elections random, this thread random, smoke detectors random, and everything random. You've defined the term so that it is useless to anyone but... well... yourself and a few creationists. You are using the words of those who think they understand natural selection, but don't have a clue. And you seem completely unable to realize this.


You are using the words of those who think they understand natural selection, but don't have a clue. And you seem completely unable to realize this.

Are you sure it is me, and not you?

Richard Dawkins implicitally uses a probabilistic treatment of natural selection in the extended phenotype. The difference is that then he is not trying for a one-sentance soundbite, but an actual exploration of the implications of different amounts of selective advantage.




Most people describe evolution as a 2 part process-- but because it is not completely random in any sense of the word, calling evolution "nonrandom" is far more accurate than calling it random to most people. In the same way that elections are nonrandom--random implies a lottery method of choosing. Non random allows you to explain how it's not like a lottery. And if you haven't understood this after years of posts on the subject, how can you ever understand it?

And what about in the evolution of citrate matabolism in E.coli?

Was this a significant difference in outcomes?

Was the important diffenence due to chance?



Whenever someone aims to clearly describe evolution, you claim that they are calling it nonrandom and that this is "wrong". The experts disagree. They would say that the way Ken Miller describes evolution is far more accurate than your silly semantics. Sorry. But that's just the way it is. You can convince yourself and others like you that you are saying something valid, but it appears to be valid only to the people on your level... even the experts you quote would not agree with you. What you are calling the same inputs are NOT the same inputs. All things that evolve have a common starting point... but once they diverge none of the inputs are truly identical. Lots of experts will say evolution is nonrandom and describe how. Of course they aren't using your definitions. None are calling evolution random. None. If we ran the tape from the beginning and EVERY input was identical... we'd expect the same output. But we cannot make every input identical. Just like every molecule in every storm would create the exacts same storm if events unfolded identically. When something different happens, it is no longer "identical". You just cannot seem to hear this.

"None. If we ran the tape from the beginning and EVERY input was identical... we'd expect the same output. But we cannot make every input identical. Just like every molecule in every storm would create the exacts same storm if events unfolded identically. When something different happens, it is no longer "identical". You just cannot seem to hear this."


This is where you are wrong.

Evolution is a nonlinear system, with really weird positive feedback loops, as long as you allow chance events to happen, these chance events will perturb the systems so that a different outcome will occur. Over time the differences will magnify.

As long as you are allowing chance events to occur, these are sufficient to cause different outcomes.

With the storm, the universe can't "know" the position and momentum of every molecule.



The key to understanding evolution is understanding how it's nonrandom. Everyone (even creationists) can tell you how it's random. Few can tell you how natural selection (which is not random --according to Miller, Dawkins, Coyne, Ayala, et. al.) leads to the appearance of design. You are included in this latter category. You are clearly not using the term nonrandom nor random the way these experts are.

The randomness or not has nothing little to do with the appearence of seeming design.



It allows more variants, because sometimes a slightly improbable trait will evolve. Sometimes it won't.

I also suspect that without randomess as an open-ended source of variation, evolution would be limited, but haven't seen if pseudorandom variation could replace this.

Do you considetr yourself to be good at explaining evolution?

Do you accept why almost every word in this is wrong, or do you think that this a good explaination of evolution?

For some life forms, they evolved to try something new or different when the old stuff isn't working...

[articulett]

I'm unsubscribing too... suck others into your silly pointless semantic game

[articulett]

Yes, I'm sure it's you and not me.

[recursive prophet]

I don't understand their answers.... so the argument fails completely.

It's a semantic debate, and totally uninteresting.
@sol.. Much as I expected from you to be honest, sol …elatus? You get detailed deconstructions of your post; models; maths; and explanations of their significance in understanding how a very complex stochastic process affects NS. But nay, with ne’er one of many specific claims by Marios and susu actually even mentioned, you dismiss all to the realm of semantics. I find myself in awe of your audacity, doubting you even read what I posted. Brass balls that large must sound like Notre Dame Cathedral at high mass if you ever jog. :cool: Word salad in the land of terse condescension, where image reigns and all logic is reducible to semantics? While I do envy your certainty, unless you can be specific in your rebuttals I won’t bother to send over your conclusion today. But thanks for all the insight shared; semantics. Got it!

@jimbob, posting on the RD thread yesterday. Think you’ll find many who will review your ideas far more constructively than has been done here, though I do thank cyborg for taking the time to explain his dismissal of your conjectures. It provoked some informative replies on RD, (above) and I really would like to read his response to all that Marios and susu have presented. I use their analysis whenever possible, as they are both far beyond me in knowledge and experience with this subject.

As we all have a limited amount of time to post online, I would suggest you don’t waste it trying to explain your thinking to those who are clearly insincere in their comments. How many times can you make a statement or ask a question, only to see it ignored in any reply? If all you get in response is more of the same ‘talking points’ and dubious appeals to authority, what makes some continue to attempt any real communication with such posters is an even bigger mystery to me than evolution. Hopefully cyborg will return soon and your earlier dialogue can be continued.

Again I would ask especially any with grad level + training in any aspect of evolutionary science to at least weigh in on this discussion, here or on the RD thread. Unless someone credible comes along soon, I for one will be ready to consider the OP finally answered. Both evolution and natural selection have been convincingly and comprehensively argued as stochastic processes. It wasn’t the answer the OPer on RDF had in mind, nor I when I first started posting there. But over time it became obvious those with the most direct experience and advanced training all saw selection as stochastic and were winning the debate. They had far more unanswered, detailed arguments go unaddressed than those arguing for a more deterministic role. I am always happy to see such perceptions changed, so if you can show me the flaws in the deductions of susu and Marios, I’d love to read them. Otherwise, this thread may have served its purpose and have little to offer me beyond this point. We’ll see, and thanks to all who have made positive contributions.

[sol invictus]

@sol.. Much as I expected from you to be honest, sol …elatus? You get detailed deconstructions of your post; models; maths; and explanations of their significance in understanding how a very complex stochastic process affects NS. But nay, with ne’er one of many specific claims by Marios and susu actually even mentioned, you dismiss all to the realm of semantics. I find myself in awe of your audacity, doubting you even read what I posted. Brass balls that large must sound like Notre Dame Cathedral at high mass if you ever jog. :cool: Word salad in the land of terse condescension, where image reigns and all logic is reducible to semantics? While I do envy your certainty, unless you can be specific in your rebuttals I won’t bother to send over your conclusion today. But thanks for all the insight shared; semantics. Got it!

You seem to be under the misapprehension that I care about this debate. I don't, because it's completely uninteresting.

Discussing what randomness is fundamentally - that's interesting. Discussing any of the myriad specifics of evolution - that's interesting. But having a moronic argument over whether the claim "evolution is random" is true - that's really boring.

I posed that question because it plainly has no answer, and whaddya know - there was no answer. If you want a more detailed response, fair enough (I inflicted this on myself by posing the question). I'll give a little bit of one here. I don't care if you post it wherever you're getting these responses from.

There simply isn´t a non-stochastic theory of evolution, so the only way to falsify stochasticity of evolution would be to falsify evolution.

Obviously untrue. Any of the huge number of simulations of evolution that run on computers using deterministic pseudo-random number algorithms falsify that immediately. Not to mention all kinds of related algorithms and processes (genetic algorithms etc.). And while those obviously don't model all aspects of the evolution of life on earth, that's equally obviously not because they are using pseudo-random rather than "true" random numbers.

Well, the ball really is on the non-stochastic side. If I´m asked what the stochastic theory of evolution looks like, I can simply point to the textbooks and 150 years of scientific papers. The other side has nothing of that sort, so far I haven´t seen a single model from that side. I´ve seen assertions, but no models. Since the term "stochastic" has a hard definition and the theory of evolution fits that definition, I don´t think I´m arguing semantics.

The computer models I mentioned satisfy the criterion for a non-stochastic model of evolution in the general sense. For evolution of life on earth the question reduces to that of fundamental randomness. Either all physical processes are stochastic or none are, and there are models in fundamental physics for both possibilities. If you want to insist that only processes that rely on randomness in some "essential" or "important" way are stochastic, you reduce the whole thing to a silly value judgment.

Look - to prove that this debate is semantic, let me present two conflicting definitions. Here you are, courtesy of google:

stochasticity - randomness: the quality of lacking any predictable order or plan

That obviously does not apply to evolution, and I can't imagine anyone (even some contributors to this thread) arguing that it does.

A stochastic process, or sometimes random process, is the counterpart to a deterministic process (or deterministic system) in probability theory. Instead of dealing with only one possible 'reality' of how the process might evolve under time (as is the case, for example, for solutions of an ordinary differential equation), in a stochastic or random process there is some indeterminacy in its future evolution described by probability distributions. This means that even if the initial condition (or starting point) is known, there are many possibilities the process might go to, but some paths are more probable and others less.

This probably does apply to evolution, although as I've said, not necessarily in the strictest sense.

The only reasonable way to have this discussion is to identify a specific phenomenon or set of phenomena on which we disagree, and discuss those (and as far as I know there are none). Anything else is useless.

[sol invictus]

I disagree:

[COLOR="Blue"]ETA: This is also Lenski's interpretation:

Maybe so, but if so there is no evidence for that in the quotes you gave.


More generally, we suggest that historical contingency is especially important when it facilitates the evolution of key innovations that are not easily evolved by gradual, cumulative selection.

Unless "historical contingency" is nonrandom...

And you persist with this false dichotomy for the thousandth time.

I'm done discussing this with you. Bye.

[cyborg]

Here´s a simple deterministic process:
In each step add 1.
You get 1;2;3...
Now for a simple stochastic one:
In each step throw a fair coin and add 1 if it is heads and 0 if it is tails (numbers in brackets: probabilities):
You get 0(.5),1(.5);0(.25),1(.5),2(.25);0(.125),1(.375),2( .375),3(.125)...
The number of possible pathways is growing exponentially with each step rather than staying constant (and 1) in the deterministic system. I can tell you the outcome for step 487 in the determinitic one (it´s 487), but my calculator quits on me when I try to figure out the probabilities for that step in the stochastic case (it´s 2-487*487!/((487-n)!n!) with n being the value we want the probability for).

This right here is ridiculous.

So somehow I should be surprised that when comparing two systems one in which there is zero branching and one in which there is non-zero branching that the non-zero branching has an exponentially growing solution space?

You have heard of the concept of abstraction right?

Here's the function in psuedo functional-code:

x = map (if flip is tails then +1 else +0) flips

This is perfectly deterministic and the result depends on the list flips.

You have chosen to compare an instance of function x to the entire class of the function x:

x [tails, tails, ...485 times... ] = 487

But any given list of flips is just as simple to calculate - just a lot harder for you to list succinctly.

Doing it for all 2^487 - 1 possible lists is going to take a long time? Really? Well I never. I don't believe I ever claimed that analysing the entire problem space was a trivial task.

But if people are going to continue to confuse the execution of an instance with the analysis of the class then they're going to continue to make silly statements.

ANY "stochastic process" will have deterministic instances - I've already proven that before. Is this really hard to get?

[jimbob]

Cyborg: I roll a pair of dice once.

I know that the score will be an integer from 2 to 12.

I can write out all the possible scores.

So the dice roll is deterministic by your argument?

[Zeuzzz]

Note that there need to be three options ("true", "false", and "not even wrong"), as there are many statements ("evolution is random" is one) which are partially true and partially false, particularly when your definition is as stupid as mijo's was.

Oh, and by the way - the answer is E and 3, as every schoolchild knows.


Na, most schoolkids, and infact probably most college students, would probably say that evoution is random. They still teach about random mutations in secondary school in evolution lessons, which pretty much confuses everyone into thinking evolution is completely random. I used to think this (I never did biology or anything at college) but have since learned it is not so online.

The choice of words used in the curriculum could certainly be better explained.

[recursive prophet]

Looking back on my observation and minor participation in the How Evolution is NOT pure random chance, [emphasis mine] on RDF, I’m reminded of Asimov’s classic short story, The Last Question. It began innocently enough as an argument between to minor techs on the Colossus 2 -or whatever-forgot name he gives the super computer that pretty much runs things. They had recently launched a satellite that would beam enough solar energy to earth to supply their needs ‘forever,’ exclaimed one tech. The other pointed out the sun’s ultimate demise, and the debate that ensued culminated with them posing a question to Colossus; can entropy be reversed? Both techs soon forgot the silly bet they’d made that night, but Colossus didn’t. It would keep working on it until the end of time, should that the answer to ‘the last question.’ (The story is short and available online)

While aspects of the Second Law have been introduced at various times in the over 3k replies on the RDF random thread, the similarity I speak of mostly concerns the occurrence of relatively insignificant persons originating questions that later snow-ball into far more complicated debates. You see it often on the boards. The user who began the random thread-Mel Z-also presented a dice analogy, which was later shown to be inaccurate on several different levels and much better models using dice and coins were presented.

From what I’ve observed it appears topics often evolve as the debate becomes more specific, and in the better ones opinions are often changed. My own was, re: the intended OP here; is NS stochastic. I now sense those such as myself that have been following the dialogue on RDF have seen enough evidence presented that it may well be time to let the thread become dormant until someone comes along with something new to say. There are links below to the start of this thread last year, and the current last page. Some very distinguished scientists, such as Milford Wolpoff, have contributed to this dialectic, along with many others working on doctorates in the field.

It’s been an interesting read, but at this juncture I’m personally quite satisfied that the combined posts made by susu, Marios, Dlx2 and mjpam and others on this subject are by far the most congruent and persuasive. I would still love to see someone go toe to toe with them and come up with a winning argument. I could use another lesson in how weaker conjectures can often prove to come down to poor presentation while being part of a larger truth. And of course such an experience would be invaluable for the ‘boys,’ in learning the true cognitive significance of uncertainty. But pontifical snarls and peremptory, unsubstantiated dismissals aren’t going to cut it on RD. Ask Articulett, the threads most active poster. She is either ignored-my preference-or ridiculed by like 85% of the regular posters there. Of course all the ‘smart people’ have left, so I guess her still mostly posting there is a sign of low self-esteem? By suspecting her own incompetence she insures herself-per her sig-she’s not really incompetent? Of course you’re not dear. You just have a lot of very negative energy that needs to be released. I’ve actually come to feel sorry for you, watching you repeat the same mantras that I suspect you say in your sleep by now. Try breathing into a paper bag a few times before hitting send, and ask yourself what is your purpose? Those who oppose you are just looking at the shadows on the cave’s wall and will never achieve your advanced perspective. We will just continue staring at the straw man’s shadow and believing he is real. Leave us in peace now, and re-join the cognoscenti. We are truly not worthy.

@sol and cyborg: Marios merely commented neither of you seem familiar with the biology involved and are confused about the terms stochastic and deterministic. The reply of susu is below, and I decided to leave in his response in the same post to David, as it is pertinent to some of the examples/models susu has presented that have gone unchallenged.

@cyborg: As to your final statement, no one is arguing you won’t see deterministic instances. The question is can you predict them, and the answer is no. The process that generates them is stochastic. You can only show they are deterministic in retrospect, and even if you could rewind the tape with the same variables you might get different results. This has been extensively examined if you could take the time to click the second link and go back a few pages. You really should enlighten those poor misguided souls on RDF. Pop open a can of whup-ass on em and show the great unwashed the path to the river. I know; not interesting and all semantics. :cool:

Beginning of How Evolution is not pure random chance, http://www.richarddawkins.net/forum/viewtopic.php?f=4&t=7428
Last page of posts on part 3. http://richarddawkins.net/forum/viewtopic.php?f=4&t=66951&p=1679853#p1679853

I had this with Dlx2 ages ago. The word "majority" is the key. All you are seeing is the fact that most genes aren't involved in any one speciation, but, having been separated by the speciation, they drift apart, near-neutral fashion, without necessarily having a significant net phenotypic impact. Some speciations may involves only drift after physical separation, but surely not "interesting" ones, that lead directly to new classes, or whatever, by virtue of a genetic innovation, encouraged by new conditions to adapt to.

WTF? What is the difference in the process of "interesting" speciations that lead to new classes? There´s none. Classes are arbitrary units and most importantly, the clustering in morphospace (remember that other discussion we had last week or so?) arises purely from common ancestry. It is replicated in the pure morphological drift models of the MBL papers. You´ve actually just thrown common ancestry out, by invoking selection as the mechanism that leads to the formation of classes (i.e. clades occupying a distinct region in morphospace).

On SCA prevalence, I don't see how you can make any sense of the time-dependence of that without taking into account taboo strength. It is you who is assuming "random mating (ie no taboos), IMO, in order to have a population-selection pressure graph that doesn't mention taboo effects.

Well, for starters I also gave an example from a yeast culture (and such cultures tend not to be very civilized). Then you are arguing that taboos make a dent big enough to explain a discrepancy of 4 orders of magnitude. How big do you figure hispanic families in the US to be, if you want to shelter the allele we´d be talking about one family encompassing 10% of hispanics. 10%!! Even then the best you can get is neutrality. Still the frequency goes up.

You are concentrating on genetic changes being unpredictable ("random", if you like), but that isn't the point, as it relates to the randomness of mutations, not of selection.

Wrong. I´m not talking about mutations at all. I´m talking about drift. I´ve noted that the rate at which the SCA allele increases through mutation is 10-8 per generation. It decreases by about 6*10-3 through selection. Together that´s a decrease of about 6*10-3. But it does go up in reality. The selection component is more than compensated for by drift.

EDIT: Selective pressure isn't on individual genes, it's on gene teams.

Wrong again.

The computer models I mentioned satisfy the criterion for a non-stochastic model of evolution in the general sense. For evolution of life on earth the question reduces to that of fundamental randomness. Either all physical processes are stochastic or none are, and there are models in fundamental physics for both possibilities.

Actually. No. Even though a few physicists work on models with underlying determinism, their models still are stocahstic. The need for stochastic models is not at all at wuestion. Now, we had the debate over t`Hooft in this thread and one thing to note is that he is not proposing non-stochastic models. He´s merely changing what is stocahstic: The standard model says that if you look at an electron it will have a stochastic value for it´s properties. t`Hoofts alternative is that the electron is composed of lots of deterministic "things" for lack of a better word and when you look at the electron you take a random sample of these things (i.e. in one case you draw one ball from an urn and it is some random shade of grey. In the other case you draw a lot of balls from the urn and they are either black or white and you measure the mean. While the balls and with them the distribution of balls in the urn is deterministic, your sample is stochastic).

This right here is ridiculous.

So somehow I should be surprised that when comparing two systems one in which there is zero branching and one in which there is non-zero branching that the non-zero branching has an exponentially growing solution space?

No, you shouldn´t be surprised. But you shouldn´t be surprised either that I might claim that the branching process is a bit more complicated than the non-branching one. That after all was being debated.

You have heard of the concept of abstraction right?

Yes.

Here's the function in psuedo functional-code:

x = map (if flip is tails then +1 else +0) flips

This is perfectly deterministic and the result depends on the list flips.

Holy ****. In that case you are really claiming that rolling a die may be stocahstic, but the number I get when I roll a die and add 1 is deterministic? x is a random variable, because it is the sum of random variables. A (finite) sum of random variables is a random variable.

You have chosen to compare an instance of function x to the entire class of the function x:

x [tails, tails, ...485 times... ] = 487

But any given list of flips is just as simple to calculate - just a lot harder for you to list succinctly.

Well, when you make predictions, which ultimately is the job of science, then you don´t have the list handy (in fact if you knew in advance a sequence of 497 coin flips before I started flipping I´d be enormously impressed).

Doing it for all 2^487 - 1 possible lists is going to take a long time? Really? Well I never. I don't believe I ever claimed that analyzing the entire problem space was a trivial task.

Actually I don´t even analyze the complete sample space, but elements of a sigma algebra in that sample space.

But if people are going to continue to confuse the execution of an instance with the analysis of the class then they're going to continue to make silly statements.

I never made any claims about the execution of an instance.

ANY "stochastic process" will have deterministic instances - I've already proven that before. Is this really hard to get?

That´s BS. Any stochastic process will have a result. But there´s no way to look back at it and say "It had to happen that way", only "It did happen that way". That´s the difference between a stochastic and a deterministic process.

[jimbob]

I disagree:

[color="Blue"]ETA: This is also Lenski's interpretation:
Maybe so, but if so there is no evidence for that in the quotes you gave.

Quote:
More generally, we suggest that historical contingency is especially important when it facilitates the evolution of key innovations that are not easily evolved by gradual, cumulative selection.

Unless "historical contingency" is nonrandom...
And you persist with this false dichotomy for the thousandth time.

I'm done discussing this with you. Bye.

What does the following mean, and how does this not mean that chance is important, and even giving a situation when it is imprtant.

How am I mistreading the paper and the aims of the experiment when this is the stated aim of the experiment:

Second, this experiment seeks to examine the repeatability of evolution by having 12 replicate populations, all of them founded from the same ancestor and maintained in the same environment. Which phenotypic and genomic aspects of evolutionary change are repeatable in this system, and which are haphazard? How can we understand the causes of parallelism and divergence of replicate lines? The issue of the predictability of evolutionary change - or lack thereof - has long been of interest. The question was well captured by the late palaeontologist Steven Jay Gould (1989) in a thought experiment: "I call this experiment 'replaying life's tape.' you press the rewind button and, making sure you thoroughly erase everything that actually happened, go back to any time and place in the past -say to the seas of the Burgess Shale. Then let the tape run again and see if the repetition looks at all like the original..." Gould went on to say, however, that "The bad news is that we can't possibly perform the experiment." Of course, we could never run an experiment on the vast temporal and spatial scales imagined by Gould. But on much smaller scales, our experiment with E.coli allows us to address the same question. We do so by allowing 12 scenarios to play out simultaneously, rather than sequentially, as the notion of replay implies, but the issue of repeatability is fundamentally the same.

Is it wrong to say that the excxperiment is attempting to address the question as to whether rerunning the tape of evolution (resetting it back to an "identical" starting condition) would always give the same output?

It has been asserted that this:
"More generally, we suggest that historical contingency is especially important when it facilitates the evolution of key innovations that are not easily evolved by gradual, cumulative selection."

Is not saying that chance is important in detrermining the outcome of an evoultionary occasion.

What is it saying then?

[sol invictus]

Actually. No. Even though a few physicists work on models with underlying determinism, their models still are stocahstic. The need for stochastic models is not at all at wuestion. Now, we had the debate over t`Hooft in this thread and one thing to note is that he is not proposing non-stochastic models. He´s merely changing what is stocahstic: The standard model says that if you look at an electron it will have a stochastic value for it´s properties. t`Hoofts alternative is that the electron is composed of lots of deterministic "things" for lack of a better word and when you look at the electron you take a random sample of these things (i.e. in one case you draw one ball from an urn and it is some random shade of grey. In the other case you draw a lot of balls from the urn and they are either black or white and you measure the mean. While the balls and with them the distribution of balls in the urn is deterministic, your sample is stochastic).


Not that this is relevant to evolution, but it is actually an interesting question. Unfortunately susu.exp's response makes it quite clear that he/she/it has no idea what she/he/it is talking about.

Off the top of my head I'll give a list of fully deterministic models of fundamental physics:

1) The "many worlds" interpretation of quantum mechanics, which is simply standard QM with the collapse postulate dropped. The Schrodinger equation is a non-singular differential equation and therefore deterministic. Incidentally this is the most widely favored interpretation among physicists, at least according to some informal polls.

2) Bohmian interpretation of QM with pilot wave etc. Fully deterministic.

3) Any of several more obscure deterministic non-local hidden variable theories.

[articulett]

Sol,

Susu's definitions elections are "pure random chance"-- he actually said that.
So his definition of random appears very loose indeed. He'd answer A on your little quiz.

RP is posting your responses at Dawkins.net with his delusional spin and trying to keep this semantic silliness going so that he can feel he's on the "smart" side... As long as he can get participants, the "mijoesque" feel like someone smart is taking them seriously, I suspect... like they are saying something worthy of respect or scientifically valid or useful.

(RP thinks Mjpam and Susu exemplify the smart side and that we are all mean to them-- so that's what you'll get for all your efforts with him as well-- more abuse, hubris, and misheard words.) He's trying to make this what he sees as a battle of experts, but his experts are what you and I and actual experts would think of as "buffoons".

Really, you are only giving him fodder for his delusion at this point.

Marios--the other person they see as smart has said that NOTHING is deterministic. That no scientists consider anything deterministic anymore.

When he said that, I put him on my "nut" list, because, at least on a macro level, EVERYTHING is deterministic--that's why we do science... every effect has a cause. We couldn't figure out how things worked if everything was random.

So you can see how their underlying belief or philosophy along with vague definitions allows them to "win" ever argument by bringing it back to the word random. Everything is random and they think it's important to drive the point home with evolution. They are as daft and disingenous as you perceive... and as impenetrable as mijo. It's the same idiotic semantic silliness, and it, apparently works to make people feel like they know something while remaining utterly clueless about how ignorant they sound.

It kind of sucks over on the Dawkins forum because it's on a thread entitled "How evolution is not pure random chance"--so some honest really smart people drop in to explain... then such people are immediately misheard, maligned, and the same nuttery you see here through years with mijo is recreated. The nutters get a tailored explanation that others would pay for but they cannot hear. They keep sucking people in with loaded questions they don't want the answer too... cut and paste arguments, straw men, allegations about things people never said, insults, experts that supposedly are saying what they are saying, and delusional spin that someone is taking them seriously... all tossed about in self important word salads complete with graphs and big words and the assertion that 50 years of biology or genetics or someone important agrees with whatever the hell their point is or whatever deep thing they are saying.

I try to keep an eye out so they don't abuse new folks who innocently post to address the OP over there, but for the record, the ones that insult your words there are akin to such people here in intelligence, honesty, education, hubris-- and the ones that you'd find intelligent and worth talking to parse your words and are frustrated with the continual shenanigans of the mijo-esque.

Please use your intelligence at skeptic wiki or on those who can understand, not on those who need you to be wrong so that they can imagine they have a clue.

Don't get drawn into the cut and paste crap... unless you have some idea of what you want to accomplish. Remember, you are being used as fodder for the Mijo game that so many here played to no avail under the assumption that he actually wanted to understand why Scientists say evolution is not random. Susu is the brightest they have.
You are wasting intelligence on the inane. They may not be creationists... but they seem to have been taught by them, and they imagine they know more than those who might remedy the ignorance.

[GreyICE]

I don't understand their answers. There are plenty of versions of evolution which are entirely deterministic - every computer simulation of it ever done, for example. Not to mention that nature itself (and therefore the evolution of life on earth) could be fundamentally non-random for all we know.... so the argument fails completely.

It's a semantic debate, and totally uninteresting.



One of the truly bizarre aspects of this is that you bring that up (and I believe you did so also in the past) as evidence that evolution is random. On the contrary, it's evidence that it is non-random, because Lenski was able to engineer the appearance and survival of several very specific mutations. All the colonies evolved some of the same traits given enough time. The fact that one colony was first on any particular one is tautological - did you expect all of them to mutate the same way at precisely the same time?

It would be interesting to know whether one can define a mathematical limit of infinite time, and if so whether all colonies would asymptote to the same sequence. But regardless of that, it is completely obvious that the mutations that persisted were non-random - on the contrary, they were adaptive and advantageous in that environment.

Sol, normally I've been avoiding this stupid debate.

But this was too painful to ignore. Ye gods.

Yes, we would expect every colony to evolve the same mutation at the same time if the system was deterministic. That's what deterministic means. We do not expect variations in travel time for objects moving at fixed velocities between two points. We do not expect that equal quantities of reactant in a chemical reaction will produce different amounts of energy. We do not expect that objects dropped in a vacuum from the same height to hit the ground at different times.

The system is betraying characteristics of both randomness and determinism. That's what stochastic IS.

How is it I come back to this stupid debate after all this time and we're STILL STUCK HERE?!?

[articulett]

Because to some people, everything with any randomness IS "random" (RP, Mijo, jim-bob)... whereas, to biologists, unless something is wholly random, we would never describe it as random, though we might describe it as "non random or not random"-- which is why Miller, when asked at the Dover trial if evolution is random-- said "no".

The "mijo-esque" think this is wrong because anything that has any randomness has been proven "random" per their definition. It's just the same old semantic silliness that mijo has been engaging in all over skeptic sites for years. Apparently it works on some. I think we should encourage people who have that definition to talk to each other since they are wasting Sol's time and everyone else who aims to clarify with their loaded questions and repeated misconstruing of what was said. They have no capability to engage in such a discussion because they cannot answer sol's simple queries as he would.

Skepticwiki could use the smart people on this forum to explain things to those who are actually interested in understanding-- rather than wasting their breath on those who cannot because of their need to prove that evolution is "random" (whatever vague thing that might mean.)

[recursive prophet]

@GreyICE: To burn the semantics straw man on RDF most-including myself-eschewed ‘random’ altogether. Stochastic is a far less ambiguous, scientific term. Yet those here clinging to random contribute to the rerun of this earlier argument, while bemoaning the recursion. If you could have followed the RDF thread you just might see things differently, but I can’t imagine you or any of the current posters here taking the time. And so it goes.

The system is betraying characteristics of both randomness and determinism. That's what stochastic IS.
You won't find any at RD disagreeing with the above statement. Naturally there are deterministic characteristics, but they can not be predicted in advance-which is what MAKES them stochastic. I think much of the confusion here has occurred because of the way stochastic and non-scientific definitions of random have been conflated. Thus we have the repeated charge that this is all primarily about semantics.

@sol: As before I left in susu’s reply to David as he makes some good points and includes some informative links. This is all getting beyond my aged grasp as a whole, and I keep coming up with questions I forgot I had already found answers for. I believe everything IS connected to everything else ultimately, and there are many paradoxes only revealed through thorough examination. The deeper you dig, the easier it becomes to lose your way. How I envy the certitude I encounter in cyberspace. Much of the discussion on RDF, as susu's reply below indicates, revolves around the role of allele drift and reproductive frequency variances. Only one poster here ever addressed susu's arguments on these factors that I posted back on page 4.(?)

I'll take that as a wind-up. It is pointless even arguing with someone who takes a minor pont about NS being stochastic and blows it up into stochastic is the "be-all and end-all".

You were the one, who has basically stated that all similarities between members of a clade are driven by selection rather than common ancestry. If you hold that, you deny the evidence for common ancestry (all of which is structural similarity with no direct adaptive value). I know that you don´t intend to say that common descent is rubbish, but your "minor point" has large consequences.

Utter nonsense. Also, you have had to pick your real examples very carefully (eg species, such as yeast, in which the breeding population is so large that more than one beneficial mutation can come along at once).

See, you are going to reject any example of drift. It´s completely pointless in adding more, but apart from yeast, E.coli (Lenski), H.sapiens, Drosophila (Ayers work). Want more:
more humans (http://linkinghub.elsevier.com/retrieve/pii/S0002929707639780)
Tomatoes, Mazie and Drosophila (http://www.genetics.org/cgi/content/full/149/4/2099#EXAMPLES)
Mice, with implications for mammals in general (http://www.columbia.edu/itc/biology/pollack/w4065/client_edit/readings/natgen14_146.pdf)
Hepatitis C (http://jvi.asm.org/cgi/content/abstract/68/8/4776)
Bachnera, the bacterial endosymbionts of aphids (http://mbe.oxfordjournals.org/cgi/content/abstract/16/1/83)

Many more, where those came from (a scholar.google search for articles on drift).

Not that this is relevant to inflation, but it is actually an interesting question. Unfortunately susu.exp's response makes it quite clear that he/she/it has no idea what she/he/it is talking about.

Off the top of my head I'll give a list of three fully deterministic models of fundamental physics:

1) The "many worlds" interpretation of quantum mechanics, which is simply standard QM with the collapse postulate dropped. The Schrodinger equation is a non-singular differential equation and therefore deterministic. Incidentally this is the most widely favored interpretation among physicists, at least according to some informal polls.

Still necessitates stochastic models. MWs is deterministic in the sense that all possible universes are realized, it is stochastic however in which universe we end up (ignoring that another we ends up in another world).

2) Bohmian interpretation of QM with pilot wave etc. Fully deterministic.

But non-local.

3) Any of several more obscure deterministic non-local hidden variable theories.

Again: Non-local. And non-local HVTs still need stochastic models, simply because they´ve got causation flipped in time, the future determines the past. But since the future´s an unknown we have to model stochastically.

@Articulett: Unsubscribed? Just stalking now are we? Not that I expected you to really drop out of this thread as you’ve so often promised, but don’t you ever feel even a bit ridiculous forever going on about what a waste of time it is debating those who oppose you both here and on RDF and trying to discourage others from participating, while YOU are the most active poster on both??? Literally chanting all the ‘smart people’ have left? Can’t we assume from this claim that YOU aren’t one of them? Do you have ANY idea how many times you’ve said that, or quoted Miller? A very clear explanation was presented as to why a scientist might say that to a general audience, but of course you never read what others write or respond directly to points they make. It’s all the same ‘elections are random,’ Miller and Dawkins said, Behe, creationists, all the experts, and these people have this mysterious drive to believe everything is random.’

How about this? I confess you’ve got it nailed, Artic. I’m obsessed with the need to believe everything is random. I suspect Marios planted a wetware virus he created in an email he sent me. It took over my OS, and started flashing subliminal messages too fast to be consciously observed on my screen. Things like RANDOMNESS RULES!!, NOTHING IS DETERMINISTIC, there IS no determiner, as all is random… YOU must determine your fate! Stuff like that. Now I’ve become one of their randombies. While I generally only use the word stochastic on the boards, as you have discovered we secretly mean random in whatever way you define it. OK?

Of course that’s not going to stop you, is it? Now that I’m getting ready to wrap up my participation in this discussion until a credible opposing argument that can stand up to cross examination by the 'random brain trust' appears, I do worry about you if both threads die out. This and the other random thread are the only ones you post on in either forum, and are where you make most of your overall posts. I can only hope for your sake jimbob or some other will find your tirades amusing. As the 2 actually posting science based arguments on RDF that NS wasn’t always stochastic have capitulated and I no longer worry about your distracting the grown-ups, I can now just laugh at how you’ve elevated projection into a comedic-farce work of art. Kudos. :clap: (Oh, and please note my new sig and in doing so refrain from any further 'imagined expertise' charges, hmm? You have enough of that type of expertise for the entire site. :rolleyes:)

[jimbob]

Because to some people, everything with any randomness IS "random" (RP, Mijo, jim-bob)... whereas, to biologists, unless something is wholly random, we would never describe it as random, though we might describe it as "non random or not random"-- which is why Miller, when asked at the Dover trial if evolution is random-- said "no".


You keep saying this, when I am actually saying that a system is random if the outcome is significantly affected by random events.



Let's take Sol's example of a smoke detector:

I used to design fire sensors and detectors before designing semiconductor devices. Anyway; fire detctors (including smoke detectors) are designed to alarm within a specified timeframe when tested with specified test fires. The detector will operate within this timeframe, so any random factors are insignificant.

Compare this with evolution: One flask of bacteria evolve citrate metabolism, whilst eleven other near-identical ones don't. I would argue that this difference is significant.

The course of evoution on Earth has been significantly altered by random events. The lack of dinosaurs is significant.

I "keep harking back" to Lenski's experiment because the aim was to see if evolution replayed itself the same way under as near-identical conditions as possible.

ETA: I am aware that rationalwiki can be edited by many people, so this isn't an appeal to authority: However it is evidence that other people take the view that historical contingency invokes chance events:


I disagree:

ETA: This is also Lenski's interpretation:
Maybe so, but if so there is no evidence for that in the quotes you gave.


More generally, we suggest that historical contingency is especially important when it facilitates the evolution of key innovations that are not easily evolved by gradual, cumulative selection.

Unless "historical contingency" is nonrandom...
And you persist with this false dichotomy for the thousandth time.

I'm done discussing this with you. Bye.

This is what rationalwiki has to say, which I agree with: (http://rationalwiki.com/wiki/Historical_contingency)

Historical contingency is an important idea in biology and evolution. It is the theory that the paths that life can evolve on are constrained by historical events that are often random. This was perhaps most succinctly described by Stephen Jay Gould, when he proposed the thought experiment of rewinding the tape of evolution backwards and replaying it. The question is: "would life evolve in the same way or in some very different manner?" This idea cuts to the heart of many issues surrounding the inevitability of humanity and determinism.

Historical events that can constrain evolutionary pathways are varied. They can include things that work at the genetic level such as genetic drift, population bottlenecks, neutral selection, exaptation, and potentiating mutations, but also external phenomena in the environment, such as climate cycles and meteorites.

Recent work by biologist Richard Lenski has looked at the role of potentiating mutations in historical contingency. The ability for multiple neutral mutations to add up and eventually allow for a beneficial mutation provides a vivid counter-example to the intelligent design propaganda of Michael Behe's book The Edge of Evolution.



Do you agree that only one strain has evolved citrate metabolism, and that this was facilitated by a mutation between 15,000 and 20,000 generations?

Do you agree that Lenski said that in this case "historical contingency" was important?

I say that "historical contingency" is a phrase that denotes randomness or more accurately: "chance events". Apparently this is incorrect, so could someone please tell me what it does mean, in the context of investigating whether replaying the tape of evolution always gives the same result?


Remember, I also say that over moderate timescales with stable environments, the random nature is often unimportant. However the exceptions are the bits that are interesting.

[jimbob]

A different question:

I define a random process as one where significant differences occur due to chance.

Does anyone disagree with this? This moves the discussion onto what is significant.

I say that evolution of citrate metabolism in one population is significant, especially when the other populations, are still as far from evolving this as the day that the experiment started.

[sol invictus]

Yes, we would expect every colony to evolve the same mutation at the same time if the system was deterministic. That's what deterministic means.

Deterministic means that if the initial conditions are absolutely identical in every way, the result will be identical. In the example I gave that means that not only are the initial conditions for the colonies are identical, so are the seeds for the pseudo-random number generator. Of course no one would run a simulation twice like that (what would be the point?), which means a perfectly deterministic simulation of those 12 colonies would be expected to produce results just like Lenski's.


We do not expect variations in travel time for objects moving at fixed velocities between two points. We do not expect that equal quantities of reactant in a chemical reaction will produce different amounts of energy. We do not expect that objects dropped in a vacuum from the same height to hit the ground at different times.

Wrong. In real life, all of those quantities will vary - because the initial conditions are never perfectly identical, the precision of the measurements is finite.

The system is betraying characteristics of both randomness and determinism. That's what stochastic IS.

I agree that evolution, in the standard picture, is stochastic according to some definitions. It is also non-stochastic according to others. That's a semantic debate with zero content.

[GreyICE]

@GreyICE: To burn the semantics straw man on RDF most-including myself-eschewed ‘random’ altogether. Stochastic is a far less ambiguous, scientific term. Yet those here clinging to random contribute to the rerun of this earlier argument, while bemoaning the recursion. If you could have followed the RDF thread you just might see things differently, but I can’t imagine you or any of the current posters here taking the time. And so it goes.

You won't find any at RD disagreeing with the above statement. Naturally there are deterministic characteristics, but they can not be predicted in advance-which is what MAKES them stochastic. I think much of the confusion here has occurred because of the way stochastic and non-scientific definitions of random have been conflated. Thus we have the repeated charge that this is all primarily about semantics.

Ye gods, I get it. This entire debate is a tribute to the false dilemma. You're arguing that if something is not a cat, it has to be a dog.

A process can be neither random NOR deterministic. Implying that a disproof of one in any way effects another is just stupid.

R: "Look, we can prove it's not a dog. Dogs have great senses of smell, 4 limbs, and are primarily scavangers and small game hunters. This evidence shows that it's clearly not a dog."
S: "Look, cats have retractable claws, prehensile tails, and are predators. This does not have those characteristics. It's not a cat."
R: "Yes, but this this and this prove it's not a dog!"
S: "Yes, and combined with this and this, I think it's pretty safe to say we're talking about a bottlenose dolphin here. The flippers, tail, and blowhole are a pretty good give-away."
R: "Well, really, a layman's description of bottlenose dolphin is cat, so you're just hung up on semantics."
S: "..."
R: "Look, you have no response to that argument. I win."


Stochastic is not random. Outcome of stochastic systems can be determined. The degree of predictability may vary (for pressure, it's going to be damn close. For weather, well, we can predict that there will be precipitation in a large area at some point in a mobile timeframe, hence the 30% chances of rain). But just saying stochastic systems are unpredictable like random ones is WRONG. This is the sort of wrong as in "Factually false. Incorrect. Has no truth, validity, or bearing on anything."

[jimbob]

GreyICE:

I am arguing that it is misleading to call evolution "random", but it is even more wrong to call it "nonrandom".

[articulett]

Stochastic is not random. Outcome of stochastic systems can be determined. The degree of predictability may vary (for pressure, it's going to be damn close. For weather, well, we can predict that there will be precipitation in a large area at some point in a mobile timeframe, hence the 30% chances of rain). But just saying stochastic systems are unpredictable like random ones is WRONG. This is the sort of wrong as in "Factually false. Incorrect. Has no truth, validity, or bearing on anything."

Bingo! Strange but true.

I'd say a liger makes a better analogy... a cross between a tiger and lion. These are interesting beasts as are their genetics, but because it isn't a lion, certain people insist that makes it a tiger. It's more accurate to say it's not a tiger and explain what it is (a hybrid)...

You can't describe what it is or how it's unlike a tiger by insisting that anything that has any tiger blood in it is a tiger.

That's what they are doing with evolution. They are insisting that it's a tiger and claiming that everyone who disagrees is calling it a lion.

Daft and dishonest-- but there you have it. And they think they can "win" this silly semantic game. That's why I think we should leave them to their own devices. Otherwise they will run around telling everyone that you are calling it a lion-- and they're lyin' :p

[mijopaalmc]

Ye gods, I get it. This entire debate is a tribute to the false dilemma. You're arguing that if something is not a cat, it has to be a dog.

A process can be neither random NOR deterministic. Implying that a disproof of one in any way effects another is just stupid.

R: "Look, we can prove it's not a dog. Dogs have great senses of smell, 4 limbs, and are primarily scavangers and small game hunters. This evidence shows that it's clearly not a dog."
S: "Look, cats have retractable claws, prehensile tails, and are predators. This does not have those characteristics. It's not a cat."
R: "Yes, but this this and this prove it's not a dog!"
S: "Yes, and combined with this and this, I think it's pretty safe to say we're talking about a bottlenose dolphin here. The flippers, tail, and blowhole are a pretty good give-away."
R: "Well, really, a layman's description of bottlenose dolphin is cat, so you're just hung up on semantics."
S: "..."
R: "Look, you have no response to that argument. I win."

These "dialogues" are really quite stupid, because, quite frankly they bear no resemblance to what is actually being argued by those who argue that evolution is a random/stochastic process.

"Random", in so far as it is a synonym of "stochastic", implies that, given a set of initial conditions, there is more than one possible outcome, where as "deterministic" implies that given a set of initial conditions, there is only one possible outcome. Therefore, since there is no number that is simultaneously equal to one and greater than one, a system cannot be both "random" and "deterministic".

[GreyICE]

These "dialogues" are really quite stupid, because, quite frankly they bear no resemblance to what is actually being argued by those who argue that evolution is a random/stochastic process.

"Random", in so far as it is a synonym of "stochastic", implies that, given a set of initial conditions, there is only one possible outcome, where as "deterministic" implies that given a set of initial conditions, there is more than one possible outcome. Therefore, since there is no number that is simultaneously equal to one and greater than one, a system cannot be both "random" and "deterministic".Can someone post an O RLY owl, because this definition of random is pure win. Apparently Harvey Dent was onto something with that coin.

I vote that we end the thread here, an argument cannot possibly get more FUBARed than this.

[mijopaalmc]

Can someone post an O RLY owl, because this definition of random is pure win. Apparently Harvey Dent was onto something with that coin.

I vote that we end the thread here, an argument cannot possibly get more FUBARed than this.

So you are capitalizing of the fact that I switched around the implications of "random" and "deterministic"?

Stochastic (http://mathworld.wolfram.com/Stochastic.html)


Stochastic is synonymous with "random." The word is of Greek origin and means "pertaining to chance" (Parzen 1962, p. 7). It is used to indicate that a particular subject is seen from point of view of randomness. Stochastic is often used as counterpart of the word "deterministic," which means that random phenomena are not involved. Therefore, stochastic models are based on random trials, while deterministic models always produce the same output for a given starting condition.

[articulett]

Can someone post an O RLY owl, because this definition of random is pure win. Apparently Harvey Dent was onto something with that coin.

I vote that we end the thread here, an argument cannot possibly get more FUBARed than this.

I think you've underestimated both the hubris and the maximum FUBARedness of those who need evolution to be "random".

[cyborg]

That´s BS. Any stochastic process will have a result. But there´s no way to look back at it and say "It had to happen that way", only "It did happen that way". That´s the difference between a stochastic and a deterministic process.

I believe this is the definition of a semantic difference.

I rest my case.

[sol invictus]

Still necessitates stochastic models. MWs is deterministic in the sense that all possible universes are realized, it is stochastic however in which universe we end up (ignoring that another we ends up in another world).

If you twist the definition of "stochastic" so much that everything is stochastic, everything is stochastic.


2) Bohmian interpretation of QM with pilot wave etc. Fully deterministic.
But non-local.

So freaking what? What does that have to do with stochastic - or does it now mean "local" too?


3) Any of several more obscure deterministic non-local hidden variable theories.
Again: Non-local. And non-local HVTs still need stochastic models, simply because they´ve got causation flipped in time, the future determines the past. But since the future´s an unknown we have to model stochastically.

"Causation flipped in time"?? Utter nonsense. These are standard causal theories which evolve according to differential equations in time. They're deterministic.

It's quite clear this guy/girl/thing is just arguing for the sake of seeing their words on the screen. I'm done with this.

[recursive prophet]

I must be in the right place seeing all the recursion going on here. Where to begin?

Susu wrote: “Still necessitates stochastic models. MWs is deterministic in the sense that all possible universes are realized, it is stochastic however in which universe we end up (ignoring that another we ends up in another world).”
If you twist the definition of "stochastic" so much that everything is stochastic, everything is stochastic.
Do those herrings you’re serving up come with a crème sauce sol? What definition of stochastic would you use that would refute susu’s comment above? This will be the third time I’ve stated my basic, working definition of stochastic.

In my reply to GreyICE on page 4; “I presented my working definition from Wiki below earlier, and none refuted it.”

"A stochastic process is one whose behavior is non-deterministic in that a state's next state is determined both by the process's predictable actions and by a random element."
It's quite clear this guy/girl/thing is just arguing for the sake of seeing their words on the screen. I'm done with this.
That’s your choice sol, and one I’m ready to make also. I’ve watched jimbob and susu make many politely expressed arguments on this thread. Instead of being responded to individually, as susu and he both do yours, you and cyborg-the only 2 here I count as serious posters-cherry pick certain aspects of their statements to disdainfully attack. And trust me I don’t assume you have any interest in this debate. YOU assume I care what interests you.

As to susu, ‘it’ is a he, working toward his doctorate in evolutionary biology in Bonn. I might note that he expresses himself pretty well considering English is a second language. Do you have any degrees in biology, sol? Fluent in any other language? Of course neither would make your opinion either right or wrong. But to say someone like susu is writing to see his words on the screen is pretty insulting given he has made a serious effort to communicate. And he does have at least 1 published peer review paper I know of. Do you have any? It wouldn’t really surprise me if you do, but not in any area of biology and I suspect that’s the point you’re missing. Local is important in biology.

I wasn’t going to bother posting your last reply on RDF, but as mijo did so here is susu’s latest rendition of the visiting orator at the asylum for the deaf and dumb.
If you twist the definition of "stochastic" so much that everything is stochastic, everything is stochastic.

Not at all. The explicity point sol has made earlier was that there were "models in fundamental physics for both possibilities" (stochastic and deterministic). The model used is stochastic, the MWH is a deterministic interpretation of that stochastic model. Even when you accept the MWH and therefore argue that there´s an ontological deterministic reality, your model(!) is stochastic.

So freaking what? What does that have to do with stochastic - or does it now mean "local" too?

Not at all, but non-local theories use stocahstic models, even if the theories are deterministic.

"Causation flipped in time"?? Utter nonsense. These are standard causal theories which evolve according to differential equations in time. They're deterministic.

Let´s look at this, shall we?
Non-locality means that causal effects can happen faster than c. And thus an event some distance away can have a causal effect on some event that will happen prior to the time of simultanety.

http://www.post-neo.com/evo/nonloc.gif

I (situated on the y axis) look at two events (the solid green circles), one of which cases the other one (black arrow). To me the one that is the cause happens after the effect (transparent circles and red arrow). In local theories you avoid causes that happen after their effects to any observer. That´s the whole point of locality. A non-local theory postulates that there are causes for aparently uncaused behaviour, but they have not as of yet occured to the observer. Thus non-local theories have to assume ignorance of the causes and therefore use stochastic methods.

I think you've underestimated both the hubris and the maximum FUBARedness of those who need evolution to be "random".

Picture Mijo and I, giving each other high fives. I feel the need…the need for ..Randide! THIS is the most random and chaotic of ALL possible worlds!! The words DETERMINISM DEMANDS A DETERMENER keep flashing on my screen. CHAOS MAKES FREE!! I now NEED to believe! It..it just has to all be random. That I was somehow possibly destined to find all this so remarkably funny at this juncture would be a recursive sleight of hand worthy of Escher. Just the thought makes me laugh again. This is good. Thanks.

I sent mijo a PM when I saw his mistake, and predicted certain users would feast on it for some time to come. I don’t know him at all, but imagine he is a social innocent. Well intentioned, but way too optimistic about the benefits of tenacity in sticking to your point with certain people. I believe he was actually surprised they would jump on his obvious ‘dyslectic moment’-most of us have them-instead of just letting it go with perhaps an aside remark. No, they add it to their ad homs stockpile, the soul of their rebuttals. In their glee, one sees why they are here. Both sad and funny, and easily verifiable if one takes the time to check out what they post and see if there is something…deterministic about their approach?


Fortunately I learned long ago not to confuse the message with the messenger, so I keep open the possibility that I simply wandered into a den of educated members of a dying sect on RDF, and that in reality it does all come down to semantics, and one’s starting assumptions. However, having standardized terminology is important in any science, and much of the debate on RDF was about which viewpoint-stochastic/non-stochastic-was the prevailing one in evolutionary biology today. This is partly why I asked the question that was the genesis of this thread. There are some similarities I’ve observed in the arguments made by sol, cyborg, my_wan, and MZ-the ‘irreducible randomness’ proponent on RDF. When time permits perhaps I’ll sift through all the ‘staw herrings’ and ad homs, and present the core argument to susu and Marios as none here can be bothered. I remain convinced there is some fundamental breakdown in communication going on, but have yet to sort out the common denominator.

[jimbob]

Just to add that I am not interested in discussing how some hypothetical version of evolution might work, but in how it actually works in biology.

I think that most people accept that random *external* events can affect the course of evolutionary history.

I also think that it has been shown to my satisfaction that the fact that organisms co-evolve within ecosystems that are affected by other organisms means that if you wait long enough, some disruptive mutation arises that changes the rules for everything else, and puts the course of evolution on a differnt path. The Long Term Evolution experiment was set up to investigate this very question, because it had been a point of contention within evolutionary biology.

In other words, in the real world, the outcome is significantly affected by random events that are inherent to the system itself (co-evolution and mutations).

I would additionally argue that as soon as you have self-replication with errors within a finite system, natural selection occurs, and that full-blown Darwinian evolution follows, including the eventual direction changes due to random events and co-evolution

[sol invictus]

I must be in the right place seeing all the recursion going on here. Where to begin?

I didn't see this before, but I'll respond to the (few) substantive points raised here.

Do those herrings you’re serving up come with a crème sauce sol? What definition of stochastic would you use that would refute susu’s comment above? This will be the third time I’ve stated my basic, working definition of stochastic.

In my reply to GreyICE on page 4; “I presented my working definition from Wiki below earlier, and none refuted it.”

"A stochastic process is one whose behavior is non-deterministic in that a state's next state is determined both by the process's predictable actions and by a random element."

I'm perfectly happy with that definition, and according to it, the MW interpretation of QM is NOT stochastic. There is no random element - I can prove that mathematically if you'd like. The apparent randomness we experience in the world arises in an interesting and profound way from an underlying model which is fully deterministic.

As to susu, ‘it’ is a he, working toward his doctorate in evolutionary biology in Bonn. I might note that he expresses himself pretty well considering English is a second language. Do you have any degrees in biology, sol?

My credentials are irrelevant, as are his. Arguments stand or fall based on their own validity, not their typer's degrees or publication history.


Fluent in any other language?

Yes, several. And? What in the world does that have to do with this?

Of course neither would make your opinion either right or wrong. But to say someone like susu is writing to see his words on the screen is pretty insulting given he has made a serious effort to communicate.

He's speaking about a topic he plainly doesn't understand as if he did. I find that very irritating.

Not at all. The explicity point sol has made earlier was that there were "models in fundamental physics for both possibilities" (stochastic and deterministic). The model used is stochastic, the MWH is a deterministic interpretation of that stochastic model. Even when you accept the MWH and therefore argue that there´s an ontological deterministic reality, your model(!) is stochastic.

False, as above. And rather incoherent.

Not at all, but non-local theories use stocahstic models, even if the theories are deterministic.

So now he comes full circle and starts over.

Please read the above statement in light of your own definition. If the model is fully deterministic - and these are - they are not stochastic. There. is. no. random. element. Locality has nothing to do with it.

Let´s look at this, shall we?
Non-locality means that causal effects can happen faster than c. And thus an event some distance away can have a causal effect on some event that will happen prior to the time of simultanety.

This is completely wrong, of course. Any theory which behaves in the manner suggested above does not exist. In the theories I referred to, the non-locality does not lead to causality violation (that criterion is among the most basic they must pass in order to be considered for more than a moment). They do not avoid it by adding randomness - that would defeat the entire purpose, since they were constructed precisely so as to avoid that. They do it in various other ways (some problematic, some not), which I would be happy to describe in a different thread where it's on topic.

[cyborg]

In other words, in the real world,

No, in other words, "in the real world" is the phrase you use to justify the conclusion you want to hold. I'm sure this appeal works for you but you don't seem to grasp that it doesn't work for me.

[recursive prophet]

@sol: Sorry this reply has taken so long sol, but I was in the humberverse and I didn’t even notice this thread had reappeared until last night. I have pasted in a reply by susu to your last post, but would really appreciate it if you could read all the comments starting with this link. (http://richarddawkins.net/forum/viewtopic.php?p=1725829#p1725829) My_wan is back, and I would like to read your take on his exchanges with susu and Marios. Just skip over my entries. If you don’t want to bother signing on there, it is very easy to use the RDF quote feature and paste it in here. Going in the other direction is a little more time consuming, but I’ll gladly make the effort to have your opinions represented in this discussion.

Again let me invite any interested in this topic to jump in, either here or on the RDF thread the link above opens. And please note if I could have changed the title to this thread it would read: Is natural selection stochastic? This subject is far more complex than it would appear at first glance, and I myself still have doubts about the answer and its relevance in evolutionary science today. Show me the light!

To SI:

I didn't see this before, but I'll respond to the (few) substantive points raised here.
A stochastic process is one whose behavior is non-deterministic in that a state's next state is determined both by the process's predictable actions and by a random element.
I'm perfectly happy with that definition, and according to it, the MW interpretation of QM is NOT stochastic. There is no random element - I can prove that mathematically if you'd like. The apparent randomness we experience in the world arises in an interesting and profound way from an underlying model which is fully deterministic.

In the Copenhagen interpretation, the mathematics of quantum mechanics allows one to predict probabilities for the occurrence of various events. In the many-worlds interpretation, all these events occur simultaneously. What meaning should be given to these probability calculations? And why do we observe, in our history, that the events with a higher computed probability seem to have occurred more often? One answer to these questions is to say that there is a probability measure on the space of all possible universes, where a possible universe is a complete path in the tree of branching universes. This is indeed what the calculations give. Then we should expect to find ourselves in a universe with a relatively high probability rather than a relatively low probability: even though all outcomes of an experiment occur, they do not occur in an equal way. As an interpretation which (like other interpretations) is consistent with the equations, it is hard to find testable predictions of MWI.

Different interpretation of the stochastic equations of QM, but still using the stocahstic equations? Probability assigned to which of the multiple worlds we´re in rather than the behaviour of particles? Seems like the model is still stochastic.

Not at all, but non-local theories use stocahstic models, even if the theories are deterministic.
So now he comes full circle and starts over.

Seriously, show me a deterministic model, rather than a deterministic interpretation. MWI uses the same model as the Copenhagen interpretation, you still get a probability for where a single electron is going to be detected in a double slit experiment. The interpretation of that probability is different, the model is not.

Please read the above statement in light of your own definition. If the model is fully deterministic - and these are - they are not stochastic. There. is. no. random. element. Locality has nothing to do with it.

Do you know what a model is?

Let´s look at this, shall we?
Non-locality means that causal effects can happen faster than c. And thus an event some distance away can have a causal effect on some event that will happen prior to the time of simultanety.
This is completely wrong, of course. Any theory which behaves in the manner suggested above does not exist. In the theories I referred to, the non-locality does not lead to causality violation (that criterion is among the most basic they must pass in order to be considered for more than a moment).

I thought my picture drove that point home. Maybe I should put it into formulae:
Event A happens in some distance xA from an observer. It causes an event B in some distance xB from the observer. The time our observer registers A is tA+xA/c and the time B is registered is tB+xB/c. tb however is tA+(xB-xA)/v, v being the velocity of the interaction between A and B.

So the time between A and B are registered is:
t=tA+(xB-xA)/v+xB/c-(tA+xA/c)
=(c-v)(xB-A)/(vc)
Which turns negative if xB<xA and |v|>c.
Since you can´t have xB>=xA for all observers, you need to ensure |v|<=c and what is a theory called in which |v|<=c holds? Local. So after My_Wan proposed hidden variable theories without hidden variables, you are now proposing non-local theories with no non-locality? You guys are really flooring me.[/quote]

[sol invictus]

I have pasted in a reply by susu to your last post, but would really appreciate it if you could read all the comments starting with this link. (http://richarddawkins.net/forum/viewtopic.php?p=1725829#p1725829) My_wan is back, and I would like to read your take on his exchanges with susu and Marios.

No thanks - one forum is more than enough.

I assume what follows are susu's replies?

Different interpretation of the stochastic equations of QM, but still using the stocahstic equations?

There are no stochasic equations in the MWI of QM.

Probability assigned to which of the multiple worlds we´re in rather than the behaviour of particles? Seems like the model is still stochastic.


Nope. Wrong. The model is a set of deterministic differential equations.

Seriously, show me a deterministic model, rather than a deterministic interpretation.

I already did. Bohm is another, as I already said. This has now gone around 540 degrees.


MWI uses the same model as the Copenhagen interpretation,

Wrong. There is no collapse - because there is no stochastic element.

you still get a probability for where a single electron is going to be detected in a double slit experiment. The interpretation of that probability is different, the model is not.

It's just the opposite.


Do you know what a model is?

Far better than you, obviously.


I thought my picture drove that point home.

I know what your picture was intended to show, and as I said, it's wrong and irrelevant. Such models do not exist - they are neither logically self-consistent nor capable of describing the world. You cannot find a two-parameter family of solutions to a 2nd order hyperbolic diff eq if it has that problem, so you throw it away.


Since you can´t have xB>=xA for all observers, you need to ensure |v|<=c and what is a theory called in which |v|<=c holds? Local.

Wrong. That is not the definition of locality. Newtonian dynamics, in which v can be arbitrarily large, is completely local.

So after My_Wan proposed hidden variable theories without hidden variables, you are now proposing non-local theories with no non-locality? You guys are really flooring me.[/QUOTE]

You don't have a clue what you're talking about.

[recursive prophet]

@sol: Well, its been a busy time on the RDF thread, so I have yet to get a reply to your last post. Hopefully you might find the exchange between susu and my_wan interesting, as it covers some of the points you made. I also threw in what I consider an excellent summary of his position by MZ, along with an analysis by Dlx2. I think he has a point with his 'shell game' analogy, yet as I pointed out in my response he assumes those who see all of evolution as stochastic know without seeing it what shell the pea is under. It's been an education, for sure, but without all the laughs humbert et al. provide. Perhaps the ancients had it right; at the end of the day all falls within the realm of philosophy.

Exactly what it says. It makes no difference whether a deterministic theory changed todays theory or not. The claim that a deterministic model entails the destruction of a random model is a false dichotomy, yet even if it did in a particular case it doesn't change the fact that both random and deterministic models can coexist.
Nope. Not as equivalent hypotheses. And generally you can get deterministic models as approximations for large N from stochastic ones or you can get stocahstic models for gross behaviour of deterministic ones. But you don´t get equivalent models.
Here you rightfully stated you can get stochastic models for gross behavior of deterministic ones, and that gross behavior can itself have deterministic behavior. So explain to me how you know when a stochastic model is not the result of gross behavior of deterministic events? To do that you must assume that what you don't know doesn't exist. Random models cannot even predict single events whether those singular events are deterministic or not.
In fact you can't generate, with any random number generator, that can't be defined by some rule after the fact.
There´s an object missing in your sentence.
Yes I did mess that sentence up pretty badly. it should have read: In fact you can't generate any number, with any random number generator, that can't be defined by some rule after the fact.
The above number can be generated by deterministic rules set determined by "-2.234323.22.3152.31.54", and ANY random series of 0 and 1 can be defined by a simpler rule the same way.
Again, that I can get any number in another way doesn´t mean jack. In fact I can generate any number in countless ways. But again, this is totally meaningless. If you want it to be science you have to assume that the rules apply constantly and then predict the future. And kick it out if it doesn´t make correct predictions.
Yes it is meaningless, but neither can you say with certainty that any random event couldn't have been predicted if you knew the initial conditions with perfect precision, which is itself impossible to know.
Not knowing the rule to that predicts a particular random series does not mean no such rule exist. Randomness measures what we don't know. The claim that deterministic randomness is an oxymoron entails the claim that what we don't know doesn't exist.
Now you´re off to lala land completely. I´ll repeat it to you: stochastic means that for the same innitial conditions there are multiple possible outcomes. Deterministic means that there is only one. 2*3 is always 6. I can multiply the two as often as I want and I will always get 6. Try that with a die.
Yet you can't know the initial conditions were the same. To know that you must know the exact state of every atom and each electronic state of those atoms in which that dice interacts with, and the dice itself. Then the calculation requires not only an extreme number of variables N, but also the calculation requires an extreme number of significant digits. Calling it the "same initial conditions" is an approximation that simply isn't technically true.
You are correct about the timeline, wrt the original ideal gas theory, yet the equivalence wasn't established till later with some modifications.
There is no equivalence.
Read below.
See, there´s something amazing about this quote. Maybe it´s the fact that it stops, where it stops rather than continuing with the following lines:
The Kinetic Theory of Gases (http://books.google.com/books?id=BXt-Ne7ytxYC&pg=PA532&ei=LS2PSZayKpCIkASjlsnIBQ) By Stephen G. Brush, Nancy S. Hall"]But the "reducing theory" was also affected: the introduction of statistical thermodynamics and in particular the recognition that irreversibility involves some kind of randomness at the molecular level27 helped to undermine the rigid determinism of Newtonian mechanics. The result was that physicists in the fist part of the 20th century were prepared to accept a new mechanics based on indeterminism
The principle of "rigid determinism" of Newtonian mechanics wasn't abandoned, it was replaced with statistical models because of the impossibility of establishing initial condition in perfect detail, or even calculating an outcome of that many variables with essentially infinite significant digits. It was the practical indeterminism, not the principle of determinism, that was replaced in the kinetic theory. They went to great lengths to validate that the statistical assumptions were valid approximations for smaller N, that could be calculated deterministically, to be secure about using the statistical method for larger N. Calling initial conditions of a system with that large N "the same" is only an approximation. It can't be considered "the same" in an absolute sense.
Oh, the reference given in your section, footnote 26 of course is Einsteins 1905 paper on Brownian motion (A.Einstein Ann.d.Phys. 17(1905)549-560 “On the movement of small particles suspended in a stationary liquid demanded by the molecular kinetic th eory of heat.”.) In this paper he shows that the deterministic classical theory of thermodynamics predicts that there is no Brownian motion, while the statistical mechanic theory predicts it.
Wenn sich die hier zu behandelnde Bewegung samt den für sie zu erwatenden Gesetzmäßigkeiten wirklich beobachten läßt, so ist die klassische Thermodynamik schon für mikroskopisch unterscheidbare Räume nicht mehr als genau gültig anzusehen [..]
If the motion discussed here is actually observable, along with the expected rules for it´s behaviour, then classical thermodynamics is only an approximation even on a scale that can be investigated with a microscope[..]
Translation: "If the movement here, together with the expected regularities for them, can be observed, then classical thermodynamics even for microscopically distinguishable spaces can no longer be exactly valid". German is a bit awkward in translation. For a full English reference see: http://lorentz.phl.jhu.edu/AnnusMirabilis/AeReserveArticles/eins_brownian.pdf (PDF)

The problem is you claim here that the classical theory of thermodynamics is deterministic. Yet it is statistical mechanic that is considered deterministic (http://plato.stanford.edu/entries/determinism-causal/#ClaMec), not classical thermodynamics. Yet they both used statistics and ensembles. What was then demonstrated was the deterministic statistical mechanics model (that's what "mechanics" in "statistical mechanics" signified). Remember it was Einstein that made the famous quote: "God does not play dice with the Universe".

So what was the difference in the two models? It wasn't that one used statistics and the other didn't, they both did. The difference was that classical thermodynamics assumed that the randomness of fluctuations was an independent property of the system. Statistical mechanics assumed the same randomness, but the randomness of the fluctuations were actually the result the parts bumping into each other. Randomness didn't have parts in the classical thermodynamics model. So the difference wasn't a difference in randomness, but how that randomness was constrained, deterministically by atoms in the Brownian motion case. Of course it can't be prove that statistical mechanics is purely deterministic either, only that it's consistent with determinism.

Ironically the randomness of QM is also limited to specific quantities, like the atoms in Brownian motion. The problem is that the quanta in QM is not quanta of parts, but quanta of properties.
Complementarity does not require adding anything to it, it only requires mutual validity. Yet it did add to it, and this addition validated the existence of atoms as pointed out in the above book quote.
No it didn´t. It was made falsifiable by Einsteins paper and falsified in turn by Jean-Baptiste Perrins experiments, remaining as an approximate special case at larger scales.
Problem is you got the deterministic theory (statistical mechanics) confused with the statistical theory (classical thermodynamics). It was therefore verified by Einsteins paper, not falsified.
I don't have to "reduce statistical mechanics to the ideal gas law" because it has already been done.
Citation?
See the Einstein's paper on Brownian motion and just try to remember that classical thermodynamics is not deterministic, statistical mechanics is the deterministic model.
That's how the existence of the atom was originally established.
False. The existence of atoms was established by a paper by Einstein and experimental findings supporting his derivations, that showed that the classical thermodynamics (in that case classical thermodynamics in fluids, not in gases) did not universally hold.
Which happens to be opposite your claims because you made the assumption that "statistical" in "statistical mechanics" means random, when actually the "mechanics" in "statistical mechanics" indicates deterministic. Look it up.

So we're now reduced to the question of whether subatomic processes are truly deterministic but mimic random functions, or whether they are truly random.

This escapes the question of whether evolution as a mechanism is random or deterministic. If evolution acts as a stochastic resampling of variation resulting from subatomic processes, then it doesn't matter whether or not the subatomic processes are or are not "random." It only matters whether the mechanisms involved in evolution are stochastic.

And they are.

This thread should seriously be allowed to die peacefully.

The point of the so-called "QM diversion"is that, as far as can be determined at this point in the history of science, we cannot know the initial conditions of an event or its outcome to to arbitrary precision. Systems therefore cannot be even theoretically deterministic, because the best we can do, by the nature of the system, is obtained a range of possible outcomes. This relates to the discussion on evolution in so far as most of those who have argued that evolution is in some ill-defined way nonrandom have posited, at some point along the way, that if we could measure conditions (both initial and final) with enough precision, evolution would become deterministic. In other words, evolution only appears random because we lack adequate knowledge of the conditions of the system.
Historically, evolutionary biologists considered Selection to be deterministic. TODAY, determinism is largely rejected by evolutionary biologists, especially population geneticists. Some still hold Selection to be deterministic, at least in some systems. The rejection of determinism arguably derives not from empirical evidence, but from a shift in philosophy, and the push for more refined mathematical approach to models and theories.

The error in this thread, is in a continued claim that determinism is not only impossible with respect to Natural Selection, but also that no evolutionary biologist thinks deterministically. These claims are based on assertions alone.

The simple truth is, we don't know.

[Dr Adequate]

A different question:

I define a random process as one where significant differences occur due to chance.

Does anyone disagree with this? Yes. That would make my life a "random process".

"Random" is just too sloppy a word for it.

[jimbob]

Agreed DrA, just that it is less wrong than saying "nonrandom"

If you have to chose a one-word answer...

[articulett]

But nobody has to choose a one-word answer... and calling one's life random is as uninformative and misleading as calling evolution random.

Besides, I don't think it's less wrong than saying nonrandom --whether you are talking "life", smoke detectors or evolution--nor does Ken Miller who writes multiple books on evolution, is considered an expert in his field, and testified at the Dover trial. If life my life is "random" than whose life isn't? Or what isn't? What process can't be construed as random via your definition?

It seems the only people who think it's "less wrong" to call such things random is, well, people like you, recursive prophet, and mijo. It's sloppy and not particular useful in describing evolution--more obfuscating than clarifying. That's why scientists don't describe things like you guys do.

[jimbob]

Articulett,

Lenski's experiment showed that you could get significantly different outcomes, when all externallu-applied differences were insignificant. The only important difference in inputs between the populations was the random mutations that occured.

Random mutations are enough to produce different outcomes. Not just in theory, but this has been observed.

ETA: All the populations showed adaptation to their environments, but they showed different adaptations.

[cyborg]

Blah, blah, blah. Ignore the reasons why it was not persuasive in the first place: if at first it doesn't persuade, repeat, repeat, repeat.

[mijopaalmc]

Blah, blah, blah. Ignore the reasons why it was not persuasive in the first place: if at first it doesn't persuade, repeat, repeat, repeat.

So why exactly is the fact that genetically identical populations evolved differently when exposed to the same environment not persuasive when arguing that evolution is a stochastic process?

[articulett]

Putting mentos in diet soda is not about proving that tan foam results.

By the same token, Lenski's data is not about showing how evolution is random. It's about observing evolution on a human time frame! We have the data to go back and determine what mutations were selected and how they accumulated.

Those desperate to see evolution as "random" repeatedly miss valuable opportunities to learn about evolution, as they imagine they are saying something useful to someone other than a creationist interested in obfuscating understanding.

[jimbob]

Putting mentos in diet soda is not about proving that tan foam results.

By the same token, Lenski's data is not about showing how evolution is random. It's about observing evolution on a human time frame! We have the data to go back and determine what mutations were selected and how they accumulated.

Those desperate to see evolution as "random" repeatedly miss valuable opportunities to learn about evolution, as they imagine they are saying something useful to someone other than a creationist interested in obfuscating understanding.

What was the aim of Lenski's experiment then?

Were the initial minor differences in the populations influential in determining that only one population would evolve the citrate metabolism?

Were the unavoidable subtle differences in the environment influential in determining that only one population would evolve the citrate metabolism?

If Lenski hadn't frozen samples at certain generations, then you could argue that maybe these differences *were* influential. However, because of the experimental design, we can see that the important event occured between 15k and 20k generations.

or from his PNAS abstract (http://www.pnas.org/content/105/23/7899.abstract)

The long-delayed and unique evolution of this function might indicate the involvement of some extremely rare mutation. Alternately, it may involve an ordinary mutation, but one whose physical occurrence or phenotypic expression is contingent on prior mutations in that population.

Note that both explanations depend on a mutation. This trait evolved due to a either a rare mutation, or to a common mutation that had its expression modulated by something unique to this population (which had to be further utations).

Later on in the abstract:

However, we observed a significantly greater tendency for later clones to evolve Cit+, indicating that some potentiating mutation arose by 20,000 generations. This potentiating change increased the mutation rate to Cit+ but did not cause generalized hypermutability. Thus, the evolution of this phenotype was contingent on the particular history of that population.

Ot wasa mutation that made the evolution of citrate+ metabolism more likely. This mutation didn't occur in the other populations, it was a random mutation...

And from an earlier paper showing that this was the aim of his experiment:

<in this paper (pdf) (http://myxo.css.msu.edu/lenski/pdf/2004,%20Plant%20Breeding%20Reviews,%20Lenski.pdf)>
Second, this experiment seeks to examine the repeatability of evolution by having 12 replicate populations, all of them founded from the same ancestor and maintained in the same environment. Which phenotypic and genomic aspects of evolutionary change are repeatable in this system, and which are haphazard? How can we understand the causes of parallelism and divergence of replicate lines? The issue of the predictability of evolutionary change - or lack thereof - has long been of interest. The question was well captured by the late palaeontologist Steven Jay Gould (1989) in a thought experiment: "I call this experiment 'replaying life's tape.' you press the rewind button and, making sure you thoroughly erase everything that actually happened, go back to any time and place in the past -say to the seas of the Burgess Shale. Then let the tape run again and see if the repetition looks at all like the original..." Gould went on to say, however, that "The bad news is that we can't possibly perform the experiment." Of course, we could never run an experiment on the vast temporal and spatial scales imagined by Gould. But on much smaller scales, our experiment with E.coli allows us to address the same question. We do so by allowing 12 scenarios to play out simultaneously, rather than sequentially, as the notion of replay implies, but the issue of repeatability is fundamentally the same.


Second, this experiment seeks to examine the repeatability of evolution by having 12 replicate populations, all of them founded from the same ancestor and maintained in the same environment.

This experiment was aimed at seeing how repeatable evolution is. Now obviously there were subtle differences in the environment, but these were not important.

I think I will leave the rest of Lenski's paper for another time, but if you read the full text (pdf here (http://www.pnas.org/content/105/23/7899.full.pdf+html)) especially the "discussion and future direction" section, you can see the author's interpretation of his experiment.

It supports what was said in the abstract.

[jimbob]

lenski's final paragraph:

In any case, our study shows that historical contingency can have a profound and lasting impact under the simplest, and thus most stringent, conditions in which initially identical populations evolve in identical environments. Even from so simple a beginning, small happenstances of history may lead populations along different evolutionary paths. A potentiated cell took the one less traveled by, and that has made all the difference.

How does that not say that random mutations can alter the course of evolution?

[GreyICE]

I'm perfectly happy with that definition, and according to it, the MW interpretation of QM is NOT stochastic. There is no random element - I can prove that mathematically if you'd like. The apparent randomness we experience in the world arises in an interesting and profound way from an underlying model which is fully deterministic.


Multi-worlds theorem introduces its own problems (for instance... everything) and is hardly well-accepted enough to be used to handwave aside a random element. The last time a deterministic theory for QM was posited it was hidden variables. It seems a fair enough statement that QM has a random element - if it is truly is NOT random, the outcome of its events is truly difficult to determine in some fundamental way that really defies analysis.

P.S. If you clone 15 bacteria in 15 petri dishes with identical nutrients, light conditions, etc - will they all evolve the same features at exactly the same time?

Yes - deterministic
No - Stochastic

Which do you think it is?

[cyborg]

P.S. If you clone 15 bacteria in 15 petri dishes with identical nutrients, light conditions, etc - will they all evolve the same features at exactly the same time?

Yes - deterministic
No - Stochastic


Correct - simplistic.

[sol invictus]

Multi-worlds theorem introduces its own problems (for instance... everything) and is hardly well-accepted enough to be used to handwave aside a random element.

It's accepted by the majority of physicists, at least according to one informal poll.

The last time a deterministic theory for QM was posited it was hidden variables.

Huh? What about MW? What do you mean, "last time"?


It seems a fair enough statement that QM has a random element - if it is truly is NOT random, the outcome of its events is truly difficult to determine in some fundamental way that really defies analysis.

So?

P.S. If you clone 15 bacteria in 15 petri dishes with identical nutrients, light conditions, etc

Impossible premise, and so the question is meaningless.

[my_wan]

P.S. If you clone 15 bacteria in 15 petri dishes with identical nutrients, light conditions, etc - will they all evolve the same features at exactly the same time?

Yes - deterministic
No - Stochastic

Which do you think it is?
What does it matter? The notion that identical conditions can be created in 2 much less 15 petri dishes is absurd.

[my_wan]

It seems recursive prophet is posting selected post of mine from RD. Here is my position statement if anyone want to take issue with it.

As posted on RD, these are the facts.

(1) A purely deterministic system can legitimately be modeled as random.
(2) We cannot prove that all systems are fundamentally deterministic.
(3) When no known deterministic model exist we cannot prove that no deterministic model exist.
(4) The empirical content of both definitions are identical.
(5) The meaning of terms are incongruent between the two definitions, causing much consternation.

So take your pick, just be cognizant of when people are using an alternate definition of randomness. The primary difference is that for the reducible definition randomness is a property of the definition, for the irreducible case randomness is a property of the system.

As a result of debate with susu, it became obvious that Einstein's paper on Brownian motion was a seminol example of where a deterministic model was validated experimentally. So the concept of determinism remains an important theoretical consideration, even when we don't presently have such a model.
http://lorentz.phl.jhu.edu/AnnusMirabilis/AeReserveArticles/eins_brownian.pdf

[articulett]

Jim-Bob... only the vaguest of people think that historical contingency makes evolution "random.".. because describing it as such makes every THING random. If a different sperm fertilized a different egg, then you wouldn't exist, nor Hitler, nor the common ancestor of all humans alive today--but that means that whatever made that sperm fertilize that egg would have to have been different too--not on a macro-level... on the level of a hiccup or smaller. The sperm that ultimately fertilizes a given egg does so in a deterministic fashion... we can't know all the inputs. But per your definitions, all life is random. You don't seem to have a clue as to how cumulative selection of the best reproducers plays a much larger role as to what evolves.

So calling everything where randomness plays a major role to Jim (or fellow fuzzy thinkers) "random", isn't something you are likely to see any serious scientist do.

That's why you guys are relegated to conversing with each other. Your definitions are so broad, that they can apply to everything. If life is random (and how can it not be given the above scenario and your "historical contingency" definition) then every THING is. What would it even mean to say something is "nonrandom" to someone who thinks that way? What is an example of a "nonrandom" process? What a silly application of definitions... and whatever for?

We already understand that per your definition, every thing is random-- it's just that people who want to communicate what evolution actually is, use much more precise definitions and explanations. You don't seem to have the education nor capability to do so. You are stuck using your definitions with people like T'ai Chi, Mijo, Behe, RP, et. al.-- not anyone who actually teaches and understands the subject. Lenski, himself, would dismiss your silliness.

[GreyICE]

What does it matter? The notion that identical conditions can be created in 2 much less 15 petri dishes is absurd.

If your argument is that the system is SO DEPENDENT on initial conditions that the fact that we use clones, the fact that we can control nutrient inflow, light conditions, size of dish, etc., to AT LEAST the level of 0.01% variation, and the fact that we can reproduce this for a very large number of dishes STILL means that the conditions have too much variation, you have to accept that whether the system is 'non-deterministic' or 'uncontrollably sensitive' is a distinction that is ONLY semantic.

You're basically jacking off to a semantic argument at this point, and thinking it profound.

[my_wan]

If your argument is that the system is SO DEPENDENT on initial conditions that the fact that we use clones, the fact that we can control nutrient inflow, light conditions, size of dish, etc., to AT LEAST the level of 0.01% variation, and the fact that we can reproduce this for a very large number of dishes STILL means that the conditions have too much variation, you have to accept that whether the system is 'non-deterministic' or 'uncontrollably sensitive' is a distinction that is ONLY semantic.

It can quiet often, for practical reasons, be only semantics. It can also be of immense theoretical value to define the deterministic underpinnings of a statistical model. The existence of the atom was established by a comparison of the non-deterministic classical thermodynamics and the deterministic statistical mechanics via Brownian motion.
http://lorentz.phl.jhu.edu/AnnusMirabilis/AeReserveArticles/eins_brownian.pdf

You're basically jacking off to a semantic argument at this point, and thinking it profound.
So if there is any jacking off, it is the notion you presented that it is ONLY semantics, so that you can pretend that ignoring and denying any definition but your own is worthy of stroking.

[jimbob]

If your argument is that the system is SO DEPENDENT on initial conditions that the fact that we use clones, the fact that we can control nutrient inflow, light conditions, size of dish, etc., to AT LEAST the level of 0.01% variation, and the fact that we can reproduce this for a very large number of dishes STILL means that the conditions have too much variation, you have to accept that whether the system is 'non-deterministic' or 'uncontrollably sensitive' is a distinction that is ONLY semantic.



Furthermore, the experimental design addressed that issue, because every 500-generations, population samples were frozen, and the researchers were able to replicate the experiment.

Below 15k-generations, in none of the replicated samples was citrate+ metabolism opbserved to evolve.

Only in the population of interest, was citrate+ observed to re-evolve in the replicated examples, and with increasing probability from 15k to 20k generations.

The only way that an extreme high-sensitivity to initial conditions could be reconcilled with these replicated experimental results would be if the "hidden" sensitising factor happened to occur only in some of the replicated experiments. Now this isperfectly possible if the important factor was an alteration in the genetic makeup of the population (due to a mutation, or combination of mutations) but that is the only plausible explaination, which brings us back to the statement that random mutations can affect the course of evolution.

[cyborg]

brings us back to the statement that random mutations can affect the course of evolution.

Which says nothing about the causality of the mutations themselves - something you simply cannot get your head around it would seem.

[jimbob]

So we have accepted that mutations can cause significant differences in outcomes.

You now say that these mutations mightn't be random.

Or are you saying that only "nonrandom mutations" can cause these differences? What is your evidence? Or is this your idea that maybe radioactive decay events aren't random?

[cyborg]

So we have accepted that mutations can cause significant differences in outcomes.

Uh, I've never said anything different.

You now say that these mutations mightn't be random.

No. I've always said that the underlying cause of mutations is irrelevant.

It's like you've managed to read a couple of hundred posts from me on this issue and you've still not managed to pick this up. I've never changed what I've said.

[jimbob]

So, random mutations can (and do) significantly affect the outcome of evolution, which includes altering the fitnes landscape for all the other organisms in the ecosystem.

How does this make it nonrandom (which is what I am arguing against).

I have said that in stable ecosystems, over moderate timescales, and with sufficiently large populations then the random aspect can be ignored, but this is only a special case.

Why do you think "nonrandom" is a beter word than "random", when "nonrandom" is incorrect, whilst "random" is merely misleading?

[cyborg]

Ugh.

Did you even read/understand what I just said?

[my_wan]

For this debate, or peoples characterizations of others claims, to make any sense whatsoever depends on what definitions are used for what.

cyborg, jimbob, etc., answer these questions.

(1) Is random and deterministic antonyms?
(2) Does randomness rule out determinism?
(3) Is your answers here model specific or do they also apply to the actual system being modeled?

Differences in how you answer these questions determines what others think you said that you didn't. This leads to accusations of false claims, lying, etc., etc. So answer them or waste your time.

[jimbob]

Well I read it as this:


So we have accepted that mutations can cause significant differences in outcomes.
Uh, I've never said anything different.

So the course of evolution is sometimes altered by random mutations, which are inherent to the system.

The fitness landscape is subject to random change. Which I now think you agree with.



You now say that these mutations mightn't be random.
No. I've always said that the underlying cause of mutations is irrelevant.

It's like you've managed to read a couple of hundred posts from me on this issue and you've still not managed to pick this up. I've never changed what I've said.

So are you saying that for a given stable ecosystem, that evolution is nonrandom? I would say that in this situation, given time, and assuming that no disruptive mutations ocur, then this would be the case. This is like saying that in situations where randomness can be ignored, randomness can be ignored.

In real life, the ecosystems are subject to self-imposed random changes, when random mutations alter the fitness landscape. I agree that these are not very common events, but over geological timescales (and a lot shorter) this is very important.

[articulett]

I think Sol should repost his 3 choice query, because it's clear people who answer the question the same way understand each other, and those who can't answer it or make up their own answer cannot. They will always hear people saying things they are not saying in order to prove their "point" while repeatedly missing what is said to an embarrassingly impenetrable extreme.

People who think it's informative or useful to call evolution "random" without caveat or further explanation need to converse with other "random" people living "random" lives to confirm their "random" theories of how best to convey understanding of well understood processes that they don't seem to understand as well as they imagine.

[GreyICE]

It can quiet often, for practical reasons, be only semantics. It can also be of immense theoretical value to define the deterministic underpinnings of a statistical model. The existence of the atom was established by a comparison of the non-deterministic classical thermodynamics and the deterministic statistical mechanics via Brownian motion.
http://lorentz.phl.jhu.edu/AnnusMirabilis/AeReserveArticles/eins_brownian.pdf No, it's of no theoretical value. Brownian motion is an observed phenomena. Explaining it is obviously of value. You are taking an observed phenomena and stating, as an article of faith that it is both deterministic and impossible for us to predict.

That's what you said. Not difficult. Not hard. Not beyond our current technology. Impossible. You are making a statement, a factual statement, about impossible things.

That's why I'm calling it semantic masturbation. You're not doing anything useful to explain anything. Stochastic is a fine word. Chaotic works perfectly well. Deterministic but impossible to actually determine is a bit wordy, but sure.

Stating that I'm right, but you don't like the language I'm using while I'm right is just semantics.

[my_wan]

No, it's of no theoretical value. Brownian motion is an observed phenomena. Explaining it is obviously of value. You are taking an observed phenomena and stating, as an article of faith that it is both deterministic and impossible for us to predict.

That's what you said. Not difficult. Not hard. Not beyond our current technology. Impossible. You are making a statement, a factual statement, about impossible things.

No, you are distorting what I said. I said we can't know it is true for every possible mechanism in the Universe. Not impossible, nor potentially without empirical meaning, for any given model. Though in principle it could be true for some particular model. When a deterministic model adds to the empirical content, which it can in some cases, we predict things like Brownian motion. It is therefore not an article of faith, it is an important theoretical consideration, sometimes with important empirical consequences.

That's why I'm calling it semantic masturbation. You're not doing anything useful to explain anything. Stochastic is a fine word. Chaotic works perfectly well. Deterministic but impossible to actually determine is a bit wordy, but sure.

Determinism was the guiding principle in the development of statistical mechanics, and this defines its differences wrt classical thermodynamics that didn't take microscopic mechanics into account. Without it there would be no quantitative explanatory model for Brownian motion. We could qualitatively claim it was microscopic collisions, but the disparate scales would make that more than a little suspect. Again, I did not say it was impossible to determine for any given case, I said it's impossible to determine for every possible case. In other words it's impossible to say everything about the entire Universe is deterministic

Stating that I'm right, but you don't like the language I'm using while I'm right is just semantics.

How right you are wrt to any given model is limited to the accuracy of that model. By rejecting the formal limitations of the model, and assuming randomness means non-deterministic, you are falsely rejecting that these formal limitations exist. What are these formal limitations? They are formalized in the notion of an ensemble.
...is an idealization consisting of a large number of mental copies (sometimes infinitely many) of a system, considered all at once, each of which represents a possible state that the real system might be in. This article treats the notion of ensembles in a mathematically rigorous fashion, although relevant physical aspects will be mentioned.
So what are these "physical aspects" to be mentioned?
Physical considerations
The ensemble formalises the notion that a physicist repeating an experiment again and again under the same macroscopic conditions, but unable to control the microscopic details, may expect to observe a range of different outcomes.

So an ensemble is a formalized method of leaving microscopic states undefined. It must remain undefined in order to maintain generality. Yet in any particular model these microscopic states presumably exist, and define the way the statistics are expressed. Undefined does not mean non-existent causal mechanisms. Yet we cannot prove that in every possible case there exist microscopic states that causally defines the statistics. We only know that in all the cases we have proved there exist causal mechanisms that define these states.

An equally serious problem is that, by rejecting these formal limitations, people in these debates are constantly claiming people made claims they didn't make, calling each other liars, et., etc. Due simply to some people are talking about mechanisms that the formal ensemble leaves undefined, as it relates to a given model, while others are rejecting the notion that anything is left undefined. Defining the unique causal mechanisms of a particular model can have empirical consequences that extend the accuracy of the model, as it did with Brownian motion.

So how right you are is contingent upon the accuracy of the model and the physical limitations of measurement not improving to a point of exposing possible inaccuracies. Those who accept these limitions as undefined, rather than non-existent, will be the ones that improve the models. Even if in some case randomness really does have non-existent causal mechansims, who want to limit themselves from discovering and improving these models simply because of an exclusive definition we chose.

[GreyICE]

No, you are distorting what I said. I said we can't know it is true for every possible mechanism in the Universe. Not impossible, nor potentially without empirical meaning, for any given model. Though in principle it could be true for some particular model. When a deterministic model adds to the empirical content, which it can in some cases, we predict things like Brownian motion. It is therefore not an article of faith, it is an important theoretical consideration, sometimes with important empirical consequences.

We're not talking about the universe, or Brownian motion. We're talking about Evolution. Which does not act deterministically.

The fact that we observe that under certain conditions (like, presence of life, generations, timeframes, pressures, etc.) evolution will occur, and can make generalizations about its nature does not make it deterministic. Pretending evolution is something as pretty and simple as "Insert tab A into slot B, wait 52,768 generations, observe mutation X" is silly.

Determinism was the guiding principle in the development of statistical mechanics, and this defines its differences wrt classical thermodynamics that didn't take microscopic mechanics into account. Without it there would be no quantitative explanatory model for Brownian motion. We could qualitatively claim it was microscopic collisions, but the disparate scales would make that more than a little suspect. Again, I did not say it was impossible to determine for any given case, I said it's impossible to determine for every possible case. In other words it's impossible to say everything about the entire Universe is deterministic No, no. You said it was impossible to say everything about our petri dishes could even be termed identical.

I've heard of relativism, but reducing the entire universe to a couple petri dishes seems a bit much.



How right you are wrt to any given model is limited to the accuracy of that model. By rejecting the formal limitations of the model, and assuming randomness means non-deterministic, you are falsely rejecting that these formal limitations exist. What are these formal limitations? They are formalized in the notion of an ensemble.
Nonsense. For non-chaotic systems, the accuracy of the inputs determines the accuracy of your knowledge of the outputs.

Chaotic systems are not so nice. We can generalize them as stochastic, deterministic but impossible to determine, what have you.

The final point remains. You seem to have latched onto a semantic point, and are insisting that evolution is deterministic but impossible to determine.

If you have any objections with the term "deterministic but impossible to determine" feel free to voice them. Otherwise I will be sticking with the more descriptive and less verbose stochastic.

P.S. Random is right out.

[my_wan]

We're not talking about the universe, or Brownian motion. We're talking about Evolution. Which does not act deterministically.

http://biocomplexity.indiana.edu/jglazier/courses/p548/files/p548_lecture4v2.ppt
Monte Carlo Methods Use Statistical Physics Techniques to Solve Problems that are Difficult or Inconvenient to Solve Deterministically.
...
Both Applications Important in Biology.

It appears that you ignored the undefined for a reason part of the formalization of ensembles in my last post.

The fact that we observe that under certain conditions (like, presence of life, generations, timeframes, pressures, etc.) evolution will occur, and can make generalizations about its nature does not make it deterministic. Pretending evolution is something as pretty and simple as "Insert tab A into slot B, wait 52,768 generations, observe mutation X" is silly.

Well duh... I made no such pretension. However, can you deny that by increasing the level of control of the micro-states, even to a tiny degree, can impart some control on variables that otherwise would have been purely random? We have even developed methods of controlling radioactive decay rates to some degree, but that's not NS is it.

No, no. You said it was impossible to say everything about our petri dishes could even be termed identical.
Oh.. Well duh... You generalized this to say I claimed it was "deterministic but impossible to actually determine"? That's a freaking stretch. Do you really claim that those petri dishes are perfectly identical? Do you really claim that control can't be increased to some small degree that has empirical effects, over and above previous runs?

I've heard of relativism, but reducing the entire universe to a couple petri dishes seems a bit much.
Your control in producing identical dishes is woefully inadequate to call them identical. It can pretty much be guaranteed that whatever control you use, it can be improved for dishes even more alike, yet never identical. Even tiny improvements can have empirical effects.

Nonsense. For non-chaotic systems, the accuracy of the inputs determines the accuracy of your knowledge of the outputs.
Yet even chaotic systems can be deterministic.
http://ptp.ipap.jp/link?PTP/64/1532/
Statistical Mechanics of Deterministic Chaos

Chaotic systems are not so nice. We can generalize them as stochastic, deterministic but impossible to determine, what have you.
Yes, and even small improvements in measurement has allowed people to beat roulette wheels.
http://www.iht.com/articles/1992/05/05/chao.php

The final point remains. You seem to have latched onto a semantic point, and are insisting that evolution is deterministic but impossible to determine.
No. I explained in detail why it is not totally semantics. Claiming this is a semantics argument to keep falsely accusing me and other of claims we did not make. Also explained in detail was this false notion that it is always impossible to determine. Neither am I insisting that evolution or anything else MUST be deterministic.

I suggest you read and respond to those explanations, rather than repeating the same false BS. If you make such naked claims about what I claimed again, I'll call your claim a bald faced lie.

If you have any objections with the term "deterministic but impossible to determine" feel free to voice them. Otherwise I will be sticking with the more descriptive and less verbose stochastic.

Duhhh.. Let's see, what did I say?
...Defining the unique causal mechanisms of a particular model can have empirical consequences that extend the accuracy of the model, as it did with Brownian motion.

So how right you are is contingent upon the accuracy of the model and the physical limitations of measurement not improving to a point of exposing possible inaccuracies. Those who accept these limitations as undefined, rather than non-existent, will be the ones that improve the models. Even if in some case randomness really does have non-existent causal mechanisms, who want to limit themselves from discovering and improving these models simply because of an exclusive definition we chose.
Where in there was "impossible to determine? And this was my particular post you responded to with that crap. Nor did I ever claim that everything is fully deterministic. I don't want to hear this claim again.

P.S. Random is right out.
Given the biocomplexity.indiana.edu (http://biocomplexity.indiana.edu/jglazier/courses/p548/files/p548_lecture4v2.ppt) reference above the Boltzmann Equation is in fact relevant, as this equation defines statistical mechanics, irrespective of field of application. This reference should be of interest then:
Deterministic Methods for the Boltzmann Equation

Are these people just masturbating themselves?

Edit: Clickable link for that last quote
http://mathema.tician.de/news.tiker.net/download/academic/talks/boltzmann-slides.pdf

[cyborg]

So the course of evolution is sometimes altered by random mutations, which are inherent to the system.

That's an entirely different statement.

The fitness landscape is subject to random change. Which I now think you agree with.

That is also an entirely different statement.

So are you saying that for a given stable ecosystem, that evolution is nonrandom?

No.

What I'm saying is pretty simple. I really don't understand how you can infer such tangential meanings from it.

[jimbob]

What are you saying?

Well I read it as this:


So we have accepted that mutations can cause significant differences in outcomes.
Uh, I've never said anything different.

So the course of evolution is sometimes altered by random mutations, which are inherent to the system.


So the course of evolution is sometimes altered by random mutations, which are inherent to the system.

That's an entirely different statement.


What are you saying?

So we have accepted that mutations can cause significant differences in outcomes. But the course of evolution isn't affected by this despite the fact that it causes different outcomes?



The fitness landscape is subject to random change. Which I now think you agree with.

That is also an entirely different statement.

OK, so you don't agree with that? I would say that the evolution of citrate+ metabolism changed the fitness landscape for other organisms within the ecosystem, why do you disagree?

So are you saying that for a given stable ecosystem, that evolution is nonrandom?

No.

What I'm saying is pretty simple. I really don't understand how you can infer such tangential meanings from it.

What you are saying *might* be pretty simple, however you prefer to allude to it, rather than state it outright.

Hey, I'd say that for a given stabe ecosystem, over moderate timescales, with sufficiently large populations, then randomness is often unimportant in evolution, even if it is still implicitly present.

[cyborg]

What are you saying?

So we have accepted that mutations can cause significant differences in outcomes. But the course of evolution isn't affected by this despite the fact that it causes different outcomes?

Again: I have not "accepted" this because I've never said anything to the contrary. I would like to know why you think otherwise. It is almost certainly because you've never understood what I've been saying and have been confusing the application of certain concepts to levels of abstraction I was not applying them to.

OK, so you don't agree with that? I would say that the evolution of citrate+ metabolism changed the fitness landscape for other organisms within the ecosystem, why do you disagree?

Since I don't disagree why do you think I do?

What you are saying *might* be pretty simple, however you prefer to allude to it, rather than state it outright.

I've stated it outright hundreds of times.

Hey, I'd say that for a given stabe ecosystem, over moderate timescales, with sufficiently large populations, then randomness is often unimportant in evolution, even if it is still implicitly present.

You do not understand what I've been saying. The overall behaviour of the system doesn't imply a particular behaviour of its constituents. I've proven this before. Either you don't want to understand or can't.

[jimbob]

For this debate, or peoples characterizations of others claims, to make any sense whatsoever depends on what definitions are used for what.

cyborg, jimbob, etc., answer these questions.

(1) Is random and deterministic antonyms?
(2) Does randomness rule out determinism?
(3) Is your answers here model specific or do they also apply to the actual system being modeled?

Differences in how you answer these questions determines what others think you said that you didn't. This leads to accusations of false claims, lying, etc., etc. So answer them or waste your time.

My_wan:

(1) Is random and deterministic antonyms?
Yes: In a deterministic system, identical inputs produce identical outputs. *similar* inputs can produce grossly dissimilar outputs, and this can still be determistic.

In a random system, identical inputs needn't produce identical outputs. For example, it really does seem as if radioactive decay is random. Furthermore, some of the double-slit (http://en.wikipedia.org/wiki/Double-slit_experiment#Quantum_version_of_experiment) experiments, it seems as if the universe itself hasn't decided on the outcome of a quantum event before it happens. In other words, these events are truely random, and not just unpredictible.


(2) Does randomness rule out determinism?
It depends on the system, but in evolution, I'd say that the course of evolution can be afffected by (say) a "disruptive mutation" that confers such selective advantage, that it alters the entire fitness landscape. This exists potentially within any evolutionary system, even though it wouldm't occur often: To use Lenski's own words:

In any case, our study shows that historical contingency can have a profound and lasting impact under the simplest, and thus most stringent, conditions in which initially identical populations evolve in identical environments. Even from so simple a beginning, small happenstances of history may lead populations along different evolutionary paths. A potentiated cell took the one less traveled by, and that has made all the difference.

" small happenstances of history may lead populations along different evolutionary paths. A potentiated cell took the one less traveled by"

In this case, the "small happenstances of history" were simply differences in mutations, as the rest had been controlled for in the experimantal design.


(3) Is your answers here model specific or do they also apply to the actual system being modeled?
I am talking about evolution as it has been observed. I would argue that it is probably inherent in any highly nonlinear system if miroscopic random events can get magnified (in a determinsitic fashion).

I would argue that if you have self-replicating systems with errors in copying, then Darwinian evolution will follow, and that if these errors in copying are random, then there is the potential for a disruptive "mutation", so any system with self-replication with random errors will have the potential for random (grossly different) outcomes.

[cyborg]

(1) Is random and deterministic antonyms?

Yes.

(2) Does randomness rule out determinism?

No.

(3) Is your answers here model specific or do they also apply to the actual system being modeled?

Indeterminable.

[GreyICE]

My_wang, can you possibly add something useful? You seem to be attacking me along the concept that evolution MUST be deterministic (I can't really tell, you've butchered my arguments long past the point of readability). But... why? Given what can more or less be described as identical inputs, we have predictable outputs, but the predictability and is marred by what can be described as a chaotic, or a random factor.

This is perfectly described by the term stochastic, which is a process that is model-able, but contains a random element that has to be accounted for.

You have posted lots and lots of links, lots and lots of wasted verbage, lots and lots of nonsense about how improvements to the modeling of the 'chaotic' factor improves our ability to predict the outcome of the system.

Why yes, yes it does. That does not make the behavior deterministic. This does not allow a set of error bars on the inputs to determine the set of error bars on the outputs.

Also, you are unfairly setting up random and deterministic as the ONLY possibilities. Systems can be neither. Your chaotic modeled as deterministic simply seems to be saying that they can model the outcome of some chaotic systems. Your second article was about a group of bloody investment bankers. Guess what, 10:1 they're out of business right now. How'd predicting the chaotic work out for them?

My final question to you: Is there any objection to using the term stochastic to describe evolution that CANNOT be summarized using the word "semantic?"

[my_wan]

jimbob,
That's a fair position. Here are the reasons you mistakenly misjudge cyborg's statements. It is not provable one way or the other if, in some cases, whether randomness is an antonym of determinism or not. We know certain processes like, ideal gas laws, can be formulated as perfectly random systems. So in that case randomness and determinism is not antonyms. Yet we don't know that this can be extended to all systems. Like the ones you mentioned. Randomness in those cases may not in fact be an antonym of determinism. Yet we must know the things that irreducible randomness says doesn't exist to prove that it doesn't exist, in those cases we just don't know.

This creates a "defensible" position that all randomness is the result underlying deterministic causes. Yet it's not any more provable, in all cases, than the contrary. So when cyborg says randomness says nothing about the causality of the mutations themselves, his position makes sense. It neither accepts nor denies that randomness and determinism are antonyms. By assuming they are definitely antonyms it gives meaning to cyborg's statements that cyborg was not intending to imply.

I don't really care what definition you choose, so long as people learn how to parse the contrary definitions and quit falsely claiming someone else claimed something they didn't. This of course goes to both sides of the debate.

[my_wan]

GreyICE,
Yes I was a bit harsh. The point is that the epistemic differences in how people choose to define randomness. These differences hinge on epistemic notions of how it relates to determinism. It is important in a debate to at least know the opponents position even if you disagree. Here are the things that can, honestly and without taking sides, be said about randomness.

(1) Some perfectly determistic models can also be modeled as purely random.
(2) When a deterministic model is not known, we can't prove one doesn't exist.
(3) Determistic causes sometimes has empirical differences wrt certain variables.
(4) Determistic causes sometimes has no empirical differences wrt certain variables.
(5) Determistic causes sometimes "might" not exist at all wrt certain variables.

That all we can say given the fundamental limitations of both math and science.

By locking onto a singular definition, and defending it as the only definition, it forces false characterizations of the statements of others useing a contrary definition. Either of the definition may or may not be true with respect to a given phenomena. It's the old correlation does not equal causation, only the correlation the lack of correlation this time. Neither does this rule say correlation does not equal causation in any singular case. I drop a glass on the floor and it breaks, then correlation does equal causation in that case.

The ensuing debate is nothing more than semantics until this is understood. Once this is understood debate is possible.

[GreyICE]

GreyICE,
Yes I was a bit harsh. The point is that the epistemic differences in how people choose to define randomness. These differences hinge on epistemic notions of how it relates to determinism. It is important in a debate to at least know the opponents position even if you disagree. Here are the things that can, honestly and without taking sides, be said about randomness.

(1) Some perfectly determistic models can also be modeled as purely random.
(2) When a deterministic model is not known, we can't prove one doesn't exist.
(3) Determistic causes sometimes has empirical differences wrt certain variables.
(4) Determistic causes sometimes has no empirical differences wrt certain variables.
(5) Determistic causes sometimes "might" not exist at all wrt certain variables.

That all we can say given the fundamental limitations of both math and science.

By locking onto a singular definition, and defending it as the only definition, it forces false characterizations of the statements of others useing a contrary definition. Either of the definition may or may not be true with respect to a given phenomena. It's the old correlation does not equal causation, only the correlation the lack of correlation this time. Neither does this rule say correlation does not equal causation in any singular case. I drop a glass on the floor and it breaks, then correlation does equal causation in that case.

The ensuing debate is nothing more than semantics until this is understood. Once this is understood debate is possible.

1) deterministic systems are ones in which no randomness is part of the future outcome of the system. Period. Models are models, they frequently contain simplifications.
2) True, and irrelevant. If one cannot perfectly predict the outcome and one cannot replicate the outcome reliably, it is fair and fine to call the system non-deterministic. If a method later crops up that allows you to call the system deterministic, that's well and good, but stating that such a method exists on zero evidence is not good reasoning.
3) Obviously. What they do not have is a random element. One does not pretend that all velocities result in equal distance traveled for equal time, but one does not pretend that distance as a function of velocity contains a random element. You are stating the obvious.
4) True. Also irrelevant.
5) Okay, so not all systems are deterministic?

Once again, there is no evidence that evolution is a deterministic system, and plenty of evidence that there are uncontrollable elements that resemble random. Stochastic is a fine term, and one I rather like. Random is a stupid term, totally wrong, and used by idiots. Chaotic works in my book, but I like chaotic systems overly much, they appeal to me.

[recursive prophet]

Interesting debates the last couple days, and I’d like to thank all the participants for contributing to my education. One thing I still wonder about is if there is anyone here that would argue-wrt the OP-that natural selection isn’t stochastic.

I must confess this extended discussion is getting quite difficult for my aging mind to comprehend. As to my_wan’s 3 questions, I would answer yes, no, and undetermined. My answers might also apply to real systems, but Cyborg may be correct in saying they are indeterminable, and I find Jimbobs answer quite compelling.

I have posted many of the comments from this thread on a similar one at RDF, and then their replies here. However, the volume of recent activity on RDF would sort of swamp the discussion here, so this time I am going to just post links on both sites. I would strongly encourage everyone to read what has been written there, especially the most recent exchanges between my_wan and susu.exp. The latter is working toward his doctorate in paleontology at Bonn University, and it appears has some views that are congruent with those of GreyICE. He also responded to Sol’s last post here, and I would really like to read both of their reactions to what he and others there have to say.

To read all my_wan wrote when he recently reentered the debate on RDF, click here (http://richarddawkins.net/forum/viewtopic.php?p=1733396#p1733396), or to jump ahead to where susu comes in with his rebuttals, use this link (http://richarddawkins.net/forum/viewtopic.php?p=1741666#p1741666). I look forward to reading more of this hybrid-enhanced dialectic, but will withhold any commentary for now as after digesting my_wan’s latest posts on RDF I’m thoroughly confused. I'm well accustomed to being the dumbest guy in the forums I choose to read, but this one is making me consider early onset dementia. :confused:

[articulett]

Evolution is deterministic... all science is on a macro scale-- we know that what unfolds is caused by what came before. When we investigate an air disaster it's because we understand that figuring out the cause, helps us prevent such things in the future. We know that the traits we see evolved because ancestors that had such traits preferentially survived and reproduced over their peers that didn't.

Understanding evolution involved understanding the way what we see unfolded-- In this sense, it is deterministic-- what happens has a cause which has a cause which has a cause...

The reason my_wan and others are forced to delineate this is because --to Mijo, RP, and Jim-Bob, stochastic = random (because stochastic processes are sometimes called "random processes"). And anything that has any randomness in it, is wholly random to them. They are not using the same definitions as others. You don't need to use the words random, deterministic, nor stochastic to describe evolution-- for evolution to occur you just need variety and an environment that selects some information to be copied over others-- that's it. My_wan has gone out of his way to be more than clear on this topic. The problem is that people have varying definitions of random such that every process can be said to have some randomness according to some definitions of random. This makes all processes "random" and nothing deterministic. It renders the terms useless for generating understanding. That is why people who cannot answer Sol's simple query the way others would should cannot participate in the same conversation. Words mean different things to them.

My_wan's not the one playing around with semantics on this topic. He's clarifying misstatements that have been attributed to him. My_wan is explaining why it's misleading and uninformative to call evolution random--he's not saying it IS deterministic-- though, at least on a macro level--everything is deterministic in science-- that is, everything is determined by what came before-- we understand the past by putting the pieces of the puzzle together, because we know that things did not happen "magically" or without physical causes. Evolution is not a random process-- it is directed by what information gets copied the best in whatever environment it finds itself in. The genes that get passed on the best DETERMINE the direction of what evolves. It's no more random than the winner of a chess tournament or an election. Calling it random is like saying the winner of a chess tournament is chosen at random. It's misleading, vague, and daft.

Allele selection is stochastic to some extent, but not always --as the paper in Current Biology illustrates http://www.bio-medicine.org/biology-news-1/Evolution-is-deterministic--not-random--biologists-conclude-from-multi-species-study-1268-1/

It really depends on how you are defining the terms and in respect to what. But evolution is mainly driven by genes and genomes that confer a survival/reproductive advantage to their owners. Stochastic really doesn't capture that understanding. It's useful for understanding drift, but drift appears to play a much smaller role than once imagined. Eyes did not evolve by drift, for example. Nor did Lenski's bacteria. Each step in the mutation provided a slight reproductive advantage to the population that had it... My_wan has a world of science supporting his statements. There are no peer reviewed scientists saying "evolution is random (full stop)" or even "evolution is stochastic" (period)-- only that some aspects can be modeled scholastically. But so can election outcomes and tons of stuff. That doesn't mean it is informative to say "elections are stochastic". The same with evolution. It's even more meaningless to say evolution is stochastic than to say elections are stochastic. It transmits no information as to what evolution actually entails, and it's misleading. It makes people feel like they are saying something intelligent--but to the educated ear, it sounds like they think they understand evolution much better than they actually do!


It's preferential survival and reproduction that is the method by which genomes evolve and thus the organisms they code for. But each organism dies the same species it is born as. It's genome has a chance to evolve based on how successful the organism is in passing on it's genes. I don't think "stochastic" captures this understanding. That is why few of the experts in the field use it to explain evolution. And none call evolution random. Perhaps few use stochastic because it makes some not-so-bright people think that they are calling evolution "random"--it obfuscates rather than clarifies. It's not "wrong", but it's about as informative as calling evolution a "noun". It's not that stochastic can't be used to explain evolution. It's just that some people are defining words so loosely and bizarrely that saying such is completely uninformative regarding the most essential aspects of understanding evolution. If you want to describe evolution to someone, I suggest you use the words of experts like Ken Miller, Richard Dawkins, Charles Darwin, etc. They are responsible to conveying understanding to millions. I don't think anyone calling evolution "random" or "stochastic" can claim such. In fact, they are describing evolution similarly to those who are well known for obfuscating understanding in order to support "intelligent design". To those who understand the topic, they sound like they haven't got a clue as the quotes from Miller, Dawkins, Ayala, Coyne, et. al. reiterate.

Many scientists feel that nothing is truly random-- but they aren't defining random as "anything to do with probabilities" the way some vague folks are. They have a much more specific definition in mind as My_wan has repeatedly tried to clarify to no avail. Mutations aren't random in that they have causes and some areas of a genome are far more prone to mutation (hot spots) than others-- and some types of mutations are far more likely than others-- but genomes have been "selected" to be this way. They are just "random"-- or rather unbiased in that they happen without regard to whether they'll benefit an organism or not. And then the environment selects. But we know why or how many changes happen even if we can't predict where or when. We understand things like non-disjunction. They have physical causes--in this way, they are NOT random. We can predict that older mothers will conceive more trisomies for example-- this isn't "random"-- but we can't predict much more specifically than that. But because we live in a deterministic world, we can discover and predict better and better and understand more and more.

[my_wan]

1) deterministic systems are ones in which no randomness is part of the future outcome of the system. Period. Models are models, they frequently contain simplifications.
Here you took a definite position and punctuated with an unequivocation. Then the second sentence you equivocated, except you said nothing about how this equivocation might effect the first sentence.
3) Obviously. What they do not have is a random element. One does not pretend that all velocities result in equal distance traveled for equal time, but one does not pretend that distance as a function of velocity contains a random element. You are stating the obvious.
Apparently I didn't make this clear enough. All 5 of those statements were specific to systems that could be modeled as random.
5) Okay, so not all systems are deterministic?
This one just makes me cringe. The point is You don't know, I don't know, nobody knows.

Consider a machine with a red and blue light, and a button. You press the button and either the red or blue light comes on. Press again and it goes off. Let's say you kept notes for which light came on every time for thousands of hours. You get a perfectly random 50/50 normal distribution of red and blue lights. Does this mean the machine randomly selects which light to turn on? Does that mean it is unpredictable and there is no way to get it to give you the same color light every time, without changing anything about what or how the machine works? No, it doesn't! With enough precision and control you can get it to turn the same light on every time. It simply switches the between red and blue every 1/1000th of a second. So once you know this and have accurate enough a clock and button presser the color you get is predetermined even 10 years from now.

So in you response to 1), how do you know when the randomness indicates the future is indeterminable and when you model simply lacks the details to predict it? The way you stated it makes this even worse. You said: "determinism means 'no' randomness is part of the future outcome". Yet the red and blue light was perfectly random yet perfectly deterministic. Lots of different deterministic phenomena exhibit this same type of randomness, ideal gas laws for instance. Of course a real gas is not so perfect as the ideal model but the randomness is the same. So we can not know if all randomness is the result of underlying deterministic parts. It would require knowledge as perfect as that some presume sky daddy possesses.

The most important thing to learn is if someone claims NS is deterministic, they did not necessarily claim it is not random, just like the red and blue lights. Insisting they did is wrong. Accusing others of claims they did not make is the primary effect. It's a waste of time, and a diversion, to insist on debating "is NS random" when the argument is actually about the definition of randomness. So what exactly is the differences in the definitions?

Always Deterministic Definition
Definition(1): Randomness and determinism is synonymous. Randomness is a formalized methodology of leaving data, generally microscopic, undefined while maintaining data about the averages and distribution of that microscopic data. This entails, by the epistemology of some, that all randomness is pseudorandom, even when, due to undefined data, it is mathematically perfectly random.

Sometimes Deterministic, Sometimes Random Definition
Definition(2): Randomness and determinism are antonyms. Yet deterministic systems can exhibit pseudorandomness due to undefined data. Depending on ones epistemology it may or may not entail that mathematically perfectly randomness can also be pseudorandom.

Always Random Definition
Definition(3): Randomness and determinism are antonyms. Anything that exhibits randomness is in fact random, not pseudorandom. Anything that exhibits pseudorandomness has a defined deterministic causal mechanism. However complex and computationally hard that mechanism is, pseudorandomness allows this mechanism to be recovered from deviations from perfect randomness. In effect the random and nonrandom elements can be separated.

Problems with Definition(1): This is philosophically defensible, yet impossible to prove. It requires perfect knowledge of the entire Universe to prove.

Problems with Definition(2): If the epistemology states that mathematically perfect randomness can be pseudorandom then this definition entails that we can't, even in principle, alway know the difference between randomness and pseudorandomness, though sometimes we can. So functionally it similar to Definition(1). If the epistemology is the converse, then our capacity to know the difference between randomness and pseudorandomness is contingent upon the sample size (largeness of N). So we can only place increasing odds on whether it is truly random, but when it is found to be pseudorandom it is verifiably so.

Problems with Definition(3): This definition entails that all systems are definitely either random or pseudorandom, and randomness entails purely random. I am less sure of the epistemological variances of the advocates. This is made more difficult when an unequivical statement like: "deterministic systems are ones in which no randomness is part of the future outcome of the system. Period.", is followed by 'potential' equivocations like: "Models are models, they frequently contain simplifications".

What can we empirically say? We can construct a perfectly idealized deterministic models. It doesn't matter whether it represents a real model of physical laws or not, it's construction entails that it is real in some sense. The idealization simply entails that we can't generalize it to all models. Yet, when we intentional leave certain parameters of this perfectly deterministic system undefined, it can exhibit pure randomness. Ideal Gas Laws are an example of this, and Statistical Mechanics is the result when we leave microscopic detail undefined. At the microscopic level it's just a 3-D billiards game.

So we can empirically say determinism can exhibit perfect randomness, if the information is limited. Yet because these models are only idealizations, we know both the epistemology and Definition(3) wrt very limited singular cases is wrong, but may in fact be correct for some systems. To prove it requires the very information that Definition(3) says doesn't exist. Empirically we are therefore left somewhere between Definition(1) and (2). Although Definition(3) may apply in a subset of systems and/or variables.

I find it easier to presume Definition(1) for the same reason I presume naturalism. This is not a truth statement though, because I am limited by methodology the same way naturalism is limited to methodological naturalism. If in particular case I am wrong, it can never be proved. In those cases where I am right it can extend the accuracy of the models and/or expand our empirical knowledge, the same way Einstein did with Brownian motion.

[jimbob]

Historical contingency may mean "random" to you-- it doesn't mean it to Lenski or anyone else. Historical contingency IS the selection process--and nobody of note would describe that as random.

Articulett, in the Long Term Evolution Experiment, the experimental design controlled for the other factors, so that random mutations were shown to be the important difference between the populations that evolved the trait, and those that didn't.

Only in one flask did a "neutral" potentiating mutation arise that later on facilitated the evolution of citrate+ metabolism.

In this experiment, the important "historical contingencies" were simply the differences in the mutations in the different populations. These were random. This means that the occurance of a random mutation allowed later evolution of a trait.

It happened in a very simple ecosystem, and would be more common in more complex ecosystems with more compex interactions.


I agree with the New Scientist article about this: (http://www.newscientist.com/article/dn14094-bacteria-make-major-evolutionary-shift-in-the-lab.html)

In the meantime, the experiment stands as proof that evolution does not always lead to the best possible outcome. Instead, a chance event can sometimes open evolutionary doors for one population that remain forever closed to other populations with different histories.


This is practically what I had written here (http://forums.randi.org/showpost.php?p=3365035&postcount=258) (note the dates..)


Selection is probabilistic, but this still leads to optimisation to environments (although the nature of the optiimisations will depend on mutations, and might be surprising, or indeed diferent if it were posible to rreun the situation, to identical conditions, as the quantum events would still be different).

That optimisation to environmental stimuli will occur is inevitible, what optimisation isn't.

Over geological timescales, I would argue that the environment itself is random. Some orbits are chaotic, should that include a potential "dionosaur killer" asteroid, then quantum events could influence asteroid strikes on earth. As organisms fill empty niches, they prevent other organisms filling slightly differnet niches; whilst lions exist, no leopard descendant will evolve into something that competes completely with lions.

[articulett]

Only to you Jim-bob. I prefer the words of actual experts to your verbiage. I'll take Miller, Dawkins, et. al. over your self appointed expertise any day.

To me, you sound like you think you understand evolution and Lenski much better than you actually do. To you, Lenski's paper is proving that evolution is "random". I don't think any actual expert would think that Lenski's paper proved that evolution is "random" any more than it "proves" evolution is a noun. You are just so bizarrely hyperfocused on the randomness that you miss the essence of what others are saying again and again and again. Like my sig article, you have no ability to recognize your own general competence in the subject and so you miss the opportunities to ameliorate your ignorance. You think you know more than those who might clue you in.

And you haven't progressed an iota in 2 years when it comes to ability to understand or engage in conversation on this topic in my opinion. Cyborg spoon feeds you the information, and you just cannot digest it. I think you need to stick with conversing with those who find you as much of an "expert" as you imagine yourself. I don't think you are any more capable of explaining evolution to anyone than Behe is, though I'm sure your confidence in your ability to do so is on par with his.

You just seem to be in your own world on this topic-- or rather in the same world as Behe, T'ai Chi, Mijo, RP, etc. I think that makes you suitable for conversation with each other, but not particularly good at conveying information to others or learning from those who understand and communicate better than you. Your goal is to prove that your way of explaining things is useful or correct or right or informative... but it's just not. You sound identical to folks like Dembski. Who can converse with someone like that on this topic? And why would any serious scientist or person who actually understood evolution do so? Why can't you perceive how much you sound like them and how little you sound like those who actually teach others and write extensively on this subject? How can you imagine you are more right or clearer than Ken Miller, for example??

[my_wan]

jimbob,
When you read a research article that uses terms like "chance event", "random", etc., you need to realize that the use of the term is not implying any singular epistemological notions of what those terms mean. They have a purely operational meaning, especially in biological sciences. Operationally the difference between what's included in the model and what the actual state of affairs are in the dish is nil, unless you can find a way to include perfect, or at least more concise, information in the model.

The mistake you make is adding your own epistemology to these terms, pretending that these researchers are using the same epistemology in the use of those terms, then accuse others of statements they didn't make because they didn't share the same epistemology. The term "chance event" was epistemologically neutral as used by that article. Operationally it simply means: an event in which it is "unknown" what the cause was, or if it even had a cause. It simply means "we don't know".

So quit pretending you know, and can tell others they are wrong, and that they said things they didn't say.

[jimbob]

Articulett,

One of Lenski's stated aims was to see whether rerunning the tape of evolution would always produce the same results every time. It won't.

I guess you disagree with the following statement:

In an evolutionary system, where the only differences are due to random events, significantly differenct outcomes can be observed. Indeed this has been observed in the Long Term Evolution Experiment, where significant differences in outcomes have been observed, and where minor differences in the initial conditions have been shown to be unimportant to this outcome.

If you do disagree with this, could you explain why only one population evolved citrate+ metabolism? What were the important differences in this population that were not due to differences in mutations?

ETA:

Do you disagree with this:

However, selection requires heritable variation generated by random mutation, and even beneficial mutations may be lost by random drift. Moreover, random and deterministic
processes become intertwined over time such that future alternatives may be contingent on the prior history of an evolving population. For example, multiple beneficial mutations will arise
in some unpredictable order, and those that are substituted first may differ from others in their effects, thus constraining some evolutionary paths while potentiating other outcomes.

[articulett]

Jim-Bob, I don't think you are capable of having a discussion on this topic because you are bent on proving to yourself that you are saying something "intelligent". Sure, it is "valid" to say evolution is a "noun"-- but this validity doesn't make focusing on the fact "intelligent" or useful. The same with your focus on "random". I consider you as off topic and bizarre as someone obsessed with proving evolution is a "noun". Really. And I don't think it's fixable. You and Mijo are the same as you were 2 years ago--the same as Behe was 2 years ago. I'm a skeptic-- I don't have faith that anything will ever change in this area.

So I am going to put you on ignore once more so that you might engage those who are using terminology the way you are and feel that it's useful or informative to describe evolution in the way you do. Remember, to me, I'm about as interested in your obsession as I would be with someone who wants to engage me in a discussion to "prove" that evolution is a noun. To you, evolution is "random". No expert would describe it as such... your twisting their words to hear them saying is much makes you look as dishonest as Behe from my perspective. If you don't want to be treated like an "intelligent design cproponentist", then you really ought to distinguish your conversation techniques from theirs.

I mean I could use your "debate techniques" to prove that Lenski's paper is proof that evolution is a "noun" and it would be on par with what you are doing. It wouldn't be "wrong"-- just obfuscating, daft, and silly. Would you engage in a discussion with a person bent on driving home the point that Lenski's paper proves that evolution is a noun?

I wouldn't either. What would be the point? Such people need to be directed towards other people who find such semantic philosphical "debates" stimulating so they don't muck up understanding of others.

[jimbob]

jimbob,
That's a fair position. Here are the reasons you mistakenly misjudge cyborg's statements. It is not provable one way or the other if, in some cases, whether randomness is an antonym of determinism or not. We know certain processes like, ideal gas laws, can be formulated as perfectly random systems. So in that case randomness and determinism is not antonyms. Yet we don't know that this can be extended to all systems. Like the ones you mentioned. Randomness in those cases may not in fact be an antonym of determinism. Yet we must know the things that irreducible randomness says doesn't exist to prove that it doesn't exist, in those cases we just don't know.

This creates a "defensible" position that all randomness is the result underlying deterministic causes. Yet it's not any more provable, in all cases, than the contrary. So when cyborg says randomness says nothing about the causality of the mutations themselves, his position makes sense. It neither accepts nor denies that randomness and determinism are antonyms. By assuming they are definitely antonyms it gives meaning to cyborg's statements that cyborg was not intending to imply.

I don't really care what definition you choose, so long as people learn how to parse the contrary definitions and quit falsely claiming someone else claimed something they didn't. This of course goes to both sides of the debate.

So when cyborg says randomness says nothing about the causality of the mutations themselves, his position makes sense. It neither accepts nor denies that randomness and determinism are antonyms.

Fine, except for:


(1) Is random and deterministic antonyms?

Yes.





jimbob,
When you read a research article that uses terms like "chance event", "random", etc., you need to realize that the use of the term is not implying any singular epistemological notions of what those terms mean. They have a purely operational meaning, especially in biological sciences. Operationally the difference between what's included in the model and what the actual state of affairs are in the dish is nil, unless you can find a way to include perfect, or at least more concise, information in the model.

The mistake you make is adding your own epistemology to these terms, pretending that these researchers are using the same epistemology in the use of those terms, then accuse others of statements they didn't make because they didn't share the same epistemology. The term "chance event" was epistemologically neutral as used by that article. Operationally it simply means: an event in which it is "unknown" what the cause was, or if it even had a cause. It simply means "we don't know".

So quit pretending you know, and can tell others they are wrong, and that they said things they didn't say.

I am familiar with statistics where you often treat multiple nonrandom events as random ones because of ease of analysis, I do that a lot in my day job. In fact you can probably argue that the bell curve is a classic distribution that you get in such a situation, where you may not be dealing with true randomness, but simply unpredictibility.

I'd argue that in the case of evolution it is more than "we don't know".

In the Long term Evolution Experiment, the important difference was traced to potentiating mutations in one population. Furthermore, one of the stated aims of the experiment was to see whether rerunning the tape of evolution always gave the same results. The experimental design accounted for extraneous factors except for mutation. This was one of the experimental aims.

Whether you consider mutations to be truly random or not, must depend on whether you consider quantum events to be truly random. I'd also say that although it isn't completely conclusive, most evidence is that quantum events are truly random.

[jimbob]

norty

[cyborg]

Whether you consider mutations to be truly random or not, must depend on whether you consider quantum events to be truly random.

No.

[jimbob]

Why?

[cyborg]

Again:

You do not understand what I've been saying. The overall behaviour of the system doesn't imply a particular behaviour of its constituents. I've proven this before. Either you don't want to understand or can't.

[Dr Adequate]

Agreed DrA, just that it is less wrong than saying "nonrandom"

If you have to chose a one-word answer... But I don't.

[my_wan]

So when cyborg says randomness says nothing about the causality of the mutations themselves, his position makes sense. It neither accepts nor denies that randomness and determinism are antonyms.

Fine, except for:

(1) Is random and deterministic antonyms?
Yes.
Yet he also said:
(2) Does randomness rule out determinism?
No.
I can't speak for him, but my take is that he made a distinction between random and randomness. Random being operationally unknown, and randomness being a numerical property that can apply to either case. Yet you chose the one that was easiest to argue. I philosophically disagree with cyborg over statements like:

P.S. If you clone 15 bacteria in 15 petri dishes with identical nutrients, light conditions, etc - will they all evolve the same features at exactly the same time?

Yes - deterministic
No - Stochastic

Which do you think it is?
Correct - simplistic.
And have argued with GreyICE at length over it. Yet cyborg also said:
No. I've always said that the underlying cause of mutations is irrelevant.
So he's not ultimately denying cause or no cause. Nor does this impose an epistemology on the use of the term "random" in the general use sense. The answer to 1) seems a stricter mathematical sense, but I don't know. I do know he is also guilty of semantically imposing on your epistemology.

I am familiar with statistics where you often treat multiple nonrandom events as random ones because of ease of analysis, I do that a lot in my day job. In fact you can probably argue that the bell curve is a classic distribution that you get in such a situation, where you may not be dealing with true randomness, but simply unpredictibility.

I'd argue that in the case of evolution it is more than "we don't know".
I don't get how "we don't know" can be more than "we don't know". It appears that you are implying that since we tried and still "we don't know", somehow that means the cause is non-existent.

In the Long term Evolution Experiment, the important difference was traced to potentiating mutations in one population. Furthermore, one of the stated aims of the experiment was to see whether rerunning the tape of evolution always gave the same results. The experimental design accounted for extraneous factors except for mutation. This was one of the experimental aims.

So by rerunning this same experiment are you assuming that both started with the same number of free radicals with the same locations and positions wrt the bacteria in all the dishes. I doubt it but it would take far more detail than even that to say a unique outcome of one dish means more than "we don't know".

Whether you consider mutations to be truly random or not, must depend on whether you consider quantum events to be truly random. I'd also say that although it isn't completely conclusive, most evidence is that quantum events are truly random.
Yes these are interesting subjects to me, and these things definitely fall in the category of "we don't know". We can trivially consider QM deterministic simply by defining the wavefunction as a physically real field. This is however cheating, because it doesn't actually provide the causal mechanisms, it just defines it to be so. It also creates a whole buttload of other theoretical issues.

[recursive prophet]

I was hoping the debate between Jimbob and my_wan would continue, though we seem to be stuck both here and on RDF with "we just don't know" so this is actually a philosophical issue.

There are many working assumptions in math and science that rely on an unproven hypothesis or two, and on the RDF sister thread this has been a key issue. Do most biologists approach evolution as primarily stochastic, or deterministic? As I have also requested there, I would like to read my_wan's answer to the 3 questions he posed here earlier, only with the term stochastic replacing random.

For new viewers, there are several links to the longer discussion at Dawkins.net that I posted on the previous page, and you'll find both JB and MW there also.

[Dr Adequate]

Ye have sown the wind ... (http://forums.randi.org/showthread.php?t=135796)

[recursive prophet]

Cute, Doc. And don't think I haven't often considered the possibility we are debating Russel's tea pot here. Perhaps I'll never know.

As you seem to have a philosophical bent, here's a question you might be able to answer for me. I've oft seen it mentioned that random evolution is the creationist argument, and the logic escapes me. If in fact our universe is deterministic, wouldn't this help make the case for a 'determiner?' Conversely, if all is chaotic, then how could one argue an omnipotent design? Intended chaos? Seems much more of a stretch that way then to assume all is preordained. If its all random, for what need have we gods?

Anyhoo, if you deign to enlighten me I really am curious about the reasoning here, and intuitively random wins for me every time. It only yields at a certain point to the realm of the stochastic. All Greek to me. :rolleyes:

[cyborg]

If in fact our universe is deterministic, wouldn't this help make the case for a 'determiner?'

Since one does not tend to ascribe personality to the impersonal machine of determinism no, not really.

Conversely, if all is chaotic, then how could one argue an omnipotent design? Intended chaos?

The choices of God are not for mortals to understand.

(It is not possible to understand random choice).

[Dr Adequate]

As you seem to have a philosophical bent, here's a question you might be able to answer for me. I've oft seen it mentioned that random evolution is the creationist argument, and the logic escapes me. The logic behind saying that this is the creationist argument is that this is, in fact, what creationists argue.

If in fact our universe is deterministic, wouldn't this help make the case for a 'determiner?' No. But since it appears not to be, your question is irrelevant.

Conversely, if all is chaotic, then how could one argue an omnipotent design? If its all random, for what need have we gods? You might be able to base an anti-theistic argument on the proposition that the universe is entirely random; but this would not be a reason for adopting the premise --- which is not, as it happens, true.

"A => B , I want to prove B, therefore A" is not a valid inference.

[recursive prophet]

Neither red nor blue, but green, eh Doc? :D Well, that seems to be the consensus opinion, though we have spent a lot of time arguing about shades. Many pages later nothing seems really resolved, so it appears you were right in your first post here about the futility of the subject.

At least on RDF Dawkins has finally entered the fray directly, and if he responds to susu's latest rebuttal I'll post it here. Otherwise, looks like we're the 'last men standing' here Doc, and the topic is done as we both now agree it's a cul de sac. Such is life. All we can learn leads to what we can not know.

[Walter Wayne]

Ye have sown the wind ... (http://forums.randi.org/showthread.php?t=135796)
And have reaped a single drive-by post.
At least on RDF Dawkins has finally entered the fray directly, ...
I must be having a bad brain day; couldn't find the post by Dawkins in the thread.

Walt

[Dr Adequate]

And have reaped a single drive-by post. :confused:

[Walter Wayne]

Given the quote, I was expecting a whirlwind at the link you provided and didn't find it.

[recursive prophet]

I must be having a bad brain day; couldn't find the post by Dawkins in the thread.

Walt
No Walt, I HAD one when I posted that reply. Sorry, I actually didn’t realize anyone was still reading this thread. Fact is I considered sending my reply to Doc in a PM so as not to bump it. I should have mentioned it wasn’t on the random thread but in the RD forum. Same questions though, and the effects of drift in small populations-which most are at some point-has been one of the big topics. And earlier, lbq and others got into the adaptationist issue. Recursive indeed. Lumpers and splitters, say I, the eternal struggle of natural science.

You must admit it’s interesting that on a thread where Richard Dawkins has made 3 substantial replies there was so little traffic. No one has commented on the astute question susu posed to Dawkins early yesterday, and after a week it accumulated only 3 pages of replies. DWFTTW got 4 in its first 24 hours.

Just decided I shall add a couple questions there. RD makes some pretty definitive statements lately on the non-randomness of selection. As both the determinists and the stochastic apostates quote mine his words, I really think he should take the time to counter all those who in any way have misinterpreted his thinking. I sent him a PM with links to the debate on selection between some very informed users way back in part one of that thread, focusing on disagreements over the parsing of his statements. Hopefully this recent activity will lead to his putting an end of any erroneous application of his position on the nature of NS.

Thread starts here. (http://richarddawkins.net/forum/viewtopic.php?f=14&t=72053)
First Dawkins response. (http://richarddawkins.net/forum/viewtopic.php?p=1745573#p1745573)
susu’s final question. (http://richarddawkins.net/forum/viewtopic.php?p=1764415#p1764415)

[GreyICE]

My wan, this debate has dead ended in pointless. If you consider the simple statement that models frequently contain false simplifications ambiguous, we cannot discuss anything intelligent. Models frequently assume, among other things:

Gravity is constant with distance above earth, and is a flat, unvaried field.
Constant force produces constant acceleration (mass is invariant with acceleration). Constant acceleration produces constant velocity changes in all frames.
Time in different relativistic frames is identical (the observer on the train and the observer at the station agree on the time)
Electromagnetism doesn't exist
Van der Waals forces don't exist
Air obeys the ideal gas laws
Steam obeys the ideal gas laws
Pressure is a constant across a surface at all times unless conditions change
Matter is homogeneous (i.e. has no atoms)
Building structural supports have no compressive load (i.e. the building has no weight, 100% of the load is torque/shear).


I have seen, used, and worked with plenty of models that contain all of these assumptions. They were, in fact, quite brilliant models, and these assumptions are quite necessary for them to operate.

That is not to say any of these assumptions are in any way true.

If you say you can model a deterministic process as random, well, you've stated nothing useful. To pretend its equivocation to say that is just silly.

[Dr Adequate]

Many pages later nothing seems really resolved, so it appears you were right in your first post here about the futility of the subject. Then apparently it has resolved something, namely that I was right.

I knew that all along.

[my_wan]

My wan, this debate has dead ended in pointless. If you consider the simple statement that models frequently contain false simplifications ambiguous, we cannot discuss anything intelligent.
Not false just simplified. Calling it false just because the model contains random elements that may or may not ultimately be deterministic undermines the formal definition of randomness itself, and demands that it mean things that the formal definition stays silent on for a reason.

If you say you can model a deterministic process as random, well, you've stated nothing useful. To pretend its equivocation to say that is just silly. It did turn out to be useful when we figured out that randomness was actually caused by collisions in a liquid. It proved the existence of the atom. I am equivocating because I can't honestly claim that all randomness in nature has causal underpinnings. If some aspect is truly random then stating so means nothing as you say. However, if something does have meaningful causal underpinnings, finding it even in theory, can have empirical consequences such as proving the existence of the atom historically.

If I don't hear any more accusations that so and so said X because they said Y, not X, I am happy without convincing anyone. So keep the definition of your choice, but I will be ruthless if I hear such characterizations of others words again.