There is a pretty interesting discussion going on in Noyau regarding the many definitions of “fitness” in evolutionary biology. It would be a shame for it to be lost in that particular venue here at TSZ. At the risk of being censored by the admins for posting too many OPs in one month I thought I’d start this thread.
Here’s my take so far:
Allan Miller was charged by phoodoo with resorting to different definitions of fitness. Allan denied the charge and when asked for a definition of fitness Allan provided one. Allan later stated that his definition only properly applied to asexual species.
Others chimed in to say that the definition of fitness depends on the context, which hardly seems to contradict what phoodoo was saying.
My own position is that fitness has its definition within a particular mathematical framework. My position is also that fitness can be defined generically but that such a definition is tautological. Special definitions of fitness are required to make the concept testable.
Here’s hoping we can move the discussion about fitness out of Noyau.
How can you tell?
phodoo:
Reminding you of your own statement:
“HOWEVER, if the question becomes different, like perhaps, “Do the bravest poker players win?” Or “Do the most conservative players win the most?” Or, “Do the players who bet the most win the most?” “or “Are the tallest poker players the best?”, then we at least have the possibility of being profound.”
… which accepted that poker players that tend to win more or win less exist and that we can study what correlates with that.
Similarly, phenotypes and genotypes can differ in their ability to survive and reproduce. And we can study which phenotypes and which genotypes correlate with that.
Do you have a garden?
See the problem yet Joe?
Joe Felsenstein,
So can we know if tongue curling affects one’s success at poker? What about fingernail shape, can we determine which shape is most beneficial for winning at poker?
Although that question is of course vacuous, unless you can define “skill level” independently of “winning”, the question you asked about which characters might affect winning at poker has, as you yourself said, “at least the possibility of being profound”.
As you yourself noted, we can, though I suspect neither of the ones in this list would be likely to show a significant effect. However the others you mentioned, “bravest”, “most conservative”, “ones who bet the most”, and “tallest”. The first three of those are good candidates.
So do you see the lack-of-problem yet, phoodoo?
Indeed! How and which plants colonise a patch of cleared ground is a clear demonstration of niche and biased chance.
Whoa whoa Joe, what do you mean by significant? Haven’t you claimed that even if there is 1/10th of a percent advantage, over time this will cause the allele to fix. So likewise shouldn’t we say even if we see ANY difference between tongue rollers and non-tongue rollers in poker, we have to conclude likewise. Over time the tongue rollers profits must increase.
So now we really can’t say if its just luck that the tongue rollers win more. Or tongue rollers perhaps have some other advantage tied with it. Or tongue rollers perhaps also tend to be from a certain region more, which also has some advantage….But ultimately, either tongue rollers are going to win more, or non-tongue rollers are going to win more or its EXACTLY even, which is most unlikely.
So the same problem exists when you try to measure fitness. One allele MUST be more than another, unless its exactly even. And you can never know why, just like you can’t know about the tongue rollers.
INSTEAD, all you can really do in both cases, whether its poker or life, is cherry pick out obvious cases, like when someone has some devastating genetic disease which makes them unlikely to survive past 10, and say, see, they have a very low fitness, obviously, don’t you think it will affect their chance of reproduction? Likewise you can say a guy who can not count, or who has never played poker before, or who is blind and can’t see the cards, must have a much less chance of winning. And because of this you suddenly extrapolate this to mean ALL or virtually all (or even most or some) alleles must contribute positively or negatively.
But you have no justification for assuming this, just like we saw with the tongue rollers. You are using your opinion about whether or not it should have an affect. But since we can’t differentiate between your opinion and luck, its meaningless.
No, of course not. I have written in many posts and comments about the probability of fixation of advantageous alleles. It can be quite small, but it’s much bigger than the probability of fixation of neutral alleles. Show me where I am supposed to have made this claim!
Let’s go back to the issue of whether fitness is a vacuous concept, used only in tautological arguments.
You have now admitted that it may be meaningful to ask which phenotypes affect fitness. (You accepted the analogy of winningness in poker, and said that we could investigate whether various phenotypes affect probability of winning, and that this “at least the possibility of being profound”). I’m glad we agree on that.
A lot of them will contribute at least a little bit, but many of them too little to either be detectable or to affect the probability of fixation of the relevant alleles. Beyond that you are just making stuff up.
You asserted that natural selection is deterministic and denied that natural selection is probabilistic. Now you admit that natural selection is probabilistic. Congratulations Joe.
What are we to make of your earlier comments?
Let’s not. One can re-define fitness in such a manner that it is not tautological and in a manner that can be tested. That’s why there are so many different definitions of fitness.
Do you have an explanation for why there are so many different definitions of fitness?
For one, because there are so many different life histories, as there are so many different life styles of different species. Does that startle you?
No, it doesn’t startle me.
So the definition of fitness is dependent upon each specific life history and the different life histories are, well, quite numerous and different. Thus there are many different definitions of fitness, each one unique and different, as each life history is unique and different. Are we good?
Is this an admission that there is no general definition of fitness that can be applied to all these different life histories?
Likewise there are an impossible number of strategies for reproducing; such that claiming you can pinpoint which traits exist because of luck, and which exist because of its central role in reproducing, is untenable.
You can only make broad staments, such as organisms that die right after they are born aren’t likely to reproduce, or organisms that are born without reproductive organs probably won’t produce. And from this, we are told we must accept that you then can say SOME trait is so universal that it affects lineage after lineage for millions of years, and THAT is why novel things arose (like eyes!)
You see, if an indentation can help you survive, even a little (You indeed have claimed it only needs to be a little), well, that indentation will keep helping generation after generation to survive, until the next luck innovation, like a liquid filled cavity comes along. And once again this liquid filled cavity will universally be useful for thousands of generations, such that those who got along without the indentation and liquid filled cavity before, can no longer compete again the new form.
This is where the whole fitness concept comes off the rails. This indentation, or this liquid filled cavity must supersede EVERY OTHER REASON for survival, or it would just eliminate like most other neutral traits. If every time there are 6 or 600 reasons for a reproduction, just like one reason a poker player wins is because his fingernail shape, another reason is his tongue rolling, another reason is his hair color, another reason is the shape of his eyes, and another reason is his counting ability, then can you really conclude that EVERYONE of those traits is being selected for? Then what about the guy who is being selected for other reasons, which also is successful?
You now have a complete stew of reasons, none of which are coherent, and capable of advancing in a stepwise fashion for greater functionality. Fitness becomes nonsense.
Not an “admission” so much as a proclamation.
For many complex life histories, you cannot come up with a single number which “is” fitness — but you can use things like age-specific birth rates, age-specific death rates, and coefficients that express the sensitivity of these to things like population density to predict how the genotype frequencies will change through time.
But if someone wants to argue that fitness is a meaningless concept, they have to show us this in the simplest models, such as annual-plant life histories. And this they cannot do.
Joe Felsenstein,
Joe, ANY study of gene frequencies is going to show some difference. Just like any study of tongue rollers in poker is going to show a difference one way or the other. Either the tongue rollers will have an advantage or a disadvantage (or the results will be exactly even-unlikely).
So any study of genes is going to show one existing more than another. You will never be able to say why, just like you won’t be able to say why, except in the most extreme cases, where instead of using numbers you just use intuition-like assuming blind poker players won’t do as well, if they can’t see any cards.
Guessing which genes are from luck and which are universal winners because of an endless strategy win over many many generations is unfeasible. There are too many ways to reproduce, and luck might be the biggest reason. It won’t create an eye, winning each round through luck.
The tallest people have the greatest height. Does that express a meaningless concept? Does that express a vacuous concept?
The “simplest models” that you refer to, which definition of fitness do they use? Do they use a definition that is applicable only to that specific model of a life history, or do they use one that is applicable to all models of any life history?
Are they general, or are they specific?
Yes, as I said myself.
They are specific, of course. I was making the point that if someone argues that fitness of genotypes, or of phenotypes, is a meaningless concept, used only in circular arguments, then they have to show that in the simplest, most easily-analyzed situations. phoodoo already acknowledged that fitness could be nontrivial,
Your response seems not to be to look closely at the simple case, but to try to wander off to total-generality-land.
I wish I understood this argument. It seems as if it is like saying that, since it may be very hard to determine which of two closely-similar weights is the heaviest, therefore weight is a vacuous concept.
Joe Felsenstein,
Seems to be all about prediction, for phoodoo. If two unmeasured strains of bacteria may have a different growth rate – or the same – that concept is meaningless, tautolological even, if you can’t predict which, if any, is the higher before you find out.
Again no, its not just a question of being hard to determine why a certain genotype reproduced (impossible actually), in EVERY generation not only are the reasons different, but there are many many reasons.
To claim that, Oh, the reason this organism reproduced is because of its tail length, and the reason this organism reproduced is because it had a better indentation on its light sensitive spot, and the reason this one reproduced is because it had better shaped fingernails for gripping, etc…is foolhardy. There are many many reason why one animal happen to reproduce. We can all understand what a detrimental event would be, which would cause something to be unable to reproduce, but to say some new, never before used trait, that a lucky accident causes to appear on some new animal, is going to CONTINUE to be the reason why an animal reproduces, for thousands of generations!, come on, that’s how you build an eye? Please.
The lucky indentation can’t be THE REASON, that many times. Someone being tall can’t be the reason someone wins at poker every time.
No, I keep them quite distinct because I understand the difference. If only evolutionists would do the same.
In the simplest most easily-analyzed situations you are using a different definition of fitness. Because that is so, you must be very careful not to equivocate. But evolutionists often do equivocate.
Could you imagine if gravity was treated the same as fitness, requiring a different definition of gravity in each specific case in order for it to be testable?
That is their problem. If I work with a theoretical population genetics model, I have to be precise about what is fitness and what isn’t fitness.
If I had to try to come up with something totally general, I’d say that fitness of a given genotype or given phenotype is a number proportional to the expected contribution of individuals of that genotype or phenotype to future generations, when we imagine that all descendants of all individuals do not differ in ability to survive or reproduce (that rather strange condition I can explain if you want). That can be made specific in simpler models, but gets complex in more complex life cycles. Those have overlapping generations and density-dependent survival and reproduction, for example.
But in simple models, such as an annual-plant life cycle, the definition of absolute fitness is clear. And there is no validity to the argument that in those cases fitness is a meaningless concept, or one only used in circular arguments.
Creationists are often quite concerned to make the case that fitness is meaningless or only of use in useless arguments. (ID advocacy would seem not to reject the meaningfulness of fitness, though many “ID theorists are actually also creationists, so they do tend to get exercised over this). They argue this right up until the Ark lands and disgorges a limited number of individuals, who then have a lot of evolving left to do.
Joe Felsenstein,
Joe,
You have repeatedly referred to the concept of fitness as “the expected contribution of individuals” whereas EXPECTED is part of the definition of what it means to be fit. This is where the theory becomes meaningless and circular. How can EXPECTED be an attribution of fit? It obviously can’t except in the mind of someone making a model.
And when you make your models (rather than reality) what are the expectations based on? When you don’t say what those expectations are based on, then of course fitness is circular. Because the models are based on what survives. So if you were measuring tallness, and you based your expectations of tallness based on what is tall, that would be pretty circular and as you call it meaningless.
So when fitness contains no other meaning other than the expected contribution based on their existing contribution, I think it easy to see why some would consider it at the very least inadequate for discussions about evolution.
The problem is you then will equivocate and say, Well no, we don’t base the expectation of contribution based on existing contribution, we base it on OTHER traits. Well, as soon as you do that, you are then changing the definition of fitness to mean those other traits are part of fitness. It would be like saying:
“We expect those that are tallest to be fast. So in order to measure the expected tallness we measure speed, and that’s how we get our measurements.”
To call that measuring tallness would not be an accurate statement. All you are really doing is making an argument about what you claim causes tallness, and we have no reason to accept your claim about what causes tallness.
Theoretical… huh? A theory about the theory…
Um, you misunderstand what is meant by “expected”. In a statistical sense, the word means, in effect, “average”. So when we have a genotype which causes the individuals bearing it to have a probability 0.7 or surviving to adulthood and causes the survivors to then have 3 offspring each, the number of newborn offspring per individual is then either 0 or 3. But the average over a large sample of them is then 0.3×0 + 0.7×3 = 2.1.
So the expectation, the average over a large sample of newborn individuals of that genotype, is 2.1.
Similarly with an average over a large sample of individuals who have a given phenotype, such as having a given value of body length.
We can estimate these averages by taking a large sample of individuals, genotyping them, and then seeing what fraction survive and how many offspring each survivor has.
Yup. It’s sad but true. Scientists keep using theory, and then comparing it to outcomes.
I know you’d rather they didn’t. Blame folks like Galileo and Newton for getting us started down that path.
Joe Felsenstein,
Again Joe, what about this is not saying, those that survived survived? Those that are tall are tall?
You can either call your idea of fitness, “the expected contribution” or the “average contribution” or whatever-you still have the same problem.
If survival is the ONLY criteria, then we really aren’t doing anything to confirm the theory of survival of the fittest. We are only confirming survival of the survivors.
Predicting the probability of the survival of the certain survivors seems different from confirming those who survive survive.
phoodoo did, earlier, accept the analogy to seeing which sorts of poker players are most likely to win.
This is simply taking a sample of poker players, characterized by some property, such a particular level of aggressiveness in their play, and then seeing how often they win. You characterize their aggressiveness, then you watch them thereafter and count what fraction of times they win (or whatever).
Similarly, with organisms, we take newborns of those that have some particular genotype at some loci we are following. Then we see whether they survive, and for those that survive, how many offspring they have. The fitness is the number of newborns produced, divided by the number of newborns of that genotype that we started with.
(… and by the way, fitness is affected by the amount of reproduction, not just by survival).
Right, and WHAT is that some property if its not fitness??
So in the poker analogy, I already challenged you to justify how you can’t call tongue rollers better poker players, if in fact more tongue rollers, in the entire history of poker playing, have won than non-tongue rollers. Or if one shape of fingernail has one most. Or if one length of pinkie has won most. Its not enough just to count and then say see, we counted.
This is the problem with your fitness model. You count what survives, and then proclaim that is what is fit. There is no other criteria other than how many. So you can not know why. You can’t know if tall poker players win more because of luck, by just counting how many tall poker players win. Because the tall poker players have OTHER traits as well! You must do more than just count, and then proclaim you have said anything other than what is, is.
Fitness can never just be a count of what is-is-or you will always just be repeating a meaningless tautology.
Joe Felsenstein,
Try to look at it like this Joe. Name a property that is fit. Big? Is that fit? Small, is that fit? Is fast fit? Is slow fit? Is smart fit? Is dumb fit? Is hairiness fit? Is hairlessness fit?
You can’t name ANY one of these things and call them fit. The only thing you can call fit is living. And why is that? Because there is no one way to survive.
You can’t use Alan’s pleading that it depends on the niche, because the niche and the animal can’t be separated. You can’t say, well, in the ocean, being a fast swimmer is fit. Is it, is fast the only way to survive in the ocean? And if I take a tuna and put it in the desert does that mean it was fit and then changed to unfit, because I moved it? So now I have to count EVERYTHING. At this exact moment, in this exact temperature, in this much concentration of salt, with this many predators next to you, you are fit at this moment. If I move you, or if an oil spill happens, or if sushi becomes more popular because of a lunch special at a California sushi chain, you are less fit. The only way I can call you fit is by you being here.
What happens when the prediction is wrong?
You find out why, did the enviroment change, bad calculation, etc., what do you do when a prediction fails?
You’re using the word wrong.
There is no such thing as a property that “is fit”. Fitness is the measure of the ability to produce successful offspring. It is a quantity and it is always relative to another measure of some other organism, or other period of time, or carriers of another allele.
You’re using the word as if you would say “name an entity that has height”.
What does that mean? Are you asking HOW TALL the entity is?
What do any of these questions even mean? They’re stated in a nonsensical way. You need to learn to use the word correctly.
What you should be asking, if I can sort of guess what you’re trying to convey, is something like “what is the fitness of a hairlessness phenotype?”. Which is another way of asking “what effect (if any), positive or negative, does being hairless have on the average number of successful offspring produced?”
When stated that way, the question at least makes sense. The answer is that it depends on context. Different phenotypes have fitness effects that are dependent on the enviromental context and the competing phenotypes. There is no universal answer. That’s why fitnesses are COMPARED, not presented as some universal state of affairs.
Nothing is “fit or unfit”. You’re asking as if the word is a substitute for “living” or “dead”.
Both because to do so would be to use the term in a nonsensical way. It’d be like calling any arbitrary object “height”.
And there it is, you’re using the word “fit” as a substitute for “alive”. Rather than the actual meaning of fitness, with refers to the relative measure of the number of successful offspring the organism has.
Your conclusion is correct, but how you arrived at it isn’t. You can’t call anything “fit”. Because that’s a nonsensical use of the term.
But it is certainly true that there are many different ways to successfully persist as a population of living, reproducing organisms.
No, you can’t say that as a general condition. But that doesn’t mean you can’t say that there are NO situations where being a fast swimmer causes higher productive success. It doesn’t have to be ALWAYS true to be SOME TIMES true.
Also your use of the word “fit” is incoherent, as you are going back and forth between using it as a substitute for “living”, which is not the same as reproductive success. It is possible to be alive, but have zero reproductive success.
No.
That’s because “the ocean” isn’t really a single thing, but a huge collection of overlapping aquatic environments with many shared and different properties. And many other species with many different behaviors and properties.
Disregarding your nonsensical use of the word “fit” as a misplaced and confused use of the concept of fitness, some times substituted for “alive”, yes it is is the case that if you take a tuna from the ocean and drop it in the desert, most of it’s phenotypes are associated with lethality. The tuna will die. Probably the vast majority of it’s adaptations to life in the ocean are of no use in the desert.
What does that even mean? Count everything? Count every thing of what?
What the flying fuck are you even trying to say here?
If Joe (or Allan) goes for Sushi, he’s less fit? What the FUCK does that even mean? Does him going for Sushi then, in your view, mean he’s somehow slightly less alive, or has he lowered his reproductive success?
Rumraket,
Please don’t go over the same stupid shit when you don’t even understand what is being discussed. Everyone knows Joe’s fitness models, and concept of fitness as “Do some phenotype have a positive, negative or neutral effect on population?” No one needs ten paragraphs to figure that part out. The question is, how does that differ from saying tall things are tall, and small things are small. Until you can tackle the deeper problem here, your shallow meandering is useless. You post is so overwhelmingly stupid, you can’t even understand that it is not the fitness of HUMANS that is being affected by sushi sales, its the fitness of tunas, for fucks sake.
Back to you Joe.
If there is no universal concepts involved in your measure of fitness, then the only universal is that survivors survive. Count how many survive now, predict how many will survive later, or take an average of survival over a certain time- it still adds up to the same thing-it can’t answer the more profound question of is survival of the fittest true. No one really is interested in knowing if survival of the survivors is true.
Newton? Galileo? Felsenstein?
Please tell me you are joking by comparing your meaningless speculations on population genetics to those two…. Or, alternatively, your sombrero might be too tight…. lol
If I were you, I’d consider doing real science by proving the law of recurrent variation wrong… Dammit ! I forgot. You can’t…You’ve never been to a real lab before… You wouldn’t know where to start…. Too bad… such a good idea….lol
J-Mac,
I said “getting us started” and that’s all us scientists.
As for the rest of J-Mac’s effusions, let’s just say that that, by now is at least number 150. (This count is explained here).
There is recent evidence of why the pop gen notion of fitness is relatively bogus.
I implicitly predicted the sort of nonsense that would emerge some years back in articles I wrote where I highlighted the absurd fact that harmful mutations in the Darwinian world can be regarded as “fit”:
Darwinian fitness is defined by the number of offspring that live to reproduce. So if blindness in cavefish help the cavefish make more babies, then “it’s not a bug[harmful mutation], it’s a feature.”
Formally the formulas of fitness look like
wA = vA fA
or
wA = (1/2) vA fA
where
wA = absolute fitness
vA = viability (ability to live)
fA = fecundity, the number of babies it can make
The human race has grown from a population from under 10,000 to around 8 billion, it’s “fitness” in the absolute sense has risen, but most geneticists will concede overall we are getting sicker. So much for the utility of the evolutionary idea of “fitness” based on reproduction rates rather than fitness based on the idea of an engineered design.
The following article points to a study that likely points out a stretch of NON-CODING DNA called uc467 (just a catalog name, don’t get hung up about names, a rose is a rose by any other name).
The actual study: http://www.cell.com/cell/pdf/S0092-8674(17)31497-6.pdf
So absolute reproductive “fitness” didn’t change but the creatures were abby-normal:
More evidence Darwinian “fitness” is bogus measure of function.
PS
Way back in 2007 when there was a push to say non-coding regions like uc467, this was the story:
http://www2.lbl.gov/Science-Articles/Archive/Genomics-ultraconserved.html
But if in the 2007 study these are the same non-coding regions in this 2018 study, we are getting a different story! 11 years later, the climate is much more friendly to saying “non-coding DNA is functional”.