Because theories in palaeontology can be hard to check (no time machine, and experiments being a bit tricky for the field), theorisation can get pretty bad. And it does. How bad?
In the next few postings, I’m going to look at three branches: ancient man, ancient dogs, and ancient dinobirds. To save you, worthy reader, from unnecessary gratuitous suspense, I’m prepared to reveal at the start that one of them has turned out better than I’d expected. It’s not good enough, but it’s not as dire as it might be. Let’s see what some of the characters from palaeoanthropology have to say for themselves…
Reviews of books of one academic by another can give unsurpassedly vivid access to the heart of a subject, but listening to conversations between notable players is pretty good too, though being face to face they tend to talk more nicely than they truly feel. We’ve had two pretty good conversations recently, one being between the two old warhorses Richard Leakey and Don Johanson, who in the past have been bitter rivals (and palaeoanth. rivalry in Africa can get very bitter indeed). Leakey has reported a TV encounter between them in the past where he was set up as the fall-guy.
But first, let’s look at the legendary Milford Wolpoff being interviewed by the great… Razib Khan. Who he? Well, he’s a pretty smart cookie, mainly into population genetics. He isn’t really a palaeontologist, but he’s ‘as good as’, one might say; indeed one might say he’s better. His famous genetics and population blog gnxp (gnxp stands for gene expression) is very rich, and is probably his main claim to fame, and I have no idea how he finds the time.
To set the scene, palaeoanth. is currently at the point where people feel the question “Out of Africa” or not? for the origin of modern man, has been settled to a first approximation. The answer is definitely ‘neither’ or ‘both’, now we know that we (or at least non-Africans) have a little Neanderthal blood in our veins if not on our hands… and we also contain the blood of other types we are only just discovering, such as the Denisovans. Chris Stringer of the London NHM, had been cast as a key “Out of Africa”n, with Wolpoff as a multiregionalist. (A 6,300 word blog posting the-paradigm-is-dead-long-live-the-paradigm from Khan details the background.)
Wolpoff seemed wiser and more knowledgeable than I’d expected, not apparently violating any important principles. It was interesting that he seemed very keen to stay on the subject of “Out of Africa or not”. He didn’t really want to talk about Ardipithecus. He says Ardi wasn’t very good at bipedalism due to the grasping foot, though he accepted it was usually bipedal on the ground. Khan, tactful but eventually insistent, finally got a number out of him. It was, interestingly, 4. Or rather “most geneticists now accept” 4 for the date in millions of years for the chimp/human split. This surprised me for being so good. (John Hawks, also a 4’ist, was, like Tim White, a student of Wolpoff, and I expect influenced him.) The other thing that struck me was his admission, over some point or other, that it would be risky career-wise to express a certain view – not because he’d disapprove, but because of others. Let there be no doubt then, that you do need to think very carefully about views you express in this field, even if you think they may be true, because of political considerations. He did say repeatedly that palaeontology was very political (for those who ever doubted it).
I had been rather cross with Wolpoff and Stringer for both getting the “Out of Africa” question wrong. Perhaps they’d both been largely victims of the high-tension surrounding the field, accidentally forced to stress views more extreme than they wanted. But I still don’t think they’d been particularly clever, since I’d had no difficulty in favouring an “Out of Africa plus mixing” solution long before they saw the light, and they’d been paid to get it right.
(I still haven’t found though any trace of Stringer even entertaining the possibility of a 4my split date, and what that means for the species involved.)
With Leakey and Johanson, the antagonists were actually talking to each other, or at least being interviewed together. Leakey reminds me of the queen: born into a role they hadn’t chosen, but fulfilling it pretty enthusiastically and stoically. He has clearly battled with Africa in many ways for many years but it hasn’t quite defeated him yet. Interestingly he says he’s now more interested in later evolutionary issues, such as why anatomically [fairly] modern humans didn’t leave Africa 200,000 years ago but only 65,000ya. He says modern populations seen today diverged about 65kya, as did modern languages, though talking itself predated that; also that Lake Turkana used to be the source of the Nile till 5-6kya – with implications for continental communication.
However, from a number of their comments it’s clear that both he and Johanson show absolutely no signs of dealing with the idea that bipedality may not have arisen with the chimp/human split, nor that the split might be 4mys old; the interviewer didn’t press them. I really do think a good scientist should be expected to answer all the salient questions, and not be allowed to dodge them if they’re ‘embarrassing’. And they shouldn’t even wait to be asked them – they should bring the vital issues up themselves and only support positions they can defend with decent arguments. But then I wish a lot of things.
But Johanson does seem to be trying hard. He does get some credit for his view that he does “…not believe in evolution – it is the best explanation so it is a fact” …or at least he gets credit for talking in terms of the best explanation. But a palaeoanthropologist especially should appreciate we can’t deal in facts. A ‘fact’ cannot be questioned, cannot be undone, any more than the root of the word, springing from doing or having done, allows that an actual past event can be changed. But in any case, deciding what is the best explanation will never provide such reliability (even though it’s all we have). But the main harm from a ‘fact’ is that people expect proofs for them… so when the best explanations don’t come with proofs, people don’t like them – despite what they believe instead, not having proof either.
Johanson’s ‘Institute of Human Origins’ produced in 2008 a slick interactive documentary [it’s in Flash, so no iPads] on human evolution, which put the split “at least 5mya” and left chimps’ and gorillas’ “ghost lineages” unexplained. It also stated that bipedality was what set us apart from the rest, and Johanson doesn’t even present the topic as a controversy.
But having said that, he does offer for consideration not one but three theories of our phylogeny (“Lineages” tab, “Related exhibits”, 3:“The Human Family Tree”). That’s good, introducing the competition between theories. (Unfortunately none of them actually cover what I consider the three or four key points, but Johanson’s own version veers pretty close in places, and Bernard Wood’s one pencils in a possible link straight from Ardipithecus ramidus kadabba to P. aethiopicus, which, as in Johanson’s tree, is shown ancestral to P. boisei. What’s more, both have Aust. africanus ancestral to P. robustus, but not necessarily ancestral to us! This is sound, progressive stuff. Also, Johanson has Kenyanthropus platyops leading to K/H rudolfensis, though he doesn’t link them to anything else at either end.)
But he is offering a clear choice of theories… and selecting the “Evidence” tab, and 3:“The Scientific Method”, we see he also takes “The” “Scientific” Method seriously. Often in the past, this has been a bad sign since the processes offered haven’t always matched what, for convenience, we might term ‘Modern Practical Popperism’. However he has made an effort, and gone into some detail as to what he means by it:
“Science is a way of knowing about the world and how it works.
The scientific method is the process of forming and testing ideas.
Through the scientific method researchers build theories about the world.”
I like this. Although that last line suggests only privileged people are allowed to do science, and we dread his understanding of “testing”, this is the best description of “T” “S” M I can recall from a working scientist. He offers a simple flow chart and says the process normally begins with an observation (though we really do need implicit theories before we can do anything, and they are sometimes important) but at least he does say “normally begins”.
Mousing round the flow chart of observation, prediction, test etc., small boxes pop up giving examples of theorisation on whether humans hunted or scavenged. Unfortunately they converge on the conversion of a ‘hypothesis’ to a ‘theory’. The problem with this is illustrated by the very conversion process he gives:
“Before it can become a theory the hypothesis must be tested further to disprove other explanations.”
A test of a hypothesis/theory might disprove that hypothesis/theory itself – but such tests won’t disprove other theories! And won’t, whether it passes or fails its own test! All a theory can do is explain the evidence better than any other known theory (and hopefully well enough to be used in a world model for whatever purposes it’s needed). His process of conversion from hypothesis to theory doesn’t make sense, and is only there to satisfy the common craving for some ‘initiation ceremony’ for our new pieces of knowledge. But in the real world there is no such categoric criterion, and certainly nothing of the sort we were taught to expect by that great imposter school geometry. (Maths, within its very simple certainties, offers a better standard of meaningful positive proof and positive evidence but it doesn’t transfer at all well to “real world” natural science. Note though, Gödel says you can’t strictly prove anything even in maths.)
And later, in three further boxes available through the “Learn More” of the Evidence tab, the principle of repeatability is stressed, though it’s not really central to a historical science. Same with the “predictable laws” he mentions, which aren’t central to historical disciplines either, but which would need to be if repeatability were vital. In palaeontology we guess what happened by imagining plausible if perhaps complex processes that may have occurred. Such sequences of events will in reality have been unique in detail and probably unrepeatable in every way. They are so far above the natural ‘laws’ of various kinds from which they are built, that it’s not very helpful to talk in terms of strict laws. (Tools and techniques, for e.g. dating rocks, are closer to those laws, but peripheral to the essence of the science of reconstructing evolutionary events.) Finally, saying of those laws (never, we now see, actually central to palaeontology) that they can be “…knowable through observation and testing”, though right, gives the impression that fairly automatic observation and testing will finally produce the theories we need. That process of testing (relatively simple but so rarely done well) ignores the essential but very non-trivial black art of hypothesis construction – the generate half of “generate and test”. Ignoring that, and ignoring that testing itself is actually non-trivial, tends to de-emphasise the brain-work needed, and minimises the value of careful, inspired, often armchair thinking so essential for good science.
Johanson isn’t ‘there’ yet but at least he’s pointing in the right direction and mobile, with his fairly promising stab at “The” “Scientific” Process/Method, his detailed examples of its use, and his offering of multiple theories. And with numerous schools of thought around, and workers like John Hawks involved, there’s reason to be optimistic. By contrast, the handling of phylogeny in two other areas of palaeontology are in the wrong place, looking in the wrong direction and set in concrete.
[ I quote some relevant Sciencepolice2010 rules violated, and invite the reader to look through this posting or the Institute of Human Origins’ interactive documentary to see where:
1) Science is the generation, judging and honing of theories which model (i.e. explain or predict) the best. Only this counts so distinctions between theory, hypothesis, conjecture etc. are artificial.
9) There are no absolute facts, proofs or truths; outside mathematics at least, they are convenient untruths that can simplify our thinking. Proof needs to demonstrate the impossibility of any competing theory, known or not, of equivalent or superior power, whereas disproof merely requires the presentation of evidence inconsistent with the one theory. Knowledge is beliefs, not facts.
12) Tests a new theory can uniquely pass are best offered, and may be needed for superiority, but not for pseudo-criteria like theory status, false ‘testability’, or truth. Insisting on a mechanism for a theory is a classic error. A theory often inspires the discovery of its mechanisms and special tests.
14) Historical disciplines (e.g. palaeontology and archaeology) often test by future discovery, not experiment. In such disciplines, evidence and demonstration may be probabilistic or qualitative, and rely on complex variously valid world models left in individual minds by diverse experiences. Typically, historical sciences theorise past events given present evidence (abduction). Inductive sciences invent laws, or describe/model structures or processes; ‘applied sciences’ deduce futures or achieve goals. Despite differences, many principles apply to all sciences, though the importance orepeatability for inductive sciences does not make it a basic principle of historical disciplines.
2 ?) The worth of a theory depends on such aspects as accuracy, generality, simplicity, and the degree to which its implications are genuine predictions (and the more surprising the better).
10 ?) Evidence is those observations not well explained by a theory, or not as well explained by one theory as by another. Positive but non-comparative evidence (explained by a theory), though useful in the black art of generation, merely progresses testing. ‘Positive evidence’ must not be trumpeted as ‘evidence for’ a theory when competing theories also explain it. Papers and journals should ensure the full landscape and pattern of posable theories are reasonably accounted for.
11 ?) An untestable theory is one which is intrinsically logically untestable, not one for which no technique for testing it is yet known to some person, or indeed anyone. Deducing the scope of implications, effects, or influences of a hypothesis (via which it might be tested) can be slow and unending. ‘Untestable’ is a rare category, not to be routinely flung at everyone else’s new theory.