Ichneumonwasp
Unregistered
- Joined
- Feb 2, 2006
- Messages
- 6,240
Randomness in Evolution: Valid and Invalid Usage
- Thread starter Wowbagger
- Start date
-
- Tags
- evolution randomness
zosima
Muse
- Joined
- Mar 1, 2008
- Messages
- 536
It seems like our opinions are starting to converge enough that our argument is getting semantic, which IMO is good, but here's my next volley. (-:
I've never read this, but when you mentioned it I looked it up, and found this quote on wikipedia:
"Most of the book's conclusions were deemed controversial at publication and some of Gould's examples were soon shown to be incorrect. However, the ultimate theme of the book is still being debated among evolutionary thinkers today."
And debated right here on JREF (-:
I've read some of his older work(like Hen's Teeth and Horse's Toe's) and I've read his last work The Structure of Evolutionary Theory. Which seemed to support a much more deterministic view of evolution.
I can see how to you think that and I don't consider that entirely unreasonable.
But it is a question of very marginal probabilities. I don't dispute that there could be a situation where the fate of an entire species depended on whether a butterfly flaps its wings, but that is incredibly unlikely, to the point that it isn't a significant consideration as far as the theory of evolution goes, especially because the study of evolution does not work on human time. It works on quite long time scales, by human standards. I think this for a couple reasons.
1. Generally if a population is in a precarious situation, where the fates of one or two individuals determine the fate of the species, they just die, the gene pool is way too narrow in such a case. Really these "chaotic systems determine fitness" arguments depend on imagining that these situations can and do happen.
2. The probability of a characteristic of just about any chaotic system on earth becomes either 1 or 0 on scales of 10-100 years, which is generally shorter than evolutionary timescales. So all different possible combinations get tried on geological time(like your wind example). Chaotic systems are only chaotic on short time scales, for example the earth's climate is moving to a very predictable higher temperatures, they don't get more chaotic with more time, but less chaotic. But I will concede that, like a gas, it becomes increasingly difficult to tell what the behavior of any specific component will be doing as time increases. (by component I mean something like the minimum component of the chaotic system, in the cases we're often discussing, a molecule).
3. The randomness of the behavior of a chaotic system is often overblown. The sort of argument whereby the global weather could be effected by the flapping of a butterflies wings, is an interesting mathematical construction, but it is a highly contingent sort of Rube Goldberg machine, where every little piece of the argument has to be in just the right place, for a significant effect to be noticed. Practically, this is never true. People like to assume that because chaotic systems can behave in ways that fluctuate wildly on the minimum behavior of a single variable that natural systems that can exhibit chaotic behavior, always will. For example, if we want our metaphorical butterfly to change the weather, all the other butterflies have to be flapping their wings just enough that our little butterfly will push the system over the tipping point. In 99.9999999% of the cases, the average butterfly flapping is enough that it is either well above the tipping point or well below. In the language of differential equations we would say that phase diagram of the system has to be centered exactlyupon a critical point, and that the probability of this occurring is essentially(but admittedly not exactly) 0.
4. Some will also assume that evolution is random and ecological systems are random by conflating our models with reality. It is true that computational our models can predict many possible outcomes and that we can't often be sure which result is correct. That is because our models are inaccurate and they have too many free variables. But given any specific set of variables they predict a single specific answer, and our inability to measure variables in the past says nothing about the nature of any phenomenon.
5. It is also common to assume a reductionist standpoint, that since quantum physics dictates the behavior of particles at very small scales or very high energies, that the whole world is somehow fundamentally random. I've already addressed this. But let me be clear, quantum phenomena really really don't have any effect on our lives. It's useful for understanding radioactive decay, on the short term(short time scale), the path of a single photon(small space scale), or the nature of the big bang(big energy scale and small time/space).
6. The fossil record tends to support this conclusion. When causes are seen so are effects, species wax then they wane, but very very rarely do they change spontaneously or become dominant in an ecosystem, lose that dominance, and then return to dominance by some luck of the draw.
7. One might argue that the probabilistic argument comes in because the species has to get just the right mutation at the time its environment is changing or it will fail. But that is the very argument that punctuated equilibrium addresses. What it says is that the rate of mutation is low and that over the time scales that things change the right mutation will never enter the gene pool right when its needed. What its saying is the mutations accumulate over the stable periods of equilibrium, in times of plenty the gene pool's standard deviation increases until most every non-fatal mutation is floating around doing its thing. Then when times become thin, those mutations can become dominant by recombining through sexual selection, which will exchange all relevant genes to create all combinations in very short periods of time. Sudden pressure is applied and given the very predictable distribution of genes in the gene pool a modified species emerges.
Here are some responses:
1. Often things like the KT impact are considered random, but something like the KT impact was anything but random. It's just hard for us to predict now, given that it happened so long ago and our techniques for probing the distant past are so
weak that we can't predict it well. So I think its fair to exclude large things like that , which are only unpredictable in retrospect.
2. If we exclude unpredictable determinate things, then we're stuck with the smaller things that you label chaotic. The behavior of individuals, winds, currents, etc...The point I was making above, said explicitly, is that as things get smaller, as their effects become more contingent, then the probability that they alter evolution on a geological timescale becomes 0. When we talk about things orders of magnitude smaller than the scale of our phenomenon, then they are orders of magnitude less likely to be significant. Thus, we can exclude them, just like we can exclude the possibility that I will instantaneously quantum tunnel through my keyboard.
3. From these previous points, we can deduce that If we don't take the argument that evolution is non-random seriously then we are basically saying complex systems that we don't currently understand are not understandable. Which I'll admit is something that I'm just not comfortable defending.
I'm not sure I understand your counter example, so I'm not going to address it. What you did not do is address the logic of my numerical argument against the "convergent evolution is probable" point you made in your previous 2 posts. I showed how if we take it to be just rolling some dice, even with very reasonable and generous numbers, it shows that convergent evolution is extremely improbable. You also did not address my argument, that if it is just a probable event, why we don't see eyes appearing and disappearing throughout history.
Another way to look at what I'm saying, is that for your theory to correspond to our world all the probabilities for all the examples of convergent evolution have to be very precisely tuned to the specific values that make your theory correct. For all the other values, my theory is correct, and for all the reasonable values, my theory is correct.
But to be fair let me state my model clearly in Bayesian terms. What I'm saying is: P(H(t)) = [1,0] (The probability of history* at a given time t is some number between 0 and 1) P(S(t)) = [1,0] (The probability of a species or trait at a given time is some number between 0 and 1) but that P(S(t)|H(t)) => 1 or 0 as t=>G (The probability of a species or trait given history approaches 1 or 0 at we reach a geological timescale) The important thing to recognize is that it is not only a model that can account for the numerical argument I gave in my previous post, but because of the decreasing probability of multiplied contingencies it necessarily excludes the alternative.
*We should note that once a major historical event occurs its probability approaches 1.
1. I will concede that humans have changed the game, I think with humans all bets are off, because the fates of many species are dependent on our inherently unstable social systems. That said, it may very well turn out that in geological time, we weren't as irregular as we would like to think, and the fact that humans may make an exception to the rule, hardly seems to be salient to the discussion at hand. That said, even if you disagree with me, we should discuss humanity in another thread, but if your only data point is humanity, I think we can agree that your argument is on flimsy footing.
2. Generally ecosystems are not wildly fluctuating, they are extremely stable, and actually have many buffers to counteract perturbation. The only time an ecosystem fluctuates is due to some external effect, and even then they follow precise rules. To claim that these systems are hopelessly chaotic, that they don't eventually reach a specific equilibrium is to disregard a large amount of good science done in epidemiology and ecology.
3. I will concede that there was nothing inevitable about human evolution if you assume the major historical events were different, but if they were the same then it was just as deterministic as a steamroller.
To Conclude:
If you mean that evolution is random because there are a lot of unpredictable historical events, things like the KT impact, or volcanic eruptions, then I think we're really saying the same thing. If however you're saying that butterflies and tiny quantum fluctuations, individual gusts of wind, and ocean currents, are significant in evolution, then you are dead wrong. Any scenario that imagines those being significant quickly is so convoluted and contingent, that its probability of being significant quickly approaches zero on a geological timescale. The view of evolution I support has major things like cometary impacts, changes in solar output, continents coming together to make predictable changes on the gene pool driving the contingent(p(a|b)) probabilities of given traits and species survival to 1 or 0. This has happened so many times over so long that people often assume that it must be random. Or sometimes people find that this deterministic view of evolution contrasts too much with their own day-to-day experience of human interaction with nature, ecology, and their own capacity to make prediction in a very complex world.
I'll end with one observation. If we are to consider an historical system that is governed by large scale events iterated over long periods of time; things like the motion of continents and cometary impacts to be random, then we might as well call geology or the evolution of the universe random. Each of these has tons of small scale "chaotic" behavior. We don't know which rock is going to fall or even which stars in a galactic intersection will actually collide(if any). We could even imagine elaborate scenarios where those things would actually matter, but the fact that we call these things deterministic, but that some people insist on calling evolution random suggests very strongly that it is human anthropocentrism and not human logic that is making us think so.
If the above doesn't convince you, jimbo, then I doubt anything will.
I've never read this, but when you mentioned it I looked it up, and found this quote on wikipedia:
"Most of the book's conclusions were deemed controversial at publication and some of Gould's examples were soon shown to be incorrect. However, the ultimate theme of the book is still being debated among evolutionary thinkers today."
And debated right here on JREF (-:
I've read some of his older work(like Hen's Teeth and Horse's Toe's) and I've read his last work The Structure of Evolutionary Theory. Which seemed to support a much more deterministic view of evolution.
Over a geological instant, all different wind directions and magnitudes and combinations therein would have been tried many different times. So unless this happened to be a very unlikely statistical setup(where the probabilities of two different populations arriving at the Galapagos was nearly equal) then it could have happened no other way.
I can see how to you think that and I don't consider that entirely unreasonable.
But it is a question of very marginal probabilities. I don't dispute that there could be a situation where the fate of an entire species depended on whether a butterfly flaps its wings, but that is incredibly unlikely, to the point that it isn't a significant consideration as far as the theory of evolution goes, especially because the study of evolution does not work on human time. It works on quite long time scales, by human standards. I think this for a couple reasons.
1. Generally if a population is in a precarious situation, where the fates of one or two individuals determine the fate of the species, they just die, the gene pool is way too narrow in such a case. Really these "chaotic systems determine fitness" arguments depend on imagining that these situations can and do happen.
2. The probability of a characteristic of just about any chaotic system on earth becomes either 1 or 0 on scales of 10-100 years, which is generally shorter than evolutionary timescales. So all different possible combinations get tried on geological time(like your wind example). Chaotic systems are only chaotic on short time scales, for example the earth's climate is moving to a very predictable higher temperatures, they don't get more chaotic with more time, but less chaotic. But I will concede that, like a gas, it becomes increasingly difficult to tell what the behavior of any specific component will be doing as time increases. (by component I mean something like the minimum component of the chaotic system, in the cases we're often discussing, a molecule).
3. The randomness of the behavior of a chaotic system is often overblown. The sort of argument whereby the global weather could be effected by the flapping of a butterflies wings, is an interesting mathematical construction, but it is a highly contingent sort of Rube Goldberg machine, where every little piece of the argument has to be in just the right place, for a significant effect to be noticed. Practically, this is never true. People like to assume that because chaotic systems can behave in ways that fluctuate wildly on the minimum behavior of a single variable that natural systems that can exhibit chaotic behavior, always will. For example, if we want our metaphorical butterfly to change the weather, all the other butterflies have to be flapping their wings just enough that our little butterfly will push the system over the tipping point. In 99.9999999% of the cases, the average butterfly flapping is enough that it is either well above the tipping point or well below. In the language of differential equations we would say that phase diagram of the system has to be centered exactlyupon a critical point, and that the probability of this occurring is essentially(but admittedly not exactly) 0.
4. Some will also assume that evolution is random and ecological systems are random by conflating our models with reality. It is true that computational our models can predict many possible outcomes and that we can't often be sure which result is correct. That is because our models are inaccurate and they have too many free variables. But given any specific set of variables they predict a single specific answer, and our inability to measure variables in the past says nothing about the nature of any phenomenon.
5. It is also common to assume a reductionist standpoint, that since quantum physics dictates the behavior of particles at very small scales or very high energies, that the whole world is somehow fundamentally random. I've already addressed this. But let me be clear, quantum phenomena really really don't have any effect on our lives. It's useful for understanding radioactive decay, on the short term(short time scale), the path of a single photon(small space scale), or the nature of the big bang(big energy scale and small time/space).
6. The fossil record tends to support this conclusion. When causes are seen so are effects, species wax then they wane, but very very rarely do they change spontaneously or become dominant in an ecosystem, lose that dominance, and then return to dominance by some luck of the draw.
7. One might argue that the probabilistic argument comes in because the species has to get just the right mutation at the time its environment is changing or it will fail. But that is the very argument that punctuated equilibrium addresses. What it says is that the rate of mutation is low and that over the time scales that things change the right mutation will never enter the gene pool right when its needed. What its saying is the mutations accumulate over the stable periods of equilibrium, in times of plenty the gene pool's standard deviation increases until most every non-fatal mutation is floating around doing its thing. Then when times become thin, those mutations can become dominant by recombining through sexual selection, which will exchange all relevant genes to create all combinations in very short periods of time. Sudden pressure is applied and given the very predictable distribution of genes in the gene pool a modified species emerges.
I get that and I'm not completely excluding the possibility, but I'll say it again: Its not significant.
Here are some responses:
1. Often things like the KT impact are considered random, but something like the KT impact was anything but random. It's just hard for us to predict now, given that it happened so long ago and our techniques for probing the distant past are so
weak that we can't predict it well. So I think its fair to exclude large things like that , which are only unpredictable in retrospect.
2. If we exclude unpredictable determinate things, then we're stuck with the smaller things that you label chaotic. The behavior of individuals, winds, currents, etc...The point I was making above, said explicitly, is that as things get smaller, as their effects become more contingent, then the probability that they alter evolution on a geological timescale becomes 0. When we talk about things orders of magnitude smaller than the scale of our phenomenon, then they are orders of magnitude less likely to be significant. Thus, we can exclude them, just like we can exclude the possibility that I will instantaneously quantum tunnel through my keyboard.
3. From these previous points, we can deduce that If we don't take the argument that evolution is non-random seriously then we are basically saying complex systems that we don't currently understand are not understandable. Which I'll admit is something that I'm just not comfortable defending.
I'm not sure I understand your counter example, so I'm not going to address it. What you did not do is address the logic of my numerical argument against the "convergent evolution is probable" point you made in your previous 2 posts. I showed how if we take it to be just rolling some dice, even with very reasonable and generous numbers, it shows that convergent evolution is extremely improbable. You also did not address my argument, that if it is just a probable event, why we don't see eyes appearing and disappearing throughout history.
Another way to look at what I'm saying, is that for your theory to correspond to our world all the probabilities for all the examples of convergent evolution have to be very precisely tuned to the specific values that make your theory correct. For all the other values, my theory is correct, and for all the reasonable values, my theory is correct.
But to be fair let me state my model clearly in Bayesian terms. What I'm saying is: P(H(t)) = [1,0] (The probability of history* at a given time t is some number between 0 and 1) P(S(t)) = [1,0] (The probability of a species or trait at a given time is some number between 0 and 1) but that P(S(t)|H(t)) => 1 or 0 as t=>G (The probability of a species or trait given history approaches 1 or 0 at we reach a geological timescale) The important thing to recognize is that it is not only a model that can account for the numerical argument I gave in my previous post, but because of the decreasing probability of multiplied contingencies it necessarily excludes the alternative.
*We should note that once a major historical event occurs its probability approaches 1.
1. I will concede that humans have changed the game, I think with humans all bets are off, because the fates of many species are dependent on our inherently unstable social systems. That said, it may very well turn out that in geological time, we weren't as irregular as we would like to think, and the fact that humans may make an exception to the rule, hardly seems to be salient to the discussion at hand. That said, even if you disagree with me, we should discuss humanity in another thread, but if your only data point is humanity, I think we can agree that your argument is on flimsy footing.
2. Generally ecosystems are not wildly fluctuating, they are extremely stable, and actually have many buffers to counteract perturbation. The only time an ecosystem fluctuates is due to some external effect, and even then they follow precise rules. To claim that these systems are hopelessly chaotic, that they don't eventually reach a specific equilibrium is to disregard a large amount of good science done in epidemiology and ecology.
3. I will concede that there was nothing inevitable about human evolution if you assume the major historical events were different, but if they were the same then it was just as deterministic as a steamroller.
To Conclude:
If you mean that evolution is random because there are a lot of unpredictable historical events, things like the KT impact, or volcanic eruptions, then I think we're really saying the same thing. If however you're saying that butterflies and tiny quantum fluctuations, individual gusts of wind, and ocean currents, are significant in evolution, then you are dead wrong. Any scenario that imagines those being significant quickly is so convoluted and contingent, that its probability of being significant quickly approaches zero on a geological timescale. The view of evolution I support has major things like cometary impacts, changes in solar output, continents coming together to make predictable changes on the gene pool driving the contingent(p(a|b)) probabilities of given traits and species survival to 1 or 0. This has happened so many times over so long that people often assume that it must be random. Or sometimes people find that this deterministic view of evolution contrasts too much with their own day-to-day experience of human interaction with nature, ecology, and their own capacity to make prediction in a very complex world.
I'll end with one observation. If we are to consider an historical system that is governed by large scale events iterated over long periods of time; things like the motion of continents and cometary impacts to be random, then we might as well call geology or the evolution of the universe random. Each of these has tons of small scale "chaotic" behavior. We don't know which rock is going to fall or even which stars in a galactic intersection will actually collide(if any). We could even imagine elaborate scenarios where those things would actually matter, but the fact that we call these things deterministic, but that some people insist on calling evolution random suggests very strongly that it is human anthropocentrism and not human logic that is making us think so.
If the above doesn't convince you, jimbo, then I doubt anything will.
mijopaalmc
Philosopher
- Joined
- Mar 10, 2007
- Messages
- 7,172
Are you claiming that PV=nRT is all you need to know about gases?
Why is statistical mechanics such an important part of basic physical chemistry?
zosima
Muse
- Joined
- Mar 1, 2008
- Messages
- 536
I was just looking at my post, and I think I wasn't clear what I mean when I said "The only time an ecosystem fluctuates is due to some external effect".
I'll agree that taken literally that statement is clearly wrong.
What I mean is that the population dynamics of an ecosystem tends to be stable in absence of perturbation. So you can view an ecosystem as an energy economy(as far as I know, it is always pyramidal), and the number of creatures(or more accurately the total biomass) on each tier of predation will maintain the same ratio over time, and those ratios are generally stable solutions to a game theoretic problem(Nash Equilibrium). Dawkins talks about this a lot in his books. As seasons vary the numbers of individuals on all levels will vary. Although on higher levels, the weights of the creatures will vary more than the number. The geographic distribution of creatures stays quite constant as well, creatures stick to their own ecosystem, that is why introduction of species is traumatic. I figured I'd better clarify before I get accused of some quackery.
I'll agree that taken literally that statement is clearly wrong.
What I mean is that the population dynamics of an ecosystem tends to be stable in absence of perturbation. So you can view an ecosystem as an energy economy(as far as I know, it is always pyramidal), and the number of creatures(or more accurately the total biomass) on each tier of predation will maintain the same ratio over time, and those ratios are generally stable solutions to a game theoretic problem(Nash Equilibrium). Dawkins talks about this a lot in his books. As seasons vary the numbers of individuals on all levels will vary. Although on higher levels, the weights of the creatures will vary more than the number. The geographic distribution of creatures stays quite constant as well, creatures stick to their own ecosystem, that is why introduction of species is traumatic. I figured I'd better clarify before I get accused of some quackery.
zosima
Muse
- Joined
- Mar 1, 2008
- Messages
- 536
At the risk of getting sucked into the other dispute on this post. PV=nRT is an approximation. You can get better results by calculating the moments of the particle energy distribution, which generally follows a Maxwell-Boltzmann distribution. That doesn't mean the behavior of the gas becomes any less reliable than the first order approximation.
Ichneumonwasp
Unregistered
- Joined
- Feb 2, 2006
- Messages
- 6,240
Good God, no. How in the world could you ever come to that conclusion? You asked for examples of how evolution can be viewed as non-random. We view smoke detectors as non-random because we do not focus on the details of how they work normally -- that they will blare in the presence of smoke is non-random. We do not view pressure-volume gas relationships as random because we do not focus on the details of how that relationship arises normally. The same is true of evolution. That is the sense in which we have always been speaking of it being non-random -- from the big vantage.
You are focusing on the organisms themselves. On the gas molecules, smoke particles and americium decay. At that level everything is random.
What we've been trying to get you to see is that there are other levels of description.
Again, what started this whole fiasco was a misreading of Dawkins. Dawkins was arguing against JWs who characterized natural 'ranodm' processes as necessarily occurring in a one-off process, all-or-none. He rightfully corrected that misperception by relating the story of Mt Improbable -- small 'random - improbable' changes that amass into one huge improbability. At the level of individual organisms and species no one that I know argues that the process is anything but chance. But when the system as a whole is examined -- like looking at a gas container -- we do not describe it as 'random' because our interest at that level is in the predictions we can make from it.
Belz...
Fiend God
And why in the blue hell do you have this definition ?
And didn't you just define evolution as non-random ?
Ichneumonwasp
Unregistered
- Joined
- Feb 2, 2006
- Messages
- 6,240
Mijo said:
I wasn't originally going to respond to this directly, but perhaps this might help?
What you are describing is more properly placed under "the struggle for existence", one of Darwin's chapters. That is not evolution.
Evolution is an abstraction. It is the grand abstraction concerning all the interactions of various individuals -- the big grand process of changing allele frequency over time. Those individuals act 'randomly', or suffer the slings and arrows of 'random chance'. But that level of description is not evolution. It is a part of what comprises evolution. Evolution is a high level abstraction.
While the quarks that comprise a billiard ball function 'randomy', when we describe the billiard ball we do not call it random.
Last brief comment -- this is not a zero-sum game.
Last edited:
Paul C. Anagnostopoulos
Nap, interrupted.
- Joined
- Aug 3, 2001
- Messages
- 19,141
Yes, yes, I agree that evolution is a stochastic process. But your definition is worthless. One organism with an identical phenotype to another might get squashed by my shoe or blown away in a tornado, but that does not negate the fact that the survival of phenotype is not purely random with respect to the environment. In fact, those two accidents might be entirely deterministic. Instead, the survival of phenotype is clearly nonrandom with respect to the environment by the very definition of the selection process. Otherwise the surviving organisms would be arbitrarily chosen and evolution would not work.Mijo said:
Another way to put it is that the survival probability distribution is highly skewed toward those individuals more fit for the current environment. The more highly skewed it is, the more we are justified in saying the process is nonrandom. Is that acceptable?
~~ Paul
Last edited:
Paul C. Anagnostopoulos
Nap, interrupted.
- Joined
- Aug 3, 2001
- Messages
- 19,141
A modest proposal
How about this:
heads 0.5, tails 0.5: pure random
heads 1.0, tails 0.0: nonrandom, deterministic
heads 0.95, tails 0.05: nonrandom with respect to whatever is causing the skew
Does that work for people?
~~ Paul
How about this:
heads 0.5, tails 0.5: pure random
heads 1.0, tails 0.0: nonrandom, deterministic
heads 0.95, tails 0.05: nonrandom with respect to whatever is causing the skew
Does that work for people?
~~ Paul
Last edited:
Ichneumonwasp
Unregistered
- Joined
- Feb 2, 2006
- Messages
- 6,240
Works for me, but I tried that strategy a few months ago and only got back -- if it's a probability distribution, then it's random.
Walter Wayne
Wayne's Words
- Joined
- Jul 26, 2002
- Messages
- 2,502
heads 0.5, tails 0.5
Pseudo-random number generators get this ratio perfect, but are not random. The sequence where each term is the opposite of the term before it also has this ratio.
heads 1.0, tails 0.0: agreed
heads 0.95, tails 0.05: it is non-random if the condition causing the skew correlates with the result. i.e. If 95% of my coin flips I begin with the coin heads-up on my finger, and all of those come up heads then it is non-random, period.
As a side note, a pet-peeve of mine is the term "(non)random with respect to" phrase. We already have terms for this in english language; dependence, when some sort of causal relation is suspected, and correlated when the physical relation is unknown.
What that doesn't address is when you have more than two outcomes possible. What if you have many outcomes with odds ranging from 1% to 15%. This is obviously skewed. In this case the most-likely result (at 15% chance) is still and unlikely result. Do you call that random?
Walt
Last edited:
Paul C. Anagnostopoulos
Nap, interrupted.
- Joined
- Aug 3, 2001
- Messages
- 19,141
Agreed. I was thinking of a situation where we know there is no deterministic process behind the coin tossing. There is a problem with knowing this, of course.Walter said:
Those phrases do eliminate the problem word, random. "Selection is dependent on the environment." However, in this thread the whole point is to rant about the word random.
If the probability distribution is largely determined by the environment, then calling it simply "random" is misleading, I think. The probability distribution is dependent on the environment.
~~ Paul
zosima
Muse
- Joined
- Mar 1, 2008
- Messages
- 536
I really like the approach of boiling it down to how we interpret coin flips, and I think the discussion ya'll are having is adding a lot clarity.
I think in populations, you don't really get things that are that discrete where the winner of this spectral coin flip ends up winning, at least not very frequently. (I wouldn't exclude it entirely)
Its more like we have equal populations(say 200) of two species(or more). In the first round, 2 of the species at 1% survive and 30 of the species at 15% survive. In round two the species @15% does much better; its percentage goes up and the species @1% does much worse; its percentage goes down. So as long as we are making a lot of coin flips, the top percentage quickly becomes dominant.
The only time, I can really think of, where you really have a single flip effecting the entire fate of a species, is like when you have a cometary impact or a volcanic eruption. For sure depending on what time of day it is when it hits, that will have a huge effect on which species die, because that will determine point of impact, but I don't see that as particularly random.
zosima
Muse
- Joined
- Mar 1, 2008
- Messages
- 536
I don't mean to imply that there are not other interpretations of the spectrum percentage you gave. I think often what happens is that, the probability ends representing a stable solution. So that the equilibrium population is 15% species 1, 13% species 2,...1% species n.
But its hard to construct a situation where we're talking about the probability of the fate of a species and have that situation be one that we think of as random.
Last edited:
I'd agree, and have been comparing natural selection to a bent dice game...
zosima
Muse
- Joined
- Mar 1, 2008
- Messages
- 536
But its not a single roll. Its like we play a dice game all day with the dice bent in your favor. By the end of the day you're going to have all my money.
Last edited:
DeiRenDopa
Master Poster
- Joined
- Feb 25, 2008
- Messages
- 2,582
Important caveat: I'm jumping into this thread late, and have not (yet) read it through, and - more importantly - have not absorbed or understood what has already been said.
Question re randomness etc: given genetic drift (and a moderately small population with a very narrow ecological niche), given that fixation one way spells extinction (no accessible genetic path to an adaptation that leads to survival/speciation) but the other leads to survival/speciation, how could this not be (or lead to) a (potentially) purely contingent history? The variation that is fixed, one way or the other, is selectively neutral.
ETA: intended principally as a question for zosima.
Question re randomness etc: given genetic drift (and a moderately small population with a very narrow ecological niche), given that fixation one way spells extinction (no accessible genetic path to an adaptation that leads to survival/speciation) but the other leads to survival/speciation, how could this not be (or lead to) a (potentially) purely contingent history? The variation that is fixed, one way or the other, is selectively neutral.
ETA: intended principally as a question for zosima.
Last edited:
Dancing David
Penultimate Amazing
Um, chance is there but i thought the key phrase was 'contingent history' (It has been a while since I have read it.) first explored in The Panda's Thumb.
mijopaalmc
Philosopher
- Joined
- Mar 10, 2007
- Messages
- 7,172
As Walter Wayne has already mentioned, statistical dependence (what you call "contingent history") does not imply non-randomness.