[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]

Re: What is biomechanics? (or, The Truth About Flying Snakes - Was: Re: science and philosophy)

Ralph wrote

>Something like (but not specifically) _Microraptor_, approximately 41 cm
>long when mature?  _Microraptor_, a basal deinonychosaur, is the only
>dinosaur yet found to be smaller than pigeon-sized _Archaeopteryx_.

Yeah, that's the sort of thing.

 Nobody is
>saying that the later _Troodon_ or _Velociraptor_ evolved into birds.  I
>with you that the immediate ancestors of birds were probably quite small.

No, I realise that no-one is saying that _Troodon_ or _Velociraptor_ evolved
into birds.  But they are saying that the families to which they belong are
the sister group of birds.  Thus they are saying that animals with basic
bodyplans similar to _Troodon_ or _Velociraptor_  are the nearest relatives
to the ancestral bird that we've been able to identify.

The point is that, if you are a biomechanist interested in the evolution of
flight in brids, these groups then become 'taxa of interest' in your
analysis.  You start to wonder if
A. Birds evolved from a 'raptor' like animal, with _Troodon_ or
_Velociraptor_  being reasonable anatomical approximations of the common
ancestor.  In my understanding, that's the current orthodoxy.
B. The common ancestor or birds and _Troodon_ and _Velociraptor_  was
basically bird-like - thus _Troodon_ and _Velociraptor_ are secondarily
flightless, as GSP is advocating.  Which sounds profound, but really just
shoves the question of the origin of flight further down the tree.
C.  The common ancestor or birds and _Troodon_ or _Velociraptor_  was an
animal quite dissimiliar from either (well, as dissimiliar as you can get
whilst still being a common ancestor - you know what I mean).  E.g., a small
scurrying thing, perhaps.

>What other approaches are cladists ignoring?  Could you be more specific?

Well, the seeing if the transformations required by the cladogram are
biomechanically realistic, for example.

>It depends on the method.  If the method is not scientific, then it should
>rejected.  Are you not rejecting cladistic methodology?

No, I'm not rejecting it, I'm challenging it.  I'd be delighted if it
emerged stronger as a result - e.g.
1.  by facing up to the issue of reductionism: reductionism is inherent in
cladistic analyses.  Now, there's nothing wrong with this per se - most
useful paradigms in biology involve a large amount of reductionism - but the
fact remains that a cladistic analysis separates a coherent organism into a
number of constituent characters.  The basic assumption of the parsimonious
analysis is then that these characters are not functionally related....I
know that everyone realises that this assumption is not quite right, but it
seems to be a necessary step that you have to take in order to use the
methodology.  Again, I'm not against this - most methodologies need to make
assumptions that their proponents may not be that happy with; biomechanics
is full of them.  Anyway, having done the analysis, how do you then deal
with this issue?  Many cladistic analyses that I have seen dodge it
entirely.  Mike Lee tried to grapple with it a while back, in a paper where
he discussed the idea of 'correlated progression' - the idea that some
characters in a matrix are 'functionally dependent on each other'.  At the
time I teased him for re-discovering the wheel (wow, you mean that the parts
of an organism actually inter-relate?  Amazing!), but at least he was having
a crack at the issue, for which he deserves credit.  I don't think that
ignoring the problem is a meaningful solution.
    Reletively recently someone was talking about this - I think it was Mike
Taylor.  it was in the context of the repeatability and accessability of
cladistic analysis; someone had stated that one of the strengths of
cladistic analysis was that it doesn't rely on any 'special knowledge', as
phylogenetics once did back in the dark times.  Anyway, Mike (I think)
chimed in and pointed out that a cladogram done by him was probably less
useful than one done by Tom, for example.
    There is an important point here.  As we all know, Tom has a huge amount
of experience with theropods - he knows them inside out.  When he's coding
his matrix he has the opportunity to slect his characters based upon that
knowledge.  No matter how much you'd like to think that science is
objective, it is performed by humans, and we are very good at imposing our
subjective opinions upon objective methodologies, whether subconsciously or
    This was really well illustrated to me by Tim Hamley once.  For those
you don't remember Tim (it was the Middle Palaeozoic of the DML!), he
was/still is doing a doctorate on an Australian procolophonid.  At the time
procolophonids were being batted around the amniote tree like a tennis ball
at the US Open - some authors were putting them in as diapsids, others as
anapsids, still more as synapsids. Tim was trying to get to bottom of this -
he collected every character matrix published on the subject and tried to
understand the coding, etc.  What he found was different authors would code
the same character, sometimes even based on the same diagram in Romer, in
different ways.  Now, the way they coded it just happened to support the
tree that they had published in their previous paper.  Curious, eh?
    BUT, I don't think this is a problem if you face up to it.  Let's go
back to the example of Tom* working through his matrix.  Humans are
fallable, yes, but if there is one thing our brains are fantastic at,
subconsciously, it's pattern recognition.  We outperform computers by orders
of magnitude.  When Tom is coding his characters, he has the opportunity to
select the characters and the character states in a way that is consistent
with the pattern he has already recognised (and started recognising, the
moment he started looking at dinosaurs) in this animals.  Sure, some of
these subconscious decisions are probably going to reinforce the prejudice
he has about the evolutionary pattern of these animals.  But some of these
decisions can equally mitigate the fact that he has chopped his animal into
little pieces - he can recognise characters that are functionally related
and choose not to over-emphasise them in the analysis. Sure, some of these
decisions he would make consciously, but I'd be very surprised if there
weren't some subconscious ones as well, which reflect his intuitive
understanding of the animals as coherent organisms. All up, this would be
what you could call the 'art' of doing good science.  Anyway, that's why I
suspect that a cladogram done by Holtz can be of more value than one done by
Taylor, for example.

For me, if this point was acknowledged more explicitly, then I'd be happier
with cladistics as a methodology.  Who knows, if we stop dodging the issue
we may even find ways of addressing it?

* Tom, mucho apologies if I'm misrepresenting you here.  I hope that it's
clear that I am using your name in vain as a hypothetical example because I
respect your work.  But in case you find being called human insulting, I
withdraw the accusation unreservedly ;-)

2. By paying more than lip service to the fact that cladistics is the first
stage in formulating phylogenetic hypotheses.  I've already laboured the
point that there other other ways of cross checking the cladogram, but I
don't see it being done that often.

3. Parsimony.  Evolution, at the scale that we can detect it, is unlikely to
be parsimonious.  Despite what Nick Pharris claims, the fact that everyone
can agree upon what parsimony is doesn't mean that you have to use it to the
extent that it is used today.  Sure, use it to identify which trees might be
most likely.  Then find independent lines of evidence to test between them.
To say the absolutely most parsimonious tree, out of several similarly
parsimonious ones, is the right one is pushing the concept further than can
be justified, I think.  I guess I'm advocating a 'weak' use of parsimony
rather than a 'strong' use, if that makes sense. Okay, I don't want to get
embroiled in this one again, I'm just trying to show that I'm not rejecting
cladistics outright.

>> What, you can cling to your belief
>-- Replace "belief" with "hypothesis," please --

No, 'belief' is the word that describes what I'm trying to say.  I'm
deliberately suggesting that many of the people on this list do not have
open minds on this subject, and the string of emails that have resulted from
my post only confirms that 'belief' on my part.  'Belief' is a word that
conveys a combination of observation, prejudice, and desire, and as such it
is the appropriate term.

>Bearing in mind the processes of fossilization, it is easy to understand
>small, hollow boned theropod skeletons were rarely preserved, particularly
>forest settings.  In addition, some time periods are very poorly
>This is a reality we must deal with, regardless of our perspectives.  The
>Liaoning and Mongolian deposits are exceptional in preserving multitudes of
>detailed fossils of small, delicate terrestrial animals.  Only in such
>remarkable settings can we glimpse the true diversity of small life forms
>once lived.  A complete understanding of the history of life on earth is
>will always be unattainable because the fossil record is far from complete.
>fossils are welcomed, even if they upset long standing phylogenies.
>closer to understanding the true history of life on earth is the whole
>isn't it?

Yes, but what I'm trying to get at is that I think that augmenting your
phylogenetic reconstruction with something like biomechanics may tell you
what sort of animal 'should' be out out there.  Something like predicting
fossils - if you can back up your phylogenetic analysis with biomechanics
then you may be in a position to say something like;

"Okay, the cladistic analysis done by X suggests that dromaeosaurs are the
closest thing to the ancestoral bird that we know of, but the biomechanical
analysis performed by Y suggests that it would be unlikely that birds
evolved from an animal that we would recognise as a dromaeosaur.  Therefore,
the common ancestor of these animals must have been something quite
different.  Further, our best guess at the moment, based on the available
phylogenetic and biomechanical evidence, is that this ancestor would have
looked like a feathered squirrel [or whatever].  We predict that one day,
someone may find an animal like this in rocks older than the Mid-Wobbly

Didn't we do something like this with whales?  Wasn't it great?  Predictive

>> Sometimes I really think Tracy's hit the nail on the head.  There are a
>> of closed minds out there.
>Scientists should be open to new ideas as new discoveries and studies come
>light.  The reality is, of course, that some people are more willing to let
>of cherished hypotheses than others.  Like it or not, scientists are only
>(and some are more "human" than others)!  ;^)

Right, and what worries me is that the false security of untested cladistic
analysis has convinced the true believers that the ancestoral bird is a
small advanced theropod in the Upper Jurassic, and has completely closed
their minds to the possibility that Protoavis or those Triassic 'bird'
prints are anything to do with the story.  When those prints were announced
the response on this list was akin to the closing of the gate of the city
walls as the barbarians come over the hills...

Not that I really think that they are necessarily anything to do with
birds - but they might be.  Ignoring data like that doesn't help the cause.

Oh well.