Hi Matt Baron,
Nice to see your work, because it shows what happens in a divergence investigation when you exclude types that lived too long afterwards, which had accumulated numerous interfering irrelevant characters. This is something I’d advocated on forums widely read (in palaeontological terms) since about 2000. It deserved to have been explained more explicitly than you did, as the size of the consequence, and its response, indicate. As this is something that I and very few others had been advocating for years, it would have been nice if you’d cited us:
The Secret Dinobird Story 3.3:
…[I]n careless hands, cladograms can be published where these two types [allosaurs and tyrannosaurs] are brought closer together than they should. In fact, they should never be included in the same cladogram. By following the principle that lines generally agreed to be separate (like the tyrannosaur and allosaur lineages) should never be analysed together, we would help prevent types probably convergent and perhaps even 80 million years apart from flocking unrealistically together and distorting the picture. Each lineage composing an obvious module of its own (e.g. “the tyrannosaurs”) should occupy its own mini-cladogram. Large confused groupings should be split into these mini-cladograms, and from each a very few early representatives should be selected. These fewer representatives, ideally close in time, should then all be analysed together in the overall cladogram without their later forms (Wagner 1998*). This would help deal with for example top predator convergence. Simulations could easily be devised to confirm the need for and effectiveness of this approach.
*Wagner, P.J. (1998) Saturation of cladistic character space – debunking the myth of the infinite. Geol. Soc. Am. Abstracts with program 30, A-326.
But thanks for observing this principle, and not inviting allosaurs and tyrannosaurs. They’ve spoiled so many parties.
I also liked the result simply because it showed there is no such thing as ‘The Cladistic Approach’. That is where you chuck everything in without thinking. There is clearly a multiplicity of ways to run and ruin a cladogram. Already this has caught Padian out – the bloke who always gets invited to write the overview for Nature’s important dinobird papers. You won’t remember but it was he who poured scorn on the possibility of finding feathered dinosaurs, and before that, guessed wrong about pterosaur bipedality. The latter was acceptable at the time, but the former showed he refused to start to consider possibilities unless they were already blatantly undeniable. This time he said that your study was justifyable because you did it the way it’s usually done – done by people like him of course – but thereby denying the nature of scientific research, where novelty is essential. Again he ignores the principle of finding theories that explain the best (and substrategies thereby implied), as the definition of science, and again he implies that it’s what his group does, not what it does, that counts.
But anyway, you didn’t do it the way it’s always been done! But more important is the indication that qualifying in palaeontology is not learning a science but learning a club. Top statistician and scientific standards crusader Andrew Gelman said recently in a blog post about a psychology clique he’s identified as pseudo-scientific: “It’s a guild, man, nuthin but an ivy-covered Chamber of Commerce. Which is fine—restraint of trade is as American as baseball, hot dogs, apple pie, and Chevrolet…. [but] Psychology is not just a club of academics, and “psychological science” is not just the name of their treehouse”, to which commenter Kane replied: “Sure about that? The evidence here is that this is precisely what psychology and psychological science, in fact, are. You (and I!) just wish that weren’t so . . .”. Padian too is clearly as groupist as he is non-scientific, and even if he’d made huge positive contributions to science instead of an overall negative, selecting one summariser time after time after time as Gee does in selecting Padian, insults moral and scientific principles.
That’s why I don’t do the subculture the honour of asking their permission to publish, and I continue to make plans for alternative channels to the Nature/Berkeley abomination. The internet is free, and sometimes free from Padian/Gee-type bias. But to ignore work just because it hasn’t kow-towed to establishment morons is Not science! That’s why you’ve no excuse for not citing me/Wagner. Remember that peer review (as opposed to simply seeking advice) is an invented irrelevance by people who don’t know what science is. Consider also that most prominent dinobird workers in USA/UK/Canada know who I am and are only pretending I don’t exist. (The publication of my book did something a bit weird, sudden and long-lasting to the posting frequencies of the dml. That couldn’t have happened if it hadn’t been noticed by the group as a whole.)
You mention the internet’s reaction. Mickey Mortimer picked holes in your coding of characters etc. but then he picks holes in everyone’s. I do hope he recodes it and produces his own cladogram. Even if he does, it won’t mean your tree is right. I remember reading Mickey’s last posting before his first on your paper. In the comments, his discussion with Marjanovic made me think: “There’s a couple of dodgy characters for you. Have either of them ever even heard of calibration?” You can’t claim scientific status if you haven’t properly compared the tool you rely on (here, cladogenesis) against known and challenging inputs, and noted its response characteristics. Mickey says bootstrapping isn’t as important as people think. In fact bootstrapping seems to be very optimistic – high bootstrapping doesn’t guarantee any validity but low boostrapping warns of likely meaninglessness. But they wouldn’t know this because they can’t think outside their social box and of course they’ve never calibrated bootstrapping or anything else involving cladograms (or anything). Mickey might have started learning about such things if he’d read my book but he hasn’t. If he had, he’d have known which birds other than penguins have no air in their bones (and why). He might even have learned that programmers start counting from 0. When Mike Keesey mentioned this on Mickey’s blog, he received no comment from anyone, not even a cheery hello. The zero issue with TNT seems to be a bit more than that, but zeroing TMK is filthy style by the other commenters. THAT’S why he never criticises their technical blunders from his IT expertise background – because he wants to avoid Being “Othered”.
The internet is set to generate opposition it seems, but they’re not expert tree-surgeons. However your tree might actually be wrong for several reasons: because you’ve wrongly excluded certain types; the root might be wrongly positioned so parts of it might flow in the wrong direction, and of course the topology could be wrong due to e.g. convergence etc.
I expect all those will apply; one piece of evidence is particularly compelling: those waisted slightly hooked pointed teeth of Eoraptor and Laquintasaura shown in your fig 2 c. And where else have we seen them oh yes Archaeopteryx. Which hypercarnivore bequeathed those? It never had a Mesozoic hypercarnivore ancestor with unwaisted teeth and neither did any ornithiscian. So you’re right to suggest the ancestors concerned weren’t specialist carnivores, but omnivory at the root is wrong too. They ate invertebrates (and miniverts) and were almost but not quite invisible to us (as if everything Triassic isn’t!). And the reason they were invisible is that they were very small… and thin boned… and fell into acid soils…
And don’t say there’s no evidence for this: the observation is that the ancestors are mysteriously absent, “Ж10.1: Evidence is those observations not well explained by a theory, or not as well explained by one theory as by another“, and this is so far the only theory that explains the absence. The absence is therefore evidence for this theory.
…Which leads us to primitive pre-feathers and fibres. The first feathers probably weren’t fibrous, fibrous structures might well have evolved more than once from their non-fibrous progenitors, sauropodomorphs might or might not have had fibrous feathers in their ancestry, and all of these are still fully consistent with the observations, as is primitive non-insulatory feathering for all archosaurs.
Such a long way still to go palaeontologically. But there’s plenty of social science of palaeontology to hack through too. The No Reply to Keesey’s comment, and Naish’s instruction in his Tetzoo blog to ignore Dave Peters (even though Peters was at least correct in rejecting the standard basal dinosaur tree, and Naish had implied disapproval of groupism in his Brexit comments)… and Mickey Mortimer doing whatever he did recently to get blocked from Dave Peters’ blog… and Padian’s claim that only what he does is science… and Gee’s refusal to allow Nature to print summaries on dinobird tree papers by anyone with IT expertise… all show that the main name of the game is still groupism (accompanied as usual by incompetence). And if you don’t believe me that groupism combined with IT/stats incompetence is a common problem, ask Gillian Tett.
But I LOVED your use of the word MINDSET!!!! 🙂