What evidence is there for evolution being non-random?

[mijopaalmc]

cyborg-

I think it needs to be pointed out that two entities' being analogous doesn't mean that they share any other point of correspondence than the one used to construct. Thus, just because biological evolution and technological development are iterative processes doesn't mean that biological evolution is non-random because technological development is.

 

[cyborg]

or pseudorandomly according to Ivor and Cyborg)

Would you please read the words I am saying and not the ones you think I am FFS?

Saying "selection is nonrandom" would mean that if one could assign a hypothetical fitness value to each organism, all those above the threshold would reproduce, and all those below wouldn't.

I didn't say it was non-random. I specifically avoided such terms. I said it is deterministic with respect to the environment. There are determinants based on the genome. There are no determinants based on the genome that select for any particular mutation. It is therefore non-deterministic.

What is different in your model of evolution from "Chimps and us evolved from a common ancestor, and the nonrandom process of divine natural selection inevitably led to the evolution of humanity, who much later made fighter aircraft in WWII that evolved into better fighters"?

I have no idea what question you are trying to ask here.

 

[cyborg]

I think it needs to be pointed out that two entities' being analogous doesn't mean that they share any other point of correspondence than the one used to construct.

Uh, duh. That's what analogies are.

Thus, just because biological evolution and technological development are iterative processes doesn't mean that biological evolution is non-random because technological development is.

If one has an unrealistic concept of both design and evolution, namely one where random means whatever it is you want it to mean and engineers are somehow magically infallible and have been completely able to determine design to implementation (hint: that's not even close to the reality of it).

 

[mijopaalmc]

If evolution is deterministic with respect to the environment, why is it modeled so well with stochastic models that take the environment into account?

 

[mijopaalmc]

Uh, duh. That's what analogies are.

And the point of correspondence is extremely weak and is overridden by all the points on which biological evolution and technological development don't correspond.

If one has an unrealistic concept of both design and evolution, namely one where random means whatever it is you want it to mean and engineers are somehow magically infallible and have been completely able to determine design to implementation (hint: that's not even close to the reality of it).

And I keep telling you the image that I have of the technological development process is not unrealistic. I do not expect engineers to be perfect or their actions to have no unintended consequences, but nice mischaracterization of my argument.

 

[jimbob]

Originally Posted by jimbob
or pseudorandomly according to Ivor and Cyborg)

Would you please read the words I am saying and not the ones you think I am FFS?

OK Cyborg, has the environment, including asteroid impacts altered randomly or pseudorandomly over the life of the Earth?

I would argue that the KT event had a significant effect on the evolution of virtually every species on Earth, and it was a "random or pseudorandom" event. I thought you agreed with those statements, sorry if I misunderstood you.

Quote:
Saying "selection is nonrandom" would mean that if one could assign a hypothetical fitness value to each organism, all those above the threshold would reproduce, and all those below wouldn't.

I didn't say it was non-random. I specifically avoided such terms. I said it is deterministic with respect to the environment. There are determinants based on the genome. There are no determinants based on the genome that select for any particular mutation. It is therefore non-deterministic.

I secifically avoided "batural selectionis random," prefering "probabilistic". The vast majority of "runts" would not reproduce, a sufficient proportion of the best adapted will.

Quote:
What is different in your model of evolution from "Chimps and us evolved from a common ancestor, and the nonrandom process of divine natural selection inevitably led to the evolution of humanity, who much later made fighter aircraft in WWII that evolved into better fighters"?

I have no idea what question you are trying to ask here.

This was in response to Articulet's contention that scientific knowledge evolves, as did fighter design in WWII. I was just comparing with her statment that natural selection is nonrandom. If the last common ancestor of chimps and humans (or any of the other common ancestors) hadn't been lucky enough to reproduce, Chimps and humans wouldn't be here. Some similar organisms might have.

A chance arrow could have meant that 8%(?) of Asian males would not have been directly descended from Ghengis Kahn (father to son). No speciation, but greater than the entire human population after Toba.

Quote:
Evolutionary algorithms: Have a determined set of criteria for defining "fitness".

So does any engineering project - a good one at least.

Engineering projects do, selective breeding does, but natural selection only has one criterion. "Did it reproduce?"

This is sufficient to explain how organisms evolved. And is brilliant in its simplicity.

For the intelligence of the designer(s): enough to define the goals, and attempt a first iteration, and learn from successes and failures.

For the intelligence of the evolutionary algorithm and the selective breeding, enough to define the goal, and select the results of each generation that are closest to that goal. It can be a very simple rule.

BTW articulett, thanks for the answer to the dog size question, what I was wondering was including other canine species (like foxes) or is the answer still the same?

 

[cyborg]

If evolution is deterministic with respect to the environment, why is it modeled so well with stochastic models that take the environment into account?

Because mutations are non-deterministic with respect to the genome.

Jesus Christ - is anyone paying attention here?

 

[cyborg]

And the point of correspondence is extremely weak and is overridden by all the points on which biological evolution and technological development don't correspond.

Um, no.

And I keep telling you the image that I have of the technological development process is not unrealistic. I do not expect engineers to be perfect or their actions to have no unintended consequences, but nice mischaracterization of my argument.

Ah so technological development must be random by your concept of it since unintended consequences are quite random.

 

[cyborg]

OK Cyborg, has the environment, including asteroid impacts altered randomly or pseudorandomly over the life of the Earth?

Neither.

I would argue that the KT event had a significant effect on the evolution of virtually every species on Earth, and it was a "random or pseudorandom" event.

Such an events have a deterministic effect on selection; if your environment includes an asteroid impact then welcome to the unlucky situation that there is no gene that could help you out.

I secifically avoided "batural selectionis random," prefering "probabilistic". The vast majority of "runts" would not reproduce, a sufficient proportion of the best adapted will.

A deterministic selection.

If the last common ancestor of chimps and humans (or any of the other common ancestors) hadn't been lucky enough to reproduce, Chimps and humans wouldn't be here. Some similar organisms might have.

Which would require the environment to have been different in some way. I never said the environment had to be 'fair' in its determinations - it may not be fair that being in location X instead of Y may or may not cause you to survive but there you go. Chess' involves more than what pieces you have; it is also where you have them. If you do not take this into account when abstracting the environment then you will miss this.

A chance arrow could have meant that 8%(?) of Asian males would not have been directly descended from Ghengis Kahn (father to son). No speciation, but greater than the entire human population after Toba.

See above.

Engineering projects do, selective breeding does, but natural selection only has one criterion. "Did it reproduce?"

Is one an impossible number?

For the intelligence of the designer(s): enough to define the goals, and attempt a first iteration, and learn from successes and failures.

There's is no learning in evolution - so why do you keep on emphasising not learning from failures? Evolution doesn't 'learn' from successes either.

 

[articulett]

Jim-Bob...
in the nozzle example--you could have weird artifacts or stupid engineers that occasionally through out models that were better and it would change nothing about the definition...the evolution...what evolution is. You are using these kinds of examples to say that selection is "probabalistic" or non-random... Do you understand how these examples are tangential to the definition as to what IS going on... moreover... this doesn't mean they've come up with the best possible design or that one of the one's they discarded might have something that worked better somewhere along the evolutionary process had they kept it. You and mijo are confusing this sort of tangential detail so that you can define evolution as random. Whether it's the nozzle or the life form--it's evolving based on information that has proven successful before. That's it. Cities evolve...everyone who comes through it is involved in the design...the people who settled first have no idea what will evolve--the people adding homes or business or communities are building on what has evolved up to the present. Obsolete buildings are destroyed. Roads are widened. Cell phone towers are added. No single person is more complex than the city itself just as no single person is more complex than the internet.
Both evolve based on what is there to build upon...what worked before...what "stuck around"... all participants are designers of the evolving entity.

All life forms today are the result of eons of successful reproductions. All ecosystems today are the result of unplanned co-evolving of environment and the organisms that inhabit them. As much as you might think people are designing things like Jet fighters--they aren't designing them from scratch--they are using info. that worked before and recombining it or altering it to see if something better results. INFORMATION...changes "randomly"--the results are then culled so that the best (or luckiest, or prickliest, or worst tasting, or most anti-biotic resistant, or horniest, or most useful) survive.

All human knowledge has evolved on knowledge accumulated by humans past and present. All life forms have evolved based on accumulated information of histories "successes".

 

[articulett]

Articulett, the nozzles fit into the "evolutionary algorithms" section above.

What is different in your model of evolution from "Chimps and us evolved from a common ancestor, and the nonrandom process of divine natural selection inevitably led to the evolution of humanity, who much later made fighter aircraft in WWII that evolved into better fighters"?

I explained above how you are mixing in a minor detail of selection to determine that it is random (or probabalistic). Fitness has nothing to do with what you or anyone else thinks is more fit. You might just be fit by living in an ocean where a meteor impact doesn't hit you.

Humans and chimps have a common ancestor. That ancestor produced progeny--one of which had descendants that evolved strength and climbing abilities and other details that made it superior in the savannas and forests where it lived. Another progeny gave rise to descendants who evolved larger brains so that strength and climbing ability became less important to survival--while knowledge and weaponry and learning from others became more important. (BTW, chimps have been demonstrated to use weapons and learn from eachother...so they've evolved some of these traits as well from the common toobox of genes we share.)

Do you think feeling pain is being fitter? It seems to be essential for many animals--it has a definite survival value in our own species...those born with a mutation that makes them unable to feel pain, die early from things like appendicitis or broken bones or infections they do not know they have. Other animals feel pain too...it seems to be very important in learning what to avoid. Pleasure seems to be very important in learning what aids survival and reproductive success (orgasms feel good because those who enjoyed them the most made sure they had the most of them and passed this ability to seek and feel that pleasure onto their progeny in the process). Having a "will" to live seems to be a good "trick" for ensuring survival too (or a fear of death).

In the nozzle example--if multiple people ran the same algorithm, you wouldn't end up with the same nozzles. You'd end up with different nozzles--that might perform similarly--you could "combine" these nozzles (like sexual reproduction) and start the algorithm going again until you had a nozzle as ideal as it could be for the materials it was working with. And all humans could then use that knowledge in their nozzle design. No single person was able to design a nozzle anywhere near as good--despite their engineering degrees. But a blind algorithmic process did what natural selection does.

And no one is denying the damn randomness--I think I was the first to bring up the fact that if Hitler's mother hiccuped during sex, the whole world could be completely different--if a different sperm would have fertilized the egg, there would be different histories with different people being born and mating and marrying and inventing and adding knowledge to the world and spawning, and we may not exist at all...we have no way of knowing--but that is irrelevant in regards to what DID happen--what was selected--who was born. It would be like saying the evolution of the nozzle was random because there were designs that could have worked better that were never tested or never thought of or because one modification eliminated the possibility of a better modification (by the way...the nozzle is really cool looking...maybe I'll post a link).

Different people could have brought different industry to your city and different businesses and a different make up of populations--but that is irrelevant as to what WAS selected--what worked best for the people there as they were passing through or deciding where to live--every aspect of our lives has randomness--from where we live, to whom we mate with and whom we spawn and what careers we have--but our lives aren't random. Selection drives the evolution of everything... the stuff that sticks...the evolution of this thread is not preplanned or designed, but it is entirely based on what is written before--what sticks...it has nothing to do with all the possible people who could have posted or all the better ways things could have been said, and everything to do with what IS said, what does appear--and all aspects of the environment that have anything to do with the appearance of these words, is "selecting". It has nothing to do with any human sense of fitness. It has everything to do with the physical aspects of the internet and the physical environment of those who read this on their computers, and then type a response. It's not "pre-planned", but it most definitely is built on what comes before--though like cyborg, I wonder if you are actually reading and assimilating what is coming before. This goes triple for mijo.

And the only point being made is that it's uninformative to call evolution random. The relative randomness of mutation is easy to understand...you are letting random components of natural selection define the whole process as random--similar to a math problem with random variables being described as being a random math problem.

And mijo, there really aren't successful models in use describing evolution as stochastic as far as I can tell--mutations and recombinations...yes--but natural selection is a pretty simple concept even though you can't seem to get it--and calling it random or modeling it with stochastic charts doesn't seem to be helping you get it and is not used by anyone in any significant way as far as I can tell. I can't imagine anyone else finding it useful; it seems to be confusing you.

 

[mijopaalmc]

Um, no.

Um....Evidence?

I have presented several points on which biological evolution does not correspond to technological development and offered reasons why these points are more pertinent to the randomness of either process than their shared iterative nature. Now, it is up to you to rebut it with evidence rather than straight out denial.

Ah so technological development must be random by your concept of it since unintended consequences are quite random.

It is the intent that matters here not its outcome, because it is the intent against the value of the outcome is judged. There is no intent behind a genetic mutation, it just happens and then is sorted out by the environment. Furthermore, the learning inherent in technological development makes the process fundamentally different than biological evolution, because the information contained in the entity can be transferred from agent to agent without having to be made manifest in an entity in technological development whereas the information has to be made manifest in an entity in order to perpetuate itself.

 

[articulett]

Um....Evidence?

I have presented several points on which biological evolution does not correspond to technological development and offered reasons why these points are more pertinent to the randomness of either process than their shared iterative nature. Now, it is up to you to rebut it with evidence rather than straight out denial.



It is the intent that matters here not its outcome, because it is the intent against the value of the outcome is judged. There is no intent behind a genetic mutation, it just happens and then is sorted out by the environment. Furthermore, the learning inherent in technological development makes the process fundamentally different than biological evolution, because the information contained in the entity can be transferred from agent to agent without having to be made manifest in an entity in technological development whereas the information has to be made manifest in an entity in order to perpetuate itself.

In the nozzle example...as in all examples including the evolution of this thread, the info. must be made manifest before being selected. Whether it's posting on this forum, or trying out new designs, or putting a new Starbucks in a city--until it is manifest, the information cannot be acted upon (selected) in any way.

The information in all examples is modified based on what "survives"--what information is manifested so as to be built upon or die out. You need to make the recipe before anything can be selected for "improvement" or modification.
It's all still INFORMATION. The "selected" info. can evolve...whether it's language, blueprints, forum threads, cities, computer viruses or the DNA that makes livestock, crops, dinosaurs, cancer, and all forms of life in existence--whether it's selected on purpose or it had some "trick" that aided it's survival--whether we find it fit or useful or purposeful or not--whether a million entities contribute to the design or blind meandering physical properties in the environment which allows something to copy itself (and thus be preferentially selected) before being destroyed.

A male praying mantis who mates for some time, has a better chance at plugging the female so that other males can't inseminate her...he also has a better chance at getting his head bit off in the process. His male offspring will inherit his risk taking staying power, but not to their own benefit...but to the benefit of the genes that get carried on in the process. Fitter? To whom? I guess his head makes a nice meal for his gestating mate and her offspring as well.

Biological selection truly only has one "goal" as it's algorithm, and what you think of as "fitness" is not that goal. The goal is having the most copies of your genome carried forward in time via living vectors (offspring). That's it. Whether the organisms know or care about this goal is irrelevant. Whether they have other goals is irrelevant.

 

[jimbob]

Back to why I would say that evolution is probabilistic.

It is possible to conceive of the following limited experiment:

a fixed amount (say 5g) of a certain strain of bacteria (E Coli, perhaps) allow it to grow in a benign environment until the mass is 1kg.

Divide into 100 sets of growth media, and at a fixed time, add a defined amount of a particular antibiotic to these media. Repeat the doseage at defined intervals.

One could then from this experiment say that 90% of the population had evolved resistance to a dosage x after a certain number of generations (or a certain time). It is highly unlikely that it would be the same time ineach case, so the time for this to develop could be described statistically, possibly by a poisson distribution.

"There is a 70% chance that this would evolve in x generations."

One could perform sufficient statistical tests to determine the truly random spread of this process, by repeating with different starting populations of the initial strain.

One could also see what the effect of controlled (non-leathal) radiation doses would be on this, and whether this reduces the time required.

I explained above how you are mixing in a minor detail of selection to determine that it is random (or probabalistic). Fitness has nothing to do with what you or anyone else thinks is more fit. You might just be fit by living in an ocean where a meteor impact doesn't hit you.

I keep harping on about this, because I am sure that mijo is fully aware of this, as am I. The whole point from my POV is that, as only those that are "fit" and lucky reproduce; the whole process is probabilistic. If you are arguing for a fully deterministic model, then in principle there is a definition of "fitness" that one could in principle calculate for any organism, and those below the threshold would not reproduce, whilst all those above it would.

Evolutionary algorithms and artificial selection do have this defined definition of fitness. In natural selection there isn't one that you can (in principle) calculate, just odds of a particular individual reproducing. These odds are heavily modulated by its traits and characteristics, but they are still odds.

I would argue that the environment changes randomly, especially over geological timeframes, so you can (in principle) make predictions about the chances of something evoloving, but not guarantee that it will within any timeframe. If the weather had been slightly different, when the parent population of Darwin's finches were blown offshore, then they wouldn't have evolved.

Biological selection truly only has one "goal" as it's algorithm, and what you think of as "fitness" is not that goal. The goal is having the most copies of your genome carried forward in time via living vectors (offspring). That's it. Whether the organisms know or care about this goal is irrelevant. Whether they have other goals is irrelevant.

That is the whole point. And mijo, is this your point too?

In engineering, there is a goal, and it is defined, it might turn out to be the wrong goal, but it is still defined.

Engineering design is not just evolutionary algorithms. They can be useful, but by their nature, the only "learning" is that "option x is better than option y", so make slight variations on option x, (it is easier to discuss asexual evolutionary algorithms, though sexual ones tend to be better).

Engineering design can assess that information that option x is better than option y, but can also say that something failed because of this, let's alter the performance of a particular parameter in a certain direction.

Maybe I want a fatter mouse. I could selectively breed for obesity, or I could say, "I know that these genes are involved in appetite and saitity signals, let's try knocking this gene out..."

Do you see the difference in the approaches?

Or from my own field

"This transistor works well, but its breakdown-voltage is higher than needed, and its on-resistance is too high, let's increase the epi-doping to reduce both". This is obtaining information from failure; that is not possible from evolution.

I would say that a meme is only an analogy, which is sometimes useful, and sometimes confusing.

But I do agree with you that evolution works at the level of the gene.

 

[cyborg]

Now, it is up to you to rebut it with evidence rather than straight out denial.

No, I don't need to rebut anything. What you need to do is grasp the concept of the analogy here.

It is the intent that matters here not its outcome, because it is the intent against the value of the outcome is judged.

Intent. What a lovely fuzzy concept.

There is no intent behind a genetic mutation, it just happens and then is sorted out by the environment.

There is no intent behind a failed prototype, it just happens and then is sorted out by the engineer.

Furthermore, the learning inherent in technological development makes the process fundamentally different than biological evolution, because the information contained in the entity can be transferred from agent to agent without having to be made manifest in an entity in technological development whereas the information has to be made manifest in an entity in order to perpetuate itself.

So in other words we're back to you have a problem with the way the information is stored and nothing whatsoever to do with how the information is changing.

So again I say:

No, I don't need to rebut anything. What you need to do is grasp the concept of the analogy here.

 

[Walter Wayne]

From here.

... And why are the people here who see unqualified statements where there aren't any the same people who have a hard time seeing how natural selection is "non-random" in comparison to mutation with regards to evolution? Hmmm... maybe someone is defending religion because they think it does no harm, not even realizing how it's harmed their ability to understand a very simple concept (Meadmaker, Walter Wayne, Mijo).

Articulett, I am responding here so as not to derail that thread.

You continue on about natural selection is non-random vehemently, and I can understand why you may be passionate about it given what you may see in the community around you.

But, at the same time, you continue to use just one definition of random (that of uniform probability), when I and others have pointed out several instances where laymen use it in other fashions (movement in monopoly, etc.). You have thrown around the straw-man that we wish to sum natural selection up as random, insinuating that our explaination leads to ambiguity. If I wish I could point out that you wish to sum it up as non-random which leads to amibguity and misinformation. I would not do that, because I know you know the discussion doesn't end there. You should know that we also know that the discussion doesn't end at "evolution is random", but you are either putting blinders on or deliberately misrepresenting us.

Why I have a hard-time seeing natural selection as non-random:
1) It isn't, except for one or possibly two particular definitions
2) It isn't just random only in the micro-scale. Because of heredity and the interaction of organisms with eachother the randomness will not simply be weeded out in numbers as it is in photons defining an electric field or gas molecules giving us a relation between pressure, volume and temperature.
3) I refuse to simplify things to the point of being wrong. Borrowing a phrase, explainations should be made as simple as possible, but no simpler.

Walt

 

[cyborg]

when I and others have pointed out several instances where laymen use it in other fashions

And as such I have pointed out that what laymen conceptualise as random is irrelevant. Which is why I provided a precise and accurate definition avoiding use of that word.

 

[Walter Wayne]

And as such I have pointed out that what laymen conceptualise as random is irrelevant. Which is why I provided a precise and accurate definition avoiding use of that word.

Since, the subject has been in part "how to get the topic across" the layman's definition is relevant in part to myself, and I believe to articulett as well.

With respect to what I have read of your argument, I am somewhat hesitant to say that natural selection is determistic with respect to environment. To take a simple example that has been brought up before, consider two eggs in a nest. The west egg is eaten by a predator before it is scared off by a parent. Exchanging its genes with its brother would not have helped it in any way, but yet it has been deemed "less fit" so to speak. The only way I can see of saying that it is deterministic with repect to the environment is to treat the west egg environment as a different environment from the other.

An over simplified example, but the point is that at times an animal can be culled while a near neighbour lives on to reproduce for no reason connected to its genes.

So while I could agree that it is deterministic based on environment, I think it rather useless if we have to define every creatures environment as different. It is at its heart true, I only have to succeed in my personal environment, not in the environment of the guy in the crowd two feet away from me. But generally we don't define environment so narrowly in conversation.

Hence, my hesitation in saying selection is determistic with respect to environment.

Walt

 

[cyborg]

The only way I can see of saying that it is deterministic with repect to the environment is to treat the west egg environment as a different environment from the other.

If position is NOT important than a fish should do as well as land as in the sea right?

I already pointed this out but since no-one is paying any attention I am not surprised people keep on banging away at the same points.

So while I could agree that it is deterministic based on environment, I think it rather useless if we have to define every creatures environment as different.

Ah, so fish should be doing as well on land as in the sea then.

It is at its heart true, I only have to succeed in my personal environment, not in the environment of the guy in the crowd two feet away from me. But generally we don't define environment so narrowly in conversation.

Um, that's not a narrow definition of environment, it's a precise one.

 

[jimbob]

Um, that's not a narrow definition of environment, it's a precise one.

But that environment varies randomly, or pseudorandomly if you are a determinist and don't accept QM randomness. Weather affects that micro environment.

When talking about populations where individuals are affected by random events, predictions can be made in terms of odds, or probabilities.