Sep 24 2011: True. All I've read is a chapter or so of 'selfish gene' (toward the end) and Stephen Jay Guold's extensive criticism of Extended Phenotype.
I realise, though, that selfish genes aren't against individual cooperation. What the selfish gene theory doesn't seem to permit is 'altruism,' that is, cooperation without some underlying selfish motive. It also dismisses organismal natural selection, group selection and species selection as different but equivalent views (with the 'necker cube' analogy) to underlying gene-level selection.
I buy Gould's argument that *causally* speaking, things cannot just be reduced to the gene level, and that Dawkins' frequent simplifying assumption 'let's talk as if there's one gene for this trait' doesn't work. (I guess I am taking Gould's word on Dawkins using this frequently.)
I guess when I say "cooperation exists" it sounds like I think Dawkins thinks it doesn't... and some other things in that paragraph are a bit extreme too. I should have stuck to just the word "altruism" and explained it as meaning cooperation outside of family unit without any mechanisms for ensuring reciprocity. (This would be encouraged by species selection, and possibly other mechanisms.)
I do plan on remedying my neglegence and reading Dawkins' work. It's just that I decided to read Gould's "Structure of Evolutionary Theory" as an introduction to the guts of the theory. And that's a 1300 page book, so it'll be a while before I move on to anything else.
Sep 22 2011: That's very interesting, Gerald. I haven't decided yet whether I buy your argument; I still think my model is replico-chauvinist (if I may coin the term) and therefore less general than it could be. You've just shown that the model isn't necessarily falsifiable... without extremely detailed knowledge of biology it appears we can explain any apparent lack of zeal for reproduction as the best solution available. But what you're saying makes perfect sense since you're not claiming the "zoomed out" prediction as part of the theory. At this point I think that theory is fine... but not as powerful, or directly predictive, as a more general formulation would be.
I think it would be difficult to untangle our disagreement here, so I hope someone else gets involved in the discussion.
Just one more point. You just referred to my little set of rules as a 'program'. In truth the rules are rather chimeric. We've been treating them sometimes as a specific program which can be executed, and at other times as a set of statements about reality which are always true, and at still other times as a set of assumptions which have to be met in order for the conclusions to hold.
The last two ways of treating the rules are equivalent, but the first one is more interesting. The rules themselves shouldn't be a program. Individual programs should either be described by the rules or not. So I can describe a program which has death, or one which doesn't, and the rules will still apply; the one without death just clearly isn't a model of biology (it still could be a physically plausible system; atoms don't die very often).
So I think the best method for comparing the effectiveness of different rulesets would be to make simulations and see if the results agree with the theory. But you're not including the long-term predictions, so I'm not sure how you'd prefer to test out the theory.
Like I said, I think we could both benefit from outside input for perspective. I'll think about your approach.
Sep 21 2011: The thing is, though, that at least in the more complex situation of light passing through a lens, we haven't actually proven how light will behave from the basic quantum equations; there are too many particles involved. So though in principle we know the laws of quantum mechanics are behind the behavior of lenses, it's really an observation; so quantum physics does nothing to explain the behavior of the light. Of course presumably that will change once someone figures out how to do the calculation.
So in order to really explain the behavior of light we need extra statements saying the light will indeed find the fastest path. And if we had a proof of that from quantum principles, that proof would be a necessary part of the physical theory of light. In the case of evolution, we do have something very much like a proof that the observed long-term behavior results from the observed short-term behavior. And to me at least, that's the whole theory. Though just listing the short-term behavior has some merit too; there are lots of special cases and technicalities... like, what sort of variation actually occurs, how much is the typical death measured by variation, etc.
Sep 21 2011: So, all I meant by predominance was that they will be more common. Imagine a system where there is no death, and the population starts out with individuals with different reproductive rates: one individual who produces 1 child per day, one who produces 2 each day, etc., with one individual at each rate all the way up to 100. Your "short view" of evolution just says they'll reproduce at those rates. The "long view" gives us a useful prediction: eventually the population will consist mostly of individuals with reproduction rate 100/day. If we throw in random death, we can say with a very high probability that it will eventually consist of only such individuals.
I think of the zooming out as necessary because without it, all we're saying is "sometimes there is such a thing as reproduction, sometimes there is such a thing as variation, sometimes there is such a thing as natural selection." Chemistry at least is referring to the elements, which we can identify and point to. I think the point of evolution is to say, "when these conditions are met, the consequence should be this long-term behavior," not just a statement of the conditions.
I guess if you look as evolution instead as an explanation for an observation (the diversity and adaptation of species) I can see where you're coming from, but seeing as this conversation is about the idea of using evolution as a wider theory rather than just explaining life, it seems important to consider the results as part of the theory.
Now about the photon... Have you heard of Maupertius' theories? There's a good book written for a popular audience named "the best of all possible worlds" by Ivar Ekeland and I think I will have to go back and reread some of the details because they're relevant to this discussion. Basically I remember Maupertius assumed photons would always find a path of "least action," which turns out to not be quite right. Instead they minimize the derivative of the action function over possible paths.
Sep 21 2011: Okay, that's a lot to reply to, and a lot of interesting stuff to reply to. I might not have time right now so I'll try and add more later.
Basically what I'm saying about the replication thing is that if we state evolution purely in terms of what's "better at duplicating", we would reach the conclusion that we'd move toward organisms which produce more and more offspring (or maybe healthier and healthier offspring). This isn't biologically supported; it often works fine to produce only a handful of children and care for them more carefully, or to produce even fewer children (per year) and have longer-lived, more durable individuals.
So if we want to talk about life, we should certainly add the fact that immortality is not biologically possible, which necessitates the strategy of reproducing. The general formulation should start out by talking purely about "pattern preservation," in the general sense that either surviving longer or copying onesself preserves the general pattern of an organism. In fact I don't know if "durability" and "reproductivity" are the only two methods of pattern preservation, but it seems like it.
Sifting is a "selection process" which meets these assumptions (meets the assumptions, or serves as an example of the principles, depending on how we want to word things...). Larger objects pass through a screen with more difficulty so over time sifting selects only the objects larger than the holes in the screen. Why shouldn't a more general theory handle such cases? (Occasionally clumps being sifted will break. This is a type of reproduction which preserves many properties of the mother-clump but not survivability, since the children can fall through the screen, and not reproducibility, since the children have less material for falling apart.)
Adding in the lack of indefinite individual survival then gets more specific, so thath we're talking about ordinary organisms.
Sep 21 2011: If we want selection we need a fourth assumption. The three assumptions I used above do lead to the conclusion that entities better at duplicating will become more common, and (assuming variation works properly, which the third assumption is supposed to guarantee) progressively better at duplicating. This is a lot like selection already. The only only thing missing is death to eliminate the losers rather than just... leave them in the dust.
So my definition of selection would be the elimination of some members of the population, combined with the other axioms, leading over time to a population of finely-tuned individuals. (Note this fourth assumption could just be *completely* random death, the logic would still work out.) Actually if I relax the requirement ... I could get rid of the fourth assumption and say natural selection is just the predominance over time of members better and better at reproducing.
(Even this simple a statement of evolution is just as misleading as the "selfish" terminology. Evolution does not inevitably lead to more and more babies. We should rather remove the idea of duplication. Surviving well into old age is just as effective as having many children; the more general idea is a bit hard to formulate though.)
Since selection is a "gradualistic" result as Darwin might say, it does require "zooming out" as you have said. It is the whole point of the definitions though.
But I don't see your point with the photon... the environment doesn't squeeze a photon toward the fastest journey, does it? The photon acquires a random, but inherently straight path when its waveform collapses. The straightness of the path isn't caused by any sort of selection or pressure or anything.
Sep 20 2011: Since the present behavior of an organism is due to a process of selection unlike the present behavior of a photon, the metaphor is much deeper. There is no desire, but there is a testing of alternatives and a weighing of advantages and disadvantages.
Besides, we all know our own brains are physical objects, and ideas like "desire" and "intelligence" are just woefully inadequate shorthands for complex physical processes. Behaviorists tried to remove talk of mental states from psychology for that very reason; it didn't work very well and now we talk about mostly-unmeasurable internal states.
Oh, but back to a point that deserves more discussion, your physical statement of evolution as "something that as the assets allowing it to be duplicated, is duplicated." I do think a little more has to be added to gain the whole theory, something like
- Whatever is better at duplicating, will duplicate more (differential proliferation)
- Whatever is better at duplicating, will produce descendants better at duplicating (constant environment)
- Whatever produces varied descendants will produce some better and worse at duplicating
That last one's awkward, but pretty necessary for real evolution... If Cronin produced a reaction which quickly hit an optimum and couldn't support the sort of arbitrary variation which even viruses do he wouldn't have created inorganic evolution. Arbitrary ability to produce varied descendants is important.
Sep 20 2011: Well, I do realize it is just a metaphor, as are any applications of terms about the mind to evolution. In fact the way I think of evolution is really that a lot of metaphors about the mind apply surprisingly well to biology, since species make 'decisions' (amongst available alleles) according to "incentives" (selective pressures) and the whole process frequently results in "designing solutions" to evolutionary "problems."
What I disagree with is simply Dawkins' assertion that selfishness is sufficient metaphor to capture all relevant evolutionary phenomena; especially if it's only applied at the lowest (genic) level. Carrying those two simplifications over to an attempted general theory seems a bad idea.
I think it's best to treat the metaphors as part of the theory since almost every evolutionary biologist ends up using them. Dawkins himself can't avoid talking about individual organisms competing even when he asserts that genes do all the competing. The fact that such a 'convenient shorthand' is often nearly necessary implies that it is a powerful part of the theory; similarly the convenient shorthand of treating evolution as some sort of (species-level) mind is nearly unavoidable and often handy.
TEDCred score: +0.10 TEDCred reflects your contribution to the TED community.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
I realise, though, that selfish genes aren't against individual cooperation. What the selfish gene theory doesn't seem to permit is 'altruism,' that is, cooperation without some underlying selfish motive. It also dismisses organismal natural selection, group selection and species selection as different but equivalent views (with the 'necker cube' analogy) to underlying gene-level selection.
I buy Gould's argument that *causally* speaking, things cannot just be reduced to the gene level, and that Dawkins' frequent simplifying assumption 'let's talk as if there's one gene for this trait' doesn't work. (I guess I am taking Gould's word on Dawkins using this frequently.)
I guess when I say "cooperation exists" it sounds like I think Dawkins thinks it doesn't... and some other things in that paragraph are a bit extreme too. I should have stuck to just the word "altruism" and explained it as meaning cooperation outside of family unit without any mechanisms for ensuring reciprocity. (This would be encouraged by species selection, and possibly other mechanisms.)
I do plan on remedying my neglegence and reading Dawkins' work. It's just that I decided to read Gould's "Structure of Evolutionary Theory" as an introduction to the guts of the theory. And that's a 1300 page book, so it'll be a while before I move on to anything else.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
I think it would be difficult to untangle our disagreement here, so I hope someone else gets involved in the discussion.
Just one more point. You just referred to my little set of rules as a 'program'. In truth the rules are rather chimeric. We've been treating them sometimes as a specific program which can be executed, and at other times as a set of statements about reality which are always true, and at still other times as a set of assumptions which have to be met in order for the conclusions to hold.
The last two ways of treating the rules are equivalent, but the first one is more interesting. The rules themselves shouldn't be a program. Individual programs should either be described by the rules or not. So I can describe a program which has death, or one which doesn't, and the rules will still apply; the one without death just clearly isn't a model of biology (it still could be a physically plausible system; atoms don't die very often).
So I think the best method for comparing the effectiveness of different rulesets would be to make simulations and see if the results agree with the theory. But you're not including the long-term predictions, so I'm not sure how you'd prefer to test out the theory.
Like I said, I think we could both benefit from outside input for perspective. I'll think about your approach.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
So in order to really explain the behavior of light we need extra statements saying the light will indeed find the fastest path. And if we had a proof of that from quantum principles, that proof would be a necessary part of the physical theory of light. In the case of evolution, we do have something very much like a proof that the observed long-term behavior results from the observed short-term behavior. And to me at least, that's the whole theory. Though just listing the short-term behavior has some merit too; there are lots of special cases and technicalities... like, what sort of variation actually occurs, how much is the typical death measured by variation, etc.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
I think of the zooming out as necessary because without it, all we're saying is "sometimes there is such a thing as reproduction, sometimes there is such a thing as variation, sometimes there is such a thing as natural selection." Chemistry at least is referring to the elements, which we can identify and point to. I think the point of evolution is to say, "when these conditions are met, the consequence should be this long-term behavior," not just a statement of the conditions.
I guess if you look as evolution instead as an explanation for an observation (the diversity and adaptation of species) I can see where you're coming from, but seeing as this conversation is about the idea of using evolution as a wider theory rather than just explaining life, it seems important to consider the results as part of the theory.
Now about the photon... Have you heard of Maupertius' theories? There's a good book written for a popular audience named "the best of all possible worlds" by Ivar Ekeland and I think I will have to go back and reread some of the details because they're relevant to this discussion. Basically I remember Maupertius assumed photons would always find a path of "least action," which turns out to not be quite right. Instead they minimize the derivative of the action function over possible paths.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
Basically what I'm saying about the replication thing is that if we state evolution purely in terms of what's "better at duplicating", we would reach the conclusion that we'd move toward organisms which produce more and more offspring (or maybe healthier and healthier offspring). This isn't biologically supported; it often works fine to produce only a handful of children and care for them more carefully, or to produce even fewer children (per year) and have longer-lived, more durable individuals.
So if we want to talk about life, we should certainly add the fact that immortality is not biologically possible, which necessitates the strategy of reproducing. The general formulation should start out by talking purely about "pattern preservation," in the general sense that either surviving longer or copying onesself preserves the general pattern of an organism. In fact I don't know if "durability" and "reproductivity" are the only two methods of pattern preservation, but it seems like it.
Sifting is a "selection process" which meets these assumptions (meets the assumptions, or serves as an example of the principles, depending on how we want to word things...). Larger objects pass through a screen with more difficulty so over time sifting selects only the objects larger than the holes in the screen. Why shouldn't a more general theory handle such cases? (Occasionally clumps being sifted will break. This is a type of reproduction which preserves many properties of the mother-clump but not survivability, since the children can fall through the screen, and not reproducibility, since the children have less material for falling apart.)
Adding in the lack of indefinite individual survival then gets more specific, so thath we're talking about ordinary organisms.
Like I said, I'll write more later.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
So my definition of selection would be the elimination of some members of the population, combined with the other axioms, leading over time to a population of finely-tuned individuals. (Note this fourth assumption could just be *completely* random death, the logic would still work out.) Actually if I relax the requirement ... I could get rid of the fourth assumption and say natural selection is just the predominance over time of members better and better at reproducing.
(Even this simple a statement of evolution is just as misleading as the "selfish" terminology. Evolution does not inevitably lead to more and more babies. We should rather remove the idea of duplication. Surviving well into old age is just as effective as having many children; the more general idea is a bit hard to formulate though.)
Since selection is a "gradualistic" result as Darwin might say, it does require "zooming out" as you have said. It is the whole point of the definitions though.
But I don't see your point with the photon... the environment doesn't squeeze a photon toward the fastest journey, does it? The photon acquires a random, but inherently straight path when its waveform collapses. The straightness of the path isn't caused by any sort of selection or pressure or anything.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
Besides, we all know our own brains are physical objects, and ideas like "desire" and "intelligence" are just woefully inadequate shorthands for complex physical processes. Behaviorists tried to remove talk of mental states from psychology for that very reason; it didn't work very well and now we talk about mostly-unmeasurable internal states.
Oh, but back to a point that deserves more discussion, your physical statement of evolution as "something that as the assets allowing it to be duplicated, is duplicated." I do think a little more has to be added to gain the whole theory, something like
- Whatever is better at duplicating, will duplicate more (differential proliferation)
- Whatever is better at duplicating, will produce descendants better at duplicating (constant environment)
- Whatever produces varied descendants will produce some better and worse at duplicating
That last one's awkward, but pretty necessary for real evolution... If Cronin produced a reaction which quickly hit an optimum and couldn't support the sort of arbitrary variation which even viruses do he wouldn't have created inorganic evolution. Arbitrary ability to produce varied descendants is important.
Anyone got a better formulation?
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
A reply on Conversation: "Selfish matter" is a catchy name for an important endeavour, but selfishness is not the foundation of evolution.
What I disagree with is simply Dawkins' assertion that selfishness is sufficient metaphor to capture all relevant evolutionary phenomena; especially if it's only applied at the lowest (genic) level. Carrying those two simplifications over to an attempted general theory seems a bad idea.
I think it's best to treat the metaphors as part of the theory since almost every evolutionary biologist ends up using them. Dawkins himself can't avoid talking about individual organisms competing even when he asserts that genes do all the competing. The fact that such a 'convenient shorthand' is often nearly necessary implies that it is a powerful part of the theory; similarly the convenient shorthand of treating evolution as some sort of (species-level) mind is nearly unavoidable and often handy.