Are evolutionary explanations "Just so stories"

Question

There's a claim that evolutionary explanations are nothin more then "just so stories", that is, for any trait we can find in nature, one could find a way to explain why it's adaptive.

For example:

"You find that people cooperate, you say, ‘Yeah, that contributes to their genes' perpetuating.’ You find that they fight, you say, ‘Sure, that’s obvious, because it means that their genes perpetuate and not somebody else's. In fact, just about anything you find, you can make up some story for it." Noam Chomsky

In some cases, the explanations are apparently based on a mathematical proof (ESS -Evolutionarily stable strategy) - but there are many degrees of freedom when choosing the exact mathematical model, which allow selecting a model that "fits" the desired results.

This question is about evolutionary explanations - how one can use evolution to explain the existence of a trait. not about proving or falsifying the theory of evolution in general

An example of a evolutionary explanation that isn't a "just so story" There are several methods of explaining altruism in evolution.

Kin Selection - by helping others we believe that are related to us, we help our shared genes to survive

Cooperation - We help others, and they help us. in the long run we earn at least as much as we lose

Handicap Principle - By helping others we prove potential mates\enemies that we are powerful enough to survive even if we help others.

Although these are explanations to the same general trait (altruism), each explanation has different predictions regarding the details of the trait - so it's possible to evaluate the validity of each explanation.

This has also been a critique from David Stove, see his book Darwinian Fairy Tales. I think the branches more likely to get into that pitfall are evolutionary psychology and sociology, which are the branches Chomsky seems to criticize in your quote. However, there are many solid evolutionary explanations that can be backed up by mathematical modelling and experimental findings as you mention yourself. — Raskolnikov, Apr 16 '11 at 19:53
I've seen it argued that the genome is pretty tightly packed just explaining our shape and physiology with the suggestions that most of evolutionary psychology could be bunk. On the other hand, one can use the language of evolution to describe the way in which cultures or societies change (a valid way to think even if it brings the much abused "meme" into the conversation). — dmckee --- ex-moderator kitten, Apr 16 '11 at 20:59
Genomic arguments such as "tightly packed" are canards in this context. The mammalian brain (and thus human) is the most plastic organ we know of. If one were to look at minuscule genomic effects and their extreme mental manifestations it is obvious that a tiny change goes a really, really long way. If this were not true, babies would be as uniform in their personalities as in their finger count, which is obviously false. — msw, Apr 17 '11 at 13:34

Jonas · Accepted Answer · 2012-04-10T13:17:27.763

27

There's a claim that evolutionary explanations are nothin more then "just so stories", that is, for any trait we can find in nature, one could find a way to explain why it's adaptive.

EDIT: That's actually two false claims!

Claim 1: For every trait we can find in nature, one could find a way to explain why it's adaptive

This claim is false for two reasons (1) Not every trait is adaptive, and (2) At the moment, it is not technically possible to test many of the explanations.

Not every trait is adaptive: Evolution, as in the change in the genetic make-up of a population, is more than "survival of the fittest". The genetic make-up of a population changes due to mutations, genetic drift, recombination, and selection. Only the latter is adaptive. The three former are non-adaptive, but can still give rise to traits that we can observe now. See for example this article for a discussion of why it can be wrong to always look for an adaptive explanation for specific traits.

Today, we simply can't know: With the exception of genetically engineered organisms, genomes, and the traits derived from them that we observe today are the product of evolution. However, in many cases, we cannot know the evolutionary mechanism that gave rise to the trait, especially in humans: Was it adaptive, i.e. the trait was so beneficial to the population and/or its individuals that they were able to produce more offspring? Or did the trait get fixed in the population by chance? (note that in e.g. bacteria, with their enormous populations, we can quite safely assume that every trait we observe is selected for). There are two problems often faced when trying to answer this question. (1) You need a large amount of genetic information on many individuals in order to statistically determine how the trait has been propagated, and (2) in many cases (sickle cell anemia is a notable exception), we are not able to really quantify the fitness advantage of a trait, and/or we simply don't know where on the genome it is encoded.

Claim 2: evolutionary explanations are nothing more than "just so stories"

This claim is false in its general scope. Yes, as mentioned above, not every trait is adaptive, and not every trait is well enough understood so that we can make a claim. Thus, some "explanations" (which should usually be labeled speculation) may be "just so stories", though many of those are simply hypotheses that are not testable yet, at least not on the level of the genome. That is likely to change as sequencing technology improves, and the $1000 genome becomes available (see here for a bit of speculation what this will do for research).

However, there are plenty of examples of evolutionary explanations involving adaptive changes, even in humans, such as sickle cell anemia, lactose persistence, or the adaptation to low oxygen at high altitude.

EDIT 2: Not all "explanations" are equal. There is speculation, hypothesis, and what I would call founded explanation. In the OP and my above post, "explanation" is normally understood as hypothesis or founded explanation, rather than speculation. "Speculation" is what you can just make up by purely using your superior intellect, but without needing to consult actual data. Within the realm of evolution, speculation is often resulting in just-so stories, since many people tend to not understand how evolution works. "Hypotheses" are what you propose when you actually consider the data; they are ideally haven't-gotten-around-to-do-the-experiment, rather than not-yet-testable, and should not be sorry-can't-test-this-ever. Founded explanation is once you successfully confirmed a hypothesis. This may still turn out to be wrong, since reality, and biology in particular, tends to be more complex than what we think, but it's at least consistent with what we know right now.

edited Apr 10 '12 at 13:17

answered Apr 17 '11 at 01:37

Jonas

3,683
31
23

1

You actually claim exactly that evolutionary explanations are "just so stories" (lack scientific base) – Ophir Yoktan Apr 17 '11 at 18:42
@Ophir Yoktan: Oops, you're right. These were two false claims in one. I've updated my answer. – Jonas Apr 17 '11 at 19:15
1

@Jonas, your answer is, at least in part, tautological, in that within it you state, "With the exception of genetically engineered organisms, every trait in every living thing we observe today is the product of evolution." It also appears to be wrong, if traits can include learned behaviours. – Apr 17 '11 at 22:26
@sampablokuper: You're right, that was sloppily formulated. Better now? – Jonas Apr 18 '11 at 00:42
@Jonas, well, you've addressed my second point. – Apr 18 '11 at 09:09
@sampablokuper: Oh, now I see what you mean. Yes, it appears a bit tautological to say that everything we see today is the product of evolution, when the accusation is that "for every trait we can find in nature, one could find a way to explain why it's adaptive", right? The "molecular" theory of evolution explains the mechanisms by which genomes change over time in populations, so unless your religion forbids the thought, whatever we see today has evolved by any of the mechanisms I mentioned. Which means it can be currently impossible to make a good claim that it was adaptive. – Jonas Apr 18 '11 at 11:52
@Jonas, it might be worth teasing out the semantics a little more. Adaptive to what?... Evolution as a whole is, arguably, an adaptive mechanism. I think you're distinguishing its sub-mechanisms and pointing out that they can affect genomes in ways that aren't direct, adaptive responses to the environments of the organisms (or, perhaps, populations) carrying those genomes. – Apr 18 '11 at 20:27
@sampablokuper: Evolution as a whole is describing changes in allele frequencies. These can, but don't have to be, adaptive (though Darwin originally described only the adaptive aspects of evolution). In this answer I tried to give a few definitions for adaptive vs. non-adaptive changes. – Jonas Apr 19 '11 at 00:43
@Jonas, with respect: (a) the "allele frequencies" sentence is a very ahistorical description of evolution, and focusing narrowly on it blinds one to the thrust of the "just-so stories" attack, which is disinterested in allele frequency; and (b) if evolution wasn't, overall, adaptive, it seems unlikely any species would survive it - unless they were specially created already fit for survival and much as we see them today, which seems as doubtful to me as it did to Darwin and surely does to you. – Apr 19 '11 at 21:40
@sampablokuper: (a) Yes, it is ahistorical. But it is modern: it is via the genome that traits are inherited (unless you want to include learned behaviors), and it is from the analysis of genomes that we can assess whether a trait is under selection. Sure, you could disregard allele frequencies and all the science around them when you talk about evolution, but then the "just-so stories" attack would end up being a "just-so" attack, because it's on (possibly equally) unscientific footing. – Jonas Apr 21 '11 at 03:48
@sampablokuper: (b) How many alleles have to be under selective pressure at any given time so that you would consider evolution "overall adaptive"? There are many alleles. The distribution of some is changing due to selective pressure, but for many others, the distribution may be evolving non-adaptively. In bacteria, with their large populations, most alleles are evolving adaptively. In mammals, with their comparatively tiny populations, it's probably the inverse - though that doesn't mean that it's not adaptive at all. – Jonas Apr 21 '11 at 03:49
@Jonas, (a) within the fairly terse context of Stack Exchange, I guess that's fair enough, even though it prevents your answer from addressing the many kinds of evolutionary explanations that have been given historically, in favour of merely defending modern genetic explanations. (b) I think you're initially conflating allele "distribution change due to selective pressure" with "adaptive allele distribution change", making your question about "How many alleles" a non-sequitur. After all, distribution can change for reasons other than selective pressure, yet still increase fitness. – Apr 21 '11 at 13:21
@sampablokuper: (a) You're of course absolutely right that especially "just-so"-stories are made up by people who don't know the more modern view of evolution. But should you evaluate the value of a theory (or anything else) by the theory itself, or by the wrong perception of the theory? – Jonas Apr 28 '11 at 17:22
@sampablokuper: (b) Evolution by non-adaptive mechanisms can, in isolated cases, result in increased fitness. However, overall, their effect will be neutral at best. – Jonas Apr 28 '11 at 17:29
@Jonas, (a) historical != wrong; this is dubious historiography, since by the same token, the latest theoretical consensus is likely to be found to be "wrong" in one regard or more as science progresses. (b) In what way does it make sense to describe an increase in fitness as a result that is "overall neutral at best"? – May 01 '11 at 03:19
3

The fact that a trait isn't adaptive in no way implies that you can't find an explanation of why it's adaptive. – Christian Jun 08 '11 at 15:32
@Christian Exactly. I would also like to add that "(2) At the moment, it is not technically possible to test many of the explanations." doesn't prevent anybody to give a suggestion for the "explanation", either. The entire Claim 1 refutation is very shaky. – ipavlic Apr 09 '12 at 22:06
@ipavlic: "Explain" and "Speculate" are two different things. You can speculate anything. To explain, however, needs more substantial information, which may not be currently accessible. – Jonas Apr 10 '12 at 11:35
@Jonas I thought that was the OP's idea - that one can "explain" ("speculate" in your terminology) away reasons why something is evolutionary beneficial in two different, mutually exclusive ways. Obviously one "explanation" ("speculation") could be wrong. If you stand by your, more precise, meaning of the term "explain", then I deem your refutation acceptable. But in that case I don't understand Claim 2 - it seems to me that you a) accept that some of the explanations are just so stories (in which case they wouldn't be explanations) b) claim that some explanations are for adaptive changes – ipavlic Apr 10 '12 at 11:45
@Jonas ...and I don't see how is b) related to the original question. All right, some explanations are for adaptive changes, but are they only possible explanations for those adaptive changes (in that case, they wouldn't be "explanations" by your definition, but "speculations")? – ipavlic Apr 10 '12 at 11:46
@ipavlic: Does my edit address your concerns? – Jonas Apr 10 '12 at 13:17

Lagerbaer · Answer 2 · 2012-04-09T16:49:14.373

Evolution itself is a fact; the theory part is the mechanism behind it all. Every trait you see in a living being has had to evolve. Therefore, it is only natural that every trait has an evolutionary explanation. See, e.g., here: http://www.talkorigins.org/faqs/comdesc/

The example you bring makes an attempt at ridiculing evolution, with the rationale being "Since it explains both cooperation and fighting, clearly it is wrong or just arbitrary". That is a fallacy. We don't cooperate and fight at the same time. The same humans that cooperated in hunting down a mammoth would later fight for a female, or for clan leadership. So what?

The claim that evolutionary explanations are "just so stories" is also taken up at talkorigins here: http://www.talkorigins.org/indexcc/CA/CA500.html

Also, evolution cannot explain every traits. Only those that we observe in the real world. If we'd found rabbits in the precambrian, for example, we'd immediately know that we've got something wrong concerning the origin of species. See here: http://www.talkorigins.org/indexcc/CA/CA211.html

Evolution also would be unable to explain traits that are so destructive and detrimental that they cannot be remedied or outweighed by other traits. But it is quite obvious that a species with these traits would go extinct in a very short time. So again, evolution is consistent here in that it can tell us why we see certain traits, and why we don't see certain others. Like the exploding Swamp Dragons in Terry Pratchet's Discworld universe...

@Lagerbaer is correct. Evolution is both a theory and a fact. The fact of evolution comes from the universal observation of evolution in all scientific fields, and the theory of evolution by natural selection explains one of the mechanisms by which it occurs. See here for a more in-depth explanation. — tak, Apr 17 '11 at 00:34

score 6 · Answer 3 · answered Apr 17 '11 at 13:27

There are "just-so" stories that are amenable to testing, but after having been tested cease to be just-so stories.

To take one of the examples cited, kin selection, or more broadly inclusive fitness, started life as a conjecture. Although you can derive the basis of inclusive fitness from population genetics whereby if the cost of an action is less then the benefit of to others de-rated by the degree of genetic relationship through common descent (which is always a fraction less than 1) and a mutation conferring that trait happens to arise, it will increase its proportional representation within the gene pool. That's a mathematical statement made in somewhat twisted prose, the formalism is simpler:

rB > C

where r is the degree of relatedness (0.5 for full siblings, 0.125 for cousins, etc.), B is the benefit and C is the cost.

Sherman's 1977 fieldwork (Science 197:1246) was seminal in testing the predictions of this model. They observed ground squirrels and alarm calls made when a predator is spotted. Not only did they find that the alarm call increased the risk of being eaten by the caller (C) but that their likelihood of making such alarm calls was highly correlated with the presence of nearby kin (r) who could benefit (B) by hiding. Conversely in the absence of nearby kin, alarm calls were less likely to be given. This is strong confirmation for the inclusive fitness model and part of the corpus that has made it no longer a just-so story from the time the eminent Sir R. A. Fisher conjectured the mathematical model in 1964.

Okay, but those are squirrels. How about humans, which is what we really care about.

In incredibly detailed and careful methodology, Daly and Wilson (1998 and 1994) showed that the predictions of inclusive fitness were demonstrated in differential rates of violence of step-parents toward their children than by natural parents. To make a long story short, infants were more than 100 times as likely to be killed at the hands of a step-parent. So there is a just-so story of the evil step-parent going back in folk-tale for milennia gets tested and is found to be a feature of today's human world.

How about the sexy son? Here's a just-so story: if women want attractive sons but good, providing cuckolded "fathers", they might just adopt a strategy whereby they increase the likelihood of being inseminated by a sexy, but non-parental man. How the hell could you test that? Baker and Bellis (1993) found a way, but it wasn't delicate. They measured sperm retention by women and their orgasmic frequency with their nominal partner and their extra-pair copulations. They looked at uterine "upsuck" in relation to female orgasm and even the ovulatory timing of extra-pair copulations. As it happens, when women do seek outside boinking, they time it to match their fertility cycle and cum harder. This fits rather nicely with the reliable 10% incidence of first-world contemporary children not being sired by their putative fathers.

Just-so stories can be tested, and continue to be. This tends to move them out of the realm of Kipling and into the realm of "we know".

The usual definition of a “just-so” story is precisely that it’s untestable. Untested conjectures are not in general just-so stories. — Konrad Rudolph, Apr 17 '11 at 13:51
Agreed. However, yesterday's "just-so story" has a frequent habit of turning into today's testable hypothesis. I'm sorry that I didn't make that more clear. — msw, Apr 17 '11 at 13:58
@msw That was precisely my point: yesterday’s just-so story cannot turn into a testable hypothesis because it’s untestable on principle. A just-so story is a logical fallacy. If a hypothesis turns out to be testable after all then it was actually never a just-so story; it was incorrectly called so. — Konrad Rudolph, Apr 17 '11 at 14:02
How does it fit nicely with the 10% incidence? Why 10% instead of 5% or 20%? A good scientific theory should give you an estimate about the size of the effect. — Christian, Jun 08 '11 at 16:29
As the "10%" was incidental to the point made and not at all a prediction, please substitute "non-zero". Based on your profile here, I expect you well understand that not every model predicts every aspect of the modeled system. — msw, Jun 11 '11 at 02:25

score 0 · Answer 4 · answered Apr 09 '12 at 19:04

No, but scientists aren't immune to spreading "just-so" stories.

Karl Popper's influential book The Structure of Scientific Revolutions would agree that that, in any area, the scientific establishment will try to shoehorn newly-discovered anomalies and shortcomings into the old worldview ("paradigm"). According to Popper (whose model is not universally accepted), the establishment will resist change even when its shortcomings seem, in retrospect, obvious. For instance, it seems odd to us that a model of planets moving along wheels-within-wheels-within-wheels was thought to be more reasonable than Kepler's model where they move along ellipses. More recently, it seems strange that the key-and-lock shapes of South America and Africa were dismissed as coincidences as recently as the 1960s, before plate tectonics became accepted.

In retrospect, the increasingly implausible defences of the old paradigm ("add another epicycle to the planetary motions!", "Mid-ocean mountain range? Just another coincidence!") might be called "just-so" stories. It's an insulting viewpoint, since it makes it sound like the defenders are willfully lying, which they're not. Our worldview (our "paradigm") is filled with unexamined assumptions and conclusions; we've invested hundreds and thousands of hours of time into understanding the way things work in this worldview and we are really good at defending it. We get caremad.

Importantly, though, the scientific ideal is accuracy, not popularity or tradition. Science does have paradigm shifts. Celestial spheres gave way to Kepler's ellipses gave way to Einstein's model. Plate tectonics are accepted.

Having said all that, is evolutionary theory in a paradigm-shifting "crisis"? No.

There is a lot of hubbub about altruism lately and the Haldane/Hamilton "kin selection" paradigm, but there's nothing about the argument that presents any kind of challenge to evolutionary theory in general.

Are evolutionary explanations "Just so stories"

4 Answers4

Linked