Feathers: induction and the scientific method

[Tyrannosauron]

There's still some lingering enthusiasm over the feathered dino debate from the "best Oviraptor" thread in the Dino toys folders, so I figured that it would be best to start a new thread here before more people get in trouble.
Here are some of the last few relevant posts that were left in that thread:

Sept 8, 2010 8:56:52 GMT paleofreak said:

Sept 7, 2010 21:25:53 GMT Tyrannosauron said:

but to say that the inferences we've drawn about dinosaurs as a result qualify as knowledge is just unscientific

It's scientific knowledge, yes, with its limitations, open to corrections. Not "dogmatic", of course.



OK. But you appear to make a big separation between non avian dinosaurs and avian dinosaurs, non avian being "mysterious". The "mystery" should be a matter of degree. We didn't observe any "living" Hesperornis or Concornis (avian), but you probably consider them less "mysterious" than Caudipteryx (non avian, initially considered as avian), and Caudipteryx less "mysterious" than Stegosaurus, wich is a *#€$~ monster :-)



Keep in mind that many birds of modern aspect and classification (like Vegavis) were already living in the Late Cretaceous. To be fair, you should compare Oviraptor with an ancient "non mysterious" bird. Bah, that's a lot of million years anyway. Forget it :-)



I don't know... but I would probably decide after considering the size and the environment of the animals to estimate the probability of hair or feather losing. Hyracotherium was a bit smaller in weight, and lived in woodlands. Oviraptor in an arid environment ? I think I would put some money on a partially naked, partially feathered Oviraptor.

I'm not saying that there isn't good evidence for nonavian dinosaurs having feathers; it's just that if you're relying on taxonomy alone, there's a greater degree of certainty in drawing conclusions about early mammals, which was the initial question.

Ok, your point is good, and I think you are right in a big part. But I'm not considering just taxonomy. I'm considering phylogeny, evolution, parsimony. Occam's razor... I think the distance between taxons or the "mystery" is not very relevant here. What is really relevant is the scientific evidence about what is the primitive, "default" feature in each case: fur in mammals, a feather covering in Maniraptora.

Sept 8, 2010 9:29:40 GMT blackdanter said:

Sept 7, 2010 20:59:13 GMT gwangi said:

I'm not going to accept that certain species didn't have feathers when there is no evidence otherwise, especially considering that their closest relatives had them. Seriously, there is no evidence that any oviraptor had scales...or dromaeosaur for that matter. There is evidence for feathers, so aren't you basically accepting evidence that isn't there? Or perhaps bald dinosaurs are more popular after all.

It remains a matter of preference. There doesn't appear to be any physical evidence that this species, Oviraptor, was feathered or scaled. Therefore there are no evidential grounds for saying that one representation with feathers is more accurate than one with scales. I've trawled for the evidence and literally what it comes down to is that Oviraptor is very similar to something else and therefore was likely feathered. That's not scientific fact based on physical evidence it's scientific theory alone and certainly wouldn't win a case in a court of law. Believe me, If the actual evidence were there I'd happily go for the feathered representation however as things stand I personally prefer the unfeathered

Sept 8, 2010 10:47:45 GMT paleofreak said:

Sept 8, 2010 9:29:40 GMT blackdanter said:

It remains a matter of preference. There doesn't appear to be any physical evidence that this species, Oviraptor, was feathered or scaled. Therefore there are no evidential grounds for saying that one representation with feathers is more accurate than one with scales.

There is also no physical evidence that Smilodon was furry or scaled or feathered, or covered with long spikes, or multicolor tentacles. Therefore there are no evidential grounds for saying that one representation is more accurate than the others. So let's draw a tentacled Smilodon

Sept 9, 2010 0:48:58 GMT gwangi said:

Sept 8, 2010 9:29:40 GMT blackdanter said:

It remains a matter of preference. There doesn't appear to be any physical evidence that this species, Oviraptor, was feathered or scaled. Therefore there are no evidential grounds for saying that one representation with feathers is more accurate than one with scales. I've trawled for the evidence and literally what it comes down to is that Oviraptor is very similar to something else and therefore was likely feathered. That's not scientific fact based on physical evidence it's scientific theory alone and certainly wouldn't win a case in a court of law. Believe me, If the actual evidence were there I'd happily go for the feathered representation however as things stand I personally prefer the unfeathered

Well of course it is a theory, (evolution is a theory too BTW) but it is the best theory we have in regards to how these animals looked. No evidential grounds? Really? Oviraptors are nestled within Oviraptorosauria along with the feathered Caudipteryx and Protarchaeopteryx among others. Quill knobs have been found on Avimimus and some species have a pygostyle at the ends of their tails. The positions of the fossils on their nests suggests a bird-like brooding posture where feathers would be beneficial. Oviraptors are also nestled within Maniraptora which contains birds and most of the other feathered dinosaurs we know about. This is evidence, evidence that allows me to theorize that oviraptors had feathers. What evidence is there that oviraptors lacked feathers? what evidential grounds? Skin impressions from unrelated hadrosaurs? If a relative of oviraptor closer to it than caudipteryx were discovered with scale impressions then I would theorize oviraptors had scales but that evidence does not exist. Scaly oviraptors are purely a matter of personal preference, there is no scientific evidence to support them.
As for the "hairless" human comparison so often made in these arguments...are you kidding me? Humans have hair all over their bodies. Shorter and lighter than in other mammals but still hair. Just look at the tops of our heads, imagine that humans never cut our hair or shaved. I don't feel I should have to tell this to fellow humans.

[Dinotoyforum]

Thanks Tyrannosauron

[Horridus]

It is worth mentioning that humans are actually really, really hairy....well, some of us are....(with apologies for horrifying everyone)

[blackdanter]

Sept 9, 2010 15:28:47 GMT Horridus said:

It is worth mentioning that humans are actually really, really hairy....well, some of us are....(with apologies for horrifying everyone)

You're drinking too much Guinness You should see where constant Guinness consumption puts hair on the ladies

[Horridus]

Sept 9, 2010 18:32:25 GMT blackdanter said:

Sept 9, 2010 15:28:47 GMT Horridus said:

It is worth mentioning that humans are actually really, really hairy....well, some of us are....(with apologies for horrifying everyone)

You're drinking too much Guinness You should see where constant Guinness consumption puts hair on the ladies

Haha! I hardly drink any Guinness. Although I do favour bitters and stout...

[Griffin]

Haha wow I had no clue such a debate was going on! Everytime I saw there was a new post under that thread I just assumed people were really enthusiastic about non-retired oviraptor toys.

[Tyrannosauron]

I'll start with the old "evolution is just a theory" canard. The phrase can be interpreted in two ways: one false, the other trivially true. If by "evolution" one means a change in gene frequencies over time, then evolution is as much a fact as any other observation about the external world: everyone uses the example of the flu, but that's because it's such an intuitive and powerful illustration of the fact. On the other hand, if by "evolution" one means change due to natural selection, then evolution is "just" a theory--because science by definition cannot do any better. As was admitted above, what some people call "scientific knowledge" is always fallible and open to revision, which is to say that it doesn't qualify as knowledge (i.e. justified true belief) at all. This is why gravity is "just" a theory; if anything, there's less debate among practicing biologists about natural selection as a mechanism of evolution than there is among practicing physicists about the mechanism of the gravitational force. Nevertheless, it would be entirely unreasonable to believe that the next item I throw up into the air won't fall down, as there are very few examples of similar beliefs being proven correct.

And then there are feathered dinosaurs. The fact is that the only evidence for the appropriateness of restoring certain dinosaur genera with feathers is taxonomic (a field that includes phylogeny, which makes use of the principle of parsimony to determine the course of evolution--so, ultimately, everything suggested so far comes down to taxonomy). Like I've said--repeatedly--it's reasonable to infer that, say, Deinonychus antirrhopus had feathers. However reasonable it may be, however, the conclusion must be due to inference from the appearance of closely-related taxa, as there's no direct evidence of feathers in that particular species. The point of the human-nonhuman ape comparison is that there are very obvious examples of how inference of appearance from one taxon to another, no matter how closely related, may fail (and calling attention to human hair is disingenuous at best: to deny the very obvious differences in integument between humans on the one hand and chimps and gorillas on the other is to substitute sophistry for observation). Heck, that sort of inference fails between different populations within a single species, as anyone who's ever seen a Mexican Hairless next to an Afghan Hound may attest. The point is that we have quite a few examples of how the sort of taxonomy-based reasoning used to justify restoring every maniraptoran dinosaur with feathers fails, and so our confidence in the accuracy of those restorations should be adjusted accordingly.

Put it another way: let's assume that there are 1000 species (a real lowball estimate) within the Maniraptora (not counting Aves). As reasonable as it is to infer that any of those species had feathers, it would be just as reasonable--given the examples discussed above--to infer that at least one did not have feathers. We can say with relative certainty that some species--Velociraptor mongoliensis, for example--will not be that exception to the rule. What we can't say with any certainty is which species *could be* that one exception. Any claim to *know* that a nonavian dinosaur species had feathers despite an absence of direct evidence is therefore objectionable on those grounds. It's fine to restore maniraptorans with feathers, but that practice has to be tempered by the understanding that there is likely to be some case in which doing so is less accurate than the alternative.

[dinoguy2]

Put it another way: let's assume that there are 1000 species (a real lowball estimate) within the Maniraptora (not counting Aves). As reasonable as it is to infer that any of those species had feathers, it would be just as reasonable--given the examples discussed above--to infer that at least one did not have feathers. We can say with relative certainty that some species--Velociraptor mongoliensis, for example--will not be that exception to the rule. What we can't say with any certainty is which species *could be* that one exception. Any claim to *know* that a nonavian dinosaur species had feathers despite an absence of direct evidence is therefore objectionable on those grounds. It's fine to restore maniraptorans with feathers, but that practice has to be tempered by the understanding that there is likely to be some case in which doing so is less accurate than the alternative.

It's interesting you put it that way. There are over 9,000 species of maniraptoran around today. By your logic, it would be safe to assume that most have feathers, and a few do not. So where are all the featherless birds? There are none. Why is it likely to be a few featherless maniraptorans among the 200 or so known species, when there isn't even a single oddball exception among the nearly 10,000 living species?

More important question: WHY have no birds lost their feathers, while so many mammals have?

Answer: Feathers are simply better at just about everything than hair. Better insulators, better display structures (can carry more shades and more complex colors, hair is incapable of being anything other than blond to red to black). Better for covering nests (which mammals don't need to do). Better at aerodynamics (useful even in flightless running birds like ostriches.

Given what we know about birds, and given that the difference between "birds" and "other maniraptorans" is 100% an arbitrary human delineation with no basis in any physical characteristics, all maniraptorans should be presumed feathered until such a mythological "exception to the rule" ever turns up.

[Horridus]

I can't imagine any case in which restoring a maniraptoran dinosaur with no feathers would be 'more accurate' than restoring it with even a small number (eg. Luis Rey's vision of Therizinosaurus), for the reasons as explained by Dinoguy2 (in a far more eloquent fashion than I could manage).

[paleofreak]

Sept 10, 2010 3:56:47 GMT Tyrannosauron said:

The fact is that the only evidence for the appropriateness of restoring certain dinosaur genera with feathers is taxonomic (a field that includes phylogeny, which makes use of the principle of parsimony to determine the course of evolution--so, ultimately, everything suggested so far comes down to taxonomy).

Err.... no. Taxonomy is classification, and phylogeny is ancestry and kinship. There are multiple ways of classify, multiple possible taxonomies, but there is only ONE TRUE phylogeny of the organisms. Phylogenetics is the study of phylogeny. Modern taxonomy is based on Phylogenetics. Taxonomy doesn't "include" phylogeny, but depends on phylogeny.

[Himmapaan]

Sept 10, 2010 10:54:25 GMT Horridus said:

I can't imagine any case in which restoring a maniraptoran dinosaur with no feathers would be 'more accurate' than restoring it with even a small number (eg. Luis Rey's vision of Therizinosaurus), for the reasons as explained by Dinoguy2 (in a far more eloquent fashion than I could manage).

Merely on the grounds that we do not yet know for certain, I suppose. Note that I speak as someone who does strongly support the inference of the presence of feathers in all maniraptorans; but I do agree that Tyrannosauron has a very good point. At best, it would be almost a kind of arrogance to truly claim it as 'real knowledge'. Yes, we can with more than reasonable deduction presume that they were all feathered, but doing so with that degree of caution would, I think, still be an admissible and admirable thing.

[Tyrannosauron]

Sept 10, 2010 4:44:08 GMT dinoguy2 said:

It's interesting you put it that way. There are over 9,000 species of maniraptoran around today. By your logic, it would be safe to assume that most have feathers, and a few do not. So where are all the featherless birds? There are none. Why is it likely to be a few featherless maniraptorans among the 200 or so known species, when there isn't even a single oddball exception among the nearly 10,000 living species?
More important question: WHY have no birds lost their feathers, while so many mammals have?
Answer: Feathers are simply better at just about everything than hair. Better insulators, better display structures (can carry more shades and more complex colors, hair is incapable of being anything other than blond to red to black). Better for covering nests (which mammals don't need to do). Better at aerodynamics (useful even in flightless running birds like ostriches.
Given what we know about birds, and given that the difference between "birds" and "other maniraptorans" is 100% an arbitrary human delineation with no basis in any physical characteristics, all maniraptorans should be presumed feathered until such a mythological "exception to the rule" ever turns up.

That did actually occur to me as I typed up my last post, but it ultimately doesn't change the point that I'm trying to make very much, for two reasons: first, it presumes that we know the function(s) that feathers served when they first evolved; second, that we've proved that no featherless avian exists. The first proposition is clearly false: the origin of feathers is slowly becoming clearer, but we still aren't sure when feathers first evolved, much less why*. The second proposition is open to the very same objections that I've been raising all along. It would be unreasonable to believe that the next species of bird that we discover will be featherless, but deductive logic--the only basis of infallible knowledge--affords us no way to ever prove that belief true. This is why developing a theory is the best that Science can do: inductive logic is necessarily fallible and therefore cannot qualify as knowledge in any strict sense of the word.

*My understanding is that some paleontologists are now speculating that feathers might have evolved within the Archosauria before the evolution of the Dinosauria; if so, it would seem reasonable to believe that at least a few dinosaur genera would prove exceptions to any proposed "feathers never disappear from a lineage" rule (although feathered sauropods would be awesome, if unlikely).

Sept 10, 2010 11:17:51 GMT paleofreak said:

Err.... no. Taxonomy is classification, and phylogeny is ancestry and kinship. There are multiple ways of classify, multiple possible taxonomies, but there is only ONE TRUE phylogeny of the organisms. Phylogenetics is the study of phylogeny. Modern taxonomy is based on Phylogenetics. Taxonomy doesn't "include" phylogeny, but depends on phylogeny.

Taxonomy is organization. As you said, there are multiple possible ways to do taxonomy: for example, Aristotle's Organon presents a taxonomy of logical propositions, Bloom developed a taxonomy of learning tasks, and a colleague of mine is working on a taxonomy of evolutionary biology publications. Phylogeny is the organization of taxa according to evolutionary history. Simple syllogism: all organization is taxonomy and phylogenetic reconstruction is organization; therefore, phylogeny is taxonomy.

[paleofreak]

Sept 10, 2010 12:54:52 GMT Tyrannosauron said:

Phylogeny is the organization of taxa according to evolutionary history.

That's phylogenetic taxonomy, not phylogeny.

Simple syllogism: all organization is taxonomy and phylogenetic reconstruction is organization; therefore, phylogeny is taxonomy.

But not all organization is taxonomy. Taxonomy is just one kind of organization, it is the practice of classification of things. When you do phylogenetic reconstruction you are not (necessarily) classificating organisms. You are looking if A is more close to B than to C, etc. Then you may classify (taxonomy) A, B, C in groups. If you make those groups based on phylogeny (and using certain rules and methods), then you do phylogenetic taxonomy.

[Tyrannosauron]

Sept 10, 2010 13:57:27 GMT paleofreak said:

Sept 10, 2010 12:54:52 GMT Tyrannosauron said:

Phylogeny is the organization of taxa according to evolutionary history.

That's phylogenetic taxonomy, not phylogeny.

Simple syllogism: all organization is taxonomy and phylogenetic reconstruction is organization; therefore, phylogeny is taxonomy.

But not all organization is taxonomy. Taxonomy is just one kind of organization, it is the practice of classification of things. When you do phylogenetic reconstruction you are not (necessarily) classificating organisms. You are looking if A is more close to B than to C, etc. Then you may classify (taxonomy) A, B, C in groups. If you make those groups based on phylogeny (and using certain rules and methods), then you do phylogenetic taxonomy.

This is ultimately semantics, although I'll concede that I probably equivocated between practice and product in my earlier posts. I don't think it changes my points.

Sept 10, 2010 11:29:39 GMT Himmapaan said:

Sept 10, 2010 10:54:25 GMT Horridus said:

I can't imagine any case in which restoring a maniraptoran dinosaur with no feathers would be 'more accurate' than restoring it with even a small number (eg. Luis Rey's vision of Therizinosaurus), for the reasons as explained by Dinoguy2 (in a far more eloquent fashion than I could manage).

Merely on the grounds that we do not yet know for certain, I suppose. Note that I speak as someone who does strongly support the inference of the presence of feathers in all maniraptorans; but I do agree that Tyrannosauron has a very good point. At best, it would be almost a kind of arrogance to truly claim it as 'real knowledge'. Yes, we can with more than reasonable deduction presume that they were all feathered, but doing so with that degree of caution would, I think, still be an admissible and admirable thing.

That's exactly it! It's not a very strong claim and I'm surprised it's being met with such resistance. One can never be completely certain about an inductive inference, although some entail a higher degree of certainty than others. We may be somewhat certain that many--perhaps most, if not all--species within the Maniraptora (if not the whole of the Dinosauria) were feathered, but however certain we are of that, the process of induction entails that there are some paleontological inferences about which we can be more certain.

[foxilized]

Sept 10, 2010 22:33:18 GMT Tyrannosauron said:

It's not a very strong claim and I'm surprised it's being met with such resistance. .

I guess the resistance won't awake for any inconsistence in your point at all. I would say it's more human resistance to accept the uncertainty. Simply because holding uncertainty is quite difficult, meanwhile grabbing a certainty -no matter if not really real- brings some sort of peace of mind. (Er, I am not giving any example of this kind of thing happening through human history, but well, some great business have been made selling "false certainties" to avoid reality).

Fact is reality is more about uncertainty and mystery. If we can accept that then we are way closer to "a truth", whatever that "truth" is.

Great thread, Tyranno.

[gwangi]

Sept 10, 2010 3:56:47 GMT Tyrannosauron said:

Any claim to *know* that a nonavian dinosaur species had feathers despite an absence of direct evidence is therefore objectionable on those grounds. It's fine to restore maniraptorans with feathers, but that practice has to be tempered by the understanding that there is likely to be some case in which doing so is less accurate than the alternative.

I don't think I've seen anyone make that claim. Like I said in my previous post I am aware that these are theories and may be wrong but based on the evidence it is the best we've got and quite convincing. Sure, oviraptors may not have been feathered but based on the evidence we have it stands a much more likely chance of being feathered then is does being scaly.