Randomness in evolution / natural selection

Get your science fix here: research, quackery, activism and all the rest
Post Reply
User avatar
jimbob
After Pie
Posts: 2469
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Randomness in evolution / natural selection

Post by jimbob » Wed Apr 21, 2021 9:19 pm

Mod note: these posts were split from a different discussion at both OPs' requests.

Slight digression but I actually think his [mod: Richard Dawkins'] popularised explanation of evolution is misleading in a way that needn't be.

Describing evolution as "the opposite of random" is not really correct and lends itself to the argument made by several ID proponents, that it's a deity setting the direction. It also opens itself to the interpretation that humanity is the current "pinnacle" of evolution as presumably something "the opposite of random" has been pushing the evolution of our ancestors towards us. I also take issue with the phrase "evolved to" for example "naked mole rats evolved to have loose skin for moving in their tunnels" which has it backwards.

If he didn't insist on having the word "opposite" I would accept it. It is the opposite of blind chance. But the actual selective landscape is significantly affected by random events - including mutations from other organisms, or the history of previous mutations for the organism's ancestors. You can't get grasslands without grass, for example. And there were hundreds of millions of years of other sophisticated plants before grasses evolved. Ecological niches open and close due to stochastic processes. Also catastrophic events.
Last edited by Bird on a Fire on Thu Apr 22, 2021 7:27 pm, edited 2 times in total.
Have you considered stupidity as an explanation

User avatar
Bird on a Fire
Light of Blast
Posts: 5736
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: Randomness in evolution / natural selection

Post by Bird on a Fire » Wed Apr 21, 2021 10:45 pm

jimbob wrote:
Wed Apr 21, 2021 9:19 pm
Slight digression but I actually think his popularised explanation of evolution is misleading in a way that needn't be.

Describing evolution as "the opposite of random" is not really correct and lends itself to the argument made by several ID proponents, that it's a deity setting the direction. It also opens itself to the interpretation that humanity is the current "pinnacle" of evolution as presumably something "the opposite of random" has been pushing the evolution of our ancestors towards us. I also take issue with the phrase "evolved to" for example "naked mole rats evolved to have loose skin for moving in their tunnels" which has it backwards.

If he didn't insist on having the word "opposite" I would accept it. It is the opposite of blind chance. But the actual selective landscape is significantly affected by random events - including mutations from other organisms, or the history of previous mutations for the organism's ancestors. You can't get grasslands without grass, for example. And there were hundreds of millions of years of other sophisticated plants before grasses evolved. Ecological niches open and close due to stochastic processes. Also catastrophic events.
Well, Dawkins did deliberately downplay the importance of random events, as he didn't think the randomness is that important: if you have enough chances, random events happen; somebody wins the lottery every week.

There was a big ideological spat in the 80s between the dominant view of natural selection as an optimization routine, and the Marxist contingent emphasizing the importance of historical contingency and environmental context. Part of it was characterised as a battle between "jerks" and "creeps" - creeps like Dawkins who think evolution creeps along at a steady, infinitesimal pace versus jerks like Stephen Jay Gould who think there's often approximate stasis, interspersed with sudden jerks of accelerated evolution, often responding to an environmental crisis (or key innovation).

I'm not quite sure what a contemporary orthodox evolutionary biologist would say, but I think the creeps largely won in terms of shifting the dominant paradigm. They also published a paper with the best paper title ever. Clearly both randomness and optimisation are important. I'd say randomness defines the parameter space over which natural selection optimises. It's probably not a coincidence that Dawkins' hands-on research focussed on micro-evolutionary behavioural changes, whereas Gould was a palaeontologist looking at shifts in gastropod bauplan.
He has the grace of a swan, the wisdom of an owl, and the eye of an eagle—ladies and gentlemen, this man is for the birds!

User avatar
basementer
Catbabel
Posts: 801
Joined: Mon Nov 11, 2019 1:03 pm
Location: Aotearoa
Contact:

Re: Randomness in evolution / natural selection

Post by basementer » Thu Apr 22, 2021 1:14 am

But this one is a strong contender IMO.
Money is just a substitute for luck anyway. - Tom Siddell

User avatar
nekomatic
Snowbonk
Posts: 502
Joined: Mon Nov 11, 2019 3:04 pm

Re: Randomness in evolution / natural selection

Post by nekomatic » Thu Apr 22, 2021 5:59 am

Those aren’t bad, but they aren’t this

User avatar
shpalman
Stummy Beige
Posts: 4180
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: Randomness in evolution / natural selection

Post by shpalman » Thu Apr 22, 2021 6:47 am

Bird on a Fire wrote:
Wed Apr 21, 2021 10:45 pm
jimbob wrote:
Wed Apr 21, 2021 9:19 pm
Slight digression but I actually think his popularised explanation of evolution is misleading in a way that needn't be.

Describing evolution as "the opposite of random" is not really correct and lends itself to the argument made by several ID proponents, that it's a deity setting the direction. It also opens itself to the interpretation that humanity is the current "pinnacle" of evolution as presumably something "the opposite of random" has been pushing the evolution of our ancestors towards us. I also take issue with the phrase "evolved to" for example "naked mole rats evolved to have loose skin for moving in their tunnels" which has it backwards.

If he didn't insist on having the word "opposite" I would accept it. It is the opposite of blind chance. But the actual selective landscape is significantly affected by random events - including mutations from other organisms, or the history of previous mutations for the organism's ancestors. You can't get grasslands without grass, for example. And there were hundreds of millions of years of other sophisticated plants before grasses evolved. Ecological niches open and close due to stochastic processes. Also catastrophic events.
Well, Dawkins did deliberately downplay the importance of random events, as he didn't think the randomness is that important: if you have enough chances, random events happen; somebody wins the lottery every week.

There was a big ideological spat in the 80s between the dominant view of natural selection as an optimization routine, and the Marxist contingent emphasizing the importance of historical contingency and environmental context. Part of it was characterised as a battle between "jerks" and "creeps" - creeps like Dawkins who think evolution creeps along at a steady, infinitesimal pace versus jerks like Stephen Jay Gould who think there's often approximate stasis, interspersed with sudden jerks of accelerated evolution, often responding to an environmental crisis (or key innovation).

I'm not quite sure what a contemporary orthodox evolutionary biologist would say, but I think the creeps largely won in terms of shifting the dominant paradigm. They also published a paper with the best paper title ever. Clearly both randomness and optimisation are important. I'd say randomness defines the parameter space over which natural selection optimises. It's probably not a coincidence that Dawkins' hands-on research focussed on micro-evolutionary behavioural changes, whereas Gould was a palaeontologist looking at shifts in gastropod bauplan.
Once you get to reasonably complex organisms, random mutations aren't that important for evolution. They're almost always deleterious. Sexual reproduction involves organisms choosing the best alleles to shuffle their own alleles together with, except of course that they base their choices on the phenotype created by the alleles of the other organism.
molto tricky

User avatar
bob sterman
Snowbonk
Posts: 474
Joined: Mon Nov 11, 2019 10:25 pm

Re: Randomness in evolution / natural selection

Post by bob sterman » Thu Apr 22, 2021 6:53 am

Bird on a Fire wrote:
Wed Apr 21, 2021 10:45 pm
It's probably not a coincidence that Dawkins' hands-on research focussed on micro-evolutionary behavioural changes, whereas Gould was a palaeontologist looking at shifts in gastropod bauplan.
Indeed - looking at the fossil record is liable to make you a jerk rather than a creep. However, since Gould's death, with the development of the genetic techniques and computing power needed for molecular cladistics - the jerks are looking fewer.

On issues away from Evolutionary Biology - Gould tended to be way less offensive than Dawkins has been in recent years. Although who knows what he would have been like on Twitter?

However, in their popular science writings - Dawkins tended to give a more accurate picture of what was going on in evolutionary biology at the time.

To quote John Maynard Smith on Gould -
"Because of the excellence of his essays, he has come to be seen by non-biologists as the preeminent evolutionary theorist. In contrast, the evolutionary biologists with whom I have discussed his work tend to see him as a man whose ideas are so confused as to be hardly worth bothering with, but as one who should not be publicly criticized because he is at least on our side against the creationists. All this would not matter, were it not that he is giving non-biologists a largely false picture of the state of evolutionary theory."

User avatar
jimbob
After Pie
Posts: 2469
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: Randomness in evolution / natural selection

Post by jimbob » Thu Apr 22, 2021 8:58 am

shpalman wrote:
Thu Apr 22, 2021 6:47 am
Bird on a Fire wrote:
Wed Apr 21, 2021 10:45 pm
jimbob wrote:
Wed Apr 21, 2021 9:19 pm
Slight digression but I actually think his popularised explanation of evolution is misleading in a way that needn't be.

Describing evolution as "the opposite of random" is not really correct and lends itself to the argument made by several ID proponents, that it's a deity setting the direction. It also opens itself to the interpretation that humanity is the current "pinnacle" of evolution as presumably something "the opposite of random" has been pushing the evolution of our ancestors towards us. I also take issue with the phrase "evolved to" for example "naked mole rats evolved to have loose skin for moving in their tunnels" which has it backwards.

If he didn't insist on having the word "opposite" I would accept it. It is the opposite of blind chance. But the actual selective landscape is significantly affected by random events - including mutations from other organisms, or the history of previous mutations for the organism's ancestors. You can't get grasslands without grass, for example. And there were hundreds of millions of years of other sophisticated plants before grasses evolved. Ecological niches open and close due to stochastic processes. Also catastrophic events.
Well, Dawkins did deliberately downplay the importance of random events, as he didn't think the randomness is that important: if you have enough chances, random events happen; somebody wins the lottery every week.

There was a big ideological spat in the 80s between the dominant view of natural selection as an optimization routine, and the Marxist contingent emphasizing the importance of historical contingency and environmental context. Part of it was characterised as a battle between "jerks" and "creeps" - creeps like Dawkins who think evolution creeps along at a steady, infinitesimal pace versus jerks like Stephen Jay Gould who think there's often approximate stasis, interspersed with sudden jerks of accelerated evolution, often responding to an environmental crisis (or key innovation).

I'm not quite sure what a contemporary orthodox evolutionary biologist would say, but I think the creeps largely won in terms of shifting the dominant paradigm. They also published a paper with the best paper title ever. Clearly both randomness and optimisation are important. I'd say randomness defines the parameter space over which natural selection optimises. It's probably not a coincidence that Dawkins' hands-on research focussed on micro-evolutionary behavioural changes, whereas Gould was a palaeontologist looking at shifts in gastropod bauplan.
Once you get to reasonably complex organisms, random mutations aren't that important for evolution. They're almost always deleterious. Sexual reproduction involves organisms choosing the best alleles to shuffle their own alleles together with, except of course that they base their choices on the phenotype created by the alleles of the other organism.
I was aware of the spat

I'd say that things like the Chicxulub impact were pretty random and had an enormous impact. And smaller, but still huge events could be even affected by weather. The Toba eruption came close to wiping out our hominid branch - would any others have survived if our ancestors hadn't got lucky? And now in the Anthropocene, serendipitous events/discoveries that affect our culture and behaviour also have huge impacts on the development of ecosystems and the organisms within them.

On the micro scale - the Long Term Evolution experiment does seem to show that in some of the repeat trials that a few random mutations allowed the E. coli to utilise a previously-inaccessible resource and that affected the whole selective landscape within those particular jars.

Yes the evolution of traits within a selective landscape is not random. And various traits are so often beneficial that they evolve independently many times. However I'd say that the selective landscape is subject to change both from within and without. Some of the change from within could be predicted, but there is also the chance for something unexpected.
Have you considered stupidity as an explanation

User avatar
Bird on a Fire
Light of Blast
Posts: 5736
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: Randomness in evolution / natural selection

Post by Bird on a Fire » Thu Apr 22, 2021 11:12 am

shpalman wrote:
Thu Apr 22, 2021 6:47 am
Bird on a Fire wrote:
Wed Apr 21, 2021 10:45 pm
jimbob wrote:
Wed Apr 21, 2021 9:19 pm
Slight digression but I actually think his popularised explanation of evolution is misleading in a way that needn't be.

Describing evolution as "the opposite of random" is not really correct and lends itself to the argument made by several ID proponents, that it's a deity setting the direction. It also opens itself to the interpretation that humanity is the current "pinnacle" of evolution as presumably something "the opposite of random" has been pushing the evolution of our ancestors towards us. I also take issue with the phrase "evolved to" for example "naked mole rats evolved to have loose skin for moving in their tunnels" which has it backwards.

If he didn't insist on having the word "opposite" I would accept it. It is the opposite of blind chance. But the actual selective landscape is significantly affected by random events - including mutations from other organisms, or the history of previous mutations for the organism's ancestors. You can't get grasslands without grass, for example. And there were hundreds of millions of years of other sophisticated plants before grasses evolved. Ecological niches open and close due to stochastic processes. Also catastrophic events.
Well, Dawkins did deliberately downplay the importance of random events, as he didn't think the randomness is that important: if you have enough chances, random events happen; somebody wins the lottery every week.

There was a big ideological spat in the 80s between the dominant view of natural selection as an optimization routine, and the Marxist contingent emphasizing the importance of historical contingency and environmental context. Part of it was characterised as a battle between "jerks" and "creeps" - creeps like Dawkins who think evolution creeps along at a steady, infinitesimal pace versus jerks like Stephen Jay Gould who think there's often approximate stasis, interspersed with sudden jerks of accelerated evolution, often responding to an environmental crisis (or key innovation).

I'm not quite sure what a contemporary orthodox evolutionary biologist would say, but I think the creeps largely won in terms of shifting the dominant paradigm. They also published a paper with the best paper title ever. Clearly both randomness and optimisation are important. I'd say randomness defines the parameter space over which natural selection optimises. It's probably not a coincidence that Dawkins' hands-on research focussed on micro-evolutionary behavioural changes, whereas Gould was a palaeontologist looking at shifts in gastropod bauplan.
Once you get to reasonably complex organisms, random mutations aren't that important for evolution. They're almost always deleterious. Sexual reproduction involves organisms choosing the best alleles to shuffle their own alleles together with, except of course that they base their choices on the phenotype created by the alleles of the other organism.
Yes, exactly, plus most adaptive changes need more than one mutation, as they result from interacting networks of multiple genes. This means that the "next step" is massively contingent on what's gone before and what's available within the genome to work with (retroviruses etc. aside). The old-fashioned adaptationist narrative was that natural selection optimised between various constraints, but generally ignored the fact that the available options are highly lineage-specific, and therefore often biogeographically restricted, etc etc.

Natural selection is more like when the A Team jerry-rig an aircraft out of whatever crap is in the shed they're locked in, than a precision-engineered Boeing factory.
He has the grace of a swan, the wisdom of an owl, and the eye of an eagle—ladies and gentlemen, this man is for the birds!

User avatar
Bird on a Fire
Light of Blast
Posts: 5736
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: Randomness in evolution / natural selection

Post by Bird on a Fire » Thu Apr 22, 2021 11:16 am

bob sterman wrote:
Thu Apr 22, 2021 6:53 am
Bird on a Fire wrote:
Wed Apr 21, 2021 10:45 pm
It's probably not a coincidence that Dawkins' hands-on research focussed on micro-evolutionary behavioural changes, whereas Gould was a palaeontologist looking at shifts in gastropod bauplan.
Indeed - looking at the fossil record is liable to make you a jerk rather than a creep. However, since Gould's death, with the development of the genetic techniques and computing power needed for molecular cladistics - the jerks are looking fewer.

On issues away from Evolutionary Biology - Gould tended to be way less offensive than Dawkins has been in recent years. Although who knows what he would have been like on Twitter?

However, in their popular science writings - Dawkins tended to give a more accurate picture of what was going on in evolutionary biology at the time.

To quote John Maynard Smith on Gould -
"Because of the excellence of his essays, he has come to be seen by non-biologists as the preeminent evolutionary theorist. In contrast, the evolutionary biologists with whom I have discussed his work tend to see him as a man whose ideas are so confused as to be hardly worth bothering with, but as one who should not be publicly criticized because he is at least on our side against the creationists. All this would not matter, were it not that he is giving non-biologists a largely false picture of the state of evolutionary theory."
For sure, I think people like Lewontin and Levins are more respected as academic theorists. Nevertheless, every evolutionary biology prof who's taught me in the last decade has made a point of pointing us towards Gould's work, so I think maybe the newer generation has rehabilitated his reputation a little bit now the egos and McCarthyism are out of the way.

JMS is pretty old school - born over 100 years ago!
He has the grace of a swan, the wisdom of an owl, and the eye of an eagle—ladies and gentlemen, this man is for the birds!

User avatar
Woodchopper
Stummy Beige
Posts: 3143
Joined: Sat Oct 12, 2019 9:05 am

Re: Randomness in evolution / natural selection

Post by Woodchopper » Thu Apr 22, 2021 11:46 am

Bird on a Fire wrote:
Thu Apr 22, 2021 11:12 am
Yes, exactly, plus most adaptive changes need more than one mutation, as they result from interacting networks of multiple genes. This means that the "next step" is massively contingent on what's gone before and what's available within the genome to work with (retroviruses etc. aside). The old-fashioned adaptationist narrative was that natural selection optimised between various constraints, but generally ignored the fact that the available options are highly lineage-specific, and therefore often biogeographically restricted, etc etc.

Natural selection is more like when the A Team jerry-rig an aircraft out of whatever crap is in the shed they're locked in, than a precision-engineered Boeing factory.
I seem to remember that a good example was from an earlier discussion about why no mammals have green fur (algae covered Sloths aside). There were several explanations and one was that the pigmentation genes can only do shades of red, brown, orange and yellow etc. A simple random mutation isn't going to produce a green furred mammal which might have an advantage if it lived in a jungle.

User avatar
jimbob
After Pie
Posts: 2469
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: Randomness in evolution / natural selection

Post by jimbob » Thu Apr 22, 2021 3:03 pm

Bird on a Fire wrote:
Thu Apr 22, 2021 11:12 am
shpalman wrote:
Thu Apr 22, 2021 6:47 am
Bird on a Fire wrote:
Wed Apr 21, 2021 10:45 pm


Well, Dawkins did deliberately downplay the importance of random events, as he didn't think the randomness is that important: if you have enough chances, random events happen; somebody wins the lottery every week.

There was a big ideological spat in the 80s between the dominant view of natural selection as an optimization routine, and the Marxist contingent emphasizing the importance of historical contingency and environmental context. Part of it was characterised as a battle between "jerks" and "creeps" - creeps like Dawkins who think evolution creeps along at a steady, infinitesimal pace versus jerks like Stephen Jay Gould who think there's often approximate stasis, interspersed with sudden jerks of accelerated evolution, often responding to an environmental crisis (or key innovation).

I'm not quite sure what a contemporary orthodox evolutionary biologist would say, but I think the creeps largely won in terms of shifting the dominant paradigm. They also published a paper with the best paper title ever. Clearly both randomness and optimisation are important. I'd say randomness defines the parameter space over which natural selection optimises. It's probably not a coincidence that Dawkins' hands-on research focussed on micro-evolutionary behavioural changes, whereas Gould was a palaeontologist looking at shifts in gastropod bauplan.
Once you get to reasonably complex organisms, random mutations aren't that important for evolution. They're almost always deleterious. Sexual reproduction involves organisms choosing the best alleles to shuffle their own alleles together with, except of course that they base their choices on the phenotype created by the alleles of the other organism.
Yes, exactly, plus most adaptive changes need more than one mutation, as they result from interacting networks of multiple genes. This means that the "next step" is massively contingent on what's gone before and what's available within the genome to work with (retroviruses etc. aside). The old-fashioned adaptationist narrative was that natural selection optimised between various constraints, but generally ignored the fact that the available options are highly lineage-specific, and therefore often biogeographically restricted, etc etc.

Natural selection is more like when the A Team jerry-rig an aircraft out of whatever crap is in the shed they're locked in, than a precision-engineered Boeing factory.

The bolded bit is why I think of it as random. The initial organism's descendants will have mutations that predispose them towards certain niches and make others less available. As long as there are descendants of lions, there would not be any opportunity for the descendants of cheetahs to become more generalist pack hunters, because they'd be outcompeted by the lions' descendants.

It's a drunkard's walk where one can explain how one got there, but unless one is talking about specific traits (antibiotic resistance being a classic one) what would happen is dependent on multiple options so is often not predictable (nor, I'd argue, predetermined).

It's like the A-team lashing something up, and sometimes they have a bulldozer, whilst other times they have a quadbike and this is subject to the previous story. Probably more like scrapheap challenge. You can start down one route, but that closes others, even though the initial direction was random.


Certainly describing evolution as "the opposite of random" is more than downplaying the role of chance events - it's denying it. Dinosaurs were top predators on land - why did mammals and reptiles replace them? It wasn't anything internal to the selective landscape in the thousands of years before most of them became extinct.
Have you considered stupidity as an explanation

User avatar
Bird on a Fire
Light of Blast
Posts: 5736
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: Randomness in evolution / natural selection

Post by Bird on a Fire » Thu Apr 22, 2021 7:23 pm

Split some posts from another thread in another subforum.
He has the grace of a swan, the wisdom of an owl, and the eye of an eagle—ladies and gentlemen, this man is for the birds!

User avatar
shpalman
Stummy Beige
Posts: 4180
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: Randomness in evolution / natural selection

Post by shpalman » Thu Apr 22, 2021 7:49 pm

jimbob wrote:
Thu Apr 22, 2021 3:03 pm
Bird on a Fire wrote:
Thu Apr 22, 2021 11:12 am
shpalman wrote:
Thu Apr 22, 2021 6:47 am

Once you get to reasonably complex organisms, random mutations aren't that important for evolution. They're almost always deleterious. Sexual reproduction involves organisms choosing the best alleles to shuffle their own alleles together with, except of course that they base their choices on the phenotype created by the alleles of the other organism.
Yes, exactly, plus most adaptive changes need more than one mutation, as they result from interacting networks of multiple genes. This means that the "next step" is massively contingent on what's gone before and what's available within the genome to work with (retroviruses etc. aside). The old-fashioned adaptationist narrative was that natural selection optimised between various constraints, but generally ignored the fact that the available options are highly lineage-specific, and therefore often biogeographically restricted, etc etc.

Natural selection is more like when the A Team jerry-rig an aircraft out of whatever crap is in the shed they're locked in, than a precision-engineered Boeing factory.
The bolded bit is why I think of it as random. The initial organism's descendants will have mutations that predispose them towards certain niches and make others less available. As long as there are descendants of lions, there would not be any opportunity for the descendants of cheetahs to become more generalist pack hunters, because they'd be outcompeted by the lions' descendants.

It's a drunkard's walk where one can explain how one got there, but unless one is talking about specific traits (antibiotic resistance being a classic one) what would happen is dependent on multiple options so is often not predictable (nor, I'd argue, predetermined).
If there's selection pressure then it's a drunkard's walk if the drunkard is on a hill, or can smell a pub or something.
jimbob wrote:
Thu Apr 22, 2021 3:03 pm
It's like the A-team lashing something up, and sometimes they have a bulldozer, whilst other times they have a quadbike and this is subject to the previous story. Probably more like scrapheap challenge. You can start down one route, but that closes others, even though the initial direction was random.
The A-Team would, of course, do the best with whatever they had available, but then they are Designers.

You'd think the bad guys would think to not always lock them up in a shed with a bunch of machinery and fertiliser.
jimbob wrote:
Thu Apr 22, 2021 3:03 pm
Certainly describing evolution as "the opposite of random" is more than downplaying the role of chance events - it's denying it. Dinosaurs were top predators on land - why did mammals and reptiles replace them? It wasn't anything internal to the selective landscape in the thousands of years before most of them became extinct.
Well, the selective landscape abruptly changed. You can call that a random event but they way evolution responded to it, such that smaller warm-blooded animals coped better with it than big terrible lizards, is to a broad degree prepost-dictable.
molto tricky

User avatar
Bird on a Fire
Light of Blast
Posts: 5736
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: Randomness in evolution / natural selection

Post by Bird on a Fire » Thu Apr 22, 2021 7:54 pm

Randomness is a key component of a lot of efficient search algorithms, as I understand it. EG the MCMC samplers for Bayesian analysis often/always (?) involve random jumps - followed by an assessment of "fitness" e.g. by likelihood.

I think both bits are important, and using different perspectives can be helpful for solving different questions (or parts of them). Dawkins himself describes the "selfish gene" perspective as being like a Necker Cube (in The Extended Phenotype, I think, which is his best book), which I guess is a meta-analogy.
He has the grace of a swan, the wisdom of an owl, and the eye of an eagle—ladies and gentlemen, this man is for the birds!

User avatar
bob sterman
Snowbonk
Posts: 474
Joined: Mon Nov 11, 2019 10:25 pm

Re: Randomness in evolution / natural selection

Post by bob sterman » Thu Apr 22, 2021 8:30 pm

Bird on a Fire wrote:
Thu Apr 22, 2021 11:16 am
JMS is pretty old school - born over 100 years ago!
Yes - but where would we be without the 20th century "fab four" of gene-centered evolutionary theory? JMS, Hamilton, Williams and Trivers?

Mixing my pop analogies slightly - but the surviving member (Trivers) would certainly be the Keith Richards of the group.

User avatar
shpalman
Stummy Beige
Posts: 4180
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: Randomness in evolution / natural selection

Post by shpalman » Thu Apr 22, 2021 8:31 pm

Bird on a Fire wrote:
Thu Apr 22, 2021 7:54 pm
Randomness is a key component of a lot of efficient search algorithms, as I understand it. EG the MCMC samplers for Bayesian analysis often/always (?) involve random jumps - followed by an assessment of "fitness" e.g. by likelihood.

I think both bits are important, and using different perspectives can be helpful for solving different questions (or parts of them). Dawkins himself describes the "selfish gene" perspective as being like a Necker Cube (in The Extended Phenotype, I think, which is his best book), which I guess is a meta-analogy.
Or, for example, "simulated annealing" which allows the trial solutions to jump out of local minima as if they were being thermally excited, and then you gradually reduce the "temperature".
molto tricky

User avatar
jimbob
After Pie
Posts: 2469
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: Randomness in evolution / natural selection

Post by jimbob » Thu Apr 22, 2021 9:22 pm

shpalman wrote:
Thu Apr 22, 2021 7:49 pm
jimbob wrote:
Thu Apr 22, 2021 3:03 pm
Bird on a Fire wrote:
Thu Apr 22, 2021 11:12 am


Yes, exactly, plus most adaptive changes need more than one mutation, as they result from interacting networks of multiple genes. This means that the "next step" is massively contingent on what's gone before and what's available within the genome to work with (retroviruses etc. aside). The old-fashioned adaptationist narrative was that natural selection optimised between various constraints, but generally ignored the fact that the available options are highly lineage-specific, and therefore often biogeographically restricted, etc etc.

Natural selection is more like when the A Team jerry-rig an aircraft out of whatever crap is in the shed they're locked in, than a precision-engineered Boeing factory.
The bolded bit is why I think of it as random. The initial organism's descendants will have mutations that predispose them towards certain niches and make others less available. As long as there are descendants of lions, there would not be any opportunity for the descendants of cheetahs to become more generalist pack hunters, because they'd be outcompeted by the lions' descendants.

It's a drunkard's walk where one can explain how one got there, but unless one is talking about specific traits (antibiotic resistance being a classic one) what would happen is dependent on multiple options so is often not predictable (nor, I'd argue, predetermined).
If there's selection pressure then it's a drunkard's walk if the drunkard is on a hill, or can smell a pub or something.
jimbob wrote:
Thu Apr 22, 2021 3:03 pm
It's like the A-team lashing something up, and sometimes they have a bulldozer, whilst other times they have a quadbike and this is subject to the previous story. Probably more like scrapheap challenge. You can start down one route, but that closes others, even though the initial direction was random.
The A-Team would, of course, do the best with whatever they had available, but then they are Designers.

You'd think the bad guys would think to not always lock them up in a shed with a bunch of machinery and fertiliser.
jimbob wrote:
Thu Apr 22, 2021 3:03 pm
Certainly describing evolution as "the opposite of random" is more than downplaying the role of chance events - it's denying it. Dinosaurs were top predators on land - why did mammals and reptiles replace them? It wasn't anything internal to the selective landscape in the thousands of years before most of them became extinct.
Well, the selective landscape abruptly changed. You can call that a random event but they way evolution responded to it, such that smaller warm-blooded animals coped better with it than big terrible lizardswarm-blooded birds with teeth, is to a broad degree prepost-dictable.
If it hit a few hours earlier or later, it would have wiped out a completely different set of burrowing animals - probably, as you say large, warm-blooded land animals (so with a high-metabolic rate and need for food) would have been doomed, but the burrowing animals would have been quite different. Likewise if it hit at a different latitude.

The subsequent selective landscape could have looked quite different almost literally depending on which direction the wind was blowing (more so with the Toba eruption about 70k years ago).

You can't say that evolution is the opposite of random, when it's so heavily affected by both unpredictable and also events that are actually random. And it's even more significant now, because truly random events do affect human behaviour (a silly example, premium bonds relying on thermal noise) and these can impact whether ecosytems get protected.
Have you considered stupidity as an explanation

User avatar
shpalman
Stummy Beige
Posts: 4180
Joined: Mon Nov 11, 2019 12:53 pm
Location: One step beyond

Re: Randomness in evolution / natural selection

Post by shpalman » Thu Apr 22, 2021 9:24 pm

Well then, this post is also random, because all those random events led up to it happening.
molto tricky

User avatar
jimbob
After Pie
Posts: 2469
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: Randomness in evolution / natural selection

Post by jimbob » Thu Apr 22, 2021 9:24 pm

shpalman wrote:
Thu Apr 22, 2021 8:31 pm
Bird on a Fire wrote:
Thu Apr 22, 2021 7:54 pm
Randomness is a key component of a lot of efficient search algorithms, as I understand it. EG the MCMC samplers for Bayesian analysis often/always (?) involve random jumps - followed by an assessment of "fitness" e.g. by likelihood.

I think both bits are important, and using different perspectives can be helpful for solving different questions (or parts of them). Dawkins himself describes the "selfish gene" perspective as being like a Necker Cube (in The Extended Phenotype, I think, which is his best book), which I guess is a meta-analogy.
Or, for example, "simulated annealing" which allows the trial solutions to jump out of local minima as if they were being thermally excited, and then you gradually reduce the "temperature".
I would argue that the local minima are subject to change due to the interactions between organisms. It's probably fastest and clearest in the evolution of hosts and pathogens
Have you considered stupidity as an explanation

User avatar
jimbob
After Pie
Posts: 2469
Joined: Mon Nov 11, 2019 4:04 pm
Location: High Peak/Manchester

Re: Randomness in evolution / natural selection

Post by jimbob » Thu Apr 22, 2021 9:29 pm

shpalman wrote:
Thu Apr 22, 2021 9:24 pm
Well then, this post is also random, because all those random events led up to it happening.
I think that if your ecosystem is highly affected on whether an asteroid hit several hours earlier or later 65-million years ago, there's a significant random effect. For less-extreme effects, the Long Term Evolution Experiment Citrate metabolism result is interesting, as not all samples have evolved this.
Have you considered stupidity as an explanation

User avatar
Bird on a Fire
Light of Blast
Posts: 5736
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: Randomness in evolution / natural selection

Post by Bird on a Fire » Fri Apr 23, 2021 12:32 am

Ecosystem stability is massively environmentally dependent too. You can have a stable system that's totally transformed by a degree of average temperature, or a cm of rainfall. Walk up a mountain to confirm.
He has the grace of a swan, the wisdom of an owl, and the eye of an eagle—ladies and gentlemen, this man is for the birds!

User avatar
Bird on a Fire
Light of Blast
Posts: 5736
Joined: Fri Oct 11, 2019 5:05 pm
Location: nadir of brie

Re: Randomness in evolution / natural selection

Post by Bird on a Fire » Fri Apr 23, 2021 12:35 am

shpalman wrote:
Thu Apr 22, 2021 8:31 pm
Bird on a Fire wrote:
Thu Apr 22, 2021 7:54 pm
Randomness is a key component of a lot of efficient search algorithms, as I understand it. EG the MCMC samplers for Bayesian analysis often/always (?) involve random jumps - followed by an assessment of "fitness" e.g. by likelihood.

I think both bits are important, and using different perspectives can be helpful for solving different questions (or parts of them). Dawkins himself describes the "selfish gene" perspective as being like a Necker Cube (in The Extended Phenotype, I think, which is his best book), which I guess is a meta-analogy.
Or, for example, "simulated annealing" which allows the trial solutions to jump out of local minima as if they were being thermally excited, and then you gradually reduce the "temperature".
Ah yes! Thanks. I keep meaning to brush up on computer science algorithms, but it's that part of my "background reading" I find least interesting.

I've even written a simulated annealing algorithm once - they described the maths first and I thought it made sense, and then I discovered I was meant to be stimulating annealing.
He has the grace of a swan, the wisdom of an owl, and the eye of an eagle—ladies and gentlemen, this man is for the birds!

User avatar
dyqik
Stummy Beige
Posts: 3706
Joined: Wed Sep 25, 2019 4:19 pm
Location: Masshole
Contact:

Re: Randomness in evolution / natural selection

Post by dyqik » Fri Apr 23, 2021 1:44 am

shpalman wrote:
Thu Apr 22, 2021 7:49 pm

If there's selection pressure then it's a drunkard's walk if the drunkard is on a hill, or can smell a pub or something.
Or is looking for his keys under random lampposts, for cases where large parts of parameter space aren't viable.

Post Reply