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

cladistics revisited (was Re: Sterna, Whales, and Evolution)

[ I'm allowing this one in violation of my "no second responses to the
  same post" rule, but since the first response should not have been
  approved based on another criterion, and since this one actually
  addresses the issues rather than advertises where the issues are
  addressed, I'm going against policy.  Be warned that I'm not likely
  to make too many mistakes like the one I made last night, so be sure
  to put any "meat" into your FIRST reply to any given message.  -- MR ]

<<Dinogeorge's reply to the following post.>>

Date:     96-02-15 18:53:09 EST
From:     mickey@lepomis.psych.upenn.edu (list moderator)
Sender:   dinosaur@lepomis.psych.upenn.edu
Reply-to: mickey@lepomis.psych.upenn.edu
To:  dinosaur@lepomis.psych.upenn.edu (Multiple recipients of

<<My replies to Mickey Rowe's comments are in double
angle-brackets. Makes it easier to type.>>

Dinogeorge@aol.com attempts a critique:

<<"Attempts a critique," indeed! Sorry if I went a little too
_fast_ for you. I've lived with this stuff for so long it's second
nature to me, and I forget that others may not understand points
that seem obvious to me. So, I'll take it real slo-o-w...>>

> Reducing your problem to a situation between just two species,
> A and B, there is only one cladogram that relates those two
> species, namely the trivial cladogram with two branches A and
> B.

It's not clear what George has in mind here, but I think it
doesn't matter because it appears to me that he's just missing
the boat.

<<What's not clear here? The concept of "trivial cladogram"?
Here's a picture of the trivial cladogram, the only possible
rooted cladogram relating two taxa:

A          B
 \        /
  \      /
   \    /
    \  /

Can you think of any others? Well, you could interchange the
letters A and B in the diagram, but that's just the same
cladogram all over again.>>

> But there are _six_ different ways A and B can be related to
> each other: (1) they can both arise from a different common
> ancestral species;

What is a "different common" ancestral species?  Sounds like an
oxymoron to me.  However, I *think* George might mean:

            ----- A
     |     |
     |      ----- C
     |      ----- D
     |     |
            ----- B

<<No, no, no, a thousand times no! Here's a picture of
common-ancestral species C splitting into the two species A and B
(time reads up in all my diagrams, not horizontally as in the

A         B
 \       /
  \     /
   \   /
    \ /

Common ancestral species C is "different" from A and B. Does this
picture make it more comprehensible?>>

> (2) B can branch off from A; 
> (3) A can branch off from B; 

In principle these cases might be distinguishable, and if that
were the case, then the name A could be given to the older
species (even after the branch point) and B to the younger
species (or vice versa).

<<Here are the pictures for cases 2 and 3, side by side:

A         B        B         A
|        /         |        /
|       /          |       /
|      /           |      /
|     /            |     /
|    /             |    /
|   /              |   /
|  /               |  /
| /                | /
|/                 |/
|                  |
|                  |
|                  |
A                  B

Okay? They're _not_ distinguishable cladistically; you get the
same cladogram--the trivial cladogram noted at the beginning of
this post--for either situation.>>

The clade would retain the generic name of the older species as
Tom pointed out to George last December.  In practive, however,
cladists recognize that we virtually never have the superfluity
of data required to allow us to name such organisms.

<<I have no problems with the names of the clades; name them what
you like. The issue I've raised here is that these are two (out
of six) different speciation situations that cannot be
distinguished by cladistic analysis alone.>>

> (4) A can become B; 
> (5) B can become A; 

<<Here are the pictures of these two situations, side by side:

B         A
|         |
|         |
|         |
|         |
|         |
|         |
A         B

Okay? A is ancestral to B; and B is ancestral to A. Again, you
get the same cladogram, and it really doesn't do justice to the
situation at all.>>

Once again, in principle we can deal with this virtually the same
way as the branching; the name for B is used for the organisms
after a certain time, and A before (or vice versa).  In practice
we never really have to make these sorts of judgement calls
because the fossil record just doesn't allow it.

> or (6) A and B are not closely related at all.

<<And here's the picture of this situation:

A         B
|         |
|         |
|         |
|         |
|         |
|         |
A         B


Since I wasn't able to parse George's (1), I could be out to
lunch, but it seems to me that in principle there's no difference
between (1) and (6).  You want to try to draw pictures, George?

<<Satisfied with the pictures? Like I said: real slo-o-w...>>

> The cladogram clearly fails to distinguish among these cases,
> and indeed only cases (1) and (6) resemble the cladogram itself
> to any extent.

<<Case (6) is like case (1), only with the common ancestral
species way down the tree somewhere, not immediately in the

For (1) and (6), you need to have other organisms in the
cladogram to see the difference.  For (2), (3), (4) and (5) you
have to be convinced that there *are* no other relevant organisms
that could fit into your cladogram, and you have to have a very
complete set of representatives of both A and B before, during
and after the transition.  The reason that cladists don't worry
about these sorts of nuances is that they appreciate that this
isn't a situation they can expect to find themselves.  The
availability (or rather lack thereof) of data just doesn't allow

<<I'm talking about hypothetical situations here, among two
supposedly closely related taxa. Cladists can get away with
cladograms as models as long as they hide behind the poverty of
the fossil record. In the cases I describe above, the fossil
record is irrelevant. We could even have time-lapse motion-
picture film taken via H. G. Wells's time machine of every single
individual of each species A and B; not only would the trivial
cladogram not work, it would be even more apparent to everyone
that it couldn't work.>>

I don't know how, exactly, this was supposed to relate to Ron's
question...  so let me go back to that.

<<It doesn't, very much. But so what?>>

ornstn@inforamp.net (Ronald Orenstein) asked:

>> Must the whole of the new radiation, no matter how
>> differentiated and diverse it becomes, be included in the stem
>> genus of which it forms a nested subset?

If you want all supraspecific taxa to be monophyletic (_sensu
stricto_), then yes.  If this really seems problematic to you, I
suggest you try to come up with an example in which you know of
"ten closely-related and very similar species" only one of which
-- one that you could identify based on known data -- left
descendents which then radiated into a myriad of forms.  You can
drive yourself nuts thinking of situations you wouldn't like, but
what's the point if those situations never come up?

<<Because they reveal the inadequacies of the methodology. And
the point is, those situations _continually_ come up, but
cladistic analysis cannot sort them out for us. Given _only_
cladistic analysis, we wouldn't even know whether they were
coming up or not! Because of the nature of cladistic analysis,
the True Phylogeny always comes out looking like a cladogram,
when that is not necessarily what the True Phylogeny really looks

Finally, since some of the dust has briefly settled (seems to me
everybody's in a bit of stunned silence), I'd like to throw
something else in here that's been bugging me about the
dinosaur/bird discussion.  Taxonomy is about naming organisms and
grouping them together.  That's it.  It's up to us to decide what
the significance of the groupings should be.  Prior to Darwin,
people thought they were intuiting the mind of God.  God had an
idea of what a fish should be, and all real fish were just
variations around the theme of the ideal fish.  Making a taxonomy
of fish was thus tantamount to seeing the ideas God had as he
made different groups.  Darwin recognized that the similarities
in organisms had nothing to do with them being variations around
a predetermined theme, but rather they arose via the environment
sculpting populations from previous common populations. What
should taxonomy represent?  How organisms appear to us?  Or
perhaps something that's independent of us.  Why should we care
that wings, or feathers, or tachykinesis or any other trait or
suite of traits seems really unique to us when we're constructing
a taxonomy? Should Darwin's realizatin that taxonomists were not
reconstructing God's thoughts cause people to develop taxonomies
that are nothing more than reconstructions of their own thoughts?

To me it seems fortunate that most taxonomists don't feel that
way.  That is, any taxon that includes all of the animals we call
"dinosaurs" should include birds as well.  Birds *do* have their
own clade.  But no group that includes both _Iguanodon_ and
_Megalosaurus_ should exclude the bird clade if we want the
clades to represent something independent of us as observers.

<<The one great quest of paleontology is to elucidate the
appearance of the True Phylogeny--the big master Family Tree of
all earth's life forms. The _topology_ of the Family Tree is
what's important; the names we give to the branches or other
sections of the Tree are for our convenience only, so that we can
refer to them with ease in our discourse. There is no
_topological_ difference between saying "birds are dinosaurs,"
"dinosaurs are birds," "birds descended from dinosaurs," or
"dinosaurs descended from birds." These are terminological
distinctions that depend on what we mean by the terms "birds" and
"dinosaurs." Unfortunately, however, these terminological
distinctions color our images of the dinosaur-bird phylogeny, and
occasionally blind us to alternative, and perhaps more accurate,
ways of visualizing what the organisms at the dinosaur-bird
transition might have looked like.

The only names allowed in cladistic taxonomies are those attached
to the nodes of a cladogram. The terminal nodes are the species
(or monophyletic groups collapsed into single nodes), the branch
nodes are the supraspecific taxa. This is inadequate because (1)
the True Phylogeny is not necessarily modeled by a cladogram (as
the simple example of six different phylogenies all yielding the
same cladogram showed), and (2) the _branches_ of a cladogram
represent organisms worthy of classification and identification,
but these organisms receive no names. Indeed, the branches of a
cladogram are where all the "evolutionary action" takes place,
where all the change happens in between the essentially static
nodes. It's crazy not to supply some way of naming those
transitional organisms, and crazier still to ignore them entirely.>>