Re: Fw: Popper and Palaeontology (was: Re: "running" elephants - locomotary analoges)

[John Conway <john_conway@mac.com>]

On Tuesday, April 22, 2003, at 08:23  PM, mjmurphy wrote:

> I think one thing useful to note here is that Popper ideas concerning 
> verification/falsifiability were developed largely with an eye on 
> physics. Maybe because physics is highly mathematised, Popper's idea's 
> tend to be very theory centric (as opposed to observation-centric).  
> There is a technical sense in which what you choose as your "data" is 
> arbitrary, and the whole notion of "gathering facts" in the absence of 
> a guiding theory is derided as irrelevant.

I'll stick my neck out here and say that it is. And not just 
irrelevant, but impossible. Even naming something is offering an 
hypothesis about it.  For example, say we want to measure the length of 
the femur of _Archaeopteryx_  - we obviously have to name the thing we 
are measuring - and by doing this we are hypothesising that the femur 
is different from other bones, and we also have to have theories about 
space, length, measurement, and so on. Now say we want to do something 
a little more complex - compare the length of the femur in  
_Archaeopteryx_ and _Microraptor_,  not only do we need the theories 
above, but we also have to have a hypothesis of homology to be able to 
identify the bone. Already we are up to our eyeballs in deeply 
theoretical concerns, just comparing the length of two bones.

> The idea, expressed by one of the Leakey's, that went something like 
> "give me a good data set over a theory any day" is quite alien to 
> Popper.

And for good reason, I'd say.

> Further, the whole idea of inductive reasoning--reasoning from 
> specific facts via some inductive method like "concomittant variation" 
> (hope thats the right spelling) to generalities--is dismissed.  It is 
> invalid (by deductive logic) to conclude from the fact that the sun 
> has risen every day for the past billion years that it will rise 
> tomorow.  In roughly the same way, it would be invalid (not just 
> speculative) to conclude from the fact that many small theropods were 
> feathered that all theropods were feathered at some point in their 
> life cycle (or something like that). In a strict sense, for Popper a 
> large portion of the argumentation that goes on on this list would 
> have to be invalid.

Logically invalid yes. Popper was fighting against a _formalisation_ of 
inductive inference. However, to hypothesise  on the basis of many 
feathered small theropods that feathers were present on all theropods 
(at some stage of their life) is perfectly popperian. Popper was not 
against making generalisations - on the contrary, he was for them - he 
was against the construction of a formal methodology for induction.

Phylogenetic bracketing can look like induction, but because it works 
just as well with two markers as with a million, it is more like an 
automated popperian generalisation machine :). We only needed 
_Sinosauropteryx_ and modern birds to believe that maniraptorans were 
feathered - the other feathered theropods just filled in the picture, 
and didn't falsify the hypothesis.

> Now, Popper is somewhat ambiguous with regards to inductive reasoning. 
> There are times when he literally sounds as though this kind of 
> reasoning doesn't exist and if it does exist people should stop doing 
> it immediately.

Yes, some times he says very clearly that there is no such thing, but 
if that is so it isn't clear what he is arguing against (actually the 
issue is a lot more involved than that - Popper didn't often make 
really obvious mistakes).

> There are also times when he compares inductive reasoning to "rules of 
> thumb"--I think he actually compares inductive reasoning to the way 
> people reason about gardening (best time to plant petunias, and so 
> on)--which, while they can be used to generate hypotheses, are 
> irrelevant to the testing of hypotheses.  However, I think in a 
> discipline like paleontology its more likely that an argument will be 
> taken as worthy of testing just because it takes a form that is 
> recognizably inductive.

Popper was unashamedly normative - he thought his method represented 
the way science _should_ be done. So it could be argued that maybe 
palaeontology would be better off if it dropped any pretence at 
inductive method.

> Cheers,
>
> The shapes of things are dumb.
> -L. Wittgenstein

(Wittgenstein once threatened Popper with a fire poker during a 
debate...)


John Conway, Palaeoartist & Protophilosopher

"Do I contradict myself? Very well, then I contradict myself, I am 
large, I contain multitudes." - Walt Whitman

Palaeoart: http://homepage.mac.com/john_conway/_palaeoart/
Palaeontology & philosophy discussion forum: 
http://clouds.proboards16.com/