Saturday, 11 May 2013

Hypothesis or Proof?

Premiering back in 1980, Carl Sagan’s Cosmos led us on an enthralling tour of our universe and took us on a crash course of science. We watched as he strolled around the heavens and our planet as he guided us through the very essence of what it meant to be human in the natural order of things. Sagan was a wonderful guide who was clearly ahead of his time, highlighting such issues as the possibilities of nuclear Armageddon and global warming and there are many quotes by Sagan that are still cited today.

Chief amongst them and most often used by palaeontologists of today, is this one:

“The absence of evidence is not the evidence of absence.”

Most people would tend to agree with this, as would I, and yet should we? On the face of it this is a very strange concept since we are suggesting that we should accept certain issues as a given despite a lack of physical proof. In other words perhaps what we should actually be saying is:

“The absence of evidence is the evidence of absence.”

Let me explain what brought this on. The well documented, and almost universal, criticism of the makers of the already much hyped Jurassic Park 4 for not having feathered dromaeosaurids is scientifically well founded since we have extensive fossils now of these dinosaurs clearly displaying such coverage. It is bizarre that the producers have seen fit not to feather their raptors and their argument that it is best to maintain a form of continuum from the previous films is a weak excuse. Perhaps it is a blind to throw us off the scent and they will indeed unleash feathered dromaeosaurs but I fear not.

Then last week the trailer for the new Walking with Dinosaurs 3D movie went on line globally and generally met with a pretty good reaction. Scott Hartman, over at Skeletal Drawing, naturally enough featured a post about it since he was involved as one of the anatomy design team for the production. One of the comments on his blog mentioned the fact that there were no shaggy gorgosaurs in the trailer and Scott suggested that the movie world is not quite ready for that yet and I tend to agree with that statement.

In the comments section I pointed out the following:

And yet, technically, the absence of shaggy gorgosaurs is correct since Tyrannosauridae is still fuzz free for now albeit phylogenetically likely. We cannot really moan about the film makers in this instance because they are depicting tyrannosaurids accurately - in as much as the fossil record permits.”

To which Scott replied:

”That doesn't necessarily follow - absence of evidence is not evidence of absence here. The phylogenetic data is really the only data we have, so it's what we should be relying on.”

Perfectly reasonable comment and yet my initial reaction was to be a little bemused. We can justly criticise film makers for not including feathered dromaeosaurs because there is ample physical proof and yet can we justify criticism of film makers for not including feathered tyrannosaurids when there is no actual physical evidence currently available?

The phrase itself demands closer scrutiny. In effect, and in the simplest of terms, the following shows a classic example:

Many people take glucosamine and chondroitin tablets to ward off the effects of osteoarthritis (OA) and yet there is no solid data that there is any form of relationship between the two ingredients and the prevention of OA. Therefore absence of evidence is clear but, in reality, those who manufacture or believe in the properties of these tablets would simply point out that just because there is no proof currently available it does not mean that the tablets do not work – therefore absence of evidence is not evidence of absence.

Now I confess that this has very little relevance in palaeontological terms but it does show the dichotomy and how easy it is to manipulate the lack of data into possibility – even probability. However, in the case of tyrannosaurids, we have a completely different picture. I suggested that since there is no physical proof of feathers or fuzz in tyrannosaurids then we should not necessarily take it as a given simply because, and as Scott pointed out, the phylogenetic data is highly indicative that they were.

So in this case, does the absence of physical evidence mean that there is no evidence for feathered tyrannosaurids at all? Well of course not – in fact virtually the whole coelurosaurian clade displays integument of one form of another and Tyrannosauridae is currently the last group devoid of physical evidence. Very important here – we are referring to tyrannosaur-ids and not tyrannosaur-oids of which there are indeed animals displaying integument of which the most famous in recent times is Yutyrannus (if, in fact, it is a tyrannosauroid).

So we have lots of phylogenetic data to support the probability of feathered tyrannosaurids but this still needs to be quantified and rigorously tested. When this is done then the dataset will provide a statistical probability of confidence in the hypothesis. If there is confidence in the dataset then you must turn the process on its head and try to prove that there is NOT a probability of feathered tyrannosaurids. This is an extremely important part of the process and lends substantial credence to the theory.

As a result, the phylogenetic signal is highly indicative that tyrannosaurids were indeed covered in some form of fuzz or feathered plumage. And yet for all that detail, all that probability – indeed all of this “evidence” – there is still no proof. And this is where we have to be a little careful since the available data is so compelling that we sometimes forget that physical evidence is required for confirmation.
BHI 6230 (Wyrex) - No feathers here...
The very essence of all science is based on this fundamental process and now the onus is on finding fossils of tyrannosaurids that may preserve impressions of fuzz/feathers. There are many specimens already in collections that need re-examining as well as those specimens that may still reside in their jackets and, of course, there are still many wonderful fossils out there awaiting discovery. And, if it exists, then we should be able to find the evidence since our detection techniques are much more advanced these days – especially with UV light.

But, for the sake of argument, what if we never find a feathered tyrannosaurid? What then? I mean we have the phylogenetic data but still no proof. So we will rerun our data, perhaps add more updated information and check and double check – but still no fossil. And then we will question ourselves again since we keep performing the science and keep coming out with the same results. And then if we still find no fossils then what? Perhaps we have to consider the possibility we may be wrong.

But this is unlikely to happen and we will continue to search for feathered tyrannosaurids because they must exist – the phylogenetic data insists it to be true. And this is the point – we must not then fall into the trap of “just because there are no feathered tyrannosaurids, this does not mean they did not exist”.

“Absence of evidence is not evidence of absence” must not become a given in science – not without the proof to back it up. Circumstantial evidence, no matter how strong, can lead to the wrong conclusion and we must all be aware of this. What this lack of feathered tyrannosaurid fossils does indicate, however, is that they are, if they exist, at the very least exceptionally rare but this, again, is not actually proof!

There are so many variables as well – sampling and preservational biases are probably the biggest issues here. It is also worth pointing out that the phylogenetic dataset also has vast gaps in it although there is still a wealth of relevant data available. As usual more fossils are required – especially when you consider how few tyrannosauroids are actually known throughout the Cretaceous.

So “absence of evidence is evidence of absence” is kind of true but, at the same time, is as equally fallible as the other. I accept the near certainty that tyrannosaurids had some form of integument but it is not enough to merely want to believe. In the end we have the data, we have the hypothesis and all we need now is the proof.


Andrea Cau said...

"Very important here – we are referring to tyrannosaur-ids and not tyrannosaur-oids of which there are indeed animals displaying integument of which the most famous in recent times is Yutyrannus (if, in fact, it is a tyrannosauroid)."
I'm a bit confused by that phrase.
Are you claiming that the presence/absence of a tegument in a clade (or the inference of tegument) may be linked to the name/rank/suffix of the clade?
For example, since Eotyrannus is a basal tyrannosauroid, then the presence of feathers in Eotyrannus is more supported than in Gorgosaurus? Both lack direct evidence of feathers, both are bracketed in the same way by feathered taxa, regardless to the name of the least inclusive clade they are referred to. The inference of tegument in the two taxa is the same.
The terms 'Tyrannosaurid' and 'Tyrannosauroid' are arbitrary and refer to TAXONOMIC units, and not on their DIAGNOSES (including the type of tegument). So, that taxonomic distinction means nothing in the context of feathers/scales distribution.
Perhaps I'm wrong in including you among them, but I've noted that many people continue to fix the presence/absence of morphological features to the name of the clade the species is referred to, following a sort of pre-Darwinian typological thinking that persists in the XXI century. Like most continue to think that Reptilia=Scaly and Aves=Feathered, although we now know that Aves is inside Reptilia and many non-Aves are feathered.

Hadiaz said...

"BHI 6230 (Wyrex) - No feathers here..."

Any idea what part of the body that skin impression is from? Just wondering.

Mark Wildman said...

Hi Andrea and, as always, thanks for the comments. My tyrannosaurid/tyrannosauroid partitioning is a lot more simplistic than that I'm afraid. As I see it there is obvious integument in Tyrannosauroidea - that is an irrefutable fact.

By that mark, then both Eotyrannus and Gorgosaurus were likely covered with integument and, again, I have no issues with that. I do take issue, however, with the predominant belief that all tyrannosauroids were covered with fuzz without the physical evidence to support it.

I do not mean to come across as being "pre-Darwinian" - I certainly have faith in cladistics and phylogenetics as much as the next man. I am very happy with the phylogenetic bracketing as you described but I cannot accept every hypothesis as fact - not without the evidence.

For now, there are too many gaps in the fossil record between the Early Cretaceous tyrannosaurs and their giant cousins in the Late Cretaceous to make such assumptions. But I do think the truth may be closer especially now that the feathered ornithomimids from Alberta have been identified and in rocks that contain the bones of tyrannosaurids.

The post is simply my way of saying that I agree with the implications but lets just wait for the proof. If this was a behavioural implication it would have got slaughtered by now.

Herman - I understand that most of the Wyrex skin impressions are from the underside of the tail. Not ideal I grant you but interesting all the same.

Andrea Cau said...

Note that I'm not endorsing any particular interpretation for tyrannosaurid integument (my personal preference is for a combination of scales and feathers, not for a simple dichotomy black vs white: scales vs feathers) or for a particular distribution of teguments around the body. I'm just noting that many follow (or challenge) a too semplicistic version of the phylogenetic bracketing. Carrano and Hutchinson (2002) on hindlimb and pelvic musculature inference in Tyrannosaurus is a very good example of how phylogenetic bracketing works: it's not just a simple series of rough statements like "since C is bracketed by A and B, then showed feature X", instead it's a rigorous methodology that allows us to classify hypotheses and reconstructions based on phylogeny, direct fossil evidence and correlates of soft tissues present in bones. Not all levels of inferences produced by the phylogenetic bracketing method have the same 'robustness', and this 'ranking' is probably the most important product of the method, more than the mere assertion of presence/absence of features.
I agree that there are many gaps in the fossil records, but this is true for ALL fossils: the skeletal, dimensional and temporal disparities between Yutyrannus and the sparrow are larger than those between Yutyrannus and Tyrannosaurus, thus I find no reason to believe that the tegument of Tyrannosaurus was very different from the one shared by Yutyrannus and the sparrow (two dinosaurs with an important feathered covering): this is the simplest explanation based on what we know, and science has to follow the simplest explanation, pending new evidence that may falsify it.
Why should we stop to use a scientific method of inference when we talk about tyrannosaurs? Are they special? Were they supernatural beings?

Mark Wildman said...

Well said Andrea - couldn't agree more. But this is the very essence of what I have been trying to explain (although you have put it a far clearer and succinct way).

"Why should we stop to use a scientific method of inference when we talk about tyrannosaurs? Are they special? Were they supernatural beings?"

Again I agree - I am not disagreeing or contradicting the science. I merely state that you cannot ASSUME anything because phylogenetic bracketing indicates the presence of any characteristic. The presence of some form of covering in tyrannosaurids is likely - probable even - but unproven. That is all I am saying.

Hadiaz said...

"Herman - I understand that most of the Wyrex skin impressions are from the underside of the tail. Not ideal I grant you but interesting all the same."

That's too bad. I was hoping they were from Wyrex's topside (IIRC, Juravenator had scales on its underside & protofeathers on its topside).

Andrea Cau said...

"The presence of some form of covering in tyrannosaurids is likely - probable even - but unproven."
The same is true for the heart, liver, eyes: no direct evidence. There are no fossil evidence of these organs for any single tyrannosauroid fossil, including Dilong and Yutyrannus.
In absence of direct evidence, the presence of these organs is UNPROVEN, as you wrote for feathers, and we assume their presence by phylogenetic bracketing. Paradoxically, based on evidence alone, and not using phylogenetic bracketing, the presence of feathers is more robust than the presence of the heart: we have direct evidence of tyrannosauroid feathers, but not just a single tyrannosauroid heart!

Mark Wildman said...

Just read your blog post Andrea - excellent and I'm glad I read it before commenting here. At first I thought your heart reference was strange but taking it into context I fully understand where you are coming from.

I guess it is easier to accept that tyrannosaurs had a heart since all animals have a heart in one form or another - but not all animals have feathers. Still, it would be nice to have that feathered tyrannosaurid turn up one day just to seal the deal. We will just have to be patient.

Thanks for taking the time to discuss and also for the very informative blog post.

Andrea Cau said...

Thank you for this post, that was the inspiration for mine!

Anonymous said...

Well done Mark--along the writing on tyrannosaurid forelimbs, you summarize equally well about the question of feathery--it is a "fuzzy" subject.

It is nice with T. rex scale impressions--then again, there is the suggestion that they were only partially feathered (like G. S. Paul, who have added feathery "wings" on some old tyrannosaurid drawings on his website... no wonder that some suggest the arms were used in sexual attraction).

On absence of evidence: have you read the Yutyrannus report on the Discover magazine blog (written by Ed Young, 4/4 2012)? It quoted T. Holtz, who said that T. rex fossils often are found in sandstone (not good for preserving fine details). It also quoted P. Sereno--he claim to have a T. rex fossil showing they were NOT scaly--but that might just be a distortion of his words?

Mark Wildman said...

Thanks for the comments again. The sandstone preservational factor IS an issue but now not as big a concern as maybe it once was. This was why I mentioned the feathered ornithomimids that were recovered from Drumheller in Alberta which were found in sandstone and, in these very same sandstones, are found tyrannosaurids so perhaps it is now only a matter of time.

I am not familiar with the Sereno claim so cannot justly comment about it. However, I suggest that the story may be just that since, if there was a feathered T.rex in existance, it would be worldwide news.

Jaime Headden said...

I'd like to note that Sagan's quote is a reference to the fallacy by induction: Implying the possibility of a thing's existence merely by the absence of any contradictory evidence. He said this in large part to fight pseudoscience and religiously-motivated policy. Today, we'd discuss things like violent videogames = terrorist kids, etc. Induction fallacy is good for some times of argument, but most of them are rhetorical, not factual. Sagan is arguing we should focus on deductive reasoning, rather than inductive reasoning.

Mark Wildman said...

Yes that's right Jaime. Part of the idea behind this post is exactly this premise and that the quote can easily be hijacked for the wrong reasons.

Anonymous said...

I think a common problem with "lacking evidence is not evidence of absence” is that it only convince those agreeing with the one who say so--and the one using the argument can dismiss it in the next moment, if it is used to support things he/she do not agree about.

One example from the "BAND"-group and "temporal paradox": (1) absence of pre-Archaeopteryx maniraptorans = evidence of absence. But (2) absence of non-dinosaurian bird ancestors in the Triassic = NOT evidence of absence.

Mark Wildman said...


Thomas Holtz said...

Got to this WAY late, but there is a critically important issue here:

Presence of feathers does not equal absence of scales.
And, the inverse: presence of scales does not equal absence of feathers.

For any given follicle, it would be one or the other. But animals can have one form of integument in one part of the body, but the other in the other part of the body.

I think people are being WAY too binary about this.

Mark Wildman said...

That's very well put Tom - the one thing I realise is that the only rational thought is to be completely open minded, use lateral thinking a lot more and lose the two dimensional black and white logic that so many people (myself included) often fall into.

Post a Comment