PDA

View Full Version : Unpredicting Evolution



Selfsim
2013-Jan-09, 10:10 PM
A recent study into the predictability, (or otherwise), of Evolutionary features, in replicated lineages (ie: descendants), was recently published:

"Predictability of evolution depends nonmonotonically on population size" (http://www.pnas.org/content/110/2/571.abstract), Lenski et al, (PNAS Dec 24th, 2012).

A press release is:
Optimal population size allows maximum predictability of evolution (http://phys.org/news/2013-01-optimal-population-size-maximum-evolution.html)


The researchers combined genetic information from a fungus with computer models which simulate the course of evolution. They concluded that, while evolution in small populations is known to be unpredictable, this is also the case for very large populations. There is an optimal population size for predicting evolutionary outcomes in every situation.So what if the numbers of these separable populations is very large?
(Eg: the populations of LiS Forum believed 'aliens', in the observable universe?)

Well, let's look at the results

The best summary of their results, is given in the graph at the beginning of the article. They quantify 'predictability', to be the measure of the repeatability, of the path or outcome of evolution, in replicated lineages. (This turns out to be entropy measures on trajectories and endpoints).

It would seem from this study that unpredictability of Evolution increases for larger population sizes and this is ...
caused by the fact that they contain more individuals with not one but two or more mutations. Since the number of beneficial mutation combinations is much higher than the number of favourable mutations themselves, however, the predictability of evolution again declines in very large populations.

So, if we consider 'Earth-like' life to be a universally very large population across the observable universe, then predictability inside that population would asymptote to zero*. For this to follow however, clearly, one would have to assume that 'replicated lineage' exists across that population, but does the also speculated 'universality of abiogenesis', provide that linkage of 'replicated lineage'?
(In Panspermia, it would, I think(??) ).

Interesting ...

Footnote:
* From this study, that is .. There are others, of course .. which we should attempt to dig up during the course of this thread.

profloater
2013-Jan-09, 10:19 PM
Hey, no it says there is an optimum population size for predictability, that's all. and why on earth do they have to use the word nonmonotonically? And noncontiguous populations would not be covered.

Selfsim
2013-Jan-09, 10:46 PM
Hey, no it says there is an optimum population size for predictability, that's all.Which is the same as saying "there is an optimum population size for unpredictability" ... and it appears to apply to the large and small ones (look at the graph) .. which is relevant to the size of the observable universe, (albeit, admittedly, provided 'replicated lineages' is also 'in-common').

Somewhere along the line, if one goes with life being universal, some sub-phenomenon must be 'in common'. Self-replication descended from some type of process, would result in a 'replicated lineage', wouldn't it?

Paul Wally
2013-Jan-09, 10:50 PM
The best summary of their results, is given in the graph at the beginning of the article. They quantify 'predictability', to be the measure of the repeatability, of the path or outcome of evolution, in replicated lineages. (This turns out to be entropy measures on trajectories and endpoints).

Note that predictability/unpredictability is used as a measure and not as an absolute determination. In other words what they are saying is that evolution is less predictable with very small and very large populations and not absolutely unpredictable.

Selfsim
2013-Jan-09, 11:01 PM
Note that predictability/unpredictability is used as a measure and not as an absolute determination. In other words what they are saying is that evolution is less predictable with very small and very large populations and not absolutely unpredictable.
Agreed .. with the condition that the predictability/unpredictability is sensitive to scale … (ie: 'small', 'intermediate' and 'large').

Astronomical scales are clearly 'large'.
:)

transreality
2013-Jan-09, 11:32 PM
yes, but the population is necessarily contiguous, that is how the determination between small and large populations are made. A large population does not include a subset of small populations, since we are talking about genetic interconnectability. Basically with regard to actual breeding populations, a very large contiguous populations means theoretically a mutation rising in a single individual, will cross with any other individual in the populations potentially, with a chance of transmission to the next generation. The larger the population then the larger the target cross section for mutation. Since these are novel alterations, as the number of mutations a population hold increase, the less predictable is its trajectory. Consider then a very small population (of two) selected and isolated from a large population above. The larger the parent population the less chance to predict the genetic makeup of the child population. This is the heart of speciation. However the next generation of the small population is relatively predictable. This predictability in effect allows meaningful trajectories of evolution based on natural selection. As this population increases in size the selective forces of the environment become more diffuse, and the number of mutations entering the population increases, with drop in predictability.

Extrapolate this to a universal size, can say very little. Except that overall selection would be so diffuse and mutation distribution so random, that the chance of predicting the attributes of one alien 'sub-population' based on the attributes of another (say our local) 'sub-population', is effectively zero

Selfsim
2013-Jan-09, 11:56 PM
Extrapolate this to a universal size, can say very little. Except that overall selection would be so diffuse and mutation distribution so random, that the chance of predicting the attributes of one alien 'sub-population' based on the attributes of another (say our local) 'sub-population', is effectively zeroWell, I wouldn't say that was saying 'very little'.

When I see threads arguing for/against the recurrence of intelligence, (as a result of 'Evolutionary trajectories'), it seems reasonable to investigate just where those trajectories end up in a large population sharing replication of certain generalised, universal 'characteristics' in common. Ie: the idea of 'Intelligence' as a consequence of the assumption that Evolution, (and presumably, abiogenesis), might be universal, carries with it the consequences of that assumption. In this case, it would seem to be diminishing predictability, (ie: increased uncertainty), as the sample space increases.

This is of course, predicted on the assumption, (as you point out), that some mechanism supporting the concept of 'contiguousness', or 'replicated lineages' exists in common. Such a generalised characteristic, would seem to be inherent in the abiogenesis assumption too, otherwise how else does 'it do its thing', everywhere?

(Just food for thought about consistency … more than anything else, in particular).

profloater
2013-Jan-10, 12:36 AM
The issue of contact between populations cannot be ducked, there is no accepted way for non contacting populations to be considered as a group. I am also reminded of the Gaia hypotheis which starts out with a single species, like the study, but shows that as the number of different species increases the population dynamics get both complex and more stable. This factor I guess is much too complex for the predictability criterion being studied..

Selfsim
2013-Jan-10, 12:55 AM
The issue of contact between populations cannot be ducked, there is no accepted way for non contacting populations to be considered as a group. Yep .. I actually agree about the 'non acceptance' of the idea that a 'non-contacting' population could be part of a physical group …

But the issue is about which theoretically abstractable, generalised characteristics might be ubiquitous within an abstracted universal 'group'.

If the cloud we call 'abiogenesis' includes some abstracted principle which results in effectively 'replicated lineages' (of subsequent chemical behaviours), then the downside of that line of argument would be that the speculated size of the population, (of life which emerges from it), would be that making predictions about its characteristics, would be diminishingly predictable.

What is the basis then, of 'intelligence' being 'likely' in that population, (and all of its consequences .. (like civilisations, structures, artifacts, etc)? After all, that is a prediction of a particular type of characteristic isn't it?

transreality
2013-Jan-10, 09:51 PM
What is the basis then, of 'intelligence' being 'likely' in that population, (and all of its consequences .. (like civilisations, structures, artifacts, etc)? After all, that is a prediction of a particular type of characteristic isn't it?

I don't think so. We could predict that alien life would require hydrogen, carbon and oxygen, perhaps require water. Beyond we might posit a metabolism, a DNA functional-analogue, or membranes as basic requirements life but as to the actual specific arrangement of atoms and molecules in those structures, or even the attributes of those structures such as size, or complexity the number of possible pathways from the simple organic atoms and molecules to these massive integrated functional systems is likely to be so vast as to make these effectively unpredictable.

We know in our own system that certain attributes of macroscopic multicellular life are useful enough that they have had multiple independent origins. This is the case for wings, eyes, legs etc. Yet in just about all those cases the function performed by that structure can also be performed by completely unrelated structures. A bat can substitute echolocation for eyes, plants distribute can seeds by 'chute structures not wings, snakes can slither etc. These convergences happen along way from the original prebiotic chemistry, obviously there is no inherent tendency of C, H, and O to make wings, eyes, legs or any other structure. Intelligence, surely, is just another highly derived attribute (like flying, moving on land, detecting things nearby, etc), probably evolveable by any number of genetically unrelated mechanisms, and substituted for by others we would not recognise as intelligence, but which do the job.

Selfsim
2013-Jan-10, 11:34 PM
I don't think so. We could predict that alien life would require hydrogen, carbon and oxygen, perhaps require water. Beyond we might posit a metabolism, a DNA functional-analogue, or membranes as basic requirements life but as to the actual specific arrangement of atoms and molecules in those structures, or even the attributes of those structures such as size, or complexity the number of possible pathways from the simple organic atoms and molecules to these massive integrated functional systems is likely to be so vast as to make these effectively unpredictable.So, its all about varying the scales of view, to fit the predictability or unpredictability curve, in order to tailor the hypothesis towards predictability? That's Ok, mind you .. so long as it ends up being practically testable the OP was more about the consequences of a particular theoretical model with chosen macro temporal and spatial scales. If we step outside that model, then anything's possible (yet again) .. and is really a different discussion theme.

I might add that your above approach seems somewhat like 'reverse engineering' from the end result, in order to arrive at mechanism which suits the need for prediction(?)
(Mind you, that would just be my opinion :) .. nothing more )


We know in our own system that certain attributes of macroscopic multicellular life are useful enough that they have had multiple independent origins.There is no prior 'knowledge' of the end result, which would facilitate the selection of a particular characteristic, to suit a particular environment in an evolving species. This is a 'post-diction' technique, which we use in our models. That's not how we see the natural environment working, however.
This is the case for wings, eyes, legs etc. Yet in just about all those cases the function performed by that structure can also be performed by completely unrelated structures. A bat can substitute echolocation for eyes, plants distribute can seeds by 'chute structures not wings, snakes can slither etc. These convergences happen along way from the original prebiotic chemistry, obviously there is no inherent tendency of C, H, and O to make wings, eyes, legs or any other structure. Intelligence, surely, is just another highly derived attribute (like flying, moving on land, detecting things nearby, etc), probably evolveable by any number of genetically unrelated mechanisms, and substituted for by others we would not recognise as intelligence, but which do the job.Hmm .. its a bit like the computer analogy the code might be different, but Macs still end up making use of windows and vice versa (ie: OS-X vs Windows). But who says 'windows' is going to be the end (best) result?

The problem with this approach, is that we slip into thinking that the end result has a purpose or intent. It therefore becomes a goal, (or end-point), of some pre-destined trajectory which is specifically shown in the OP model/study to be not so that was the surprise (for them).

Not all animals share what we'd call 'intelligence' .. and yet we all share a common ancestry why not assume our particular flavour of intelligence was just what we happened to end up with, giving us the ability to observe it in ourselves?
(Ie: whatever happened to the Anthropic Principle (AP) ?)

I'm not saying you're speculation is not valid its just as valid as any other coming from an imposed philosophy .. like the AP.

The outcome of the study in the OP, empirically overturned our expectation of predictability in large (connected) populations. The simulation made the difference (as opposed to a philosophical position driving the creation of a hypotheses).

Paul Wally
2013-Jan-11, 12:28 AM
Astronomical scales are clearly 'large'.
:)

I don't understand. How do you manage to extrapolate a study done on a particular Earth life-form to astronomical scales? If this study/experiment was at all realistic there had to be causal interactions within the population, causal interactions between the population and other populations of other species, causal interactions between the population and the physical environment on Earth, and also the mutation rate applicable to Earth-life. There is no causal connection between a population on Earth and population on a distant planet, with a different abiogenesis, different other species, different environment and probably different mechanisms of heredity, not to mention mutation.

TooMany
2013-Jan-11, 01:09 AM
So, its all about varying the scales of view, to fit the predictability or unpredictability curve, in order to tailor the hypothesis towards predictability? That's Ok, mind you .. so long as it ends up being practically testable the OP was more about the consequences of a particular theoretical model with chosen macro temporal and spatial scales. If we step outside that model, then anything's possible (yet again) .. and is really a different discussion theme.

I might add that your above approach seems somewhat like 'reverse engineering' from the end result, in order to arrive at mechanism which suits the need for prediction(?)
(Mind you, that would just be my opinion :) .. nothing more )

There is no prior 'knowledge' of the end result, which would facilitate the selection of a particular characteristic, to suit a particular environment in an evolving species. This is a 'post-diction' technique, which we use in our models. That's not how we see the natural environment working, however. Hmm .. its a bit like the computer analogy the code might be different, but Macs still end up making use of windows and vice versa (ie: OS-X vs Windows). But who says 'windows' is going to be the end (best) result?

The problem with this approach, is that we slip into thinking that the end result has a purpose or intent. It therefore becomes a goal, (or end-point), of some pre-destined trajectory which is specifically shown in the OP model/study to be not so that was the surprise (for them).


It doesn't matter that at any time evolution is a divergent process (unpredictable). What matters more in the long run are the opportunities for survival that are available. Eyes have evolved many times, but we don't understand that as a purpose or intent at all. Rather it is the existence of a benefit, derived from the development of a sense organ that exploits light, that guides evolution toward eyes.



Not all animals share what we'd call 'intelligence' .. and yet we all share a common ancestry why not assume our particular flavour of intelligence was just what we happened to end up with, giving us the ability to observe it in ourselves?
(Ie: whatever happened to the Anthropic Principle (AP) ?)

I'm not saying you're speculation is not valid its just as valid as any other coming from an imposed philosophy .. like the AP.

The outcome of the study in the OP, empirically overturned our expectation of predictability in large (connected) populations. The simulation made the difference (as opposed to a philosophical position driving the creation of a hypotheses).

It does not and cannot overturn the fact that evolution exploits available paths that provide benefits. The way in which it does this is unpredictable, but it eventually does. In the case of brains, we know (or at least strongly believe) that the brains of different lines have developed independently over time. How much it develops depends on the value of the brain to survival. We see a number of "smart" mammal species (e.g. dolphins, bears, dogs, primates).

What is needed, for intelligence to reach the level it has in humans, is a benefit of that intelligence. That benefit is tool making (and society). This is why the most intelligent animals descended from the apes. The apes developed dexterous hands, good binocular and color vision, setting the stage for the benefits of tool use to drive evolution toward additional improvements in the powers of abstract reasoning. Ultimately that advantage of tool use is so great that we dominate the earth.

On some other planet the path to intelligence could be very different. But the advantage of intelligence exists independently of the random process of evolution and that's the reason that intelligence emerges.

Selfsim
2013-Jan-11, 11:59 PM
It doesn't matter that at any time evolution is a divergent process (unpredictable). What matters more in the long run are the opportunities for survival that are available. Eyes have evolved many times, but we don't understand that as a purpose or intent at all. Rather it is the existence of a benefit, derived from the development of a sense organ that exploits light, that guides evolution toward eyes.I don't think I have a problem with any of that .. but whether eyes are ubiquitous throughout an observable universe, would have to be 'unknown'. Until we have some evidence that eyes can develop elsewhere, we can't predict accurately from the theory alone. The uncertainties outweigh the precision of such a prediction.

It does not and cannot overturn the fact that evolution exploits available paths that provide benefits. The way in which it does this is unpredictable, but it eventually does. In the case of brains, we know (or at least strongly believe) that the brains of different lines have developed independently over time. How much it develops depends on the value of the brain to survival. We see a number of "smart" mammal species (e.g. dolphins, bears, dogs, primates). How can the origin of brains, across multiple species, be decoupled from the Last Common Ancestor's inherited genes?


What is needed, for intelligence to reach the level it has in humans, is a benefit of that intelligence. That benefit is tool making (and society). This is why the most intelligent animals descended from the apes. The apes developed dexterous hands, good binocular and color vision, setting the stage for the benefits of tool use to drive evolution toward additional improvements in the powers of abstract reasoning. Ultimately that advantage of tool use is so great that we dominate the earth.This is an hypothesis.
There are others such as:
- allows more effective foraging, especially for learning and remembering where and when fruits ripen;
- allows more successful functioning in the complex social groups that primates form. This could be intertwined with the development of language.
(The latter appears to be more favoured in Evolutionary Biology circles).

Either way, this rationale is still a hypothesis .. and would surely remain to be tested?

On some other planet the path to intelligence could be very different. But the advantage of intelligence exists independently of the random process of evolution and that's the reason that intelligence emerges.I still don't see how it can be decoupled from the common LUCA genetics?

Selfsim
2013-Jan-12, 12:11 AM
I don't understand. How do you manage to extrapolate a study done on a particular Earth life-form to astronomical scales? If this study/experiment was at all realistic there had to be causal interactions within the population, causal interactions between the population and other populations of other species, causal interactions between the population and the physical environment on Earth, and also the mutation rate applicable to Earth-life. There is no causal connection between a population on Earth and population on a distant planet, with a different abiogenesis, different other species, different environment and probably different mechanisms of heredity, not to mention mutation.Tee hee ... one has to use one's imagination and abstract some mechanism which constitutes 'inherited lineage'. I know its a stretch but abiogenesis might be speculated to be that mechanism(??)
Anway, the argument which seems to make intelligence ubquitous throughout the speculated 'alien' population must make the same 'leaps of faith' ... (see TooMany's lines of argument).

All I'm trying to point out is that if one invokes a linkage by suggesting things like eyes and intelligent brains might be inevitable outcomes of Evolution throughout our speculated 'alien' population, then one must also accept the unpredictability aspect, as both invoke the same, (albeit abstracted, generalised) 'common lineage' feature in the model, resulting in commonality of a feature, we so far only know, exists on Earth(?)

TooMany
2013-Jan-12, 12:55 AM
I don't think I have a problem with any of that .. but whether eyes are ubiquitous throughout an observable universe, would have to be 'unknown'. Until we have some evidence that eyes can develop elsewhere, we can't predict accurately from the theory alone. The uncertainties outweigh the precision of such a prediction.


You are still missing the point. The opportunity for eyes is available where there is light. By chance some light sensitivity of some sort develops in an organism and gives it an advantage, e.g. in seeking food or escaping predators. While the process of evolution at each step is purely accidental, those accidents are selected by what brings advantage in the environment. Thus the evolution of eyes should not be understood as some random accident or unpredictable outcome.

It is common in evolution for evolved things to atrophy (become vestigial) or be lost altogether by a line because it's environment is changed. Thus you may in the deep have forms in which eyes have disappeared. Some types of bat's are blind because they have moved into life style where eyes are not useful.

There is every reason to believe that life evolving on another planet, similar to earth, would also yield many species with eyes. We cannot predict exactly what those alien species we be like, but we can assert with confidence that eyes will develop.



How can the origin of brains, across multiple species, be decoupled from the Last Common Ancestor's inherited genes?


Are you trying to say that there is a single origin of nervous systems? I'm not sure whether that is true or not, but what is certain is that in each branch that subsequently developed, the nervous system also developed independently. There is not a single thread of nervous system development, there are many branches. For example the brains of bats and dolphins have likely evolved organizations of nerves to handle echo-location.

There are many other coordinating/signaling systems in biology. For example the complex hormone system in humans. Bacteria have senses and transmit information both internally and externally to the colony. The organization of nerves in a jellyfish is entirely different from that in chordates. I would guess that electric/synapse nerves have been most effective because of the response time.



This is an hypothesis.
There are others such as:
- allows more effective foraging, especially for learning and remembering where and when fruits ripen;
- allows more successful functioning in the complex social groups that primates form. This could be intertwined with the development of language.
(The latter appears to be more favoured in Evolutionary Biology circles).


Yes all of that is true too. I picked on tool making because it has led to such a huge advantage. Dogs have complex social groups and chimps remember where and when food ripens.



Either way, this rationale is still a hypothesis .. and would surely remain to be tested?

Nothing is proven 100% to absolutely certainty until it is observed. A very big surprise would be that evolution only works on Earth. No scientists would even consider such a negative hypothesis because it doesn't make sense. But still, it's true that it has not be proven.



I still don't see how it can be decoupled from the common LUCA genetics?


Who's trying to decouple it and why? Some sort of system encoding instructions is itself an example of an extremely basic advantage in propagation of any life form. We do not yet know whether the emergence of RNA is some preferred chemical path or whether other configurations are also workable. If you consider the very definition of life then it seems clear that the development of an information system that assures more or less exact replication is a huge advantage over some process that is less sophisticated or more haphazard in it's means of self-reproduction.

Selfsim
2013-Jan-12, 02:18 AM
The opportunity for eyes is available where there is light.Or .. the opportunity for eyes was always there, encoded into the very first self-replicating molecules, (or environs, etc), which was onpassed with each subsequent generation. The environment and natural selection explains its emergence across the entire population .. which still shares the descendant results of that initial encoding.
Thus the evolution of eyes should not be understood as some random accident or unpredictable outcome.I would agree that it most definitely couldn't be explained as a purely random 'accident'. (That's not how Evolution works anyway .. we both agree about that, I think).

It is common in evolution for evolved things to atrophy (become vestigial) or be lost altogether by a line because it's environment is changed. Thus you may in the deep have forms in which eyes have disappeared. Some types of bat's are blind because they have moved into life style where eyes are not useful.And some things which disappear, reappear again when the habitat of that particular species reverts back to what seems to have been ithe original habitat

The information is never 'lost'. It simply doesn't get expressed, (or it turns up somewhere else) the information persists in some way, shape or form, however. (Which is why I'm saying that the view that the LUCA possessed all of this 'potential', implies that Evolution has merely shaped what might have been there all along. I don't think we can decouple that initial information from Evolution, until we find something elsewhere with say, 'eyes' or an 'intelligent brain'. (Or make some other abiogenesis related discovery).

That light is ubiquitous, might be necessary for Evolution/Nat Selection to do its thing, and to give some particular eyed species an advantage on light-bathed objects, but it isn't sufficient (I'm pretty sure we can't grow an eye by shining light on a petrie dish for a few hundred million years).

There is every reason to believe that life evolving on another planet, similar to earth, would also yield many species with eyes. We cannot predict exactly what those alien species we be like, but we can assert with confidence that eyes will develop.The confidence with which you assert this, is entirely dependent on evidence from Earth's own Theory of Evolution, supported by Earth's own fossil and genetic record, given that we all share a common ancestor, from which everything subsequently descended.

Are you trying to say that there is a single origin of nervous systems? I'm not sure whether that is true or not, but what is certain is that in each branch that subsequently developed, the nervous system also developed independently. There is not a single thread of nervous system development, there are many branches. For example the brains of bats and dolphins have likely evolved organizations of nerves to handle echo-location.

There are many other coordinating/signaling systems in biology. For example the complex hormone system in humans. Bacteria have senses and transmit information both internally and externally to the colony. The organization of nerves in a jellyfish is entirely different from that in chordates. I would guess that electric/synapse nerves have been most effective because of the response time.Err yep!

Nothing is proven 100% to absolutely certainty until it is observed. A very big surprise would be that evolution only works on Earth. No scientists would even consider such a negative hypothesis because it doesn't make sense. But still, it's true that it has not be proven.Its not so much about 'proof' its about the nature of the supporting evidence and in this case, its all come from, Earth, (as far as we think we know, that is) ..
Oh and 'surprises'? .. well .. is there some rule that says they're outlawed?

The point is that what you call 'negative hypotheses', are equally valid, ... until there's evidence to the contrary ...(they're just the 'flip-side' of the same coin). It just seems that we've become so focused on looking for what we want to believe, there's always a surprise in store when the 'negative hypothesis' demonstrates its equal validity.

Who's trying to decouple it and why? It has to be decoupled in order to support the assertion of universality, (outside of our LUCA lineage). We can't just say that Evolution will inevitably result in intelligent brains, eyes, noses, tool-using hands, etc, outside of that group .. but we can use it to develop tests for a hypothesis of that. Those tests, (executed beyond that LUCA lineage), haven't returned results yet.

Theoretical modelling has shown us that unpredictability of Evolution measures, (entropy based), across large populations sharing common inherited lineage, increase with the size of that population.

If you consider the very definition of life then it seems clear that the development of an information system that assures more or less exact replication is a huge advantage over some process that is less sophisticated or more haphazard in it's means of self-reproduction.We don't know that yet. There may be other ways around the 'need' for self replication .. particularly if initial abundances or longevities, (for eg), are significant evolutionary factors.

Paul Wally
2013-Jan-12, 12:58 PM
Tee hee ... one has to use one's imagination and abstract some mechanism which constitutes 'inherited lineage'. I know its a stretch but abiogenesis might be speculated to be that mechanism(??)
Anway, the argument which seems to make intelligence ubquitous throughout the speculated 'alien' population must make the same 'leaps of faith' ... (see TooMany's lines of argument).

All I'm trying to point out is that if one invokes a linkage by suggesting things like eyes and intelligent brains might be inevitable outcomes of Evolution throughout our speculated 'alien' population, then one must also accept the unpredictability aspect, as both invoke the same, (albeit abstracted, generalised) 'common lineage' feature in the model, resulting in commonality of a feature, we so far only know, exists on Earth(?)

What bothers me was your use of the term "astronomical scales". I don't see where that fits into our discussion. An alien population will be a different population with a different independent inherited lineage from an Earth species and it will also be a different lineage from other populations on other planets. So any population anywhere will be limited by resources wherever they are.

We all know evolution is unpredictable. The question is how unpredictable. Mutations are incremental and this leads to evolution being a largely continuous process. The kinds of discontinuities necessary for complete unpredictability is simply not evident. If evolution was completely unpredictable it wouldn't even be possible for us to distinguish any lineages within the fossil records. It wouldn't for instance be possible for us to look for missing evolutionary links.

transreality
2013-Jan-13, 10:52 PM
Or .. the opportunity for eyes was always there, encoded into the very first self-replicating molecules, (or environs, etc), which was onpassed with each subsequent generation. The environment and natural selection explains its emergence across the entire population .. which still shares the descendant results of that initial encoding....

And some things which disappear, reappear again when the habitat of that particular species reverts back to what seems to have been ithe original habitat …

The information is never 'lost'. It simply doesn't get expressed, (or it turns up somewhere else) … the information persists in some way, shape or form, however. (Which is why I'm saying that the view that the LUCA possessed all of this 'potential', implies that Evolution has merely shaped what might have been there all along. I don't think we can decouple that initial information from Evolution, until we find something elsewhere with say, 'eyes' or an 'intelligent brain'. (Or make some other abiogenesis related discovery).

That light is ubiquitous, might be necessary for Evolution/Nat Selection to do its thing, and to give some particular eyed species an advantage on light-bathed objects, but it isn't sufficient … (I'm pretty sure we can't grow an eye by shining light on a petrie dish for a few hundred million years).



evolution simply does not work that way.

The LUCA was necessarily a very simple organism. Above, did you not notice I was attempting to differentiate between the functions and the particular structures that might evolve to perform those functions. Yet here you seem to be claiming that LUCA's genome possesses the codes for particular structures, long before the function performed by that ever becomes relevant to the organism.

A function is to detect light that is a common factor in the organisms environment. So we may expect that function to arise again and again. One solution is the eye. One type of eye is the compound eye of insects, another type of eye is the camera eye of cephalopods, another type of eye is the different camera eye of vertebrates. Long after LUCA, the early of ancestors of each of these lineages lacked eyes; LUCA lack eyes. The structures of each of these lineages are genetically unrelated, they use different structures, different mechanisms, different genes. Somehow, really, do you think that little archaean LUCA, that is not only newly experimenting with replication, but also possesses the 'codes' for at least three different eye types hardwired in its molecules...That would be absurd.

The mechanisms for how a genome increases complexity over time, how it replicates genes, then coopts those copies as the basis for novel genes due to natural selection, over generations, and over species, how genes with weak selection for their products decline into functional junk are reasonably well-known. Yes, 'information' (in some meaningful evolutionary sense) can be lost, and can be newly created by the process of evolution.

If a gene is no longer required it is not put into an archive until required again. It remains in circulation, accumulating mutations like all DNA, if selection is weak those mutations will eventually prevent its function, it turns to junk. If that function is required again by some distant ancestor, too bad, the junk cannot be repaired since it can no longer produce a functional product, to which selection can be applied. Instead another active gene or a mutationally produced copy of an active gene may produce a product that can evolve towards that function, and if it does there is absolutely no requirement except any inherited legacy, for that product to have structural similarity to the product produced by the distant ancestor, for the same function.

Evolution is not constrained to a subset of possibilities derived from the 'potentialities' of LUCA. Evolution can produce novel structures in response to selection for functions that have not been encountered by the ancestral organisms.

Since these functions result from the ecological response to the environment, and many aspects of the evironment do not directly stem from the biology, but from universal aspects such as light, gravity, time, we can expect those functions to be applicable variously across the universe, but certainly we cannot predict (or even constrain the possibilities of) the specific structures evolved by any of those organisms in response to that functional requirement.

Selfsim
2013-Jan-13, 11:44 PM
evolution simply does not work that way.

The LUCA was necessarily a very simple organism. Above, did you not notice I was attempting to differentiate between the functions and the particular structures that might evolve to perform those functions. Yet here you seem to be claiming that LUCA's genome possesses the codes for particular structures, long before the function performed by that ever becomes relevant to the organism. Well I did also say "environs etc" .. meaning that information has a tricky way of increasing (or decreasing) in an organism, when external factors are taken into account .. (Ie: somewhat more specifically, I concur with the consensus view that evidence points to informational content of the LUCA being less overall, than what we presently see across all species … and that species 'information' can go up and down with subsequent generations .. and that that means not much in particular, and that this is not particularly surprising ..).


A function is to detect light that is a common factor in the organisms environment. So we may expect that function to arise again and again. One solution is the eye. One type of eye is the compound eye of insects, another type of eye is the camera eye of cephalopods, another type of eye is the different camera eye of vertebrates. Long after LUCA, the early of ancestors of each of these lineages lacked eyes; LUCA lack eyes. The structures of each of these lineages are genetically unrelated, they use different structures, different mechanisms, different genes. Somehow, really, do you think that little archaean LUCA, that is not only newly experimenting with replication, but also possesses the 'codes' for at least three different eye types hardwired in its molecules...That would be absurd.I agree.
And I think that there may also be other ways which light perception could develop which doesn't necessarily result in what we presently define as being 'eyes'.


The mechanisms for how a genome increases complexity over time, how it replicates genes, then coopts those copies as the basis for novel genes due to natural selection, over generations, and over species, how genes with weak selection for their products decline into functional junk are reasonably well-known. Yes, 'information' (in some meaningful evolutionary sense) can be lost, and can be newly created by the process of evolution.

If a gene is no longer required it is not put into an archive until required again. It remains in circulation, accumulating mutations like all DNA, if selection is weak those mutations will eventually prevent its function, it turns to junk. If that function is required again by some distant ancestor, too bad, the junk cannot be repaired since it can no longer produce a functional product, to which selection can be applied. Instead another active gene or a mutationally produced copy of an active gene may produce a product that can evolve towards that function, and if it does there is absolutely no requirement except any inherited legacy, for that product to have structural similarity to the product produced by the distant ancestor, for the same function.

Evolution is not constrained to a subset of possibilities derived from the 'potentialities' of LUCA. Evolution can produce novel structures in response to selection for functions that have not been encountered by the ancestral organisms. Then it would seem that something more fundamental than inherited genetics is at play .. I propose its fundamental thermodynamics .. which allows for a vast array of other possibilities (permutations and constraints), beyond specific inheritance via genetics shaped by specifically, a competition for resources.


Since these functions result from the ecological response to the environment, and many aspects of the evironment do not directly stem from the biology, but from universal aspects such as light, gravity, time, we can expect those functions to be applicable variously across the universe, but certainly we cannot predict (or even constrain the possibilities of) the specific structures evolved by any of those organisms in response to that functional requirement.Agreed.
I think we've ended up in roughly the same place.
:)

TooMany
2013-Jan-14, 12:37 AM
evolution simply does not work that way.

The LUCA was necessarily a very simple organism. Above, did you not notice I was attempting to differentiate between the functions and the particular structures that might evolve to perform those functions. Yet here you seem to be claiming that LUCA's genome possesses the codes for particular structures, long before the function performed by that ever becomes relevant to the organism.

A function is to detect light that is a common factor in the organisms environment. So we may expect that function to arise again and again. One solution is the eye. One type of eye is the compound eye of insects, another type of eye is the camera eye of cephalopods, another type of eye is the different camera eye of vertebrates. Long after LUCA, the early of ancestors of each of these lineages lacked eyes; LUCA lack eyes. The structures of each of these lineages are genetically unrelated, they use different structures, different mechanisms, different genes. Somehow, really, do you think that little archaean LUCA, that is not only newly experimenting with replication, but also possesses the 'codes' for at least three different eye types hardwired in its molecules...That would be absurd.

The mechanisms for how a genome increases complexity over time, how it replicates genes, then coopts those copies as the basis for novel genes due to natural selection, over generations, and over species, how genes with weak selection for their products decline into functional junk are reasonably well-known. Yes, 'information' (in some meaningful evolutionary sense) can be lost, and can be newly created by the process of evolution.

If a gene is no longer required it is not put into an archive until required again. It remains in circulation, accumulating mutations like all DNA, if selection is weak those mutations will eventually prevent its function, it turns to junk. If that function is required again by some distant ancestor, too bad, the junk cannot be repaired since it can no longer produce a functional product, to which selection can be applied. Instead another active gene or a mutationally produced copy of an active gene may produce a product that can evolve towards that function, and if it does there is absolutely no requirement except any inherited legacy, for that product to have structural similarity to the product produced by the distant ancestor, for the same function.

Evolution is not constrained to a subset of possibilities derived from the 'potentialities' of LUCA. Evolution can produce novel structures in response to selection for functions that have not been encountered by the ancestral organisms.

Since these functions result from the ecological response to the environment, and many aspects of the evironment do not directly stem from the biology, but from universal aspects such as light, gravity, time, we can expect those functions to be applicable variously across the universe, but certainly we cannot predict (or even constrain the possibilities of) the specific structures evolved by any of those organisms in response to that functional requirement.

You explained it very well. Thanks, I was getting tired from trying and so didn't respond to that post.

profloater
2013-Jan-14, 09:58 AM
This business of predictability of evolution is really a parlour game for maths modellers because we cannot rerun the experiment. It is also different from the questions about how life began,..... blob of fat in a pond, I read this week (New Scientist).

Selfsim
2013-Jan-14, 08:56 PM
This business of predictability of evolution is really a parlour game for maths modellers because we cannot rerun the experiment. It is also different from the questions about how life began,.....The question might be different, but there is clearly a causal relationship between Abiogenesis and Evolution.

One could say that abiogenesis caused (or initiated) Evolution (or at least, that's how science has defined it).

transreality
2013-Jan-14, 10:42 PM
The question might be different, but there is clearly a causal relationship between Abiogenesis and Evolution.

One could say that abiogenesis caused (or initiated) Evolution (or at least, that's how science has defined it).

Or it could be the other way round...

TooMany
2013-Jan-15, 07:12 PM
Or it could be the other way round...

And most likely is. Otherwise it's quite hard to explain how anything as complicated as a bacterium could exist. Something that metabolizes and rebuilds copies of itself using a coding system is already highly-evolved. Evolution must apply in some form right from the start, otherwise what we see now would be an impossible accident.

Concerning your other comment, I agree that the viability of life, even in extremes changes in environment, is an established fact. Life is not a fragile thing snuffed out by the slightest changes in conditions. I suspect that abiogenesis shares this robustness.

transreality
2013-Jan-15, 10:45 PM
And most likely is. Otherwise it's quite hard to explain how anything as complicated as a bacterium could exist. Something that metabolizes and rebuilds copies of itself using a coding system is already highly-evolved. Evolution must apply in some form right from the start, otherwise what we see now would be an impossible accident.

This is what I am thinking. By the time abiogenesis has occurred a series of complicated systems must be in place, for metabolism and for replication. The first living cell already has a membrane, DNA, RNA transciption, Ribosome factories for protein synthesis, Glucose metabolism, ion pumps that maintain the interior of the cell relative to the exterior. Can these individual systems function by themselves, can they combine perhaps through parasitism and endosymbiosis. Perhaps the appearance of globular membranes can form of themselves environments into which these other component systems can evolve, driven not by the requirement to keep a composite creature alive, but in their own competition for a serve of the primordial soup.

Selfsim
2013-Jan-15, 11:05 PM
Or it could be the other way round...I think Evolution is the biological consequence of Thermodynamics Law constraining bio-complexity, over time.

(Darwin's work preceeded Boltzmann's, also, didn't it? That might explain the perception that Evolution might be more fundamental ... Darwin was not aware of Thermodynamics Law, because that came later with Boltzmann).

So, if Abiogenesis defines the initial phases of Evolution then, its not so surprising that we'd see the (undefined) Abiogenesis process, as also being a direct consequence of Thermodynamics Law, constraining organic chemistry over time. As molecular complexity increases, Thermodynamics Law characteristics becomes more subtle, the non-linear interactions cause structure to appear, and so we then notice the similarity amongst those structures and categorise 'species'. This then is used in explaining those similar observations, but only as they apply to biology. This would also explain why Evolution happens to be well modelled by non-linear algorithms. Its still thermodynamics physics that's at its fundamental core though.

And there's still a difference between Evolution and Abiogenesis. The dominating aspects in the initial phases of Abiogenesis would 'most likely'* be along the principles of non-equilibrium thermodynamics (http://en.wikipedia.org/wiki/Non-equilibrium_thermodynamics) because:

Most systems found in nature are not in thermodynamic equilibrium; for they are changing or can be triggered to change over time, and are continuously and discontinuously subject to flux of matter and energy to and from other systems and to chemical reactions. Non-equilibrium thermodynamics is concerned with transport processes and with the rates of chemical reactions.... and further ...
Many natural systems still today remain beyond the scope of currently known macroscopic thermodynamic methods.
Those 'methods' are the methods I've been describing in the other thread on 'Chaos Theory and Life Outcome' , (amongst others):

The thermodynamic study of non-equilibrium systems requires more general concepts than are dealt with by equilibrium thermodynamics. One fundamental difference between equilibrium thermodynamics and non-equilibrium thermodynamics lies in the behaviour of inhomogeneous systems, which require for their study knowledge of rates of reaction which are not considered in equilibrium thermodynamics of homogeneous systems... The latter of which, we are all more familiar with, because that's what we've been educated in ...

So, Paul Wally's comment about Earth not being a thermodynamically isolated system, would be consistent with this reasoning, because non-equilibirum behaviours are a consequence of open systems. But we are talking about Abiogenesis and Evolution phenomenon with the full weight of evidence coming solely from only Earth's known past if we were to speak of the broader universe, the 'weight' of this evidence diminishes with considerations of the observational evidence of increasing diversity amongst the increasing set of diversity of conditions, and an increasing number of objects in the population set.

This is not to say that Earth has to always be modelled as an open system though. Take for example a query wishing to examine 'life' distributions across the entirety of the obs. universe. In this case, we are surely compelled to consider Earth, (and its life), as a sub-component of that model, in which it is largely thermodynamically isolated from others? Well, if we wish to include statements about the structures resulting from those non-linearities, and draw generalisations about them, we might then be forced into having to include the relationships between Earth and the Sun, and then say that's an isolated sub-system within the obs. universe model. Doing that, then forces us to again, into having to confront the non-linearity behaviours resulting from the interactions of Earth and Sun and thereby, we must also assume these to be applicable elsewhere (in a Copernican, volume limited Universe) .. one can't have one, without having the other and so all generalised system behaviours, (ie: oscillating, under-damped, over-damped, critical, far from equilibrium, etc .. then come back into consideration again). The specifics however, carry less 'weighting' as these are offset by the increasing diversities.

Its all about consistency in the model under discussion. We can't just pick certain generalised characteristics, whilst ignoring others, (which we know to represent the majority of systems we see both on Earth, and elsewhere in the universe). The empirical evidence drawn from Earth however, diminishes in 'weight' because of the corresponding increase in weight of empirical 'unknowns'. (Basically uncertainty increases, as the volume increase).

If our models tell us that certain traits across large populations, become unpredictable due to an ever-increasing level of interactions between them, (which is what genetic inheritance gives us), then the same phenomenon persists, (perhaps in a more virtualised manner than specifically, genetic inheritance and it comes from the Abiogenesis phase of Evolution, which we reasonably posit, exhibits the characteristics described in the Laws of Thermodynamics). Non-homogeneity, is just as much a part of a volume limited universe, as homogeneity is, and so is non-linear behaviours, caused by complex interactions. In fact, we suspect non-linearity outweighs linearity, based on our own experience to date.

Footnote:
* I dislike using the term; 'most likely' ... but this seems to be acceptable practice in this forum, so I use the term as an appeal to 'standard LiS Forum practices'. (At least I'm attempting to justify 'why' I'm asserting the 'most likely' catch-cry.

TooMany
2013-Jan-15, 11:52 PM
Footnote:
* I dislike using the term; 'most likely' ... but this seems to be acceptable practice in this forum, so I use the term as an appeal to 'standard LiS Forum practices'. (At least I'm attempting to justify 'why' I'm asserting the 'most likely' catch-cry.

The concept of 'most likely' is fundamental to the way nervous systems work. This is quite different from most computer systems and far more powerful. Whether you want to admit it or not, you use 'most-likely' every waking moment of your life. Just try demanding certainly before you do anything. You will never make it to work in the morning. Rigorous mathematics and science are not the same. We've discussed that before I believe.

Selfsim
2013-Jan-16, 12:25 AM
The concept of 'most likely' is fundamental to the way nervous systems work. This is quite different from most computer systems and far more powerful. Whether you want to admit it or not, you use 'most-likely' every waking moment of your life. Just try demanding certainly before you do anything. You will never make it to work in the morning. Rigorous mathematics and science are not the same. We've discussed that before I believe.The point was to at least, recognise and acknowledge that the argument presented, was dependent on perception from a particular chosen perspective .. as opposed to incontrovertible fact. I also recognise that the perspectives I'm coming from change .. but the reasons underlying those perspectives, hasn't it doesn't 'need' to yet.

Your use of the term 'powerful' contains such undistinguished perceptions, (for eg), and actually emphasises their importance for you .. one therefore doesn't necessarily have to accept the following argument, as these perceptions may not necessarily be of importance to others, or the generality of all situations. We can make choices as to whether or not, to follow arguments based on individual 'intuition'.

When it comes to accepting something as: 'as close to reality' as our models can take us, intuition is also not mandatory, because we'll never really get from intuition, what reality is, anyway.

TooMany
2013-Jan-18, 07:18 PM
The point was to at least, recognise and acknowledge that the argument presented, was dependent on perception from a particular chosen perspective .. as opposed to incontrovertible fact. I also recognise that the perspectives I'm coming from change .. but the reasons underlying those perspectives, hasn't … it doesn't 'need' to … yet.


With your emphasis on "incontrovertible fact" as what really matters, I'm puzzled that you pursue these speculations. Is it to tell us that it's speculative? I think we know that. That point is that there are reasonable speculations and less reasonable ones. That's what we are talking about, not incontrovertible fact.



Your use of the term 'powerful' contains such undistinguished perceptions, (for eg), and actually emphasises their importance for you .. one therefore doesn't necessarily have to accept the following argument, as these perceptions may not necessarily be of importance to others, or the generality of all situations. We can make choices as to whether or not, to follow arguments based on individual 'intuition'.


It's how we deal with the world around us and how we approach science. It's power lies not in my perception or bias but in the accomplishments that have arisen from our non-computer-like minds.

You are perfectly free to negate all speculation until an actual observation is made. However, this is not what science does. For example, the events during first second of the history of the Universe are speculative. We cannot actually observe them, and yet there is an entire science called cosmology built around it.



When it comes to accepting something as: 'as close to reality' as our models can take us, intuition is also not mandatory, because we'll never really get from intuition, what reality is, anyway.

What we get from our intuition are hypotheses (guesses) about how things work that allow us to predict, even though we do not actually know.

theloniusmonkey
2013-Feb-26, 12:57 PM
Well, I wouldn't say that was saying 'very little'.

When I see threads arguing for/against the recurrence of intelligence, (as a result of 'Evolutionary trajectories'), it seems reasonable to investigate just where those trajectories end up in a large population sharing replication of certain generalised, universal 'characteristics' in common. Ie: the idea of 'Intelligence' as a consequence of the assumption that Evolution, (and presumably, abiogenesis), might be universal, carries with it the consequences of that assumption. In this case, it would seem to be diminishing predictability, (ie: increased uncertainty), as the sample space increases.

This is of course, predicted on the assumption, (as you point out), that some mechanism supporting the concept of 'contiguousness', or 'replicated lineages' exists in common. Such a generalised characteristic, would seem to be inherent in the abiogenesis assumption too, otherwise how else does 'it do its thing', everywhere?

(Just food for thought about consistency more than anything else, in particular).

At its most fundamental level intelligence is self awareness and living systems are fundamentally intelligent. They are self aware. This is what most essentially defines them as living, as being biologic systems. All biological systems think and so are self aware. This is what it means to be alive.

Noclevername
2013-Feb-28, 05:41 PM
At its most fundamental level intelligence is self awareness and living systems are fundamentally intelligent. They are self aware. This is what most essentially defines them as living, as being biologic systems. All biological systems think and so are self aware. This is what it means to be alive.

[citation needed]