Methods and characters (Re: Quetzalcoatlus northropi!!!)

[Cal King]

In article <34CA5DB0...@saluki-mail.siu.edu>,
matt...@saluki-mail.siu.edu (Matthew Fain) wrote...

>Cal King wrote:
>
> In article <34CA14B0...@saluki-mail.siu.edu>,
> matt...@saluki-mail.siu.edu (Matthew Fain) wrote...
> >
> >Many happen to be experts on morphology and the evolution of those
characte
>rs
> >for their particular group.
>
>> But cladistic methodology forbids them from applying their expertise in
the
>> choice of and analysis of characters because doing so would be
"subjective."
>
>Absolutely untrue. There is nothing subjective about applying the most
>likely explanation to all the available evidence.

Untrue about what? Are you saying that cladists can apply their expertise to
weed out questionable characters a priori? That is not necessarily culpable
or wrong, as Cronquist noted, but it does compromise the objectivity claim by
the cladists.

> > You have yet to present a viable alternative model of character
evolution
> >that systematists may use to infer phylogenetic trees from morphological
> >characters.
>
>> I have. Use only synapomorphs and make damn sure that they ARE or at
least
>> likely to be synapomorphs by using character analysis OF the characters
bef
>ore
>> one embarks on an analysis ON the characters, while keeping in mind that
>
>
>This *is* pretty much the method.

Absolutely untrue. Cladists assume a priori that ALL of their characters are
synapomorphs. They only perform a test of polarity (but not of homology)
before the analysis. I can dig up old posts to illustrate my point if you
want.

>Thank you for demonstrating your knowledge in this area. Some mutations are
>more likely than others at specific points in the genome.

True, and never disputed. That is why not all molecules can be equally
informative.

> Most *mutations* are certainly not neutral at the DNA sequence
>level, but those that are deleterious are weeded out by selection.

Those that are neutral are fixated, correct?

> Molecular
>distances track *substitution* rates, not mutation rates, another very
>important
>distinction which I encourage you to read and think about.

Since only neutral or adaptive changes are fixated, substitution rate is, in
effect, mutation rate.

>Again, thanks for demonstrating your knowledge on the history of the subject.
>Cladists were ensconced in systematics well before the '80s,

Yeah, but hardly anybody was doing cladistics until the 80's, after the threat
of obsolescence caused by molecular techniques in the late 70's. In fact, one
famous American herpetologist committed suicide because he thought that there
is no future for him as a systematist and that everything would be done using
molecular techniques in the future, so the story goes.

[Matthew Fain]

Cal King wrote:

In article <34CA14B0...@saluki-mail.siu.edu>,
matt...@saluki-mail.siu.edu (Matthew Fain) wrote...
>
>Many happen to be experts on morphology and the evolution of those characters
>for their particular group.

> But cladistic methodology forbids them from applying their expertise in the
> choice of and analysis of characters because doing so would be "subjective."

Absolutely untrue. There is nothing subjective about applying the most likely
explanation to all the available evidence.

> You have yet to present a viable alternative model of character evolution

>that systematists may use to infer phylogenetic trees from morphological
>characters.

> I have. Use only synapomorphs and make damn sure that they ARE or at least
> likely to be synapomorphs by using character analysis OF the characters before
> one embarks on an analysis ON the characters, while keeping in mind that

This *is* pretty much the method. Determine which characters are more likely to
be representing phylogenetic correlation, and which are more likely to have been
subject to adaptive change.

> Because often mutations at the molecular level are random and neutral and thus
> molecular distance can be a more accurate gauge of divergence time than
> morphological distance. That is why the "gene tree" (inferred from molecular

Thank you for demonstrating your knowledge in this area. Some mutations are more

likely than others at specific points in the genome. Mutations are random with
respect to their future benefit or detriment to the organism. Be careful in using
the word "random". Most *mutations* are certainly not neutral at the DNA sequence
level, but those that are deleterious are weeded out by selection. Molecular

distances track *substitution* rates, not mutation rates, another very important

distinction which I encourage you to read and think about. Base substitution, for
example, is a simple enough process that models can take into account the
selective forces acting, and thus these characters retain utility until they are
saturated beyond the point that the model can account for.

> data) is not congruent with the "organismal tree" (inferred from morphological
> characters). BTW, such excitement is not new. Molecular methods have been
> available since the 70's. Cladists (who are mostly morphologists in training)
> climbed on board of the cladistic bandwagon in the 80's just as molecular
> methods are threatening morphological techniques with obsolescence. It is
> amusing that many morphologists cum cladists are just now turning to molecular
> methods themselves.

Again, thanks for demonstrating your knowledge on the history of the subject.

Cladists were ensconced in systematics well before the '80s, and during the '70s
they had nothing like the sophistication we currently have wrt to molecular data
and analytical techniques. Nowhere did I even imply that the excitement over
phylogenetic analysis of molecules was new.

[Matthew Fain]

Cal King wrote:

> In article <34CA5DB0...@saluki-mail.siu.edu>,
> matt...@saluki-mail.siu.edu (Matthew Fain) wrote...
>
> >Cal King wrote:
> >

> > Most *mutations* are certainly not neutral at the DNA sequence

> >level, but those that are deleterious are weeded out by selection.
>

> Those that are neutral are fixated, correct?
>

Not necessarily. One thing you *have* to understand is that "neutral" can be very
context-dependent. Changes in a locus that have a small selective advantage such
that they will be swept to fixation in a large population, may well be neutral in
a small population (drift effects). If a selectively advantageous mutation pops
up, it will be swept to fixation (ie, be substituted) much more rapidly than
neutral mutations. The other thing to remember is that it is much more likely
that a neutral allele will be lost than that it will be fixed. Avg. time to loss
of a neutral allele destined to be lost = (2Ne/N)ln(2N) generations. Avg. time to
fixation of a neutral allele destined to be fixed = 4Ne generations (Kimura
1983). (Ne is effective population size; N is census size.) Neutral mutations
that eventually substitute take a very long time, while those that will be lost do
so relatively quickly. Result: The prob. that a new neutral allele will be lost
>> prob. same will substitute.

> > Molecular
> >distances track *substitution* rates, not mutation rates, another very
> >important
> >distinction which I encourage you to read and think about.
>

> Since only neutral or adaptive changes are fixated, substitution rate is, in
> effect, mutation rate.
>

Again, this is a common misconception about mutation vs. substitution and
application of the neutral theory. Time to rediscover the application of
population genetics to molecular systematics. The substitution rate will always
be < mutation rate. Changes in the third position of a codon will often be
neutral which will allow them to persist and fix more often than mutations at
second positions, for which mutations will often be deleterious, and thus
eliminated by purifying selection. This has nothing to do with the intrinsic
mutation rate; everything to do with the substitution rate. This is also the
reason sites such as 3rd positions saturate more rapidly and lose their
phylogenetic information over long enough periods of time.

Regards,

Matthew Fain

[Peter Nyikos]

Matthew Fain <matt...@saluki-mail.siu.edu> writes:

>Cal King wrote:

> In article <34CA14B0...@saluki-mail.siu.edu>,
> matt...@saluki-mail.siu.edu (Matthew Fain) wrote...
> >

> >Many happen to be experts on morphology and the evolution of those characters
> >for their particular group.

>> But cladistic methodology forbids them from applying their expertise in the
>> choice of and analysis of characters because doing so would be "subjective."

>Absolutely untrue. There is nothing subjective about applying the most likely
>explanation to all the available evidence.

The available evidence may be strongly biased and
anyone who attaches confidence to "the most likely explanation"
is making a subjective decision. See the alpha hemoglobin
data in

Fushitani K, Higashiyama K, Moriyama EN, Imai K, Hosokawa K,
Evolution of Amniote Hb alpha Globin Sequences
Mol Biol Evol 1996 Sep;13(7):1039-1043

and my reply to Harshman on "Molecular methods (was: Benton's book)".

The authors did not bat an eye at the conclusion that birds
are more closely related to mammals than to lizards. They
seemed to accept it as a matter of course.

> > You have yet to present a viable alternative model of character evolution
> >that systematists may use to infer phylogenetic trees from morphological
> >characters.

>> I have. Use only synapomorphs and make damn sure that they ARE or at least
>> likely to be synapomorphs by using character analysis OF the characters before
>> one embarks on an analysis ON the characters, while keeping in mind that

>This *is* pretty much the method. Determine which characters are more likely to
>be representing phylogenetic correlation, and which are more likely to have been
>subject to adaptive change.

Thereby making a "subjective" decision? Tom DiBenedetto
calls EVERY method that doesn't give equal weight to
ALL characters "subjective".

That's the Charybdis you are in danger of if you avoid the
Scylla of poorly chosen samples of characters:

Another reason that phylogenies proposed by some cladists
may run counter to long-accepted phylogenetic hypotheses
is the matter of character selection. Little attention has been
paid to how particular characters and taxa are chosen for
cladistic analysis. This problem is obscured in many works
on Henningian methodology by the use of hypothetical examples
and the use of letters and numbers to represent characters
and taxa. Study of the writings of some cladists suggest
that the choice of characters is essentially arbitrary.
Lovtrup (1985), giving his own examples and citing from other
recent papers, listed 22 characters that unite birds and
mammals, 14 that establish crocodiles as the sister-group
of a combined taxon including birds and mammals, and 16 that
unite chelonians and thecondontians. These are chosen from
the nearly limitless number of characters exhibited by these
groups. The choice of other characters has led previous authors
to entirely different interpretations of the relationships of
these groups. Clearly, there must be some objective criteria
for selecting characters for analysis.
--Robert L. Carroll and Philip J. Currie, "The Early
Radiation of Diapsid Reptiles," pp. 354-424 in:
_Origins of the Higher Groups of Tetrapods: Controversy and Consensus_,
ed. by Hans-Peter Schultze and Linda Trueb, co. 1991, Cornell
University Press, at p. 359.

[...]

>Cladists were ensconced in systematics well before the '80s, and during the '70s
>they had nothing like the sophistication we currently have wrt to molecular data
>and analytical techniques.

That sophistication may be there in the Fushitani et al paper,
but they lose sight of the forest for the trees.

Peter Nyikos -- standard disclaimer --
Professor, Dept. of Mathematics
University of South Carolina
Columbia, SC 29208

[Matt Fain]

Peter Nyikos wrote in message <6al5s0$8...@redwood.cs.sc.edu>...

>
>The authors did not bat an eye at the conclusion that birds
>are more closely related to mammals than to lizards. They
>seemed to accept it as a matter of course.
>

And it is these sorts of conclusions without reference to other
evidence that got molecular systematists into so much trouble with
more traditional systematists (by which I suppose I mean here
morphologists; there is no longer any clear demarcation between
these two "camps").

>Thereby making a "subjective" decision? Tom DiBenedetto
>calls EVERY method that doesn't give equal weight to
>ALL characters "subjective".
>

It's not a subjective decision if the evidence clearly allows you to
make the assumption of greater or less weight. Besides the methods
used are generally clearly presented by the author of the study, so
there's no need for guesswork. (If this isn't the case, the work
shouldn't even have made it into peer-reviewed print.) If you have
a problem with DiBenedetto's approach that's one thing. Do keep in
mind that he's not the only cladist out there, and not even all
cladists would agree with his opinions.

> these groups. Clearly, there must be some objective criteria
> for selecting characters for analysis.

Yes, analysis of phylogenetic information for the characters.

>>Cladists were ensconced in systematics well before the '80s, and
during the '70s
>>they had nothing like the sophistication we currently have wrt to
molecular data
>>and analytical techniques.
>
>That sophistication may be there in the Fushitani et al paper,
>but they lose sight of the forest for the trees.

Not having seen the paper, I can't say whether their analysis was
"sophisticated" or not. You still haven't answered my question as
to whether their true focus was on evolution of the species or
evolution of the molecules. It DOES make a difference.

Regards,

Matthew Fain

[Peter Nyikos]

"Matt Fain" <matt...@siu.edu> writes:

>Peter Nyikos wrote in message <6al5s0$8...@redwood.cs.sc.edu>...
>>
>>The authors did not bat an eye at the conclusion that birds
>>are more closely related to mammals than to lizards. They
>>seemed to accept it as a matter of course.
>>

>And it is these sorts of conclusions without reference to other
>evidence that got molecular systematists into so much trouble with
>more traditional systematists (by which I suppose I mean here
>morphologists; there is no longer any clear demarcation between
>these two "camps").

[...]

>> these groups. Clearly, there must be some objective criteria
>> for selecting characters for analysis.

>Yes, analysis of phylogenetic information for the characters.

That's like pulling oneself up by the bootstraps (no pun
intended): how do you arrive at the phylogenetic information,
if not by prior analysis of characters?

>>>Cladists were ensconced in systematics well before the '80s, and
>during the '70s
>>>they had nothing like the sophistication we currently have wrt to
>molecular data
>>>and analytical techniques.
>>
>>That sophistication may be there in the Fushitani et al paper,
>>but they lose sight of the forest for the trees.

>Not having seen the paper, I can't say whether their analysis was
>"sophisticated" or not. You still haven't answered my question as
>to whether their true focus was on evolution of the species or
>evolution of the molecules. It DOES make a difference.

Their true focus was on evolution of the species. Evolution
of the molecules was a sideshow, one to which they devoted
less than 10% of their effort, and it shows.

[Cal King]

In article <6angij$n...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter
Nyikos) wrote...

>
>"Matt Fain" <matt...@siu.edu> writes:
>
>>Peter Nyikos wrote in message <6al5s0$8...@redwood.cs.sc.edu>...
>>>
>>>The authors did not bat an eye at the conclusion that birds
>>>are more closely related to mammals than to lizards. They
>>>seemed to accept it as a matter of course.

To be fair, it is what their data shows, so they have to accept it. Whether
it conflicts other data sets based on different characters or different
molecules is something that has to be resolved.

>>And it is these sorts of conclusions without reference to other
>>evidence that got molecular systematists into so much trouble with
>>more traditional systematists (by which I suppose I mean here
>>morphologists; there is no longer any clear demarcation between
>>these two "camps").

I do agree that molecular biologists who have little training in systematics
often misinterpret data and thus often present laughably inaccurate
conclusions. Gregory Paul, the author of Predatory Dinosaurs of the World,
nevertheless noted (p.175) that cladograms were often contradicted by DNA-DNA
hybridization studies. He suggested that either "the best cladograms are
grossly in error" or DNA studies are wrong or misleading. Paul concluded that
cladograms "map character patterns, not necessarily true relationships." I
agree. Since morphological evolution is highly variable in rate, some members
of a group that is lumped together by a "synapomorph" may have closer temporal
relationships with members of other groups (also delimited by morphology
alone) than they do with members of their own clade.

Take the following cladogram based on cladistically analyzed mtDNA data by
Lopez and Maxson (1995, Biochem. Systematics and Ecol. 23[5]:487-505, fig.2)

___ Coluber constrictor priapis
|
___| ____ Coluber constrictor
|_____|
|____ Masticophis flagellum

Morphologically the 2 populations of Coluber were thought to be the same
species by most workers. Yet molecular data reveal a hidden paraphyletic
relationship with another genus (also delimited by morphology alone). Since
evolution occurs by speciation (higher taxa do not beget higher taxa), imagine
this sort of unequal evolutionary rate writ large and duplicated on every
single branch of the tree of life. Any effort to group organisms on the basis
of morphology would result in paraphyletic taxa because highly divergent crown
species would most likely be excluded, and what is thought to be a synapomorph
that unites a genus like Coluber is more often than not a symplesiomorph in
reality. The same pattern of evolution is duplicated in DNA studies of the
genus Clemmys (Bickham, J.W. 1996 Molecular Systematics of the genus Clemmys
and the intergeneric relationships of Emydid Turtles. Herpetologica
52(1):89-97.) and the same pattern is most likely duplicated in most branches
of the tree of life.

G.L. Stebbins was correct that "Fickle Dame Nature has very different ends in
view from that of creating neat hierarchies of species and genera, which
naturalists can file away tidily in cabinets with the least possible trouble"
(Stebbins, 1950. Variation and evolution in plants. New York & London, p. 28).
G.G. Simpson was also correct when he observed that "Evolution does not work
in a straight line, but the minds of some scientists do."

[Cal King]

In article <34d130c4...@news.su.se>, michae...@nrm.se (Mike Noren)
wrote...
>
>Thusly "Matt Fain" <matt...@siu.edu> spake unto :
>
>: >Thereby making a "subjective" decision? Tom DiBenedetto

>: >calls EVERY method that doesn't give equal weight to
>: >ALL characters "subjective".
>:
>: It's not a subjective decision if the evidence clearly allows you to
>: make the assumption of greater or less weight. Besides the methods
>: used are generally clearly presented by the author of the study, so
>: there's no need for guesswork. (If this isn't the case, the work
>: shouldn't even have made it into peer-reviewed print.) If you have
>: a problem with DiBenedetto's approach that's one thing. Do keep in
>: mind that he's not the only cladist out there, and not even all
>: cladists would agree with his opinions.
>

>If any, including DiBenedetto, does; Peter is probably misremembering.
>What he is likely referring to, although badly garbled, is that I, not
>DiBenedetto, have said that THE VALUE of a-priori weights, NOT the
>methods used to evaluate them, are arbitrary and impossible to
>determine in an objective fashion, and that using equal weights
>therefore removes a source of subjectiveness.
>
>: Matthew Fain
>
>
>Michael Norén, Doctoral student, Tel: Int +46 (0)8 6664236
>Swedish Museum of Natural History, Fax: Int +46 (0)8 6664125
>Dept. of Invertebrate Zoology
>P.O.B. 50007
>S-104 05 Stockholm, Sweden

Giving all characters equal weight is a subjective decision in itself.

[Peter Nyikos]

get...@nobull.net (Cal King) writes:

>In article <6angij$n...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter
>Nyikos) wrote...
>>
>>"Matt Fain" <matt...@siu.edu> writes:
>>
>>>Peter Nyikos wrote in message <6al5s0$8...@redwood.cs.sc.edu>...
>>>>
>>>>The authors did not bat an eye at the conclusion that birds
>>>>are more closely related to mammals than to lizards. They
>>>>seemed to accept it as a matter of course.

>To be fair, it is what their data shows, so they have to accept it. Whether
>it conflicts other data sets based on different characters or different
>molecules is something that has to be resolved.

They could, however, have remarked on the singular nature of
the conclusions. They devoted quite a lot of space to the
controversy over whether various kinds of molecular data
supported grouping mammals as the sister group of Aves even
over crocodilia.

And even there, they did a slipshod job because they did not
explicitly mention earlier studies using alpha-hemoglobin which
did just that. Nor could you even infer from what they wrote
that such studies existed.

I haven't chased down the references yet, but my source for
these earlier studies is:

Michael J. Benton, "Amniote Phylogeny", in:

_Origins of the Higher Groups of Tetrapods: Controversy and Consensus_,
ed. by Hans-Peter Schultze and Linda Trueb, co. 1991, Cornell

University Press, pp. 317-330, at p. 323.

Benton refers to four studies, at least one of which found the
following grouping for alpha hemoglobin:

[Crocodilian [Bird [Mammal]]]

and some unspecified subset of the four found the following for Myoglobin:

[[[[Turtle] Lizard] Crocodilian] [Bird [Mammal]]]

or

[[[[Turtle] Crocodilian] Lizard] [Bird [Mammal]]]

Beta-hemoglobin fared a little better--one study actually
agreed with the fossil evidence, but the other two did not,
including one which agreed that birds are more closely
related to mammals than to crocodiles.

>>>And it is these sorts of conclusions without reference to other
>>>evidence that got molecular systematists into so much trouble with
>>>more traditional systematists (by which I suppose I mean here
>>>morphologists; there is no longer any clear demarcation between
>>>these two "camps").

>I do agree that molecular biologists who have little training in systematics
>often misinterpret data and thus often present laughably inaccurate
>conclusions. Gregory Paul, the author of Predatory Dinosaurs of the World,
>nevertheless noted (p.175) that cladograms were often contradicted by DNA-DNA
>hybridization studies. He suggested that either "the best cladograms are
>grossly in error" or DNA studies are wrong or misleading. Paul concluded that
>cladograms "map character patterns, not necessarily true relationships." I
>agree.

But the grouping

[Mammal [Lizard [Crocodilian [Bird]]]]

has been a mainstay of traditional systematics all thru this
century AFAIK, including well before the advent of cladistics.

[Peter Nyikos]

get...@nobull.net (Cal King) writes:

>In article <34d130c4...@news.su.se>, michae...@nrm.se (Mike Noren)
>wrote...
>>
>>Thusly "Matt Fain" <matt...@siu.edu> spake unto :
>>
>>: >Thereby making a "subjective" decision? Tom DiBenedetto
>>: >calls EVERY method that doesn't give equal weight to
>>: >ALL characters "subjective".
>>:
>>: It's not a subjective decision if the evidence clearly allows you to
>>: make the assumption of greater or less weight. Besides the methods
>>: used are generally clearly presented by the author of the study, so
>>: there's no need for guesswork. (If this isn't the case, the work
>>: shouldn't even have made it into peer-reviewed print.) If you have
>>: a problem with DiBenedetto's approach that's one thing. Do keep in
>>: mind that he's not the only cladist out there, and not even all
>>: cladists would agree with his opinions.
>>
>>If any, including DiBenedetto, does; Peter is probably misremembering.

DiBenedetto supported Noren on this one IIRC.

>>What he is likely referring to, although badly garbled, is that I, not
>>DiBenedetto, have said that THE VALUE of a-priori weights, NOT the
>>methods used to evaluate them, are arbitrary and impossible to
>>determine in an objective fashion, and that using equal weights
>>therefore removes a source of subjectiveness.

Well, I didn't want to bring Noren back into the discussion since
he and I stopped following up to each other over half a year ago,
but it is true that he was the most vociferous proponent of
"equal weighing of anatomical characters" while Tom, less
vociferously, supported equal weighing of ALL characters,
molecular as well as anatomical, IIRC.

Noren actually relented to the extent of opting for two
kinds of characters, morphological and molecular, and using
the consensus tree for the two, and Tom may have gone along
with that too.

>>: Matthew Fain
>>
>>
>>Michael Norn, Doctoral student, Tel: Int +46 (0)8 6664236

>>Swedish Museum of Natural History, Fax: Int +46 (0)8 6664125
>>Dept. of Invertebrate Zoology
>>P.O.B. 50007
>>S-104 05 Stockholm, Sweden

>Giving all characters equal weight is a subjective decision in itself.

I kept telling Noren this, but it was like beating my head
against a stone wall.

Peter Nyikos -- standard disclaimer --

University of South Carolina

[Tom DiBenedetto]

Several people have written:

>>> Tom DiBenedetto
>>>: >calls EVERY method that doesn't give equal weight to
>>>: >ALL characters "subjective".

Unless there is proposed some objective weighting scheme. I know of
none.

> Tom, less
>vociferously, supported equal weighing of ALL characters,
>molecular as well as anatomical, IIRC.

yes, because Tom sees characters as hypotheses of homology, with
inherintly equal (potential) validity, and subject to testing against
eachother. Not as outputs from an evolutionary model which can be
accorded probailistic scores.

>Noren actually relented to the extent of opting for two
>kinds of characters, morphological and molecular, and using
>the consensus tree for the two, and Tom may have gone along
>with that too.

no, Tom favors using a total evidence approach. It is one of the
distinct advantages of cladistic/homology/parsimony approahces that
historical evidence from all parts of the organism can be assessed.

>>Giving all characters equal weight is a subjective decision in itself.

Only within a probailistic approach, wherein it could be argued that
equal weighting is but one particular weighting scheme. My perspective
is fundamentlaly different.

Tom DiBenedetto td...@umich.edu
Museum of Zoology http://www-personal.umich.edu/~tdib
University of Michigan 734-647-2192

[Tom DiBenedetto]

Mike Noren wrote:

>Thusly td...@umich.edu (Tom DiBenedetto) spake unto all:
>
>: >>> Tom DiBenedetto

>: >>>: >calls EVERY method that doesn't give equal weight to
>: >>>: >ALL characters "subjective".
>:
>: Unless there is proposed some objective weighting scheme. I know of
>: none.
>

>METHOD, Tom? Don't you mean 'a-priori weighting scheme'? Method, to
>me, in this context, would be e.g. parsimony analysis.

Hi Mike, you unscrupulous cladomaniac you, :)
so what is this METHOD, TOM stuff,,,???
that was sweet little Peter (I think) who wrote the first phrase
above, and my response follows. Scheme is the word alright!

>How do you feel about a-posteriori weighting schemes (successive
>weighting)? I've tried various aposteriori weighting schemes, and must
>say the result has been varied.

I am not convinced. I am not comfortable with all the talk about
"reliability" of characters. I am reluctant to go along with the
notion that a character state which seems to change 5 times is more
likely to change a sixth, relative to the liklihood of one which
changes once changing a second time. Seems like a liklihood argument
to me :). Homoplasies are independant. I dont wish to make my
phylogeny contingent on an assumption that there is a force regulating
the rate of character changes in such a stark manner. But I'll admit
that the arguments pro are much stronger than they are for aprioir
weighting. Have you ever pushed SF on this?

[Tom DiBenedetto]

Mike Noren wrote:

>BTW, if you don't already know - you get about as much abuse from
>Peter as I do. Try doing a search for your name in talk.origins.

Uggh,,I have actually never even read that group. Knowing it is
populated by Nyikos types,,,why would I? He (and his soulmate "cal
-april fool-king" have done so much to inhibit intellegent respectful
discussion on these groups, the last thing I want to do is march into
their lair.

>Maybe you should read what you reply to... ;-) Peters claim, included
>at the top, was that you claimed that the METHOD was subjective, and
>you seemed to agree. I didn't think you wanted to go on record as
>saying that, not to a person likely to re-post it ad infinitum.

I see your point. Unfourtunatly if he doesnt have a direct quote which
he can distort and repost, he will simply mis-characterize my thoughts
in his own words. I have given up asking him to stop that. I just rest
on the assumption that everyone can see him for what he is, and judge
his assertions on that basis.

re. successive weighting
: one amplifies ANY signal
>present in the data -including homoplasy- and get increased resolution
>at the price of decreased accuracy.

I guess it is the impulse to pretend one has the complete answer, when
in fact the data doesnt really have that to offer. Remember how the
anonymous weenie was ragging on Darrell Frost because he had the
routine honesty to present a set of equally parsimonious trees? I
guess honesty is a crime in some peoples eyes.....

>BTW, I was at an interesting seminar recently, where a researcher
>presented preliminary evidence, based on simulations, that Goloboff
>weighting gets more accurate than Farris weighting as homoplasy
>increases.

hmmm,,we have a compatability type here who has come up with a more
precise method of devising Goloboff-type cost curves - nice work if
you are into all that,,,,I aint.

ciao...

[Peter Nyikos]

td...@umich.edu (Tom DiBenedetto) writes:

>Several people have written:

[Nyikos:]

>>>> Tom DiBenedetto
>>>>: >calls EVERY method that doesn't give equal weight to
>>>>: >ALL characters "subjective".

>Unless there is proposed some objective weighting scheme. I know of
>none.

Nice to know I got it right, at least in practice if not in
principle.

[Nyikos, again:]

>> Tom, less
>>vociferously, supported equal weighing of ALL characters,
>>molecular as well as anatomical, IIRC.

>yes, because Tom sees characters as hypotheses of homology, with
>inherintly equal (potential) validity, and subject to testing against
>eachother. Not as outputs from an evolutionary model which can be
>accorded probailistic scores.

That word "potential" hides a multitude of sins. Sure,
having histidine at E7 in adult alpha hemoglobin is "potentially"
as valid as having the male be the heterogamete (as in humans)
rather than the female (as in birds) but I pity the poor
cladist who treats them as *prima facie* of equal importance.

>>Noren actually relented to the extent of opting for two
>>kinds of characters, morphological and molecular, and using
>>the consensus tree for the two, and Tom may have gone along
>>with that too.

>no, Tom favors using a total evidence approach.

Possible Application: a pair of changes from histidine to glutamine at E7 is
no less likely to be a homoplasy *a priori* than a pair of changes
from the male being the heterogamete to the female being the
heterogamete.

It is one of the
>distinct advantages of cladistic/homology/parsimony approahces that
>historical evidence from all parts of the organism can be assessed.

And given equal weight, since cladistic philosophies are stuck
on a junior high school concept of what Ockham's Razor is all about.

Or is there a cladist in this ng who dares incur the displeasure
of the illustrious DiBenedetto by saying that the "total evidence"
approach is naive?

>>>Giving all characters equal weight is a subjective decision in itself.

>Only within a probailistic approach, wherein it could be argued that
>equal weighting is but one particular weighting scheme. My perspective
>is fundamentlaly different.

Yes, see Possible Application above. Not a subjective application,
no siree.

It's objective; this I know
For Ockham's Razor tells me so.

;-)

Peter Nyikos -- standard disclaimer --

[Tom DiBenedetto]

Peter Nyikos wrote:

>having histidine at E7 in adult alpha hemoglobin is "potentially"
>as valid as having the male be the heterogamete (as in humans)
>rather than the female (as in birds)

yes, very good, because the "validity" refers tp whether the instances
of the similar state are actually historically homologous.

>but I pity the poor
>cladist who treats them as *prima facie* of equal importance.

I wonder what you mean by "importance". The goal here is to identify
heritable characters at the level at which they arose in a lineage. If
you correctly identify one, it is a correct synapomorphy. The test of
congruence is a test of hypotheses against eachother; it is not an
assessment of structural or physiological "importance". There are
plenty of molecular characters that are monomorphic across a far wider
range of taxa than are some heterogametic patterns.

>>Tom favors using a total evidence approach.
>
>Possible Application: a pair of changes from histidine to glutamine at E7 is
>no less likely to be a homoplasy *a priori* than a pair of changes
>from the male being the heterogamete to the female being the
>heterogamete.

very good Peter.

>And given equal weight, since cladistic philosophies are stuck
>on a junior high school concept of what Ockham's Razor is all about.

You just cant resist these little junior high school insults can you
Mr. Professor (you who have a junior high school concept of
systematics, and proud of it).

>Or is there a cladist in this ng who dares incur the displeasure
>of the illustrious DiBenedetto by saying that the "total evidence"
>approach is naive?

Or perhaps my fellow cladists who disagree might express their
disagreement in a somewhat more substantive manner.

>>>>Giving all characters equal weight is a subjective decision in itself.
>
>>Only within a probailistic approach, wherein it could be argued that
>>equal weighting is but one particular weighting scheme. My perspective
>>is fundamentlaly different.
>
>Yes, see Possible Application above. Not a subjective application,
>no siree.

no, it isnt actually

> It's objective; this I know
> For Ockham's Razor tells me so.

and it really doesnt have anything to do with Occams razor. If you
read your own "possible application" you should be able to see that it
refers to an unwillingness to make a priori judgements about the
relative liklihoods of evolutionary transformations. This has
nothing to do with parsimony per se. Those cladists who approve of
weighting do not thereby abandon parsimony.

[Peter Nyikos]

td...@umich.edu (Tom DiBenedetto) writes:

>Mike Noren wrote:

>>Thusly td...@umich.edu (Tom DiBenedetto) spake unto all:
>>

>>: >>> Tom DiBenedetto

>>: >>>: >calls EVERY method that doesn't give equal weight to
>>: >>>: >ALL characters "subjective".
>>:
>>: Unless there is proposed some objective weighting scheme. I know of
>>: none.
>>

>>METHOD, Tom? Don't you mean 'a-priori weighting scheme'? Method, to
>>me, in this context, would be e.g. parsimony analysis.

>Hi Mike, you unscrupulous cladomaniac you, :)

Cladomaniac: A person who wants to ram cladistic classification
down the throats of systematists, AND deliberately uses unscrupulous methods
to achieve this end.

Mike qualifies beyond a reasonable doubt. You, Tom, I'm willing to give
the benefit of the doubt.

You are certainly a cladophile (one who wants to see paraphyletic
taxa banned) but your misrepresentations of me MIGHT be due
to innocent misunderstandings rather than malice as is true
beyond a reasonable doubt where Mike is concerned.

Peter Nyikos -- standard disclaimer --

University of South Carolina

[Peter Nyikos]

td...@umich.edu (Tom DiBenedetto) writes:

>Mike Noren wrote:

>>BTW, if you don't already know - you get about as much abuse from
>>Peter as I do. Try doing a search for your name in talk.origins.

>Uggh,,I have actually never even read that group. Knowing it is
>populated by Nyikos types,,,

Wrong. It has one or two Nyikos types and all too many anti-Nyikos
types who use every dirty debating trick in the book, and then
some, to discredit me.

>why would I? He (and his soulmate "cal
>-april fool-king" have done so much to inhibit intellegent respectful
>discussion on these groups,

On the contrary, it is the anti-Nyikos types who do that,
thereby goading me to say things about them that are almost
perfectly like the things you are saying about me here, Tom.

Thus, they have goaded me to be like you in this respect,
and have fooled you into thinking that my nasty remarks
about them are qualitatively worse than your nasty remarks about me.

the last thing I want to do is march into
>their lair.

Fortunately, the only person in this newsgroup who is at home
in such a den of iniquity is Mike Noren himself, unless you are
hiding some of your REALLY nasty qualities from us, Tom.

>>Maybe you should read what you reply to... ;-) Peters claim, included
>>at the top, was that you claimed that the METHOD was subjective, and
>>you seemed to agree. I didn't think you wanted to go on record as
>>saying that, not to a person likely to re-post it ad infinitum.

>I see your point. Unfourtunatly if he doesnt have a direct quote which
>he can distort and repost, he will simply mis-characterize my thoughts
>in his own words.

Hmmm... this is EXACTLY the kind of statement that has earned
for me the undying enmity of all too many people in talk.origins.
The difference is, I have much better grounds for making such
statements than you have here.

Note, Tom, that it even seemed to MIKE that you agreed with him.
Now, YOU seem to think that Mike does not have a deceitful bone
in his body. Why, then, do you attribute my misunderstanding
(if misunderstanding it was) to malice?

> I have given up asking him to stop that.

Since I never started doing it deliberately, it's a good thing
you DID stop.

I just rest
>on the assumption that everyone can see him for what he is, and judge
>his assertions on that basis.

Too bad you can't seem to see me for what I am, Tom: a tireless
(if too outspoken in your eyes) crusader for justice and fair play.

One would think that my willingness to abide by a dual system
of classification (cladistic and Linnean side by side like
the Dewey Decimal and Library of Congress systems) would help
you understand my commitment to fair play, but perhaps you
think anyone who does not wish to abolish paraphyletic taxa
despite your
unanswerable :-)
arguments is the unreasonable kind.

>re. successive weighting
>: one amplifies ANY signal
>>present in the data -including homoplasy- and get increased resolution
>>at the price of decreased accuracy.

>I guess it is the impulse to pretend one has the complete answer, when
>in fact the data doesnt really have that to offer. Remember how the
>anonymous weenie was ragging on Darrell Frost because he had the
>routine honesty to present a set of equally parsimonious trees? I
>guess honesty is a crime in some peoples eyes.....

Gosh, I sure hope you can see that in my eyes, it is DIShonesty
that is the crime.

[Cal King]

In article <6b7dgl$9...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter
Nyikos) wrote...

>But the grouping

>
> [Mammal [Lizard [Crocodilian [Bird]]]]
>
>has been a mainstay of traditional systematics all thru this
>century AFAIK, including well before the advent of cladistics.
>

>Peter Nyikos -- standard disclaimer --

The cynic may argue that some cladists "formalized" this set of established
relationships in a cladistic analysis in order to demonstrate the power of
cladistic analysis, whereas a truly *objective* analysis may have yielded
results similar to Gardiner's (1982) hypothesis that birds are more closely
related to mammals than they are to crocs.

Pritchard (1994 Herpet. Rev.) may be quite correct that:

"Wiser cladists simply dump computer output that seems wrong or goes against
their grain, and try again until the results parallel their preconceptions! I
have seen them do it."

He also observed (Ibid.):

"In truth, characters can be selected, or described in carefully chosen words,
to prove anything one wishes to prove, and this has often been done."

In conclusion, I would not put too much stock in the congruence between
cladistic analysis and long standing hypotheses of amniote relationships
based on traditional analysis. I would look instead for congruence between
fossil and biochemical evidence or congruence among different sets of
molecular data.

[Harry Erwin]

Peter Nyikos <nyi...@math.scarolina.edu> wrote:

> td...@umich.edu (Tom DiBenedetto) writes:
>
> >It is one of the
> >distinct advantages of cladistic/homology/parsimony approahces that
> >historical evidence from all parts of the organism can be assessed.
>

> And given equal weight, since cladistic philosophies are stuck
> on a junior high school concept of what Ockham's Razor is all about.
>

> Or is there a cladist in this ng who dares incur the displeasure
> of the illustrious DiBenedetto by saying that the "total evidence"
> approach is naive?
>

> >>>Giving all characters equal weight is a subjective decision in itself.
>
> >Only within a probailistic approach, wherein it could be argued that
> >equal weighting is but one particular weighting scheme. My perspective
> >is fundamentlaly different.
>
> Yes, see Possible Application above. Not a subjective application,
> no siree.
>

> It's objective; this I know
> For Ockham's Razor tells me so.
>

> ;-)
>

If you're using continuous characters, you can't give them equal weight,
because some characters naturally vary much more than others. Nasty
problem. I've seen several solutions:

1. Be arbitrary (_really_ doesn't work)
2. Look for clustering of values and assign each cluster an ordinal
3. Take the total range and slice it into equal segments
4. Scale the weights to spread out the species maximally
5. etc.

Something I tried was to recursively scale things so ancestor-descendant
relationships between chronospecies corresponded to an ordinal change of
no more than one in any character. I can't claim it worked any better
than the standard approaches.

In any case the fact that behind any binary or ordinal character lies a
continuous character suggests that giving all characters equal weight is
really begging the question.

--
Harry Erwin, her...@gmu.edu, http://osf1.gmu.edu/~herwin, Senior
Software Analyst for the FAA, PhD candidate modeling how bats
echolocate and lecturer for CS 211 (data structures and advanced C++).

[Peter Nyikos]

get...@nobull.net (Cal King) writes:

>In article <6b7dgl$9...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter
>Nyikos) wrote...

>>But the grouping
>>
>> [Mammal [Lizard [Crocodilian [Bird]]]]
>>
>>has been a mainstay of traditional systematics all thru this
>>century AFAIK, including well before the advent of cladistics.
>>
>>Peter Nyikos -- standard disclaimer --

>The cynic may argue that some cladists "formalized" this set of established
>relationships in a cladistic analysis in order to demonstrate the power of
>cladistic analysis, whereas a truly *objective* analysis may have yielded
>results similar to Gardiner's (1982) hypothesis that birds are more closely
>related to mammals than they are to crocs.

I very strongly doubt it. AFAIK Gardinier is not taken seriously by
systematists, except maybe Lovtrup:

Another reason that phylogenies proposed by some cladists
may run counter to long-accepted phylogenetic hypotheses
is the matter of character selection. Little attention has been
paid to how particular characters and taxa are chosen for
cladistic analysis. This problem is obscured in many works
on Henningian methodology by the use of hypothetical examples
and the use of letters and numbers to represent characters
and taxa. Study of the writings of some cladists suggest
that the choice of characters is essentially arbitrary.
Lovtrup (1985), giving his own examples and citing from other
recent papers, listed 22 characters that unite birds and
mammals, 14 that establish crocodiles as the sister-group
of a combined taxon including birds and mammals, and 16 that
unite chelonians and thecondontians. These are chosen from
the nearly limitless number of characters exhibited by these
groups. The choice of other characters has led previous authors
to entirely different interpretations of the relationships of

these groups. Clearly, there must be some objective criteria
for selecting characters for analysis.

--Robert L. Carroll and Philip J. Currie, "The Early

Radiation of Diapsid Reptiles," pp. 354-424 in:

_Origins of the Higher Groups of Tetrapods: Controversy and Consensus_,
ed. by Hans-Peter Schultze and Linda Trueb, co. 1991, Cornell

University Press, at p. 359.

>Pritchard (1994 Herpet. Rev.) may be quite correct that:

>"Wiser cladists simply dump computer output that seems wrong or goes against
>their grain, and try again until the results parallel their preconceptions! I
>have seen them do it."

>He also observed (Ibid.):

>"In truth, characters can be selected, or described in carefully chosen words,
>to prove anything one wishes to prove, and this has often been done."

Yes...and isn't that exactly what Gardinier did?

>In conclusion, I would not put too much stock in the congruence between
>cladistic analysis and long standing hypotheses of amniote relationships
>based on traditional analysis. I would look instead for congruence between
>fossil and biochemical evidence or congruence among different sets of
>molecular data.

I still give the fossil data pride of place when it is available.
Did Gardinier?

Peter Nyikos -- standard disclaimer --

[Cal King]

In article <6bo17s$f...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter
Nyikos) wrote:
>get...@nobull.net (Cal King) writes:

Interesting quote, Mayr and Ashlock also cited Gardiner's work as illustration
that it is NOT methodology but the choice of characters that really matters in
phylogenetic analysis, and as support for their claim that there really isn't
much difference between cladistic analysis and traditional methods, with the
notable exception of classificatory convention.

"Choice of Characters

Since two taxa may differ by an almost unlimited number of characters, the
working taxonomist is forced to make a very restricted selection. The ultimate
^^^^^^^^^^^^
classification may depend entirely on this selection. For instance Rosen et
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
al. (1981) on the basis of the characters they selected, came to the
conclusion that the Dipnoi (lungfishes) are the sister group of the tetrapods
while Holmes (1985), partly on the basis of different characters concluded
that the traditional classification according to which some group of
rhipidistians is the sister group of the tetrapods is far better documented.
The difficulty lay in the correct identification of the synapomorphies.
Similarly. the claim that birds and mammals are sister groups was based on a
strict cladistic analysis (Gardiner 1982) that used a very different set of
characters than did the traditional classification in which the birds are
derived from archosaurian ancestors and the mammals from the therapsids, two
very different groups of reptiles; this classification was confirmed by Kemp
(1988) and Gauthier, Kluge and Rowe (1988). Nor has strictly cladistic
analysis been able to establish the phylogeny of the lice, Mallophaga and
Anoplura (Lyal 1985 versus Kim et al. 1987. These cases show that a cladistic
analysis does not guarantee a correct cladogram unless it is based on valid
apomorphies."--Mayr and Ashlock (1991) [emphasis mine]

Hence it is not the method, but the characters, that really counts.

>>Pritchard (1994 Herpet. Rev.) may be quite correct that:

>>"Wiser cladists simply dump computer output that seems wrong or goes against
>>their grain, and try again until the results parallel their preconceptions!
>>I have seen them do it."
>
>>He also observed (Ibid.):
>
>>"In truth, characters can be selected, or described in carefully chosen
>>words, to prove anything one wishes to prove, and this has often been done."
>
>Yes...and isn't that exactly what Gardinier did?

Not quite. Gardiner may have chosen his characters randomly. In which case
he is being "objective." But his analysis shows that "objectivity" does not
guarantee success, something that Gould (1991 Paleobiology) argued. On the
other hand, Gauthier, by CHOOSING the same characters that Ostrom had used to
unite bird and theropod (what a coincidence!), showed that a cladistic
analysis of these same characters yielded the same conclusions as Ostrom.

As further illustration, Gould (1991. Paleobiology 17[4]:414) criticized
Briggs and Fortey's (1989, Sci. 246:241-243) cladogram for "... the surprising
arrangement of crustaceans at the primitive and trilobites at the most derived
position." A subsequent analysis by these authors (Wills et al. 1994,
Paleobiology 20[2]: 93-130) did "correct" this anomaly but the C.I. for this
more recent study (<0.3) is even lower than the one in 1989 (<.4).

These examples show that subjective involvement by systematists, even in
cladistic analysis, can indeed reduce "anomalies" such as those of Gardiner
and Briggs and Fortey. There is the danger of course, that there is too much
subjectivity and we get exactly what Pritchard had warned, i.e. results that
fit a systematist's preconception.

[Peter Nyikos]

I have added s.b.p. to the newsgroups.

get...@nobull.net (Cal King) writes:

>In article <6bo17s$f...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter
>Nyikos) wrote:
>>get...@nobull.net (Cal King) writes:

>>>The cynic may argue that some cladists "formalized" this set of established
>>>relationships in a cladistic analysis in order to demonstrate the power of
>>>cladistic analysis, whereas a truly *objective* analysis may have yielded
>>>results similar to Gardiner's (1982) hypothesis that birds are more closely
>>>related to mammals than they are to crocs.
>>
>>I very strongly doubt it. AFAIK Gardinier is not taken seriously by
>>systematists, except maybe Lovtrup:

And Lovtrup, according to Carroll, used only extant organisms.
The rationale is that soft parts don't fossilize and so
we lose a lot of characters that could falsify a sister-group
relationship. But Carroll looks askance at this reliance
on extant organisms alone, for reasons I gave in an earlier
post today on these two newsgroups, in
"Darwin on Classifications (was birds and crocs)"

Well, the concepts of plesimorphy and apomorphy were certainly
well known to the traditionalists, under the better known
terms "primitive" and "derived" or even "advanced".

By the way, I found another possible reason for recognizing Synapsida
as a clade, besides the ones given earlier of the double condyle
and the epinomous condition of the lower temporal fenestra
(in humans, the space under the zygomatic arch). This in turn
is important because of the implications it has for mammals
having diverged from birds before the squamates (lizards, snakes)
did. It's the chapter by Nicholas Hotton III in the same book
cited above. [Little tidbit: "Synapsids" below means "Synapsida minus
Mammalia".]

Synapsids conform to the traditional definition of
reptiles in having several bones behind the dentary in
the lower jaw, one of which articulates with the quadrate
bone in the skull, and in having a single ossicle, the
stapes, in the middle ear. In pelycosaurs, the stapes
approaches the quadrate, and during life probably was
connected with it by cartilage (Romer and Price, 1940;
Brinkman and Eberth, 1983). This is probably the primitive
condition, because it is shared by captorhinomorphs and
other primitive nonsynapsid reptiles. In turtles and
advanced diapsid reptiles, the contact between stapes
and quadrate is lost or much reduced, and the slender
stapes points toward a concavity at the back of the quadrate,
which in most living forms supports a tympanic membrane.
Therapsids differ from living reptiles because the
condition of the pelycosaurs is exaggerated in therapsids;
thus, tha stapes is in direct contact with the quadrate,
bone to bone. This articulation resembles a diarthosis,
homologous to the joint between the mammalian stapes
and incus, the latter being the homologue of the reptilian
quadrate. The direct contact between stapes and quadrate
is a primary feature by which therapsids are defined
as mammal-like. [pp. 600-1]

This has long been known and much talked about by classical
systematists; note the reference to a joint 1940 paper by Romer.

Hotton goes on to point out that not only the temporal fenestrae
but also the muscles that move the jaw are different in
modern reptiles and synapsids/mammals; they are attached differently.

>"Choice of Characters

>Since two taxa may differ by an almost unlimited number of characters, the
>working taxonomist is forced to make a very restricted selection.

You'd think high-speed supercomputers would have made this obsolete,
but just think of the near-impossibility of putting whole genomes
into a cladistic analysis. Yet Tom DiBenedetto clings to
the "total evidence" theory. I wonder who else does.

>The ultimate

>classification may depend entirely on this selection. For instance Rosen et
>^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
>al. (1981) on the basis of the characters they selected, came to the
>conclusion that the Dipnoi (lungfishes) are the sister group of the tetrapods
>while Holmes (1985), partly on the basis of different characters concluded
>that the traditional classification according to which some group of
>rhipidistians is the sister group of the tetrapods is far better documented.

Yeah, but Eusthenopteron, that old standby, has been dethroned
and now panderichthyan fishes are the prime candidates.

>The difficulty lay in the correct identification of the synapomorphies.
>Similarly. the claim that birds and mammals are sister groups was based on a
>strict cladistic analysis (Gardiner 1982) that used a very different set of
>characters than did the traditional classification in which the birds are
>derived from archosaurian ancestors and the mammals from the therapsids, two
>very different groups of reptiles; this classification was confirmed by Kemp
>(1988) and Gauthier, Kluge and Rowe (1988).

The *traditional* classification was, yes.

>apomorphies."--Mayr and Ashlock (1991) [emphasis mine]

Peter Nyikos -- standard disclaimer --

[Jason Anderson]

Reply attempt take two...

***************************
Cal King wrote:
<snip>

> Interesting quote, Mayr and Ashlock also cited Gardiner's work as illustration
> that it is NOT methodology but the choice of characters that really matters in
> phylogenetic analysis, and as support for their claim that there really isn't
> much difference between cladistic analysis and traditional methods, with the
> notable exception of classificatory convention.
>
> "Choice of Characters
>
> Since two taxa may differ by an almost unlimited number of characters, the
> working taxonomist is forced to make a very restricted selection. The ultimate
> ^^^^^^^^^^^^
> classification may depend entirely on this selection. For instance Rosen et
> ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> al. (1981) on the basis of the characters they selected, came to the
> conclusion that the Dipnoi (lungfishes) are the sister group of the tetrapods
> while Holmes (1985), partly on the basis of different characters concluded

^^^^^^

> that the traditional classification according to which some group of
> rhipidistians is the sister group of the tetrapods is far better documented.

Actually, the characters come from the chosen taxa. Rosen et al. (1981)
only included *extant* taxa-choices there being Dipnoans or
_Latimeria_, as sister-group to tetrapods. The computer is *always*
going to output a tree. When fossil taxa (and their often plesiomorphic
character distribution) are included Osteolepoformes appear as
sister-group to tetrapods. The issue here is included taxa. See below.

> The difficulty lay in the correct identification of the synapomorphies.

No. Given the taxa analysed, synapomorphies *were* correctly
identified-but those synapomorphies were applicable at a higher
level-Choanata. They left out the immediate sister-group because they
are only known from fossils. Patterson is renowned for his dislike of
fossils in cladistics.

> Similarly. the claim that birds and mammals are sister groups was based on a
> strict cladistic analysis (Gardiner 1982) that used a very different set of
> characters than did the traditional classification in which the birds are
> derived from archosaurian ancestors and the mammals from the therapsids, two
> very different groups of reptiles; this classification was confirmed by Kemp
> (1988) and Gauthier, Kluge and Rowe (1988).

Again, you miss the point. Gardiner also only included extant taxa.
Gauthier, Kluge and Rowe demonstrated that if fossils are excluded from
the analysis Gardiner's tree is recovered but, when fossils (especially
certain "therapsids") are included the tree topology changes to a more
"traditional" view-using the same characters as Gardiner. That the
plesiomorphic distribution of characters found in fossils is of great
importance to cladistics is one of the basic take-home messages of this
paper. You have read it?

BTW, the concept of "Homothermia" predates cladistics by about 100
years.

> Nor has strictly cladistic
> analysis been able to establish the phylogeny of the lice, Mallophaga and
> Anoplura (Lyal 1985 versus Kim et al. 1987. These cases show that a cladistic
> analysis does not guarantee a correct cladogram unless it is based on valid
> apomorphies."--Mayr and Ashlock (1991) [emphasis mine]
>

"Correct cladogram"? Correct in the eyes of Mayr and Ashlock, "Cal" or
who? Lakatos said that we will never know the *capital-t* Truth because
of the ontologic level upon which science operates. And yes, there are
many interesting questions left in the field of systematics.

> Hence it is not the method, but the characters, that really counts.
>

Hence it is the included taxa (and their distribution of characters)
that really count.

Jason

[Cal King]

In article <6bt4bb$j...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter Nyikos) wrote:
>I have added s.b.p. to the newsgroups.

>By the way, I found another possible reason for recognizing Synapsida

The evidence demonstrates that mammals are descended from a different group of
reptiles than the turtles and other diapsid reptiles, but it cannot tell us
whether squamates (or actually their ancestors) or the synapsids were the
first group to have branched off of the amniotes.

>>"Choice of Characters
>
>>Since two taxa may differ by an almost unlimited number of characters, the
>>working taxonomist is forced to make a very restricted selection.
>
>You'd think high-speed supercomputers would have made this obsolete,
>but just think of the near-impossibility of putting whole genomes
>into a cladistic analysis.

Computers can help in the analysis ON the characters, but they cannot help the
cladist weed out phylogenetic noise. To illustrate, I quote Greg Paul, who
wrote in p. 174-175 of "Predatoray Dinosaurs of the World" :

"Cladistics is a recently developed methodology, with ... many disagreements
over how it should be practiced. I follow a pragmatic approach, one that
falls back on common sense when all else fails."

Paul continued, "Quality is important as well as quantity.... After tabulating
the shared derived characters, the possibility of convergence and parallelism
are assessed. There is no simple formula for this, and a good deal of
judgment is involved. Bias in the characters chosen and the tremendous gaps
in our knowledge make things worse. Indeed, some recent studies have come up
with bizarre and misleading results because they looked at too few characters.
So, the more data the better. Yet the greater the number of species and
characters looked at, the less the human mind is able to take it all in and
weithg the various factors. Even computers are strained when assessing a
modest number of characters and taxa."

Computer technology has advanced by leaps and bounds since Paul wrote his
book, but probably not enough to permit the analysis of a vast number of
characters. Besides, as Paul pointed out, quality is just as important as
quantity. Signal can in fact be swamped by noise.

> Yet Tom DiBenedetto clings to
>the "total evidence" theory. I wonder who else does.

Total evidence is a nice way to "resolve" conflicts among data sets. Even
though there may be strongly conflicting results, just present a single
consensus cladogram and say it is constructed from a consideration of the
"total evidence." To give a recent example, even though Burke and Feduccia
provided strong evidence that bird hands are not homologous with theropod
hands, it is but one piece of evidence. Using the total evidence approach,
this one piece of evidence is given no more weight than any other pieces, and
since the hand evidence is outnumbered by other evidence, it can safely be
ignored and the "total evidence" still says birds are theropods. There is the
danger of course that a large number of homoplasies can outweigh an important
synapomorph. Hence, using "total evidence," the thylacine would be considered
a close relative of the placental wolf because the number of "synapomorphs"
would outnumber the contradictory evidence. Of course cladists cannot get
away with calling lumping the thylacine and wolf in a clade to the exclusion
of, say the ungulates. But they are doing precisely that with bird and
theropod, and they have done it before with owl and hawk and loon and grebe
(see Feduccia's critique of Cracraft's analyses in his 1996 book).

[Chris Brochu, from the Jan. Dinosaur List archives]:

"I agree entirely. This is why, in a phylogenetic analysis, *all* of the
relevant information is considered simultaneously. The dinosaurian nature
of birds is supported by several independent analyses relying on a large
body of data, not on single characters that, as you state, can be
misleading in the wrong context."

He is speaking of course of Burke and Feduccia's hand evidence. All of the
evidence, despite the hand evidence, still purportedly support a bird-theropod
link.

"The ultimate classification may depend entirely on this selection." --Mayr
and Ashlock.

As Mayr and Ashlock said, the outcome of any analysis depends on the choice of
characters. If we choose a lot of phylogenetic noise in cladistic analysis
and then use the "total evidence approach" we can definitely come up with a
wolf-thylacine clade, an owl-hawk clade, a loon-grebe clade and a
bird-theropod clade.

[Tom DiBenedetto]

Cal King wrote:

>Computer technology has advanced by leaps and bounds since Paul wrote his
>book, but probably not enough to permit the analysis of a vast number of
>characters.

Computers can handle vast numbers of characters. The limiting factor
is taxa. Beyond 17 or 18 taxa, it is impossible to examine every
posssible tree. By the time you get up into the 20's (number of taxa),
you end up with more possible topologies than there are molecules in
the universe.

>> Yet Tom DiBenedetto clings to
>>the "total evidence" theory. I wonder who else does.

Anyone who accepts as a basic principle the notion that relevant
evidence should not be ignored.

>Total evidence is a nice way to "resolve" conflicts among data sets. Even
>though there may be strongly conflicting results, just present a single
>consensus cladogram and say it is constructed from a consideration of the
>"total evidence."

Sorry "cal", but you dont understand the issues at all. There are two
senses in which "total evidence' is used in systematics. In a broad
sense (the sense I tend to use it in) it is a principle of insisting
that one consider all relevant evidence which bears on a question,
especially if one presumes to propose the "best" answer to the
question. In a more narrow sense, the term "total evidence" refers to
a postion (most prominently advocated by Kluge) in the dabate over
combining evidence from different sources into a single matrix, or
partioning data and using consensus techniques.
Thus your characterization of "total evidence" as "just present a
single consensus cladogram" is completely confusing the two positions
in that debate.

> Hence, using "total evidence," the thylacine would be considered
>a close relative of the placental wolf because the number of "synapomorphs"
>would outnumber the contradictory evidence.

Could you refer us to the analysis which proposed this? Or is this
just another false charge of yours?

[Matt Fain]

Tom DiBenedetto wrote in message
<34e446d9...@news.itd.umich.edu>...

>Cal King wrote:
>
>> Hence, using "total evidence," the thylacine would be considered
>>a close relative of the placental wolf because the number of
"synapomorphs"
>>would outnumber the contradictory evidence.
>

>Could you refer us to the analysis which proposed this? Or is this
>just another false charge of yours?
>
>
>

Absolutely false, it is. The number of characters that place the
thylacine within marsupials is far greater than the homoplastic
characters that would place it with the placental wolf. If he'd
said South American borhyaenids, he would have had more of a case,
since thylacines were allied borhyaenids based mostly on dentition.
Further molecular and morphological work demonstrates the
convergence in that character set. Poorly thought out example, I'm
afraid.

Matthew Fain

[Cal King]

In article <6bvim9$7...@saluki-news.it.siu.edu>, Matt Fain (matt...@siu.edu)
says...

>Absolutely false, it is. The number of characters that place the
>thylacine within marsupials is far greater than the homoplastic
>characters that would place it with the placental wolf.

That would depend on which characters one *chooses*. I can definitely choose
a set of characters that shows the thylacine to be more closely related to the
placental wolf than it is to, say, the kangaroo. BTW, you deleted my owl-hawk
and loon-grebe examples. Those must be well thought out examples of how
much more important the choice of characters is to methodology in the outcome
of an analysis.

[Jason Anderson]

> >
> "Correct cladogram"? Correct in the eyes of Mayr and Ashlock, "Cal" or
> who? Lakatos said that we will never know the *capital-t* Truth because

^^^^^^^

Um, that would be Popper. I was daydreaming about parsimony as a result
of the discussions on other threads, about how parsimony is one of the
criteria of Lakatos for comparing compeating hypotheses, while writing
the above. His name "popped" out without my realizing it. At least he
was a student of Poppers... :)

Jason

[Tom DiBenedetto]

Cal King wrote:

> Matt Fain says...

>
>>Absolutely false, it is. The number of characters that place the
>>thylacine within marsupials is far greater than the homoplastic
>>characters that would place it with the placental wolf.

>That would depend on which characters one *chooses*.

Gee "cal", I am so sorry for having somehow forgot to mention, in any
of my posts, the notion of using all available evidence.

[Cal King]

In article <34e36440...@news.itd.umich.edu>, td...@umich.edu wrote:
>Cal King wrote:
>
>> Matt Fain says...
>>
>>>Absolutely false, it is. The number of characters that place the
>>>thylacine within marsupials is far greater than the homoplastic
>>>characters that would place it with the placental wolf.
>
>>That would depend on which characters one *chooses*.
>
>Gee "cal", I am so sorry for having somehow forgot to mention, in any
>of my posts, the notion of using all available evidence.
>
>
>Tom DiBenedetto td...@umich.edu

Show me a cladistic analysis in which ALL (100%) available characters are
used.

[Cal King]

In article <34E3086B.72C4@no_spam.mcgill.ca>, Jason Anderson <jasona@no_spam.mcgill.ca> wrote:

>Actually, the characters come from the chosen taxa.

Of course, there would be no characters without organisms. Given a set of
organisms, the choice of characters and the judgment of their phylogenetic
value is also important.

>The computer is *always* going to output a tree.

Or more likely, a number of different trees. Does the computer make a better
phylogeneticist than the human?

>> The difficulty lay in the correct identification of the synapomorphies.
>

>No. Given the taxa analysed, synapomorphies *were* correctly
>identified-

Ah, but what would happen if homoplasies are misidentified as synapomorphs.
Would we then end up with something like a thylacine-placental wolf clade?

> Patterson is renowned for his dislike of fossils in cladistics.

He is not alone. Fossils often lack a large number of characters found
in extant organims. Since cladists like to use a large number of characters,
they can either dump the characters or dump the fossils.

>> Nor has strictly cladistic
>> analysis been able to establish the phylogeny of the lice, Mallophaga and
>> Anoplura (Lyal 1985 versus Kim et al. 1987. These cases show that a cladistic
>> analysis does not guarantee a correct cladogram unless it is based on valid
>> apomorphies."--Mayr and Ashlock (1991) [emphasis mine]
>>

>"Correct cladogram"? Correct in the eyes of Mayr and Ashlock, "Cal" or
>who?

How about the collective judgments of biologists world wide? When there is no
or little controversy, we assume that there is a consensus and that the tree
is probably correct. That doesn't mean that such a tree is The Truth, of
course.

>> Hence it is not the method, but the characters, that really counts.
>>

>Hence it is the included taxa (and their distribution of characters)
>that really count.
>
>Jason

It is actually the correct identification of which characters are most
informative of phylogenetic relationships that really counts. A computer
cannot make these judgments and may never be able to do so. Homoplasies in,
erroneous hypotheses out.

[Matt Fain]

Cal King wrote in message <6bvkvb$c...@bgtnsc02.worldnet.att.net>...

>In article <6bvim9$7...@saluki-news.it.siu.edu>, Matt Fain
(matt...@siu.edu)

>says...
>
>>Absolutely false, it is. The number of characters that place the
>>thylacine within marsupials is far greater than the homoplastic
>>characters that would place it with the placental wolf.
>

>That would depend on which characters one *chooses*. I can
definitely choose
>a set of characters that shows the thylacine to be more closely
related to the
>placental wolf than it is to, say, the kangaroo. BTW, you deleted
my owl-hawk
>and loon-grebe examples. Those must be well thought out examples
of how
>much more important the choice of characters is to methodology in
the outcome
>of an analysis.
>

WOW! Cal has done it again folks! Turned something I didn't say,
and that cannot be inferred from what I wrote, and turned it into
something to support his largely unfounded statements. Not to
mention that caricature of cladistic analysis. What a piece of
work! Let's guess at Cal's real avocation: hmmm...politician or
perhaps sensationalist journalism?

[Cal King]

In article <6c0145$b...@saluki-news.it.siu.edu>, matt...@siu.edu (Matt Fain)
wrote...

I have no idea what you are talking about. If you want to engage in ad
hominem attacks, that is your prerogative, but ad hominem attacks really add
nothing to the discussion. I suspect you have simply run out of arguments.

[Matt Fain]

Cal King wrote in message <6c05oo$k...@bgtnsc02.worldnet.att.net>...

>In article <6c0145$b...@saluki-news.it.siu.edu>, matt...@siu.edu
(Matt Fain)
>wrote...
>>
>>Cal King wrote in message
<6bvkvb$c...@bgtnsc02.worldnet.att.net>...
>>>In article <6bvim9$7...@saluki-news.it.siu.edu>, Matt Fain
>>(matt...@siu.edu)
>>>says...
>>>
>>>>Absolutely false, it is. The number of characters that place
the
>>>>thylacine within marsupials is far greater than the homoplastic
>>>>characters that would place it with the placental wolf.
>>>
>>>That would depend on which characters one *chooses*. I can
>>definitely choose
>>>a set of characters that shows the thylacine to be more closely
>>related to the
>>>placental wolf than it is to, say, the kangaroo. BTW, you
deleted
>>my owl-hawk
>>>and loon-grebe examples. Those must be well thought out examples
>>of how
>>>much more important the choice of characters is to methodology in
>>the outcome
>>>of an analysis.
>>>

Come on Cal. Do you actually think you're fooling anyone here?
Look at the above. I deleted your bird examples, so they "must be
well thought out"? What the hell kind of logic is that? Usually
you're going in so many directions at once, that WHEN and IF I
decide to add my 2c, I focus on ONE thing out of the MANY that are
erroneous. Another example: Just the other day you decided I
couldn't distinguish phenetics and evolutionary systematics on the
basis of my following statement, "Darwin recognized the importance
of phylogeny, and I'll leave it at that." Again, where the hell did
that come from? It appears that you're the one that has run out of
arguments if you have to resort to such unscrupulous chicanery to
bolster your position. Your weird twists of logic add nothing to
the discussion. I rest my case. If you haven't considered
politics, perhaps you should.

[Jason Anderson]

Cal King wrote:
>
> In article <34E3086B.72C4@no_spam.mcgill.ca>, Jason Anderson <jasona@no_spam.mcgill.ca> wrote:
>
> >Actually, the characters come from the chosen taxa.
>
> Of course, there would be no characters without organisms. Given a set of
> organisms, the choice of characters and the judgment of their phylogenetic
> value is also important.
>

If you run an analysis to test the "traditional" hypothesis of which
sarcop shares the closest common ancestor with tetrapods: lungfish,
coelecanth or osteolepiform (or even _Eustenopteron_ if you would
prefer), but then limit your included taxa to only living lungfish and
coelecanths, no possible arrangement of characters nor approaches here
advocated (your "creative" choice, Nyikos' lyrical artistry or
DiBennidito's shotgun total evidence) can possibly find any relationship
whatsoever with our friend _Eustenopteron_ (or _Osteolepis_ or
_Elpistostega_ or _Panderichthys_or...). Period. Impossible. Nada.
No no no. The null hypothesis has not been tested, and the tree can
only show relationships on a higher taxonomic level (which is ok if that
is the point of the analysis).

So, given a set of organisms, choice of characters *is* important-but
don't expect to find tetrapod sysnapomorphies if only one tetrapod is
included, because they would be autapomorphic and phylogenically
uninformative, and tossed out of the matrix.

> >The computer is *always* going to output a tree.
>
> Or more likely, a number of different trees.

Maybe, maybe not. Irrelevent nonetheless.

> Does the computer make a better
> phylogeneticist than the human?
>

A computer can evaluate all possible tree topologies while I make a cup
of tea, while by hand I would get it done too, but my advisor probably
wouldn't like my taking 20 years to finish my thesis ;)

> >> The difficulty lay in the correct identification of the synapomorphies.
> >

> >No. Given the taxa analysed, synapomorphies *were* correctly
> >identified-
>
> Ah, but what would happen if homoplasies are misidentified as synapomorphs.
> Would we then end up with something like a thylacine-placental wolf clade?
>

Homoplasies are not identified *except* in light of a tree. However, if
thylacines and wolves are the only mammals included in the analysis (say
a whole bunch of "reptiles" were also included-doesn't matter which) I
would hope that clade would be well supported-because they would share
the most recent common ancestor.

> > Patterson is renowned for his dislike of fossils in cladistics.
>
> He is not alone. Fossils often lack a large number of characters found
> in extant organims. Since cladists like to use a large number of characters,
> they can either dump the characters or dump the fossils.
>

Don't underestimate the informativeness of fossil data. And here I
assume you are referring to "cladists" as "cladists who work on strictly
extant animals-perhaps for molecular methods"? Look at Brochu's matrix
in the Syst. Bio. paper-he likes a large matrix but kept the fossils.
Your generality is only applicable to a specific subset of cladist, and
not even a majority, IMO.

> >> Nor has strictly cladistic
> >> analysis been able to establish the phylogeny of the lice, Mallophaga and
> >> Anoplura (Lyal 1985 versus Kim et al. 1987. These cases show that a cladistic
> >> analysis does not guarantee a correct cladogram unless it is based on valid
> >> apomorphies."--Mayr and Ashlock (1991) [emphasis mine]
> >>

> >"Correct cladogram"? Correct in the eyes of Mayr and Ashlock, "Cal" or
> >who?
>
> How about the collective judgments of biologists world wide? When there is no
> or little controversy, we assume that there is a consensus and that the tree
> is probably correct. That doesn't mean that such a tree is The Truth, of
> course.
>

The "collective judgments of biologists worldwide" has left you behind
already. You are fighting a battle already lost-like a soldier isolated
from his company on an island in the Pacific. Not that criticism or
refinement is not a good thing, mind you; just flip throug a copy of
Cladistics or Syst. Bio. sometime. (or Paleobiology or...)

> >> Hence it is not the method, but the characters, that really counts.
> >>

> >Hence it is the included taxa (and their distribution of characters)
> >that really count.
> >
> >Jason
>
> It is actually the correct identification of which characters are most
> informative of phylogenetic relationships that really counts. A computer
> cannot make these judgments and may never be able to do so. Homoplasies in,
> erroneous hypotheses out.

Again, how does one decide a priori what is a homoplasy-unless one
already knows what one wants the tree to look like?

[Cal King]

Simple logic, if you think the thylacine example is not well thought out, and
deleted the bird examples without commentor trace, then of course the bird
examples must be well thought out.

[Matt Fain]

Cal King wrote in message <6c1uup$l...@bgtnsc02.worldnet.att.net>...

Simply flawed logic, I'd say. If you don't understand how that is
so, find yourself a good undergraduate logic textbook. What I think
about your thylacine example has no necessary bearing on what I
think about your bird example. Or are you claiming that not only do
you know the mind of Darwin, but you know my mind too? I note that
you very predictably deleted the rest of my statement that
demonstrated your slippery ways quite nicely. I repeat, what a
piece of work you are! As you have done before, here is the missing
text:
(may I never have to stoop to your level again)

--------------------------------------------------------------------
-begin Cal's snip

Usually
you're going in so many directions at once, that WHEN and IF I
decide to add my 2c, I focus on ONE thing out of the MANY that are
erroneous. Another example: Just the other day you decided I
couldn't distinguish phenetics and evolutionary systematics on the
basis of my following statement, "Darwin recognized the importance
of phylogeny, and I'll leave it at that." Again, where the hell did
that come from? It appears that you're the one that has run out of
arguments if you have to resort to such unscrupulous chicanery to
bolster your position. Your weird twists of logic add nothing to
the discussion. I rest my case. If you haven't considered
politics, perhaps you should.

--------------------------------------------------------------------
-end Cal's snip

[Cal King]

In article <6c22pe$d...@saluki-news.it.siu.edu>, "Matt Fain" <matt...@siu.edu> wrote:

>(may I never have to stoop to your level again)

You are making ad hominem attacks your forté. That is too bad. They make it
difficult to respond because there is no point in responding to pure personal
attacks.

Here is what Peter Nyikos said to you in a post dated 1/30/98

==================
[P. Nyikos]:
Funny, someone over in s.b.e. said similar nasty things to
you, about your negative attitude and why he won't respond
to you until you get more civil.

I almost posted something saying, "Matthew is a diamond in
the rough, so don't let his rough edges turn you away" and then
I realized I still don't know enough about you to make
such an assessment.

Anyway, I'll gladly dialogue with you in a very civil way
over these issues. They are too fascinating to risk getting
lost in a cloud of personal slights."
==================

I agree with Prof. Nyikos. I think your personal attacks really don't add
anything to the discussion.

[bull.net]

In article <1d43nno.154...@pool-207-205-219-102.pitb.grid.net>,
her...@gmu.edu (Harry Erwin) wrote:

>If you're using continuous characters, you can't give them equal weight,
>because some characters naturally vary much more than others. Nasty
>problem. I've seen several solutions:
>
>1. Be arbitrary (_really_ doesn't work)
>2. Look for clustering of values and assign each cluster an ordinal
>3. Take the total range and slice it into equal segments
>4. Scale the weights to spread out the species maximally
>5. etc.

6. "[use] arbitrary (or prejudicial) wording ... to force complexly varying
characters into 'A or B' dichotomies." (Pritchard 1994, Herepetol. Rev.) while
simultaneously masquerading such practice as objectivity.

[Cal King]

In article <34E465F9.5BFD@no_spam.mcgill.ca>, Jason Anderson

<jasona@no_spam.mcgill.ca> wrote:
>Cal King wrote:
>>
>> In article <34E3086B.72C4@no_spam.mcgill.ca>, Jason Anderson
> <jasona@no_spam.mcgill.ca> wrote:
>>
>> >Actually, the characters come from the chosen taxa.
>>
>> Of course, there would be no characters without organisms. Given a set of
>> organisms, the choice of characters and the judgment of their phylogenetic
>> value is also important.

>If you run an analysis to test the "traditional" hypothesis of which
>sarcop shares the closest common ancestor with tetrapods: lungfish,
>coelecanth or osteolepiform (or even _Eustenopteron_ if you would
>prefer), but then limit your included taxa to only living lungfish and
>coelecanths, no possible arrangement of characters nor approaches here
>advocated (your "creative" choice, Nyikos' lyrical artistry or
>DiBennidito's shotgun total evidence) can possibly find any relationship
>whatsoever with our friend _Eustenopteron_ (or _Osteolepis_ or
>_Elpistostega_ or _Panderichthys_or...). Period. Impossible. Nada.

Those cladists who exclude fossils for the pragmatic reason that they lack
most of the characters used in the analysis should definitely pay attention.
I do not advocate the exclusion of fossils in analyis BTW.

>No no no. The null hypothesis has not been tested, and the tree can
>only show relationships on a higher taxonomic level (which is ok if that
>is the point of the analysis).
>
>So, given a set of organisms, choice of characters *is* important-but
>don't expect to find tetrapod sysnapomorphies if only one tetrapod is
>included, because they would be autapomorphic and phylogenically
>uninformative, and tossed out of the matrix.

Of course I do not advocate the use of as few taxa as possible either.

>> >The computer is *always* going to output a tree.
>>
>> Or more likely, a number of different trees.
>
>Maybe, maybe not. Irrelevent nonetheless.

If the computer outputs one tree, then it is easy. But if it outputs, say,
100 equally parsimonious trees, then there is a problem. BTW, this is not an
arbitrary number, but from real world cladistic analyses:

"Using 61 characters for these 49 families, subjectively polarized according
to their own views, they produced 'more than 100 equally parsimonious
solutions' with 225 steps in the cladogram. Some alternative assumptions
about polarity, following Burger’s views (1981) on early angiosperm evolution,
also led to more than 100 equally parsimonious cladograms, in this case with
226 steps." (Cronquist 1987. Bot. Rev.)

"...Frost (1992) was able to develop 36 equally parsimonious trees for his
data on the genus Tropidurus, and found no fewer than 26,588 'unrejected
trees' for the same data set. The traditional systematist must be forgiven
for wondering how systematics are advanced by such a plurality of
interpretations." (Pritchard 1994 Herpetol. Rev.)

>> Does the computer make a better
>> phylogeneticist than the human?
>>
>A computer can evaluate all possible tree topologies while I make a cup
>of tea, while by hand I would get it done too, but my advisor probably
>wouldn't like my taking 20 years to finish my thesis ;)

But is the computer better at picking out the most likely topology than the
trained taxonomist who has spent his/her career studying the group of
organisms he/she is specializing on? What should a systematist do if a
computer says a cladogram which says that birds are ancestral to thecodonts is
more parsimonious than another cladogram that shows the reverse? ;)

>> >> The difficulty lay in the correct identification of the synapomorphies.
>> >
>> >No. Given the taxa analysed, synapomorphies *were* correctly
>> >identified-
>>
>> Ah, but what would happen if homoplasies are misidentified as synapomorphs.
>> Would we then end up with something like a thylacine-placental wolf clade?
>>
>
>Homoplasies are not identified *except* in light of a tree.

That is a common claim among cladists, but it is demonstrably false. They eye
of the octopus is one example. It is not homologous to the eyes of other
animals and a tree cannot tell us anything about the origin of the octopus'
eye. The analog nature of the octopus' eye was revealed through comparative
anatomical studies, long before anyone has ever heard of cladistics. Another
example is the giant panda's opposable "thumb." It is not a real digit at
all, but is formed instead from the radial sesamoid. Hence it is not a
homolog of the vertebrate thumb at all. This fact was revealed again by
comparative anatomy, long before systematists started arguing whether the
panda is a bear or a raccoon. OTOH, if we use a tree such as Cracraft's
analysis of the owls and the hawks, we may conclude that their similarities
are homologs, when in fact these similarities have been regarded as
convergences for most of the 20th century.

Carroll and Dong (1991) noted, "The problem is to establish whether or not the
large number of derived characters uniting the Hupehsuchia with ichthyosaurs,
despite the great number of derived similarities, raises the possibilites that
skeletal features that are common to secondarily aquatic reptiles might have
evolved convergently in each of these groups." (Feduccia 1996:61) Carroll
and Dong could not establish homology even though they had a tree.

Using presumed synapomorphies to construct a tree and then using the tree to
identify synapomorphies amounts to a circular argument.

>However, if
>thylacines and wolves are the only mammals included in the analysis (say
>a whole bunch of "reptiles" were also included-doesn't matter which) I
>would hope that clade would be well supported-because they would share
>the most recent common ancestor.

Actually, one can use the similarities between thylacine and wolf to call them
a clade, to the exclusion of, say, cows. I will put myself out on a limb and
say that birds and maniraptoran theropods are the archosaurian analogy of the
thylacince and the wolf.

>> >> Hence it is not the method, but the characters, that really counts.
>> >>
>> >Hence it is the included taxa (and their distribution of characters)
>> >that really count.
>> >
>> >Jason
>>
>> It is actually the correct identification of which characters are most
>> informative of phylogenetic relationships that really counts. A computer
>> cannot make these judgments and may never be able to do so. Homoplasies >>
in, erroneous hypotheses out.
>
>Again, how does one decide a priori what is a homoplasy-unless one
>already knows what one wants the tree to look like?

As every well trained evolutionary systematist knows, it isn't easy to tell
convergences from synapomorphs; this is indeed the most difficult part of
systematic analysis. It is even harder to tell parallelism from synapomorphs,
a point that Mayr has emphasized. As Mayr and Ashlock pointed out, because of
the assumptions of cladistics, it is of paramount importance that cladistic
analysis uses ONLY synapomorphies. One of the ways to determine homology a
priori is comparative anatomy. Another way, which is seeing increasing use,
is developmental biology (e.g. Burke and Feduccia's embryological study of the
bird manus). A third way is to study the fossils. There are other ways to
tentatively determine homology but they are of course not foolproof.
Characters that are adaptive to particular ways of life, for example, are more
likely to be convergences, as Darwin and his contemporaries have pointed out
(Origin, ch. 13). Hence it is better to perform some character analysis to
determine character goodness and to discard characters that are suspect prior
to systematic analysis than to assume a priori that all characters chosen are
synapomorphies, then plunge into an analysis that assumes all characters are
of equal importance in the evolution of a group of animals. Better character
analysis makes for better systematic analysis. In other words, quality is
more important than quantity.

[Tom DiBenedetto]

Cal King wrote:

>>Homoplasies are not identified *except* in light of a tree.

>That is a common claim among cladists, but it is demonstrably false. They eye
>of the octopus is one example. It is not homologous to the eyes of other
>animals and a tree cannot tell us anything about the origin of the octopus'
>eye.

It is not demonstrably false. Homoplasy is a term which refers to a
character found to be incongruent in a cladistic analysis. I realize
that in the sloppy terminology so prevelant amongst those who dont
study methodology very deeply, it has come to be synonomous with
convergence. Thus we see people referring to the octopus eye as a
homoplasy, even though no one ever codes it as a homology in a matrix.
Why not just call it a convergence? You are correct that the
convergent nature of the octopus eye was discovered long before
cladistics. That is why no cladist would code it as homologous with
the vertebrate eye in a matrix. As I have tried to explain many times,
cladists who study morphology are trained to be expert comparative
anatomists. A matrix is a set of homology hypotheses which emerge from
detailed study of the organism. It is only if a percieved similarity
survives anatomical study (as the similarity between octopus and
vertebrate eyes do NOT) that they are coded as homologous in the
matrix and are then subject to falsification by the test of
congruence.
To use the eye example as an analogy though, were someone to code the
octopus eye as homologous with the vertebrate eye, and were they to
also construct a decent matrix with all evidence included, then the
resulting tree would most certainly speak to the issue of octopus eye
origin. It identify it as a homoplasy and indicate precisely where it
arose.

[Jason Anderson]

Cal King wrote:
>
> >> >The computer is *always* going to output a tree.
> >>
> >> Or more likely, a number of different trees.
> >
> >Maybe, maybe not. Irrelevent nonetheless.
>
> If the computer outputs one tree, then it is easy. But if it outputs, say,
> 100 equally parsimonious trees, then there is a problem. BTW, this is not an
> arbitrary number, but from real world cladistic analyses:
>

<Snip quotes of reviews by Prichard and Cronquist>

And some studies, not cited by you, are much better resolved. That
there are still areas needing work-lots actually-is irrelevant to the
present discussion. At least the uncertainty is out in the open, unlike
final statements of phylogney made in the past by figures of authority.

> >> Does the computer make a better
> >> phylogeneticist than the human?
> >>
> >A computer can evaluate all possible tree topologies while I make a cup
> >of tea, while by hand I would get it done too, but my advisor probably
> >wouldn't like my taking 20 years to finish my thesis ;)
>
> But is the computer better at picking out the most likely topology than the
> trained taxonomist who has spent his/her career studying the group of
> organisms he/she is specializing on? What should a systematist do if a
> computer says a cladogram which says that birds are ancestral to thecodonts is
> more parsimonious than another cladogram that shows the reverse? ;)
>

Re-examine the matrix and characters. If it all stands the scrutiny,
publish the study and let the "biological community as a whole" fight it
out. BTW, that would render "Class Aves" paraphyletic ;)

> >
> >Homoplasies are not identified *except* in light of a tree.
>
> That is a common claim among cladists, but it is demonstrably false. They eye
> of the octopus is one example. It is not homologous to the eyes of other
> animals and a tree cannot tell us anything about the origin of the octopus'
> eye. The analog nature of the octopus' eye was revealed through comparative
> anatomical studies, long before anyone has ever heard of cladistics.

See Tom DiBennidito's reply. Because of anatomical study, nobody except
your characature of a cladist would score the molluscan eye as a
synapomorphy.

> Another
> example is the giant panda's opposable "thumb." It is not a real digit at
> all, but is formed instead from the radial sesamoid. Hence it is not a
> homolog of the vertebrate thumb at all.

Correct. Nobody would score it as "thumb (+)" either, except your
strawman cladist.

> This fact was revealed again by
> comparative anatomy, long before systematists started arguing whether the
> panda is a bear or a raccoon.

FYI, cladists are among the few remaining comparative anatomists left in
biology. Also considered by cladists are issues of functional
morphology, development and molecular evolution (where applicable-not in
fossils, obviously).

> OTOH, if we use a tree such as Cracraft's
> analysis of the owls and the hawks, we may conclude that their similarities
> are homologs, when in fact these similarities have been regarded as
> convergences for most of the 20th century.
>

Right. People also used to think that babies existed in a fully
developed, if tiny, form within the sperm. And that the earth was the
centre of the universe. Whether Cracraft was correct or not aside,
duration that ideas are held is not a sound criterion for accuracy.

> Carroll and Dong (1991) noted, "The problem is to establish whether or not the
> large number of derived characters uniting the Hupehsuchia with ichthyosaurs,
> despite the great number of derived similarities, raises the possibilites that
> skeletal features that are common to secondarily aquatic reptiles might have
> evolved convergently in each of these groups." (Feduccia 1996:61) Carroll
> and Dong could not establish homology even though they had a tree.
>

Similar problem for limbless tetrapods, as I brough up earlier. One in
these cases can only present results (and reservations)-and if it is
corroborated by further studies (especially including new, more
plesiomorphic fossils) confidence increases. This problem is not
restricted to cladistics, BTW, but all science.

> Using presumed synapomorphies to construct a tree and then using the tree to
> identify synapomorphies amounts to a circular argument.
>

Again, one uses *synapomorphies* to construct the tree. The tree
interprets *homology* from *homoplaisy*.

> >However, if
> >thylacines and wolves are the only mammals included in the analysis (say
> >a whole bunch of "reptiles" were also included-doesn't matter which) I
> >would hope that clade would be well supported-because they would share
> >the most recent common ancestor.
>
> Actually, one can use the similarities between thylacine and wolf to call them
> a clade, to the exclusion of, say, cows.

And your matrix would be soundly picked apart.

> I will put myself out on a limb and
> say that birds and maniraptoran theropods are the archosaurian analogy of the
> thylacince and the wolf.
>

So you support a "gliding" instead of "ground up" origin for flight, eh?
:)

> >Again, how does one decide a priori what is a homoplasy-unless one
> >already knows what one wants the tree to look like?
>
> As every well trained evolutionary systematist knows,

(A common rhetorical technique designed to portray myself as "poorly
trainned")

> it isn't easy to tell
> convergences from synapomorphs; this is indeed the most difficult part of
> systematic analysis. It is even harder to tell parallelism from synapomorphs,
> a point that Mayr has emphasized. As Mayr and Ashlock pointed out, because of
> the assumptions of cladistics, it is of paramount importance that cladistic
> analysis uses ONLY synapomorphies.

A basic tennant of cladistics too.

> One of the ways to determine homology a
> priori is comparative anatomy.

See above, and earlier posts. Comparative anatomy is what most time is
spent on-before a matrix is built. One establishes *synapomorphies* in
this manner-using a rigorous series of tests. The final test, though,
is the tree-it tells *homology* from false homology-in light of all
other homologies. Things like the cephalopod eye would not be included
in the analysis because it is not a synapomorphy with the vertebrate
eye-it fails the test (a couple, truth be told) of similarity.

> Another way, which is seeing increasing use,
> is developmental biology (e.g. Burke and Feduccia's embryological study of the
> bird manus).

This study is flawed-their identification of digital number was based
upon a preconcieved notion-which was to be tested. Another
identification of digital number which is equally likely is fully
consistant with the evidence which comes from theropod fossils. This
issue is far from settled.

> A third way is to study the fossils. There are other ways to
> tentatively determine homology but they are of course not foolproof.

> Characters that are adaptive to particular ways of life, for example, are more
> likely to be convergences, as Darwin and his contemporaries have pointed out
> (Origin, ch. 13).

But animals already adapted to a particular niche are more likely to
remain in that nich-so these characters could also indicate common
ancestry. How to tell the two possibilities apart a priori is what is
difficult (and ill-advised, IMHO).

> Hence it is better to perform some character analysis to
> determine character goodness and to discard characters that are suspect prior
> to systematic analysis than to assume a priori that all characters chosen are
> synapomorphies, then plunge into an analysis that assumes all characters are
> of equal importance in the evolution of a group of animals. Better character
> analysis makes for better systematic analysis. In other words, quality is
> more important than quantity.

I agree :o. But what you seem to want to refuse to acknowledge is that
characters ARE analysed before included in a matrix. Your idea of how
cladistics works is not how cladistics is actually done, so your
criticism of cladistics in this thread is moot.

But we've already covered this ground, haven't we?

[Cal King]

In article <34E86102.6218@no_spam.mcgill.ca>, Jason Anderson

<jasona@no_spam.mcgill.ca> wrote:
>Cal King wrote:
>>
>> >> >The computer is *always* going to output a tree.
>> >>
>> >> Or more likely, a number of different trees.
>> >
>> >Maybe, maybe not. Irrelevent nonetheless.
>>
>> If the computer outputs one tree, then it is easy. But if it outputs, say,
>> 100 equally parsimonious trees, then there is a problem. BTW, this is not
an
>> arbitrary number, but from real world cladistic analyses:
>>
><Snip quotes of reviews by Prichard and Cronquist>
>
>And some studies, not cited by you, are much better resolved. That
>there are still areas needing work-lots actually-is irrelevant to the
>present discussion. At least the uncertainty is out in the open, unlike
>final statements of phylogney made in the past by figures of authority.

Cladists make too much of explicitness (or lack thereof) in some (definitely
not all) traditional systematic studies. Yes it is desirable to have the
workers spell out the reasons for their particular conclusions, but as long as
a hypothesis is falsifiable, then explicitness is not required. Desirable but
not required. Besides, there is no objective criterion that would allow one
to choose a particular cladogram for presentation if there are, say, one
hundred equally parsimonious cladograms. Cronquist's review shows that there
was no explicit reason given when only one cladogram was chosen, so cladists
can be non-explicit too. Besides, is a cladogram that requires, say, 223
steps really is inferior to one that requires 222? Is the longer one
necessarily false?

>> >Homoplasies are not identified *except* in light of a tree.

>> That is a common claim among cladists, but it is demonstrably false. They
> eye
>> of the octopus is one example. It is not homologous to the eyes of other
>> animals and a tree cannot tell us anything about the origin of the octopus'
>> eye. The analog nature of the octopus' eye was revealed through
comparative
>> anatomical studies, long before anyone has ever heard of cladistics.

>See Tom DiBennidito's reply. Because of anatomical study, nobody except
>your characature of a cladist would score the molluscan eye as a
>synapomorphy.

That was *my* point. Anatomical studies can reveal homoplasies, which is
contradictory to your claim that "Homoplasies are not identified *except* in
light of a tree."

>> Another

>> example is the giant panda's opposable "thumb." It is not a real digit at
>> all, but is formed instead from the radial sesamoid. Hence it is not a
>> homolog of the vertebrate thumb at all.
>
>Correct. Nobody would score it as "thumb (+)" either, except your
>strawman cladist.

Funny, only strawman cladists make mistakes according to you. Cracraft, who
grouped the owls and hawks as sister taxa, must have been a "strawman
cladist." I think Gardiner must have been a strawman cladist according to
you. For that matter, the cladists who consider theropods to be sister taxon
to the birds may well be strawman cladists, if Burke and Feduccia's (1997)
claim that birds have digits II-III-IV is correct.

>> This fact was revealed again by
>> comparative anatomy, long before systematists started arguing whether the
>> panda is a bear or a raccoon.
>
>FYI, cladists are among the few remaining comparative anatomists left in
>biology. Also considered by cladists are issues of functional
>morphology, development and molecular evolution (where applicable-not in
>fossils, obviously).

That is nonsense. Many comparative anatomists (e.g. Feduccia) are defintely
not cladists. Zhou (not a cladist) used comparative anatomy to call into
question Perle et al.'s identification of Mononykus as a bird. Zhou
specifically used functional morphological arguments to point out the
characters used by Perle et al. are likely adaptations for digging and
therefore likely not homologous to similar characters found in birds, which
are adaptations for flight.

>> Carroll and Dong (1991) noted, "The problem is to establish whether or not
> the
>> large number of derived characters uniting the Hupehsuchia with
ichthyosaurs,
>> despite the great number of derived similarities, raises the possibilites
> that
>> skeletal features that are common to secondarily aquatic reptiles might
have
>> evolved convergently in each of these groups." (Feduccia 1996:61) Carroll
>> and Dong could not establish homology even though they had a tree.
>>
>Similar problem for limbless tetrapods, as I brough up earlier. One in
>these cases can only present results (and reservations)-and if it is
>corroborated by further studies (especially including new, more
>plesiomorphic fossils) confidence increases. This problem is not
>restricted to cladistics, BTW, but all science.

The point is that even with a tree, one can still NOT say for sure whether
morphological similarities are in fact synapomorphies, thus again
contradicting your earlier claim that "Homoplasies are not identified *except*

in light of a tree."

>> Using presumed synapomorphies to construct a tree and then using the tree

to
>> identify synapomorphies amounts to a circular argument.
>>
>Again, one uses *synapomorphies* to construct the tree. The tree
>interprets *homology* from *homoplaisy*.

But since according to you, "Homoplasies are not identified *except* in light
of a tree," how can we find synapomorphies to construct the tree that could
then be used to identify homoplasies?

>> Actually, one can use the similarities between thylacine and wolf to call
>> them a clade, to the exclusion of, say, cows.
>
>And your matrix would be soundly picked apart.

How? The same way the evolutionary systematists picked apart the
bird-theropod link? But that would not be using cladistic methodology, which
BTW is fine by me. I am not saying that the thylacine forms a clade with the
wolf to the exclusion of ungulates. I am saying that such a cladogram is
demonstrably possible because cladists are constrained by their methodology
from evaluating characters a priori. If we then apply the recently popular
cladistic axiom that only a more parsimonious cladogram can replace an earlier
cladogram, then the thylacine-wolf cladogram may never be refuted. Unless of
course we use the evolutionary systematists' technique and give more weight to
characters which reveal the thylacine's marsupial ancestry.

>> >Again, how does one decide a priori what is a homoplasy-unless one
>> >already knows what one wants the tree to look like?
>>
>> As every well trained evolutionary systematist knows,
>
>(A common rhetorical technique designed to portray myself as "poorly
>trainned")

I had no idea you consider yourself an "evolutionary" systematist. Cladists
usually don't consider themselves evolutionary systematists. In fact,
cladists claim that their methodology is devoid of the assumption of, and
works regardless of, any model of evolution. May be they don't consider
hybridogenesis a model of evolution.

>> it isn't easy to tell
>> convergences from synapomorphs; this is indeed the most difficult part of
>> systematic analysis. It is even harder to tell parallelism from
> synapomorphs,
>> a point that Mayr has emphasized. As Mayr and Ashlock pointed out, because
> of
>> the assumptions of cladistics, it is of paramount importance that cladistic
>> analysis uses ONLY synapomorphies.
>
>A basic tennant of cladistics too.

Unfortunately, there is no way to implement it if cladists are to adhere to
their claim of "objectivity," which precludes them from discarding problematic
characters prior to an analysis.

>> One of the ways to determine homology a priori is comparative anatomy.

>See above, and earlier posts. Comparative anatomy is what most time is
>spent on-before a matrix is built. One establishes *synapomorphies* in
>this manner-using a rigorous series of tests.

So, according to you, homoplasies are now identifiable without a tree.

>The final test, though,
>is the tree-it tells *homology* from false homology-in light of all
>other homologies.

But Carroll and Dong showed that the tree was no test for homology in the case
of their study. The tree was no test for Cracraft's study and it was no test
for Gardiner's.

>Things like the cephalopod eye would not be included
>in the analysis because it is not a synapomorphy with the vertebrate
>eye-it fails the test (a couple, truth be told) of similarity.

It fails the test because the pre-cladism comparative anatomists got to it
first, before the cladists. It would have passed the test of superficial
similarity by the cladists had the cladists got to it first. For example, one
of the characters that passed the cladist test of homology is the
bird-theropod manus, but comparative anatomists like Burke and Feduccia are
refuting the cladists' claim of homology using comparative anatomical and
embryological evidence.

Since cladists choose scores of characters for their analysis, if they spend
time doing comparative anatomical analysis for every one of those characters,
they would have churned out a lot fewer papers than their traditional
couterparts. Besides, it has been a rather common practice for cladists to
"re-analyze" published data without ever examining the organims themselves.

>> Another way, which is seeing increasing use,
>> is developmental biology (e.g. Burke and Feduccia's embryological study of
>> the bird manus).
>
>This study is flawed-their identification of digital number was based
>upon a preconcieved notion-which was to be tested.

Flawed? Is there any published study in existence or being planned that would
contradict their claim?

>> Hence it is better to perform some character analysis to
>> determine character goodness and to discard characters that are suspect
prior
>> to systematic analysis than to assume a priori that all characters chosen
are
>> synapomorphies, then plunge into an analysis that assumes all characters
are
>> of equal importance in the evolution of a group of animals. Better
character
>> analysis makes for better systematic analysis. In other words, quality is
>> more important than quantity.
>
>I agree :o. But what you seem to want to refuse to acknowledge is that
>characters ARE analysed before included in a matrix.

I am not the only making this claim. For example, Feduccia (1996, Origin and
Evolution of the Birds, p. 68) writes, "Gauthier (1986) provided a detailed
character analysis (using eighty-four characters of varying importance) of the
entire archosaur assemblage, with emphasis on the saurischians and bird
origins. However, although Gauthier performed an analysis on the characters,
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
he never analyzed the characters themselves."
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

Apparently character analysis means different things to the cladists than to
the comparative anatomists. Cladists mean an analysis ON the characters, not
OF them; comparative anatomists mean an analysis OF the characters.

> Your idea of how
>cladistics works is not how cladistics is actually done,

As I demonstrated, this is not just my idea, but an idea shared by many. For
example, Gauthier (1986), according to Feduccia, certainly did not "establish
*synapomorphies* ... using a rigorous series of tests."

"Ernst Mayr (1981) used Cracraft's analysis of loon and grebe phylogeny to
point out the deficiencies of cladistic analysis in discerning homologous from
nonhomologous characters." (Feduccia 1996: 62)

"The case of the association of loons and grebes is of particular interest
because it follows a general historical pattern involving cladistic analysis
-- that is, a reversion to classifications of the past. ...after Stolpe's
famous paper of 1935, most authors accepted the view that the similarities
between loons and grebes ... were caused by convergence. Then cladistic
analysis, generally unable to deal with massive convergence, led to the groups
being placed back together as a monophyletic assemblage." (Ibid.)

That cladists have unwittingly repeated the mistakes of past systematists is a
point shared by Mayr and Ashlock (1991), who write, "Rebounding from the
extreme of phenetics, the cladists unwittingly returned to the ideal of
pre-Linnaean taxonomy: facile diagnosis." The fact that many cladistic
classifications have come to resemble erroneous classifications of the past
(particularly those of the 19th century or earlier) is thus no accident.

[Jason Anderson]

And around we go...

Cal King wrote:
> ><Snip quotes of reviews by Prichard and Cronquist>
> >
> >And some studies, not cited by you, are much better resolved. That
> >there are still areas needing work-lots actually-is irrelevant to the
> >present discussion. At least the uncertainty is out in the open, unlike
> >final statements of phylogney made in the past by figures of authority.
>
> Cladists make too much of explicitness (or lack thereof) in some (definitely
> not all) traditional systematic studies. Yes it is desirable to have the
> workers spell out the reasons for their particular conclusions, but as long as
> a hypothesis is falsifiable, then explicitness is not required.

Go back a few years in the literature. Jarvik, Moodie, Romer, Watson,
etc. etc. etc. all had their own classifications for, essentially, the
same groups. All of the classifications hinged on what characters the
authors thought most "important". Now all characters are out in the
open-we know what we are fighting over.

> Desirable but
> not required. Besides, there is no objective criterion that would allow one
> to choose a particular cladogram for presentation if there are, say, one
> hundred equally parsimonious cladograms.

Geez this is getting annoying. If one felt that presenting the results
from such a study was worthwhile one would present them ALL in a
consensus. If one says "I got 100 most parsimonious trees, but tree 77
is correct" they would be criticized.

> Cronquist's review shows that there
> was no explicit reason given when only one cladogram was chosen, so cladists
> can be non-explicit too. Besides, is a cladogram that requires, say, 223
> steps really is inferior to one that requires 222? Is the longer one
> necessarily false?

Under the operating assumptions, yes. That does not prevent one from
discussing the reasons one may have to think that a longer tree (more
likely though *many* more trees than at 222 steps) is preferrable. Same
in statistics- if you want to accept a result as significant at
alpha=15% one must justify oneself.

> >> >Homoplasies are not identified *except* in light of a tree.

<snip>

> >See Tom DiBennidito's reply. Because of anatomical study, nobody except
> >your characature of a cladist would score the molluscan eye as a
> >synapomorphy.
>
> That was *my* point. Anatomical studies can reveal homoplasies, which is
> contradictory to your claim that "Homoplasies are not identified *except* in
> light of a tree."
>

LISTEN TO ME. You have *not* identified a homoplaisy-you have failed to
identify a synapomorphy. You may not see a difference but there is a
big one. Homoplaisy is a false homology-it meets all tests of homology
(you are familiar with these, yes?) but the test of congruence, in other
words, is it not congruent with all other homologies AS IDENTIFIED BY
THE TREE. Your example does not meet the test of similarity-it does not
get far enough in the character analysis to be identified as a
synapomorphy in the first place.

> >> Another
> >> example is the giant panda's opposable "thumb." It is not a real digit at
> >> all, but is formed instead from the radial sesamoid. Hence it is not a
> >> homolog of the vertebrate thumb at all.
> >
> >Correct. Nobody would score it as "thumb (+)" either, except your
> >strawman cladist.
>
> Funny, only strawman cladists make mistakes according to you.

No, but your "example" is almost laughable.

> Cracraft, who
> grouped the owls and hawks as sister taxa, must have been a "strawman
> cladist."

Your words. Why is that grouping so obviously wrong to you, anyway?

> I think Gardiner must have been a strawman cladist according to
> you.

Strawman? Strawman is a rhetorical technique you love to employ.
Gardiner was just wrong.

> For that matter, the cladists who consider theropods to be sister taxon
> to the birds may well be strawman cladists, if Burke and Feduccia's (1997)
> claim that birds have digits II-III-IV is correct.
>

Keep you hat on-THIS issue is far from dead. Of course, you've already
made up your mind.

> >> This fact was revealed again by
> >> comparative anatomy, long before systematists started arguing whether the
> >> panda is a bear or a raccoon.
> >
> >FYI, cladists are among the few remaining comparative anatomists left in
> >biology. Also considered by cladists are issues of functional
> >morphology, development and molecular evolution (where applicable-not in
> >fossils, obviously).
>
> That is nonsense. Many comparative anatomists (e.g. Feduccia) are defintely

> not cladists.<snip>

Where did I say *only* cladists were comparative anatomists? Really,
you must read more carefully.

>
> >> Carroll and Dong (1991) noted, "The problem is to establish whether or not
> > the
> >> large number of derived characters uniting the Hupehsuchia with
> ichthyosaurs,
> >> despite the great number of derived similarities, raises the possibilites
> > that
> >> skeletal features that are common to secondarily aquatic reptiles might
> have
> >> evolved convergently in each of these groups." (Feduccia 1996:61) Carroll
> >> and Dong could not establish homology even though they had a tree.
> >>
> >Similar problem for limbless tetrapods, as I brough up earlier. One in
> >these cases can only present results (and reservations)-and if it is
> >corroborated by further studies (especially including new, more
> >plesiomorphic fossils) confidence increases. This problem is not
> >restricted to cladistics, BTW, but all science.
>
> The point is that even with a tree, one can still NOT say for sure whether
> morphological similarities are in fact synapomorphies, thus again
> contradicting your earlier claim that "Homoplasies are not identified *except*
> in light of a tree."
>

Key here is "for sure". Nothing is "for sure" in science. I always
keep my mind open for new evidence and will change my mind in the face
of it. Your use of synapomorphy is wrong-here you want "Homology".

> >> Using presumed synapomorphies to construct a tree and then using the tree
> to
> >> identify synapomorphies amounts to a circular argument.
> >>
> >Again, one uses *synapomorphies* to construct the tree. The tree
> >interprets *homology* from *homoplaisy*.
>
> But since according to you, "Homoplasies are not identified *except* in light
> of a tree," how can we find synapomorphies to construct the tree that could
> then be used to identify homoplasies?
>

(And around and around and around...)

Synapomorphy and homology are not the same thing.

<snip the wolf bit>

> >> >Again, how does one decide a priori what is a homoplasy-unless one
> >> >already knows what one wants the tree to look like?
> >>
> >> As every well trained evolutionary systematist knows,
> >
> >(A common rhetorical technique designed to portray myself as "poorly
> >trainned")
>
> I had no idea you consider yourself an "evolutionary" systematist. Cladists
> usually don't consider themselves evolutionary systematists. In fact,
> cladists claim that their methodology is devoid of the assumption of, and
> works regardless of, any model of evolution. May be they don't consider
> hybridogenesis a model of evolution.
>

No, *pattern* cladists-not all-do as you describe.

Look, I'm out of time. We are back in areas already argued over. Bye.

[Tom DiBenedetto]

Cal King wrote:

>>See Tom DiBennidito's reply. Because of anatomical study, nobody except
>>your characature of a cladist would score the molluscan eye as a
>>synapomorphy.

>That was *my* point. Anatomical studies can reveal homoplasies, which is
>contradictory to your claim that "Homoplasies are not identified *except* in
>light of a tree."

No "cal", you still dont get it. Jason did not say "convergences are
not identified except in light of a tree". Had he said that, your
rebuttal would be justified. He said homoplasy. There is a difference
between the two terms. Think about it.

>The point is that even with a tree, one can still NOT say for sure whether
>morphological similarities are in fact synapomorphies, thus again
>contradicting your earlier claim that "Homoplasies are not identified *except*
>in light of a tree."

Wrong again. Synapomorphy is the term for hypothesized homologies
which survive the test of congruence. Thus the tree indicates
synapomorphies. Whether they are therefore, true homologies is another
issue. You are right that we can still not be sure, but that relates
to the inherent limitations of science,,,a concept it seems you are
perfectly willing to accept when it relates to cladistics, but not to
methods which you prefer.

>But since according to you, "Homoplasies are not identified *except* in light
>of a tree," how can we find synapomorphies to construct the tree that could
>then be used to identify homoplasies?

We use similarities that are proposed as homologies, then tested as
best as is possible according to the standards of the particular field
in which the character is developed (e.g. comparative anatomy). When
the character has passed these tests, they are coded into a matrix as
putative homologies, or putative synapomorphies. Those that survive
the cladstic analysis are accepted as homologous, or synapomorphous.

> cladists are constrained by their methodology
>from evaluating characters a priori.

What on earth can this possibly mean? How can you say such a thing?
Have you never even looked at a cladistic analysis in your life? Does
truth mean NOTHING to you?

> comparative anatomists like Burke and Feduccia are
>refuting the cladists' claim of homology using comparative anatomical and
>embryological evidence.

This sounds to me like a dispute amongst comparative anatomists. Is
there some strict correlation between the position one takes on the
nature of the manus and ones systematic philosophy? If so, why? What
is it about the details of the manus which compels a cladist to see it
one way and a "darwinian systematist' to see it another? Not knowing
the details of this situation, I would presume that it is rather
coincidental that the protagonists have different systematic
philosophies.

>Since cladists choose scores of characters for their analysis, if they spend
>time doing comparative anatomical analysis for every one of those characters,
>they would have churned out a lot fewer papers than their traditional
>couterparts.

I can assure you "cal" that every cladist I know does detailed
anatomical work on every single character proposed. Granted, I dont
know all cladists, and I have the good fortune to be at an institution
in which cladistic methods are taken seriously and high standards
maintained. It is silly on your part to try to find any possible
example of poor work in order to disparage a method which you have
personal, irrational reason to oppose.

> Besides, it has been a rather common practice for cladists to
>"re-analyze" published data without ever examining the organims themselves.

This is not common practice. It seems to be common practice for you to
state outrageous lies about other scientists who happen to use mehtods
you cant quite understand.

>Apparently character analysis means different things to the cladists than to
>the comparative anatomists. Cladists mean an analysis ON the characters, not
>OF them; comparative anatomists mean an analysis OF the characters.

Once again "cal" you are making this up, and it is obvious.

[Cal King]

In article <34e9a15c....@news.itd.umich.edu>,
td...@umich.edu (Tom DiBenedetto) wrote:
>Path:
wn8!bgtf1!worldnet.att.net!news-out.internetmci.com!newsfeed.internetmci.com!1
31.103.1.114!news1.chicago.iagnet.net!qual.net!iagnet.net!newsxfer3.itd.umich.
edu!newsxfer.itd.umich.edu!news.itd.umich.edu!not-for-mail
>From: td...@umich.edu (Tom DiBenedetto)
>Newsgroups: sci.bio.systematics
>Subject: Re: Methods and characters (Re: Quetzalcoatlus northropi!!!)
>Organization: University of Michigan Museum of Zoology
>Reply-To: td...@umich.edu
>Message-ID: <34e9a15c....@news.itd.umich.edu>
>References: <6a55ku$e...@bgtnsc01.worldnet.att.net>
<34C6739A...@umail.umd.edu> <6a779m$k...@mtinsc03.worldnet.att.net>
<1d391bc <34CA5DB0...@saluki-mail.siu.edu>
<6al5s0$8...@redwood.cs.sc.edu> <6alo66$g...@saluki-news.it.siu.edu>
<6angij$n...@redwood.cs.sc.edu> <6avil6$4...@bgtnsc03.worldnet.att.net>
<6b7dgl$9...@redwood.cs.sc.edu> <6bh15e$g...@bgtnsc03.worldnet.att.net>
<34E86102.6218@no_spam.mcgill.ca> <6cb3fh$8...@bgtnsc03.worldnet.att.net>
>X-Newsreader: Forte Agent 1.5/32.452
>MIME-Version: 1.0
>Lines: 109
>Date: Tue, 17 Feb 1998 15:05:31 GMT
>NNTP-Posting-Host: 141.211.110.20
>NNTP-Posting-Date: Tue, 17 Feb 1998 10:05:31 EST
>Status: N

>
>Cal King wrote:
>
>>>See Tom DiBennidito's reply. Because of anatomical study, nobody except
>>>your characature of a cladist would score the molluscan eye as a
>>>synapomorphy.
>
>>That was *my* point. Anatomical studies can reveal homoplasies, which is
>>contradictory to your claim that "Homoplasies are not identified *except* in
>>light of a tree."
>

>No "cal", you still dont get it. Jason did not say "convergences are
>not identified except in light of a tree". Had he said that, your
>rebuttal would be justified. He said homoplasy. There is a difference
>between the two terms. Think about it.

That is not the point. Convergences are a category of what Hennig called a
homoplasy (=convergence + reversal + parallelism). As such Jason was in fact
saying that "convergences, reversals and parallelisms are not identified
except in light of a tree."

>>The point is that even with a tree, one can still NOT say for sure whether
>>morphological similarities are in fact synapomorphies, thus again
>>contradicting your earlier claim that "Homoplasies are not identified
*except*
>>in light of a tree."
>

>Wrong again. Synapomorphy is the term for hypothesized homologies
>which survive the test of congruence. Thus the tree indicates
>synapomorphies.

Carroll and Dong had a tree, so presumably their "hypothesized homologies
survived the test of congruence." But they STILL could not be sure that the
characters they used were in fact homologies. The tree is therefore NOT a
test for homology. Homology is a test for the tree.

>Whether they are therefore, true homologies is another
>issue.

Actually it is THE issue. Trees are demonstrably wrong if the characters used
to construct or "recover" them are in fact homoplasies, not synapomorphies.

>You are right that we can still not be sure, but that relates
>to the inherent limitations of science

It relates to the inherent weakness of cladistic methodology, not science.

"Using cladistic methodology, ... it is all but impossible to ascertain which
characters are homoplasies ... and which are homologous...." (Feduccia
1996:62)

"Sibley and Ahlquist further state that 'the grebes have no close living
relatives (551) and that 'the errors in Cracraft's reconstruction of the
phylogeny of diving birds are due to the difficulties of interpreting
morphological characters, not to the principles he used as the basis for his
analysis." (Ibid.)

The weakness of cladistics lies in the use of a large number of characters and
organisms which, according to the paleontologist Greg Paul, makes it all the
more difficult to evaluate character goodness. Further difficulties arise
from the cladistic worship of "objectivity," which precludes the cladist from
giving more weight to some characters and discarding others that are JUDGED to
be problematic. Judgment is prohibited in cladistic analysis because it would
introduce subjectivity into the analysis.

>>But since according to you, "Homoplasies are not identified *except* in
light
>>of a tree," how can we find synapomorphies to construct the tree that could
>>then be used to identify homoplasies?
>

>We use similarities that are proposed as homologies, then tested as
>best as is possible according to the standards of the particular field
>in which the character is developed (e.g. comparative anatomy).

Sure, using available information is fine. But what about actually doing the
investigation yourself? Aren't cladists concerned with the goodness of their
characters enough that they will do some character analysis of their own prior
to phylogenetic analysis using the characters? According to Feduccia,
Gauthier (1986) did not do any character analysis prior to his cladistic
analysis.

>When the character has passed these tests,

Alas, not all characters used by the cladists have seen the benefit of a
priori testing by comparative anatomists. It is the use of these untested
characters that has gotten cladists in trouble. Some characters that the
cladists used have seen a posteriori homology testing (e.g. Burke and
Feduccia's data on the bird manus and Zhou's analysis of Mononykus' bird-like
characters) and the results are not well liked by the cladists.

>they are coded into a matrix as
>putative homologies, or putative synapomorphies. Those that survive
>the cladstic analysis are accepted as homologous, or synapomorphous.

Unfortunately, this makes for a circular argument. Cracraft's analysis showed
that this "test" is fallible and that convergences can go undetected.

>> cladists are constrained by their methodology
>>from evaluating characters a priori.
>

>What on earth can this possibly mean? How can you say such a thing?
>Have you never even looked at a cladistic analysis in your life? Does
>truth mean NOTHING to you?

The truth is that cladists admit to this shortcoming of their methodology.
Kurt Schwenk (a cladist) wrote,

"In the face of no opposing evidence, we assume that characters are
independent and hope that, in any case, enough 'good' characters will outweigh
such hidden redundancy. Since we cannot judge the quality of a character
(i.e., its 'phylogenetic signal') except in light of a phylogeny, we can never
test our assumptions directly, only indirectly by their congruence with other
character sets." (Schwenk, Kurt 1994 Systematics And Subjectivity: The
Phylogeny And Classification Of Iguanian Lizards Revisited. Herpetol. Rev.
25[2]:53-57)

Pritchard agreed, "It is in fact the traditional taxonomist of today, rather
than the cladist, who uses judgment to note which characters are likely to be
fundamental ones and which are recent, plastic, or adaptive ones, and who
emphasizes the former and rejects the latter. A cladist, on the other hand,
constrained by a ban on using 'judgment' as to the quality of characters, and
excluding most or all fossils on the usual ground that they don’t reveal all
the characters of interest, might easily end up with a 'Pachydermata' clade,
or some other grouping of species with certain superficial similarities."
(Pritchard, P.C.H. 1994 Cladism: The Great Delusion. Herpetol. Rev.
25[3]:103-110)

Cracraft's analysis of the loons and grebes is a real life example of a
"Pachydermata" clade because cladists are forbidden to use judgment (which is
"subjective" and is thus anathematic to cladism) to reject problematic
characters.

>> comparative anatomists like Burke and Feduccia are
>>refuting the cladists' claim of homology using comparative anatomical and
>>embryological evidence.
>

>This sounds to me like a dispute amongst comparative anatomists. Is
>there some strict correlation between the position one takes on the
>nature of the manus and ones systematic philosophy? If so, why? What
>is it about the details of the manus which compels a cladist to see it
>one way and a "darwinian systematist' to see it another? Not knowing
>the details of this situation, I would presume that it is rather
>coincidental that the protagonists have different systematic
>philosophies.

It is a consensus that theropods have digits I-II-III and lost digits IV and V
in their manus. Birds also have three digits in their manus. Gauthier
assumed that birds also have digits I-II-III and used this particular
"synapomorph" along with other supposed synapomorphies to perform a
cladistic analysis. He came to the conclusion that birds and theropods form a
clade. Burke and Feduccia (1997, Science), using embryological evidence, show
that the bird manuss most likely has digits II-III-IV, and thus the bird
manual digits are not the same as the theropod digits, calling into question
the cladists' conclusions that birds are theropods. This example once again
illustrates how easily cladists can be fooled by convergences because they
are forbidden by their methodology from paying attention to character goodness
prior to the analysis and they only rely on the results of their analysis as a
test for character goodness.

>I can assure you "cal" that every cladist I know does detailed
>anatomical work on every single character proposed.

Do you know Gauthier and Cracraft?

>It is silly on your part to try to find any possible
>example of poor work in order to disparage a method which you have
>personal, irrational reason to oppose.

My "irrational" reason for opposing cladism is that cladistics methodology
will result in erroneous phylogenetic conclusions by methodological design.

>> Besides, it has been a rather common practice for cladists to
>>"re-analyze" published data without ever examining the organims themselves.

>This is not common practice. It seems to be common practice for you to

>state outrageous lies about other scientists who happen to use mehtods
>you cant quite understand.

It is common practice to reanalyze data. It is also common practice for you
to label as "lies" those facts you don't like to hear.

>>Apparently character analysis means different things to the cladists than to
>>the comparative anatomists. Cladists mean an analysis ON the characters,
not
>>OF them; comparative anatomists mean an analysis OF the characters.
>

>Once again "cal" you are making this up, and it is obvious.
>
>
>Tom DiBenedetto td...@umich.edu

I gave all those references. You, or anyone else can check them to see if I
made anything up. OTOH, can you give me any evidence to support your claim
that cladists routinely perform comparative anatomical studies on all or even
most of the characters they use for phylogenetic analysis?

[Peter Nyikos]

[posted and mailed, with request for a CC of any reply]

Jason Anderson <jasona@no_spam.mcgill.ca> writes:

>Cal King wrote:

[Tom DiB?]

>> >Homoplasies are not identified *except* in light of a tree.

>> That is a common claim among cladists, but it is demonstrably false. They eye
>> of the octopus is one example. It is not homologous to the eyes of other
>> animals and a tree cannot tell us anything about the origin of the octopus'
>> eye. The analog nature of the octopus' eye was revealed through comparative
>> anatomical studies, long before anyone has ever heard of cladistics.

>See Tom DiBennidito's reply. Because of anatomical study, nobody except
>your characature of a cladist would score the molluscan eye as a
>synapomorphy.

See my reply to Tom, and note the allusion to an earlier argument
regarding the pollex. Speaking of which...

>> Another
>> example is the giant panda's opposable "thumb." It is not a real digit at
>> all, but is formed instead from the radial sesamoid. Hence it is not a
>> homolog of the vertebrate thumb at all.

>Correct. Nobody would score it as "thumb (+)" either, except your
>strawman cladist.

That's only because a careful anatomist could tell it is NOT a pollex.
In the example I gave Tom about a year ago, I very carefully
used the word "pollex" to remove all ambiguity about something
an expert anatomist might identify as such.

I said that I would take a cladogram that has the pollex
disappearing and re-appearing at successive nodes several times
as wrong even if it was the most parsimonious tree; and that
I would regard that feature as falsifying not only the tree
itself but also the very principle of maximum parsimony.

Tom, of course, wouldn't buy that.

>> This fact was revealed again by
>> comparative anatomy, long before systematists started arguing whether the
>> panda is a bear or a raccoon.

>FYI, cladists are among the few remaining comparative anatomists left in
>biology.

How do you know?

>> OTOH, if we use a tree such as Cracraft's
>> analysis of the owls and the hawks, we may conclude that their similarities
>> are homologs, when in fact these similarities have been regarded as
>> convergences for most of the 20th century.
>>
>Right. People also used to think that babies existed in a fully
>developed, if tiny, form within the sperm. And that the earth was the
>centre of the universe. Whether Cracraft was correct or not aside,
>duration that ideas are held is not a sound criterion for accuracy.

Do I detect a note of skepticism about hawks and owls forming a
polyphyletic taxon?

[...]

>> Using presumed synapomorphies to construct a tree and then using the tree to
>> identify synapomorphies amounts to a circular argument.
>>
>Again, one uses *synapomorphies* to construct the tree. The tree
>interprets *homology* from *homoplaisy*.

Synapomorphy *is* homology, isn't it?

[...]

>> >Again, how does one decide a priori what is a homoplasy-unless one
>> >already knows what one wants the tree to look like?
>>
>> As every well trained evolutionary systematist knows,

>(A common rhetorical technique designed to portray myself as "poorly
>trainned")

Don't be so trigger-happy. It looks to me like Cal was defending
evolutionary systematists, not attacking cladists or you as "poorly
trained".

[...]

>> One of the ways to determine homology a
>> priori is comparative anatomy.

>See above, and earlier posts. Comparative anatomy is what most time is
>spent on-before a matrix is built. One establishes *synapomorphies* in
>this manner-using a rigorous series of tests.

In other words, one looks through the literature to see what
other cladists have done, and takes their *conclusions* about
homology as starting points for what should be synapomorphic, no?

The final test, though,
>is the tree-it tells *homology* from false homology-in light of all
>other homologies. Things like the cephalopod eye would not be included
>in the analysis because it is not a synapomorphy with the vertebrate
>eye-it fails the test (a couple, truth be told) of similarity.

Similarity is a relative concept. You aren't seriously claiming
that cladists will score any two things that show the slightest
dissimilarity as different *a priori* do you? That wouldn't even
allow for individual variation within a litter of puppies.

>> Another way, which is seeing increasing use,
>> is developmental biology (e.g. Burke and Feduccia's embryological study of the
>> bird manus).

>This study is flawed-their identification of digital number was based
>upon a preconcieved notion-which was to be tested.

After seeing Chris Brochu argue about this example, I don't think
I'll agree with you without you going to a lot of detail about it.

Another
>identification of digital number which is equally likely

In what? Birds or theropods?

is fully
>consistant with the evidence which comes from theropod fossils. This
>issue is far from settled.

I'll go along with THAT much.

[...]

>> Hence it is better to perform some character analysis to
>> determine character goodness and to discard characters that are suspect prior
>> to systematic analysis than to assume a priori that all characters chosen are
>> synapomorphies, then plunge into an analysis that assumes all characters are
>> of equal importance in the evolution of a group of animals. Better character
>> analysis makes for better systematic analysis. In other words, quality is
>> more important than quantity.

>I agree :o. But what you seem to want to refuse to acknowledge is that
>characters ARE analysed before included in a matrix.

+++++++++++++++++++++++ dogmatic cladistic posting mode on

Either the analysis took the form of earlier cladistic use
of trees or else it was subjective.

+++++++++++++++++++++++ dogmatic cladistic posting mode off

Your idea of how
>cladistics works is not how cladistics is actually done, so your
>criticism of cladistics in this thread is moot.

>But we've already covered this ground, haven't we?

Not with me, you haven't.

[Peter Nyikos]

td...@umich.edu (Tom DiBenedetto) writes:

>Cal King wrote:

[Tom?]

>>>Homoplasies are not identified *except* in light of a tree.

>>That is a common claim among cladists, but it is demonstrably false. They eye
>>of the octopus is one example. It is not homologous to the eyes of other
>>animals and a tree cannot tell us anything about the origin of the octopus'
>>eye.

>It is not demonstrably false. Homoplasy is a term which refers to a

>character found to be incongruent in a cladistic analysis.

Is that the ONLY use of the term in all of systematics? And what
does "found to be" mean in this context? That a single analysis
using a small select set of organisms (hardly a group!) came
up with a most parsimonious tree in which the character shows
up as arising twice independently?

I realize
>that in the sloppy terminology so prevelant amongst those who dont
>study methodology very deeply, it has come to be synonomous with
>convergence.

I would like to hear from other systematists, notably Felsenstein,
on this issue.

Thus we see people referring to the octopus eye as a
>homoplasy, even though no one ever codes it as a homology in a matrix.

That's because they know better--or do they? Didn't the
illustrious Tom DiBenedetto once say that cladists bring
NO preconceived notions to character analysis?

>Why not just call it a convergence? You are correct that the

>convergent nature of the octopus eye was discovered long before
>cladistics.

How? Not by cladistic methods, eh?

That is why no cladist would code it as homologous with
>the vertebrate eye in a matrix. As I have tried to explain many times,
>cladists who study morphology are trained to be expert comparative
>anatomists.

I don't recall you ever trying to explain that. I do recall
you, on the other hand, saying that glutamine at the E7 site
of alpha hemoglobin is given the same weight in a cladistic
analysis as the question of whether the male or the female
is the heterogamete. Each is just one (1) character, no more,
no less.

Further back, I brought up the hypothetical
example of a vertebrate pollex appearing and disappearing
several times in a single branch of a cladogram,
and said that I would take that as falsifying the
cladogram, while your reaction was a smug, "Who knows?
You just might learn something."

A matrix is a set of homology hypotheses which emerge from
>detailed study of the organism. It is only if a percieved similarity
>survives anatomical study (as the similarity between octopus and
>vertebrate eyes do NOT)

In other words, you rely on the subjective judgment of anatomists.
Funny, I thought you and Mike Noren were the ones who claimed
cladistics was fully objective.

>that they are coded as homologous in the
>matrix and are then subject to falsification by the test of
>congruence.

As is so often the case, "falsification" means "it didn't show
up as a single character on the most parsimonious tree," eh? Can
ANY amount of anatomical insight falsify the most parsimonious
tree in your eyes?

>To use the eye example as an analogy though, were someone to code the
>octopus eye as homologous with the vertebrate eye, and were they to
>also construct a decent matrix with all evidence included, then the
>resulting tree would most certainly speak to the issue of octopus eye
>origin. It identify it as a homoplasy and indicate precisely where it
>arose.

Your deep faith in the most parsimonious tree is touching, especially
in the light of my hypothetical pollex example and your reaction to it.

[Tom DiBenedetto]

Cal King wrote:

>>No "cal", you still dont get it. Jason did not say "convergences are
>>not identified except in light of a tree". Had he said that, your
>>rebuttal would be justified. He said homoplasy. There is a difference
>>between the two terms. Think about it.
>
>That is not the point. Convergences are a category of what Hennig called a
>homoplasy (=convergence + reversal + parallelism). As such Jason was in fact
>saying that "convergences, reversals and parallelisms are not identified
>except in light of a tree."

Listen "cal", this is not really all that difficult. The underlying
currency here is the concept of similarity. Bird wings and bat wings
can be considered similar to a casual observer. To the extent that one
considers them similar, and then considers the phylogeny, they can
then be considered convergent. If, however, one were to examine them
closely, one would conclude that they are not all that similar after
all. Then they would not be considered convergent, because convergence
is a qualifier to the concept of similarity.
Detailed anatomical examinations can cause one to change ones original
judgement that two structures are similar. Morphological cladists
spend the great majority of their research time dealing with such
issues. In this sense we can say that convergences are identified
before cladistic analysis. But in fact, since very often we do not
have any idea of what the phylogeny is before cladistic analysis, it
is technically proper to say that we are not really distinguish
convergences, but rather eliminating cases of superficial similarity.
A "convergence" implies that the phylogeny is known. The phylogeny is
not known until after the analysis. In that sense, convergences are
identified on the phylogeny.

>>Wrong again. Synapomorphy is the term for hypothesized homologies
>>which survive the test of congruence. Thus the tree indicates
>>synapomorphies.
>
>Carroll and Dong had a tree, so presumably their "hypothesized homologies
>survived the test of congruence." But they STILL could not be sure that the
>characters they used were in fact homologies. The tree is therefore NOT a
>test for homology. Homology is a test for the tree.

Wrong again. Characters which are synapomorphous on a tree are
interpreted as homologies. The fact that one cannot be sure is a
functin of the fact that one can never be sure of a historical
inference. Within the context of the methodology though, they are
accepted as homologies.
To claim homology as a test of the tree is nonsense. It would imply
some independant knowledge of what is and what is not homology.
Homology is similarity due to descent; one needs a phylogeny in order
to identify homology.

>>Whether they are therefore, true homologies is another
>>issue.
>
>Actually it is THE issue. Trees are demonstrably wrong if the characters used
>to construct or "recover" them are in fact homoplasies, not synapomorphies.

obviously, But there are no other methods for determining homology
other than carreful study of the character, and testing under the
expectation of congruence (i.e. cladistic methodology). Unless you
have a private line to god.....perhaps that is why you speak so
authoritativly about things you dont seem to understand.

>>You are right that we can still not be sure, but that relates
>>to the inherent limitations of science

>It relates to the inherent weakness of cladistic methodology, not science.

I see, You are immune to the limiations of the rest of us mere
mortals.

>"Using cladistic methodology, ... it is all but impossible to ascertain which
>characters are homoplasies ... and which are homologous...." (Feduccia
>1996:62)

Feduccia is wrong here. Can you deal with the fact that someone can be
wrong even though he says somethings that gets you off?

>The weakness of cladistics lies in the use of a large number of characters and
>organisms which,

Great! Run with this argumetn "cal"! Lets use :LESS evidence,,,that
should give us better results.

> according to the paleontologist Greg Paul, makes it all the
>more difficult to evaluate character goodness.

yeah,,,it gets so confusing when you have to deal with so many
things......

> Judgment is prohibited in cladistic analysis because it would
>introduce subjectivity into the analysis.

What systematics really needs is to develop results which bear the
stamp of judgements such as yours.....sorry, but I decline. Your
"judgement" cant even allow you to get the methodology straight.

> According to Feduccia, Gauthier (1986) did not do any character analysis
>prior to his cladistic analysis.

Who cares what Feduccia said? Did you read Gauthier's work? Are you
making this claim?

>>they are coded into a matrix as
>>putative homologies, or putative synapomorphies. Those that survive
>>the cladstic analysis are accepted as homologous, or synapomorphous.

>Unfortunately, this makes for a circular argument. Cracraft's analysis showed
>that this "test" is fallible and that convergences can go undetected.

There is no circularity involved whatsoever. And there is neither the
assertion that the results are guaranteed to be correct (nor are such
guarantees available to any scientist,,,only people like you, I guess)

Now as to the issue of cladists analyzing characters:
You begin with this clear falsehood.

>>> cladists are constrained by their methodology
>>>from evaluating characters a priori.

>>What on earth can this possibly mean? How can you say such a thing?
>>Have you never even looked at a cladistic analysis in your life? Does
>>truth mean NOTHING to you?

You then claim that someone you call a cladist admits to this:

>The truth is that cladists admit to this shortcoming of their methodology.
>Kurt Schwenk (a cladist) wrote,

>"In the face of no opposing evidence, we assume that characters are
>independent and hope that, in any case, enough 'good' characters will outweigh
>such hidden redundancy. Since we cannot judge the quality of a character
>(i.e., its 'phylogenetic signal') except in light of a phylogeny, we can never
>test our assumptions directly, only indirectly by their congruence with other
>character sets." (Schwenk, Kurt 1994 Systematics And Subjectivity: The
>Phylogeny And Classification Of Iguanian Lizards Revisited. Herpetol. Rev.
>25[2]:53-57)

And of course, the quote says nothing about the lie you proposed.
Do you think that the rest of us out here cant read?

>Pritchard agreed, "It is in fact the traditional taxonomist of today, rather
>than the cladist, who uses judgment to note which characters are likely to be
>fundamental ones and which are recent, plastic, or adaptive ones, and who
>emphasizes the former and rejects the latter.

And you dont see this as a massive source of unaccountable error?????

>A cladist, on the other hand,
>constrained by a ban on using 'judgment' as to the quality of characters,

For judgement, substitute a priori presumed knowledge which has turned
out often to be incorrect.

>and uding most or all fossils on the usual ground that they don’t reveal all
>the characters of interest,

False. Most cladists I know use as much fossil evidence as they can.

>might easily end up with a 'Pachydermata' clade,
>or some other grouping of species with certain superficial similarities."

And how many volumes would it take to list all of the unnatural groups
proposed by people following Pritchards methods????

>It is a consensus that theropods have digits I-II-III and lost digits IV and V
>in their manus. Birds also have three digits in their manus. Gauthier
>assumed that birds also have digits I-II-III and used this particular
>"synapomorph" along with other supposed synapomorphies to perform a
>cladistic analysis.

You ahve failed to address my question. How is this anything but a
dispute amongst comparative anatomists? What is it about bird digits
that cause a cladists to identify them one way and a darwinian
another? Until you provide a rational answer, I will be forced to
assume that their anatomical dispute has nothing to do with their
systematic disputes. It seems that the two groups of workers could
agree on the anatomy, and probably end up agreeing on the phylogeny,
or take opposite views of the anatomy and end up with eachothers
phylogeny. Why do you confound the two issues?

> Burke and Feduccia (1997, Science), using embryological evidence, show
>that the bird manuss most likely has digits II-III-IV, and thus the bird
>manual digits are not the same as the theropod digits, calling into question
>the cladists' conclusions that birds are theropods.

And if they are right, then Gauthier is wrong about the anatomy. What
does this have to do with cladistics? If he is wrong, they he woudl
recode the character and probably get a different result.

>This example once again
>illustrates how easily cladists can be fooled by convergences because they
>are forbidden by their methodology from paying attention to character goodness

gosh you go to such simpleminded extremes in order to disparage
people. This is a nonsense argument. Obviously both workers have
examined the bird manus. They interpret what they see differently.
This is an anatomical dispute. It has nothing to do with how they will
conduct a phylogentic analysis later,,once the matrix is built.

>>I can assure you "cal" that every cladist I know does detailed
>>anatomical work on every single character proposed.

>Do you know Gauthier and Cracraft?

Met them both. I dont follow the details of Gauthiers work (different
group) but I did read his dissertation, and it entailed a massive
amount of anatomical work. I would go so far as to say that he has
certainly looked at these characters for hundreds of more hours than
you have.

>, can you give me any evidence to support your claim
>that cladists routinely perform comparative anatomical studies on all or even
>most of the characters they use for phylogenetic analysis?

yeah "cal", why dont you read some systematic works instead of
scouring the litereature to find some nasty quote which some turkey
made about cladistics, and then posting it over and over again. You
think that makes you a serious thinker?
Or why dont you go to a museum or a university and meet some
systematists? Come to Michigan and I'll show you my office, with the
hundred boxes of skeletons,. Go visit real people who do real work.
Listen to them, Read their papers. It might actually help to look
people in the eye. It usually is harder to make slanderous charges
agaionst people you actually know. Maybe you will get a sense for what
the real world is like.

[Tom DiBenedetto]

Peter Nyikos wrote:

> A matrix is a set of homology hypotheses which emerge from
>>detailed study of the organism. It is only if a percieved similarity
>>survives anatomical study (as the similarity between octopus and
>>vertebrate eyes do NOT)
>
>In other words, you rely on the subjective judgment of anatomists.
>Funny, I thought you and Mike Noren were the ones who claimed
>cladistics was fully objective.

huh?

>Can ANY amount of anatomical insight falsify the most parsimonious
>tree in your eyes?

After all these years I still cant figure you out Peter. Is this
serious? I guess all I can do is ignore my suspicions and forge ahead.
To put it simply. Morphological cladists must be expert anatomists. We
study the anatomy according to the standards of the field of anatomy.
Cladistics per se is a sytematic method. It does not tell me how to do
anatomy. Cladistic principles tell me how to deal with characters once
they are defined. They are defined by Tom the anatomist, not Tom the
systematist. The "objectivity" of cladistics has to do with systematic
procedures. Character defintions are presented, they are made clear.
Any weighting which is done is done so explicitly, and must be
defended. The matrix is presented, as well as the algorithm
identified. In a properly done analysis, any reader of the paper
should be able to retrace the authors steps and arrive at the same
conclusion, or have every arguable issue laid out, and the authors
reasons for taking certain positions identifiable. This is in contrast
to the "evolutioanry systematists" who use subjective "judgement",
somethiong which does not allow for an "objective" resolution.

[Mike Noren]

Thusly nyi...@math.scarolina.edu (Peter Nyikos) spake unto all:

(snip)
: In other words, you rely on the subjective judgment of anatomists.

: Funny, I thought you and Mike Noren were the ones who claimed
: cladistics was fully objective.

You wouldn't mind supplying a quote to back your claim that I've
stated that "cladistics" was "fully objective", would you? You see, as
is too often the case when you paraphrase my opinions or claims, I
don't recognise them as mine.

Or, if you are unable to document that I've actually stated the above,
I would appreciate a retraction.

BTW - I've never seen Tom DiBenedetto claim this either.

: Peter Nyikos

Michael Norén, Doctoral student, Tel: Int +46 (0)8 6664236
Swedish Museum of Natural History, Fax: Int +46 (0)8 6664125
Dept. of Invertebrate Zoology
P.O.B. 50007
S-104 05 Stockholm, Sweden

[Cal King]

In article <34e9d882....@news.itd.umich.edu>, td...@umich.edu wrote:

>Detailed anatomical examinations can cause one to change ones original
>judgement that two structures are similar.

Detailed anatomical examinations of the bird manus by Burke and Feduccia has
not caused the cladists to change their original judgment that it is
homologous to the theropod manus.

>Morphological cladists
>spend the great majority of their research time dealing with such
>issues.

In the case of the bird manus, it was the non-cladists Burke and Feduccia who
actually spent the time dealing with the issue of homology. The cladists
themselves ASSUMED homology a priori, as their "null hypothesis."

> In this sense we can say that convergences are identified
>before cladistic analysis.

Ah, but unfortunately homology is often assumed a priori in cladistic
analysis. Homology is their null hypothesis. But few cladists bother to
analyze their characters a priori (for doing so would interject subjectivity
to the analysis), so they elect instead to use the results of the analysis as
test of homology a posteriori.

>But in fact, since very often we do not
>have any idea of what the phylogeny is before cladistic analysis, it
>is technically proper to say that we are not really distinguish
>convergences, but rather eliminating cases of superficial similarity.

As I have stressed, homology assessment doesn't have to rely on a tree.
Comparative anatomy and embryological evidence are often good enough evidence
of convergence.

>A "convergence" implies that the phylogeny is known.

A convergence implies that the 2 structures are different enough in their
details so that it is probably not of common origin.

>The phylogeny is not known until after the analysis.

It may not even be known after the analysis, because of the uncertainty over
whether the synapomorphies are in fact homologous, as Carroll and Dong wrote
in their paper.

>Wrong again. Characters which are synapomorphous on a tree are
>interpreted as homologies.

More of the same circular argument. How do we know that the characters are
homologous if the tree is constructed from these same characters with the
assumption that the characters are homologous?

>To claim homology as a test of the tree is nonsense.

To claim otherwise is nonsense. Hennig said that relationships can only be
inferred from true synapomorphs. Hence a tree cannot be correct unless the
characters are in fact homologous.

>It would imply
>some independant knowledge of what is and what is not homology.

Yes, it is called Dollo's principle, which states that evolution is not likely
to occur in exactly the same minute details in two independent lineages.

>Homology is similarity due to descent; one needs a phylogeny in order
>to identify homology.

Nonsense, using Dollo's principle, we can show that two characters are more
likely to be homologous than not if they are identical in every minute detail.
The identity in the genetic code among disparate organisms on earth, for
example, is evidence of homology of the genetic code, even if the details of
the branching orders among all the organisms may never be fully known.

> But there are no other methods for determining homology
>other than carreful study of the character, and testing under the
>expectation of congruence (i.e. cladistic methodology).

Consensus cladograms make a mockery of the claim that cladistics is a test of
congruence. Conflicting cladograms are often melted into consensus
cladograms, instead of being used to test for homologies through congruence.

> Unless you
>have a private line to god.....perhaps that is why you speak so
>authoritativly about things you dont seem to understand.

I won't even dignify that nonsensical statement with a response.

>>"Using cladistic methodology, ... it is all but impossible to ascertain which
>>characters are homoplasies ... and which are homologous...." (Feduccia
>>1996:62)
>
>Feduccia is wrong here.

According to you, the sole arbiter of the truth?

>>The weakness of cladistics lies in the use of a large number of characters and
>>organisms which,
>
>Great! Run with this argumetn "cal"! Lets use :LESS evidence,,,that
>should give us better results.

It is quality, not quantity that counts. A lot of phylogenetic noise is not
going to get the cladist anywhere close to the historical truth.

>> According to Feduccia, Gauthier (1986) did not do any character analysis
>>prior to his cladistic analysis.
>
>Who cares what Feduccia said?

I imagine you don't.

>You then claim that someone you call a cladist admits to this:

Why, do you have reason to doubt my claim that Schwenk is a cladist? Is he
actually committing blasphemy against cladism, hence he is no longer
considered a cladist?

"...I would not be honest if I said I agreed with everything Kurt has written
in this paper. I do not know if Kurt thinks of himself as a cladist or not.
Nevertheless, I am pretty sure that Kurt and I are still on speaking
terms."--D. Frost

Apparently Frost shared your dislike of what Kurt Schwenk said in his 1994
paper.

>>"In the face of no opposing evidence, we assume that characters are
>>independent and hope that, in any case, enough 'good' characters will outweigh
>>such hidden redundancy. Since we cannot judge the quality of a character
>>(i.e., its 'phylogenetic signal') except in light of a phylogeny, we can never
>>test our assumptions directly, only indirectly by their congruence with other
>>character sets." (Schwenk, Kurt 1994 Systematics And Subjectivity: The
>>Phylogeny And Classification Of Iguanian Lizards Revisited. Herpetol. Rev.
>>25[2]:53-57)
>
>And of course, the quote says nothing about the lie you proposed.

What lie? You keep accusing other people of lying, but you give no supporting
evidence. Is it because you have appointed yourself the sole arbiter of the
truth?

>False. Most cladists I know use as much fossil evidence as they can.

Which means very little fossil evidence since most fossils lack a large number
of the characters of extant organisms and must be discarded.

>You ahve failed to address my question. How is this anything but a
>dispute amongst comparative anatomists?

It is a dispute between those who actually studied the bird manus and theropod
manus in detail (aka the comparative anatomists like Burke, Feduccia and
Hinchliffe) and those who assume that they are homologous because they are
similar superficially (aka the cladists).

> What is it about bird digits
>that cause a cladists to identify them one way and a darwinian
>another?

It is not the bird digits, but the level of care used in ascertaining homology
which distinguishes an evolutionary systematist from a cladist. The
best evolutionary systematists try to make damn sure that his characters are
in fact synapomorphies. The cladists simply assume homology a priori.

>> Burke and Feduccia (1997, Science), using embryological evidence, show
>>that the bird manuss most likely has digits II-III-IV, and thus the bird
>>manual digits are not the same as the theropod digits, calling into question
>>the cladists' conclusions that birds are theropods.
>
>And if they are right, then Gauthier is wrong about the anatomy.

And the homology.

> What does this have to do with cladistics?

It has a great deal to do with cladistics. Cladists do not take the time
necessary to ascertain the homology of the characters they use.

>If he is wrong, they he woudl
>recode the character and probably get a different result.

Apparently he could not tell whether he was wrong, even though he had a tree.
Once again, it falsifies the cladist's claim that homology cannot falsify a
tree. The fact that Gauthier, according to you, has to recode the character
is evidence that Burke and Feduccia's data can falsify a tree. Homology can
indeed falsify a tree. Here we have you admitting to that fact.

>gosh you go to such simpleminded extremes in order to disparage
>people. This is a nonsense argument. Obviously both workers have
>examined the bird manus.

It is not obvious to me. Feduccia said Gauthier didn't analyze the characters
he used. Did you read Gauthier's paper? How can you say that it is obvious
that Gauthier analyzed the bird manus?

>This is an anatomical dispute. It has nothing to do with how they will
>conduct a phylogentic analysis later

It has everything to do with the phylogenetic analysis. Homoplasies cannot be
used to infer relationships. Hence this case is an illustration of the lack
of cladistic rigor in the most important part of a phylogenetic
analysis--character analysis.

>>, can you give me any evidence to support your claim
>>that cladists routinely perform comparative anatomical studies on all or even
>>most of the characters they use for phylogenetic analysis?
>
>yeah "cal", why dont you read some systematic works instead of
>scouring the litereature to find some nasty quote which some turkey

^^^^^^
>made about cladistics,

I read some of them and I don't find any evidence of character analysis by the
cladists. Can you give me specific cladistic papers in which there is
evidence of careful character analysis? As usual, your post seems to be
extremely high on the heat/light ratio. A lot of heated words but no
references for illumination.

>It usually is harder to make slanderous charges
>agaionst people you actually know.

Yeah, I would like to see you look those people whom you called "turkey"
above in the eye.

[Tom DiBenedetto]

Cal King wrote:

>Detailed anatomical examinations of the bird manus by Burke and Feduccia has
>not caused the cladists to change their original judgment that it is
>homologous to the theropod manus.

So what? This is still an anatomical dispute. It is not because they
are cladists that they are unconvinced by Burke and Feduccia, it is
because they see the anatomy differently.

>>Morphological cladists
>>spend the great majority of their research time dealing with such
>>issues.

>In the case of the bird manus, it was the non-cladists Burke and Feduccia who
>actually spent the time dealing with the issue of homology. The cladists
>themselves ASSUMED homology a priori, as their "null hypothesis."

No one has claimed that you need to be a cladist in order to be a
comparative anatomist. In fact this reinforces my point that the two
are separate issues. Your remark about an assumption of homology is
ridiculous. The character is assumed homologus (as oppposed to
convergent) for the purpose of cladistic analysis, AFTER it has been
concluded that it is homologous in anatomical study.

>> In this sense we can say that convergences are identified
>>before cladistic analysis.

>Ah, but unfortunately homology is often assumed a priori in cladistic
>analysis. Homology is their null hypothesis.

Listen "cal". I dont understand why you are content to rest in
permanent confusion on this issue. Is the ability to make nasty and
disparaging comments about a group of scientists so very important to
you that you will cling to mistaken ideas and destroy your own
credibility rather than understand what is going on? Homology is
hypothesized after long and extensive anatomical study. It is not
assumed at any point prior to cladistic analysis. Once all the
anatomical work is done, the characters defined, the matrix
constructed,,,then, for the purpose of the subsequent systematic
analysis, homology is "assumed". IT is not a null hypothesis. It is a
hypothesis which is tested. You must provisionally accept a hypothesis
for the purpose of testing it.

> But few cladists bother to
>analyze their characters a priori (for doing so would interject subjectivity
>to the analysis),

As I have said several times. this is a baldface lie.What is your
motivation here? Are you here merely to vent your demons, or do you
have any desire to learn and to participate in science? Do you think
you are talking to a bunch of idiots? We know what we do, and what our
colleagues do.

>so they elect instead to use the results of the analysis as
>test of homology a posteriori.

blah, blah,. blah

>>A "convergence" implies that the phylogeny is known.

>A convergence implies that the 2 structures are different enough in their
>details so that it is probably not of common origin.

And similar enough so that one could be confused. Are the tail flukes
of cetaceans convergent with fish fins? Would anyone be confused to
the point of considering them "similar"? If they are not judged
similar, they can hardly be considered a convergent similarity.

>>The phylogeny is not known until after the analysis.

>It may not even be known after the analysis, because of the uncertainty over
>whether the synapomorphies are in fact homologous, as Carroll and Dong wrote
>in their paper.

And as everyone knows. It seems only to be you who believes that
ultimate truth is knowable, and ironically, by the least rigorous
methods available.

>>Wrong again. Characters which are synapomorphous on a tree are
>>interpreted as homologies.

>More of the same circular argument. How do we know that the characters are
>homologous if the tree is constructed from these same characters with the
>assumption that the characters are homologous?

"Circular' is a nice insult to be applied to an argument, but you
sholdnt cry wolf with it. You have not demonstrated any circularity,
for their is none. Homology is hypothesized on the basis of anatomical
study, Homology hypotheses are combined and tested for congruence.
Non-congruent homology hypotheses are falsified; we cannot continue to
hypothesize their homology for they fail to match our expectation of
congruence. Those that are congruent have passed that test and are
further corroborated as homologies. No circularity whatsoever.

>>To claim homology as a test of the tree is nonsense.

>To claim otherwise is nonsense. Hennig said that relationships can only be
>inferred from true synapomorphs. Hence a tree cannot be correct unless the
>characters are in fact homologous.

So? How does that make homology a test of the tree? What concept of
homology do you propose that is independant of the tree such that it
could test the tree? The only possibility I see is an ahistorical
concept of homology, which would put you firmly back into the
eighteenth century.

>>It would imply
>>some independant knowledge of what is and what is not homology.

>Yes, it is called Dollo's principle, which states that evolution is not likely
>to occur in exactly the same minute details in two independent lineages.

And of what use is this in the general exercise of moving from a
matrix of characters to a tree?

>>Homology is similarity due to descent; one needs a phylogeny in order
>>to identify homology.

>Nonsense, using Dollo's principle, we can show that two characters are more
>likely to be homologous than not if they are identical in every minute detail.

yeah, that is why we code them as homologous.

> The identity in the genetic code among disparate organisms on earth, for
>example, is evidence of homology of the genetic code, even if the details of
>the branching orders among all the organisms may never be fully known.

obviously, "cal", that is why we accept the genetic code as a
homology. We would code it as such.

>> But there are no other methods for determining homology
>>other than carreful study of the character, and testing under the
>>expectation of congruence (i.e. cladistic methodology).

>Consensus cladograms make a mockery of the claim that cladistics is a test of
>congruence. Conflicting cladograms are often melted into consensus
>cladograms, instead of being used to test for homologies through congruence.

huh? Consensus cladograms are merely representations of those nodes
which are unambiguously present in the data. The polytomies in a
consensus cladogram merely indicate that there is conflict in the
data. How could you possibly justify asserting a particular set of
relationships when the characters you have defined are not consistent?
Doing anything but presenting a consensus would be to ignore the data.

> Most cladists I know use as much fossil evidence as they can.

>Which means very little fossil evidence since most fossils lack a large number
>of the characters of extant organisms and must be discarded.

Why must they be discarded? The characters which they have are used,
those they lack are coded as missing data. Why would you want to throw
out evidence?

>>You ahve failed to address my question. How is this anything but a
>>dispute amongst comparative anatomists?

>It is a dispute between those who actually studied the bird manus and theropod
>manus in detail (aka the comparative anatomists like Burke, Feduccia and
>Hinchliffe) and those who assume that they are homologous because they are
>similar superficially (aka the cladists).

And once again, on what grounds can you claim that the cladists didnt
study the manus as much as anyone else? What makes you feel justified
in making wild accusations against someone simply because they come to
a different conclusion than the people you seem to worship?

>> What is it about bird digits
>>that cause a cladists to identify them one way and a darwinian
>>another?

>It is not the bird digits, but the level of care used in ascertaining homology
>which distinguishes an evolutionary systematist from a cladist. The
>best evolutionary systematists try to make damn sure that his characters are
>in fact synapomorphies. The cladists simply assume homology a priori.

A nice fantasy, but completely false.

>It has a great deal to do with cladistics. Cladists do not take the time
>necessary to ascertain the homology of the characters they use.

Once again, what is your real motivation in endlessly repeating things
which are slanderous, false, and easily disproven were you to make a
simple effort to investigate what real people are really doing?

[Jason Anderson]

Peter Nyikos wrote:
>
> [posted and mailed, with request for a CC of any reply]

(BTW, replace spam trap with bio1.lan to reply)

>
> Jason Anderson <jasona@no_spam.mcgill.ca> writes:
>
> >Cal King wrote:
>
> [Tom DiB?]

(Naw, me)

> >> >Homoplasies are not identified *except* in light of a tree.
>
> >> That is a common claim among cladists, but it is demonstrably false.

<snip>

> >See Tom DiBennidito's reply.

<snip>

> >> Another
> >> example is the giant panda's opposable "thumb." It is not a real digit at
> >> all, but is formed instead from the radial sesamoid. Hence it is not a
> >> homolog of the vertebrate thumb at all.
>
> >Correct. Nobody would score it as "thumb (+)" either, except your
> >strawman cladist.
>
> That's only because a careful anatomist could tell it is NOT a pollex.
> In the example I gave Tom about a year ago, I very carefully
> used the word "pollex" to remove all ambiguity about something
> an expert anatomist might identify as such.
>
> I said that I would take a cladogram that has the pollex
> disappearing and re-appearing at successive nodes several times
> as wrong even if it was the most parsimonious tree; and that
> I would regard that feature as falsifying not only the tree
> itself but also the very principle of maximum parsimony.
>

You are free to do so if you wish. Why would this character be so
crucial as to "falsify...the tree [and] the principle of maximum
parsimony" anyway? Podial elements are notorious for disappearing and
reappearing evolutionarily. Many secondarily aquatic tetrapods cease
ossifying podials, or ossify them at increasingly delayed times of
onset. Others, notably ichthyosaurs, mosasaurs and cetaceans *increase*
ossification to lock the elements into rigid paddles (while adding
phalanges at a startling rate). If you can wait a bit I can check to
see if any lineage first delays podial ossification (say _Hupesuchus_)
while later increasing ossification.

Mostly I get the feeling that you strongly subscribe to Dollo's
principle. Is this correct?

> Tom, of course, wouldn't buy that.
>

I'm sure. I don't either.

> >> This fact was revealed again by
> >> comparative anatomy, long before systematists started arguing whether the
> >> panda is a bear or a raccoon.
>
> >FYI, cladists are among the few remaining comparative anatomists left in
> >biology.
>
> How do you know?
>

Well, comparative anatomy as a taught subject, believe it or not, is
actually in many institutions being weeded out of the undergraduate
cirriculum. Nobody wants (or can) teach it because the focus (and
money) has moved to molecular biology.

Comparative anatomy is still being used by functional morphologists,
systematists and paleontologists. Of these groups, systematists and
paleontologists are most concerned with finding evolutionary
relationships-and cladistics is the dominant paradigm for doing so.
Functional morphologists are also using phylogenies to reconstruct the
evolution of structures and functional units-these again are based upon
cladistic work predominantly. Hold outs for evolutionary systematics or
other techniques are few and becoming fewer every year.

> >> OTOH, if we use a tree such as Cracraft's
> >> analysis of the owls and the hawks, we may conclude that their similarities
> >> are homologs, when in fact these similarities have been regarded as
> >> convergences for most of the 20th century.
> >>
> >Right. People also used to think that babies existed in a fully
> >developed, if tiny, form within the sperm. And that the earth was the
> >centre of the universe. Whether Cracraft was correct or not aside,
> >duration that ideas are held is not a sound criterion for accuracy.
>
> Do I detect a note of skepticism about hawks and owls forming a
> polyphyletic taxon?
>

You detect intellectual honesty-I don't know much about relationships
among living birds, I don't have time to read about it right now to
contribute to a lower-priority persuit like posting on usenet, so I will
not speak towards Craycraft's results specifically-only the general
point about how our knowledge progresses. Flawed studies are flawed
studies-cladistic, phenetic or otherwise.

> [...]
>
> >> Using presumed synapomorphies to construct a tree and then using the tree to
> >> identify synapomorphies amounts to a circular argument.
> >>
> >Again, one uses *synapomorphies* to construct the tree. The tree
> >interprets *homology* from *homoplaisy*.
>
> Synapomorphy *is* homology, isn't it?
>

No. Synapomorphy is a "shared derived trait". Homology is "same organ,
regardless of form or function (Owen)".

One examines the animals under study. One notices a character, which
seems to be present in one group but different in another group of study
animals (or plants, etc...). One must now apply a series of tests to
establish homology. The first set is collectively referred to as "the
test of similarity". Is it in a similar place? Does it relate with
similar other organs/bones/etc. Does it scale geonetrically? Does it
form from similar embryologic tissues? Is it similar to other
structures, but appear at an earlier stage developmentally?

If it passes similarity, it must next pass the "test of consilience" Is
the structure under study and the structure with which one wishes to
establish homology present in the same organism? (Say wings and arms in
angels) If so, the two structures cannot be homologous.

If the structure passes all of these tests, one has an apomorphy-a
shared trait. Assuming the outgroup has a primitive state it will be a
synapomorphy-shared *derived* trait. Is it a homology? No, because
there is one further test to go-the test of congruence-or, how does it
stand up compared with all of the *other* shared derived traits? How to
tell? Run the matrix of all synapomorphies through the computer (or do
it by hand if you like). If the character is congruent with the tree,
it is a homology. If it requires additional ad hoc statements to
explain away an incongruent distribution (statements of reversal or
convergence) then it is a homoplaisy-a false homology.

They are not interchangable terms.

> [...]
>
> >> >Again, how does one decide a priori what is a homoplasy-unless one
> >> >already knows what one wants the tree to look like?
> >>
> >> As every well trained evolutionary systematist knows,
>
> >(A common rhetorical technique designed to portray myself as "poorly
> >trainned")
>
> Don't be so trigger-happy. It looks to me like Cal was defending
> evolutionary systematists, not attacking cladists or you as "poorly
> trained".

Fine, I'll reserve judgement for now.

>
> [...]
>
> >> One of the ways to determine homology a
> >> priori is comparative anatomy.
>
> >See above, and earlier posts. Comparative anatomy is what most time is
> >spent on-before a matrix is built. One establishes *synapomorphies* in
> >this manner-using a rigorous series of tests.
>
> In other words, one looks through the literature to see what
> other cladists have done, and takes their *conclusions* about
> homology as starting points for what should be synapomorphic, no?
>

No, see above. I right now am taking "what other cladists have done"
and am working, character by character, to assure myself that they are
correct. I disagree with the authors on several major points (so far).
At the same time I am reexamining as many specimens as possible to fing
*new* characters and assure myself that previous descriptive work was
done properly. The result will be a different matrix with different
characters (some new, others removed or redefined), different character
coding and different operating assumptions, because my views of
evolution are different. All of the above will be *explicitly*
discussed so "the community of biologists as a whole" may draw their own
conclusions as to the validity of my (and previous author's) choices.
Mind you, all of this stems from a revision, with *detailed* anatomic
description and alpha taxonomy, of a previously poorly known group.
Cladistic analysis is only one facet of a much larger study. Other
areas interest me more (generating life-history data from fossils, for
one).

> The final test, though,
> >is the tree-it tells *homology* from false homology-in light of all
> >other homologies. Things like the cephalopod eye would not be included
> >in the analysis because it is not a synapomorphy with the vertebrate
> >eye-it fails the test (a couple, truth be told) of similarity.
>
> Similarity is a relative concept. You aren't seriously claiming
> that cladists will score any two things that show the slightest
> dissimilarity as different *a priori* do you? That wouldn't even
> allow for individual variation within a litter of puppies.
>

Octopod eyes are structurally and developmentally different.
Intraspecific variation is accounted for in the alpha taxonomy.
Remember that cladistics needs *well defined* OTUs to operate properly.
But as DiBenedetto (sorry for previous misspellings but this newsreader
won't allow me to switch windows without dumping the message) pointed
out, even *if* it were coded as a synapomorphy the analysis should show
it to be homoplasious.

> >> Another way, which is seeing increasing use,
> >> is developmental biology (e.g. Burke and Feduccia's embryological study of the
> >> bird manus).
>
> >This study is flawed-their identification of digital number was based
> >upon a preconcieved notion-which was to be tested.
>
> After seeing Chris Brochu argue about this example, I don't think
> I'll agree with you without you going to a lot of detail about it.
>

Agreed.

> Another
> >identification of digital number which is equally likely
>
> In what? Birds or theropods?
>

Birds. Since Feduccia and workers did not see all digits in their devo
study, which digit was most anterior- I, II, III- could not be
established will certainty=100%. Their choice of digital homology was
colored by their desired outcome, **IMHO**. Again, I don't want to
persue this further.

> [...]
>
> >> Hence it is better to perform some character analysis to
> >> determine character goodness and to discard characters that are suspect prior
> >> to systematic analysis than to assume a priori that all characters chosen are
> >> synapomorphies, then plunge into an analysis that assumes all characters are
> >> of equal importance in the evolution of a group of animals. Better character
> >> analysis makes for better systematic analysis. In other words, quality is
> >> more important than quantity.
>
> >I agree :o. But what you seem to want to refuse to acknowledge is that
> >characters ARE analysed before included in a matrix.
>
> +++++++++++++++++++++++ dogmatic cladistic posting mode on
>
> Either the analysis took the form of earlier cladistic use
> of trees or else it was subjective.
>
> +++++++++++++++++++++++ dogmatic cladistic posting mode off
>

????????????
You are posting "dogmatic cladistic" nonsense, I'm afraid.

> Your idea of how
> >cladistics works is not how cladistics is actually done, so your
> >criticism of cladistics in this thread is moot.
>
> >But we've already covered this ground, haven't we?
>
> Not with me, you haven't.
>

[Insert dramatic soundtrack]

[Tom DiBenedetto]

Jason Anderson wrote:

>> Synapomorphy *is* homology, isn't it?

>No. Synapomorphy is a "shared derived trait". Homology is "same organ,
>regardless of form or function (Owen)".

>If the structure passes all of these tests, one has an apomorphy-a

>shared trait. Assuming the outgroup has a primitive state it will be a
>synapomorphy-shared *derived* trait. Is it a homology? No, because
>there is one further test to go-the test of congruence

.................

> If the character is congruent with the tree,
>it is a homology. If it requires additional ad hoc statements to
>explain away an incongruent distribution (statements of reversal or
>convergence) then it is a homoplaisy-a false homology.

>They are not interchangable terms.

This raises an interesting terminological dispute. I disagree with the
precise way in which you are using these terms. I sense that you are
following the definitions proposed by Sober in that lousy 1988 book,
and used by Kluge and a few others. There are many (most?) cladists
who use the terms differently.
Following Hennig (who coined the terms), synapomorphy is indeed
"shared derived trait". and he defined the term in contrast to
plesiomorphy, and both in reference to the two sides of an
evolutionary transformation within a monophyletic group. Thus the
terms are ontological; they were clearly intended to refer to the
reality which we are trying to reconstruct. As such, they should be
applied to a result; ie characters on a tree, not character matches
prior to analysis. I think this is futher reinforced by the fact that
"apo" refers to derived, and this is something which can only be
determined on a rooted tree. There is nothing in the
character-defintion phase (the anatomical descriptions) which imbues a
character with the quality of being derived or primitive.
When considering the basis for grouping, Hennig uses the tripartite
relationship between apomorphy, plesiomorphy and convergence. Once
again, this indicates clearly that the terms are meant to refer to
results on a tree, not to proposals prior to analysis.

Sober has acknowledged (in the book) that his definitions are
non-standard, and Kluge as well has outlined why he thinks it
important to use the terms in ways other than Hennig intended (pers.
comm). I agree that his goals are important (developing parallel
ontological and epistomological concepts), but I dont think this
particular formulation is very helpful. As I am sure you are aware,
there are other cladists who use the terms in line with Hennig's
intentions and who conclude that synapomorphy is equivalant to
homology (e.g. Patterson).

The question of how the concept of homology relates to the concept of
synapomorphy is, I think, as follows. Homology is, as you point out, a
term derived from comparative anatomy, and its meaning was formalized
by Owen in the days before the general acceptance of evolution. The
key word though, is "same" (the "same" organ....). With the acceptance
of the notion of evolution, I think it fair to say that "sameness" was
generally seen as being present by descent, i.e. the identity of a
character was a function of its origin and persistance in the process
of descent through reproduction. This is of course equivalent to the
cladistic concept of the identity of taxa; their identity is not a
function of subjective assesments of distinctiveness as the darwinians
would have it, but rather is strictly a function of historical
descent. So homology is also an ontological term; I think we agree on
this. It is precisely because both synapomorphy and homology are
ontological terms that some have proposed redefining one as an
epistomological term. But I really dont think it would make much sense
to use apomorphy for such a term.

So, my view is that when we study organisms, and we note similarities,
and we test these similarities, we are building toward a hypothesis of
homology. If the character has survived these tests, they are
homologies in the Owenian sense, but I would agree with you that they
need to be subjected to historical testing (the test of congruence)
before we can accept them as homologies in the modern sense. But the
tree is what specifies the range over which we will accept them as
"shared", and the rooted tree species whether they are apo or plesio.
So for me, synapomorphy refers to homologies; characters on a tree.
Before analysis they are putative homologies, and their identity as
apomorphy, plesiomorphy or convergence has yet to be determined.

[Cal King]

In article <34eaecd2....@news.itd.umich.edu>,

td...@umich.edu (Tom DiBenedetto) wrote:
>Path:

wn7!bgtf1!worldnet.att.net!news-out.internetmci.com!newsfeed.internetmci.com!1
31.103.1.116!news2.chicago.iagnet.net!qual.net!iagnet.net!newsxfer3.itd.umich.

edu!news.itd.umich.edu!not-for-mail
>From: td...@umich.edu (Tom DiBenedetto)

>Newsgroups: sci.bio.systematics,sci.bio.paleontology

>Subject: Re: Methods and characters (Re: Quetzalcoatlus northropi!!!)
>Organization: University of Michigan Museum of Zoology
>Reply-To: td...@umich.edu

>Message-ID: <34eaecd2....@news.itd.umich.edu>
>References: <34CA5DB0...@saluki-mail.siu.edu>

<6al5s0$8...@redwood.cs.sc.edu> <6alo66$g...@saluki-news.it.siu.edu>
<6angij$n...@redwood.cs.sc.edu> <6avil6$4...@bgtnsc03.worldnet.att.net>
<6b7dgl$9...@redwood.cs.sc.edu> <6bh15e$g...@bgtnsc03.worldnet.att.net>
<34E86102.6218@no_spam.mcgill.ca> <6cb3fh$8...@bgtnsc03.worldnet.att.net>

<34e9a15c....@news.itd.umich.edu> <6cci5a$h...@bgtnsc03.worldnet.att.net>
<34e9d882....@news.itd.umich.edu> <6cdj0o$6...@bgtnsc02.worldnet.att.net>

>X-Newsreader: Forte Agent 1.5/32.452
>MIME-Version: 1.0

>Lines: 220
>Date: Wed, 18 Feb 1998 15:06:42 GMT
>NNTP-Posting-Host: 141.211.110.20
>NNTP-Posting-Date: Wed, 18 Feb 1998 10:06:42 EST
>Xref: wn7 sci.bio.systematics:3005 sci.bio.paleontology:15450
>Status: N

>
>Cal King wrote:
>
>>Detailed anatomical examinations of the bird manus by Burke and Feduccia has
>>not caused the cladists to change their original judgment that it is
>>homologous to the theropod manus.
>
>So what? This is still an anatomical dispute. It is not because they
>are cladists that they are unconvinced by Burke and Feduccia, it is
>because they see the anatomy differently.

It is not just an anatomical dispute. It is also a methodological and
a homology dispute. According to R. Hinchliffe (1997, Science 278:597), the
use of embryological evidence "...represents a different methodology in
ascertaining homology from that adopted by many paleontologists, who use
multiple synapomorphies." Of course, Feduccia (1996, Origin and Evolution of
Birds) stated that paleontologists have voted overwhelmingly for cladistics as
a method of analysis. In fact, the bird-dinosaur link has become a virtual
paleontological-cladistic dogma.

Hinchliffe continues, "This convincing evidence of II-III-IV wing digit
identity will not be to the liking of the cladistic supporters of a dinosaur
origin of birds. For them, it introduces the possibility of convergence
(rather than common origin) as an explanation of the similarities between the
structure of the forelimb (and, indeed, of other structures) of theropods and
the wing of Archaeopteryx." (Ibid.)

Not only is the developmental evidence contradictory to the dogma of
the bird-theropod link, it calls into question the fundamental soundness of
cladistic methodoloy.

"L.B. Halstead (1982) has criticized cladistic methodology for the apparent
precision and respectability that cladistics confers by its cladograms on what
is in reality no more than speculation. He points out that the procedures of
selecting synapomrophies, inferring polarity, and resolving conflicts are
subjective and unreliable." (Feduccia 1996:59)

The bird manus evidence shows that cladists, because of the constraints of
their methodology have a great deal of difficulty distinguishing
synapomorphies from homoplasies.

>No one has claimed that you need to be a cladist in order to be a
>comparative anatomist.

But cladism discourages the comparative anatomist from interjecting judgment
into his analysis. Problematic characters, for example, cannot be removed
from the analysis without introducing subjectivity.

>Your remark about an assumption of homology is
>ridiculous. The character is assumed homologus (as oppposed to
>convergent) for the purpose of cladistic analysis, AFTER it has been
>concluded that it is homologous in anatomical study.

What study?

> Homology is hypothesized after long and extensive anatomical study.

Show me evidence of such studies. Your statements are again high on the
heat/light ratio. A lot of heated words saying I am wrong, but no published
evidence to back up your claim, followed by a barrage of ad hominem that
deserves silence as a response.

>Homology is hypothesized on the basis of anatomical
>study, Homology hypotheses are combined and tested for congruence.

Again, show me published evidence of such studies prior to cladistic analysis.
Feduccia said that Gauthier (1986) did not analyze his characters. Care to
show published evidence to disprove Feduccia?

>>>To claim homology as a test of the tree is nonsense.
>
>>To claim otherwise is nonsense. Hennig said that relationships can only be
>>inferred from true synapomorphs. Hence a tree cannot be correct unless the
>>characters are in fact homologous.
>
>So? How does that make homology a test of the tree?

If the assumption of homology can be falsified, then the tree that is
constructed from these characters would then be falsified.

>>Consensus cladograms make a mockery of the claim that cladistics is a test
of
>>congruence. Conflicting cladograms are often melted into consensus
>>cladograms, instead of being used to test for homologies through congruence.
>
>huh? Consensus cladograms are merely representations of those nodes
>which are unambiguously present in the data.

And those nodes that show conflict are neatly swept under the rug.

>The polytomies in a
>consensus cladogram merely indicate that there is conflict in the
>data.

Not all consensus cladograms show polytomies. Some cladists reject polytomies
altogether. Hence the test of congruence for homologies do not exist. They
don't exist because conflicting cladograms are melted into consensus
cladograms, instead of being used to falsify homology.

>> Most cladists I know use as much fossil evidence as they can.
>
>>Which means very little fossil evidence since most fossils lack a large
number
>>of the characters of extant organisms and must be discarded.
>
>Why must they be discarded?

Because the cladists think they lack most of the characters of extant
organisms.

>The characters which they have are used,
>those they lack are coded as missing data. Why would you want to throw
>out evidence?

I don't, but cladists who reject the use of fossils do throw out
evidence.

>And once again, on what grounds can you claim that the cladists didnt
>study the manus as much as anyone else?

What evidence do you have that they did? Feduccia said Gauthier didn't.
That is my evidence. So what is your evidence?

>Once again, what is your real motivation in endlessly repeating things
>which are slanderous, false, and easily disproven were you to make a
>simple effort to investigate what real people are really doing?
>
>Tom DiBenedetto td...@umich.edu

I see you have once again indulged in ad hominem but you still have not given
a single reference to support any of your points.

[Cal King]

In article <34EAF811.45FC@no_spam.mcgill.ca>, Jason Anderson
<jasona@no_spam.mcgill.ca> wrote:

>Well, comparative anatomy as a taught subject, believe it or not, is
>actually in many institutions being weeded out of the undergraduate
>cirriculum. Nobody wants (or can) teach it because the focus (and
>money) has moved to molecular biology.

I believe it. According to the late Angus d'A Bellairs (1989, Book Review:
Phylogeneric relationships of the lizard families. Essays commemorating
Charles L. Camp. Herpetol. Jour. 1[8]:379), "University libraries, at least
in the UK, where 'classical' zoology is everywhere in the retreat, are
unlikely to buy it, and it is beyond the pocket of most of us, ill-paid
academics." Indeed, classical, descriptive anatomy is in retreat even in the
UK. That is because while scholarly and informative, descriptive anatomy is
devoid of meaning unless it is put into an evolutionary or functional context.
d'A Bellairs continued the criticism of this book:

"The emphasis is on characters of systematic interest rather than upon
functional anatomy, a concomitant of the cladistic approach which will lead
morphology into even greater disrepute than it currently 'enjoys' among other
kinds of biologists." Again, more evidence that classical descriptive
morphology is being phased out, and it was not well received even when put
into a purportedly phylogenetic context.

>Comparative anatomy is still being used by functional morphologists,
>systematists and paleontologists. Of these groups, systematists and
>paleontologists are most concerned with finding evolutionary
>relationships-and cladistics is the dominant paradigm for doing so.
>Functional morphologists are also using phylogenies to reconstruct the
>evolution of structures and functional units-these again are based upon
>cladistic work predominantly. Hold outs for evolutionary systematics or
>other techniques are few and becoming fewer every year.

I imagine it is because of the "publish or perish" factor working along with
the cladistic bandwagon:

"At least in zoological journals, few recent papers reviewing the
classification of particular taxa have appeared which do not culminate in a
'cladistic analysis'; indeed, I suspect that many present editors of such
journals would tend to reject submitted systematic papers which failed to do
this." (Crowson, R. 1991. J. Nat. Hist. 25:3-5)

Perhaps the cladists have managed to become editors of many scientific
journals and are thus in a position to reject non-cladistic papers. Whatever
the reason, if a systematist cannot get his work published unless he uses
cladistics, then sooner or later, the reality of needing to publish will force
most to become cladists. This is a rather unfortunate trend because fashion,
not scientific rigor, becomes the driving force in scientific research.

>No, see above. I right now am taking "what other cladists have done"
>and am working, character by character, to assure myself that they are
>correct. I disagree with the authors on several major points (so far).
>At the same time I am reexamining as many specimens as possible to fing
>*new* characters and assure myself that previous descriptive work was
>done properly. The result will be a different matrix with different
>characters (some new, others removed or redefined), different character
>coding and different operating assumptions, because my views of
>evolution are different.

In other words, different cladists can come up with totally different opinions
of which characters are synapomorphs and of course how many synapomorphies
there are among the same group of organisms. Synapomorphies, in other words,
are indeed in the eyes of the beholder, as Feduccia claimed.

> All of the above will be *explicitly*
>discussed so "the community of biologists as a whole" may draw their own
>conclusions as to the validity of my (and previous author's) choices.

But nevertheless these are YOUR choices, based on YOUR subjective judgment.

>Birds. Since Feduccia and workers did not see all digits in their devo
>study, which digit was most anterior- I, II, III- could not be
>established will certainty=100%. Their choice of digital homology was
>colored by their desired outcome, **IMHO**. Again, I don't want to
>persue this further.

Or perhaps your rejection of their data is colored by YOUR desired outcome.
As I recall, you already indoctrinated your 5 year old daughter into saying
that "birds are dinosaurs." :)

[Cal King]

In article <34ebfded....@news.itd.umich.edu>, td...@umich.edu wrote:

>So for me, synapomorphy refers to homologies; characters on a tree.
>Before analysis they are putative homologies, and their identity as
>apomorphy, plesiomorphy or convergence has yet to be determined.
>
>Tom DiBenedetto td...@umich.edu

The same old circular argument. Characters are assumed to be synapomorphs so
they can be used to construct a tree and then the tree is used to confirm that
the characters are synapomorphs. It is this sort of methodology that has led
to such demonstrably false results as an owl-hawk clade and a loon-grebe
clade. DNA hybridization, for example, shows that loons are not at all
closely related to the grebes.

[Cal King]

In article <34eae52c....@news.itd.umich.edu>, td...@umich.edu wrote:
>Peter Nyikos wrote:

>> A matrix is a set of homology hypotheses which emerge from
>>>detailed study of the organism. It is only if a percieved similarity
>>>survives anatomical study (as the similarity between octopus and
>>>vertebrate eyes do NOT)
>>

>>In other words, you rely on the subjective judgment of anatomists.
>>Funny, I thought you and Mike Noren were the ones who claimed
>>cladistics was fully objective.
>

>huh?
>
>>Can ANY amount of anatomical insight falsify the most parsimonious
>>tree in your eyes?
>
>After all these years I still cant figure you out Peter. Is this
>serious? I guess all I can do is ignore my suspicions and forge ahead.
>To put it simply. Morphological cladists must be expert anatomists.

Actually many cladists simply take published data and reanalyze it. The
following paper is an example, in which a cladist takes the published data of
a pheneticist and re-analyzed it cladistically.

Channing, A. 1989 A re-evaluation of the phylogeny of old world treefrogs.
South African J. Zool. 24(2):116-131

>We
>study the anatomy according to the standards of the field of anatomy.
>Cladistics per se is a sytematic method. It does not tell me how to do
>anatomy. Cladistic principles tell me how to deal with characters once
>they are defined. They are defined by Tom the anatomist, not Tom the
>systematist. The "objectivity" of cladistics has to do with systematic
>procedures.

The identification of synapomorphies by Tom the anatomist requires subjective
judgment. The more knowledge the anatomist applies, the more subjective this
procedure becomes. Subjectivity is not wrong, but it contradicts the
cladist's claim of objectivity.

>Character defintions are presented, they are made clear.
>Any weighting which is done is done so explicitly, and must be
>defended. The matrix is presented, as well as the algorithm
>identified. In a properly done analysis, any reader of the paper
>should be able to retrace the authors steps and arrive at the same
>conclusion, or have every arguable issue laid out, and the authors
>reasons for taking certain positions identifiable. This is in contrast
>to the "evolutioanry systematists" who use subjective "judgement",
>somethiong which does not allow for an "objective" resolution.

Who claims that phylogeny reconstruction is an experimental science? As long
as the results can be falsified, then it is scientific. Cladists have been
claiming that they falsify earlier systematic studies, thus proving that those
studies must be scientific. OTOH, cladists claim that only a more
parsimonious cladogram can replace an earlier one. Thus Burke and Feduccia's
embryological evidence, even if correct, ostensibly cannot falsify the
cladist's hypothesis of the bird-theropod link, because Burke and Feduccia
have not performed a cladistic analysis of their own to produce a more
parsimonious cladogram. Now tell me, why is cladistics more scientific than
evolutionary systematics, even though it is harder to falsify, if we must play
by the cladist's idiosynchratic rules?

[Cal King]

In article <34ea36ea...@news.su.se>, Mike Noren (michae...@nrm.se)
says...

>You wouldn't mind supplying a quote to back your claim that I've
>stated that "cladistics" was "fully objective", would you? You see, as
>is too often the case when you paraphrase my opinions or claims, I
>don't recognise them as mine.
>
>Or, if you are unable to document that I've actually stated the above,
>I would appreciate a retraction.
>
>BTW - I've never seen Tom DiBenedetto claim this either.

I think I will save this post so I will have proof that cladists are claiming
that their method is NOT FULLY OBJECTIVE, something that evolutionary
systematists have claimed all along.

[Jason Anderson]

Tom DiBenedetto wrote:

>
> Jason Anderson wrote:
>
> >> Synapomorphy *is* homology, isn't it?
>
> >No. Synapomorphy is a "shared derived trait". Homology is "same organ,
> >regardless of form or function (Owen)".
>

<snip>

> >They are not interchangable terms.
>

> This raises an interesting terminological dispute. I disagree with the
> precise way in which you are using these terms. I sense that you are
> following the definitions proposed by Sober in that lousy 1988 book,
> and used by Kluge and a few others.

Yes. Actually for fun I almost directly quoted from a comparative
anatomy coursepack I had as an u-grad writen by Kulge and Fink. I was
hoping you'd recognize it :)

> Following Hennig (who coined the terms), synapomorphy is indeed
> "shared derived trait". and he defined the term in contrast to
> plesiomorphy, and both in reference to the two sides of an
> evolutionary transformation within a monophyletic group. Thus the
> terms are ontological; they were clearly intended to refer to the
> reality which we are trying to reconstruct. As such, they should be
> applied to a result; ie characters on a tree, not character matches
> prior to analysis.

<snip>

> I think this is futher reinforced by the fact that
> "apo" refers to derived, and this is something which can only be
> determined on a rooted tree. There is nothing in the
> character-defintion phase (the anatomical descriptions) which imbues a
> character with the quality of being derived or primitive.

Yes. For clarity in the present debate I included knowledge of the
character state of the outgroup before the analysis. But one can have
the outgroup in mind while looking at taxa-I continually fight over
proper outgroups, outgroup orders, etc., while doing so.

> When considering the basis for grouping, Hennig uses the tripartite
> relationship between apomorphy, plesiomorphy and convergence. Once
> again, this indicates clearly that the terms are meant to refer to
> results on a tree, not to proposals prior to analysis.
>
> Sober has acknowledged (in the book) that his definitions are
> non-standard, and Kluge as well has outlined why he thinks it
> important to use the terms in ways other than Hennig intended (pers.
> comm). I agree that his goals are important (developing parallel
> ontological and epistomological concepts), but I dont think this
> particular formulation is very helpful. As I am sure you are aware,
> there are other cladists who use the terms in line with Hennig's
> intentions and who conclude that synapomorphy is equivalant to
> homology (e.g. Patterson).
>

But before the analysis is run one cannot make statements about the
evolutionary origin of a shared character. I do see your point. But,
one does not usually talk about characters before analysis, except in
informal settings like usenet, which is where this is really an issue,
IMO.

Isn't part of Patterson's interchangable use of synapomorphy and
homology part of his methodological world view? He would conclude, "so
we don't need to use the term homology" as it is a word associated with
evolution-which he wants to remove from cladistics.

> The question of how the concept of homology relates to the concept of
> synapomorphy is, I think, as follows. Homology is, as you point out, a
> term derived from comparative anatomy, and its meaning was formalized
> by Owen in the days before the general acceptance of evolution. The
> key word though, is "same" (the "same" organ....). With the acceptance
> of the notion of evolution, I think it fair to say that "sameness" was
> generally seen as being present by descent, i.e. the identity of a
> character was a function of its origin and persistance in the process
> of descent through reproduction. This is of course equivalent to the
> cladistic concept of the identity of taxa; their identity is not a
> function of subjective assesments of distinctiveness as the darwinians
> would have it, but rather is strictly a function of historical
> descent. So homology is also an ontological term; I think we agree on
> this.

Yes, but I would add that homology is a more absolute statement about
common origin via evolution. A character can be shared, it can be
derived (WRT tha outgroup or other higher taxon) but not be the result
of shared ancestry.

> It is precisely because both synapomorphy and homology are
> ontological terms that some have proposed redefining one as an
> epistomological term. But I really dont think it would make much sense
> to use apomorphy for such a term.
>

Huh. I'll have to chew on this for a while, and I'd love to see other
opinions. Redefining previous terms is, as "cal" has pointed out
numerous times, not uncommon in the least, but is no justification for
further obscuring the issue.

> So, my view is that when we study organisms, and we note similarities,
> and we test these similarities, we are building toward a hypothesis of
> homology. If the character has survived these tests, they are
> homologies in the Owenian sense, but I would agree with you that they
> need to be subjected to historical testing (the test of congruence)
> before we can accept them as homologies in the modern sense. But the
> tree is what specifies the range over which we will accept them as
> "shared", and the rooted tree species whether they are apo or plesio.

> So for me, synapomorphy refers to homologies; characters on a tree.
> Before analysis they are putative homologies, and their identity as
> apomorphy, plesiomorphy or convergence has yet to be determined.
>

Putative homologies or hypothesised homologies don't really roll off the
tongue, though. I agree that the times when it is important to make a
distinction about our level of knowledge of a character is very
limited-to teaching and theoretical discussions-but I still see a need
to distinguish between pre- and postanalysis statements of shared
derived character states. Following you here- before the analysis we'd
have, what, synmorphies? :)

Good discussion. Any other input?

[Jason Anderson]

Cal King wrote:
>
> In article <34EAF811.45FC@no_spam.mcgill.ca>, Jason Anderson
> <jasona@no_spam.mcgill.ca> wrote:
>
<snip>

> Perhaps the cladists have managed to become editors of many scientific
> journals and are thus in a position to reject non-cladistic papers. Whatever
> the reason, if a systematist cannot get his work published unless he uses
> cladistics, then sooner or later, the reality of needing to publish will force
> most to become cladists. This is a rather unfortunate trend because fashion,
> not scientific rigor, becomes the driving force in scientific research.
>

PERHAPS little green men from mars run the United Nations. PERHAPS you
are really Arnold Kluge, screening potential submitters to Cladistics
for idealogical purity. PERHAPS I'll accidentally eat poorly prepared
Fugu at the Sushi bar for lunch-making this my last post.

> >No, see above. I right now am taking "what other cladists have done"
> >and am working, character by character, to assure myself that they are
> >correct. I disagree with the authors on several major points (so far).
> >At the same time I am reexamining as many specimens as possible to fing
> >*new* characters and assure myself that previous descriptive work was
> >done properly. The result will be a different matrix with different
> >characters (some new, others removed or redefined), different character
> >coding and different operating assumptions, because my views of
> >evolution are different.
>

> In other words, different cladists can come up with totally different opinions
> of which characters are synapomorphs and of course how many synapomorphies
> there are among the same group of organisms. Synapomorphies, in other words,
> are indeed in the eyes of the beholder, as Feduccia claimed.
>

Or which state are applicable, or which characters need clearer
definitions. All of this flows from anatomical study-and there are
better anatomists than others. I trust myself the most, because my
reputation depends on my interpretations. Feduccia is himself under the
same axe. Cladistics is objective in its *analysis* of your anatomical
study-it does not care what tree is desired. Can feduccia make the same
claim?

> > All of the above will be *explicitly*
> >discussed so "the community of biologists as a whole" may draw their own
> >conclusions as to the validity of my (and previous author's) choices.
>

> But nevertheless these are YOUR choices, based on YOUR subjective judgment.
>

MY knowledge. MY work. My understanding.

> Or perhaps your rejection of their data is colored by YOUR desired outcome.

Perhaps. Perhaps we are about to learn something new about development.

> As I recall, you already indoctrinated your 5 year old daughter into saying
> that "birds are dinosaurs." :)

Um, not me. No kids, the dog is more than I can handle.

[Tom DiBenedetto]

Cal King wrote:

>>So for me, synapomorphy refers to homologies; characters on a tree.
>>Before analysis they are putative homologies, and their identity as
>>apomorphy, plesiomorphy or convergence has yet to be determined.

>The same old circular argument. Characters are assumed to be synapomorphs so

>they can be used to construct a tree

No, "cal", rather the same old problem of yours reading the English
language. I stated rather explicitly that synapomorphy is determined
on the tree.

>and then the tree is used to confirm that the characters are synapomorphs.

You still have failed to demonstrate any circuarity. It does take more
than mere assertion you know. Your argument is like saying that it
would be circular to assert (for the purpose of testing) that an
element is gold, then to test that assertion, and to conclude on the
basis of the test that it really is gold. You simply have no
understanding of what circularity is. It is rather obvious that you
are using the charge as a despartate attempt to say something
negative, rahter than to understand.

[Tom DiBenedetto]

Cal King wrote:

>The identification of synapomorphies by Tom the anatomist requires subjective
>judgment. The more knowledge the anatomist applies, the more subjective this
>procedure becomes. Subjectivity is not wrong, but it contradicts the
>cladist's claim of objectivity.

First of all, please use the terminology properly. Tom the anatomist
does not identify synapomorphies. He identifies similarities, which if
they survive testing by the standards of the field of anatomy, are put
forth as hypotheses of homology.
You seem to be linking "knowledge" with subjectivity, as if it were
the case that the more knowledgable an anatomist I become, the more
subjective my work is. This makes it pretty clear to me that you are
not a scientist. You have no concept of what scientific knowledge is
all about. Using the highest standards of anatomical knowledge would
render my work less subjective, not more.
The subjectivity, the "judgement" which the darwinian systematists
used, was a cover for their unwillingness to lay out explicit,
testable (read "scientific") justifications for their decisions. Why
were birds raised to "class" level, and not cetaceans? Why not snakes?
Why were birds raised to "class" level and not some other rank? What
are the principles by which such specific decisions are made? The
answer of course is "judgement", i.e. "shutup kid, dont ask questions.
I know how this should go because I've been doing this for fourty
years". IOW, authoritarianism.

>Who claims that phylogeny reconstruction is an experimental science?

huh? no one. your point?

> As long
>as the results can be falsified, then it is scientific.

Very good "cal". Now tell us how to falsify the opinion of an
authority.

>Cladists have been
>claiming that they falsify earlier systematic studies, thus proving that those
>studies must be scientific.

No we dont. Once again, you make this up so as to keep on with your
flawed argument. We have not falsified earlier studies. They were
inherently unfalsifiable. We have replaced them with falsifiable
(scientific) results.

> OTOH, cladists claim that only a more
>parsimonious cladogram can replace an earlier one.

Thats right "cal". Only a riigorous analysis which yields different
results can falsify a rigorous analysis.

>Thus Burke and Feduccia's
>embryological evidence, even if correct, ostensibly cannot falsify the
>cladist's hypothesis of the bird-theropod link, because Burke and Feduccia
>have not performed a cladistic analysis of their own to produce a more
>parsimonious cladogram.

Exactly. Their evidence, by itself, could falsify the homology
hypothesis of the manus configuration. That is all. It is one
character. If that falsification, when coded properly and considreed
with all the other evidence. causes the overall hypothesis to come out
differntly, then it will have falsified the cladogram, and we will
have learned somehting new. But until this one character is considered
in the context of the rest of the evidence, it is simply a
reinterpretation of one character.

[Cal King]

In article <34ec6692....@news.itd.umich.edu>, td...@umich.edu (Tom
DiBenedetto) wrote...

>
>Cal King wrote:
>
>
>>>So for me, synapomorphy refers to homologies; characters on a tree.
>>>Before analysis they are putative homologies, and their identity as
>>>apomorphy, plesiomorphy or convergence has yet to be determined.
>
>>The same old circular argument. Characters are assumed to be synapomorphs
so
>
>>they can be used to construct a tree
>
>No, "cal", rather the same old problem of yours reading the English
>language. I stated rather explicitly that synapomorphy is determined
>on the tree.

How do we get the tree, by using the "synapomorphs" to either "recover" or
"construct" the tree. Different authors use different terms on how they get
the tree.

>>and then the tree is used to confirm that the characters are synapomorphs.
>
>You still have failed to demonstrate any circuarity. It does take more
>than mere assertion you know. Your argument is like saying that it
>would be circular to assert (for the purpose of testing) that an
>element is gold, then to test that assertion, and to conclude on the
>basis of the test that it really is gold.

It would be a circular argument if the presumed gold element being tested is
tested using the presumed gold. That is exactly what cladists do. They
assume homology for the characters with which they construct the tree, and
then use the tree so constructed as evidence that the characters are
homologies.

>You simply have no
>understanding of what circularity is. It is rather obvious that you
>are using the charge as a despartate attempt to say something
>negative, rahter than to understand.
>
>Tom DiBenedetto td...@umich.edu

I understand that even cladists are (desperately) distancing themselves from
Cracraft's hypotheses that loons are closely related to grebes and that owls
are closely related to hawks, even though Cracraft followed cladistic
procedures when he used presumed synapomorphs to construct his trees and is
sticking by the claim that those characters he used are in fact
synapomorphies. Cracraft's analyses is reason why many systematists are
turning to more reliable developmental evidence to test hypotheses of
homology.

[Tom DiBenedetto]

Cal King wrote:

>. Your argument is like saying that it
>>would be circular to assert (for the purpose of testing) that an
>>element is gold, then to test that assertion, and to conclude on the
>>basis of the test that it really is gold.

>It would be a circular argument if the presumed gold element being tested is
>tested using the presumed gold.

huh? I dont follow,
A homology is a similarity due to descent. If I claim that two
characters are homologous, and then I deduce, from the principle of
descent, a qualtiy which these characters should have if they truly
are homologous (congruence), and then I test to see if that quality is
present, and then draw a conclusion from the result,,,there is no
circularity.

>That is exactly what cladists do. They
>assume homology for the characters with which they construct the tree, and
>then use the tree so constructed as evidence that the characters are
>homologies.

Yeah, but you are missing the whole point of testing fro congruence.
And another thing. Trees are not "constructed". Trees are "selected".
There is a finite number of possible topologies for any given number
of taxa, and if feasable, every one is examined to see how well it can
explain the data. The one which does best is chosen. I dont know where
you get the idea that trees are "constructed".

[Cal King]

In article <34EC6779.3DDC@no_spam.mcgill.ca>, jasona@no_spam.mcgill.ca (Jason
Anderson) wrote...

[Cal King]:

>> As I recall, you already indoctrinated your 5 year old daughter into saying
>> that "birds are dinosaurs." :)

[Jason Anderson}

>Um, not me. No kids, the dog is more than I can handle.

Here is something I dug up using Deja News:
======================================
In article <34904135.340B@no_spam.mcgill.ca>, jasona@no_spam.mcgill.ca (Jason
Anderson) wrote...

> even children are now "getting" cladistics.
>I challenge you to ask any dino-mad 5 year old where birds come from:
>"Birds are Dinosaurs!"
=======================================

Are there 2 Jason Andersons in McGill University? One of them has a 5 year
old who is mad about dinosaurs and says that "Bird are dinosaurs!" and the
other Jason Anderson who has no child.

>Cal King wrote:

>> Perhaps the cladists have managed to become editors of many scientific
>> journals and are thus in a position to reject non-cladistic papers.
Whateve
>r
>> the reason, if a systematist cannot get his work published unless he uses
>> cladistics, then sooner or later, the reality of needing to publish will
for
>ce
>> most to become cladists. This is a rather unfortunate trend because
fashion
>,
>> not scientific rigor, becomes the driving force in scientific research.
>>
>PERHAPS little green men from mars run the United Nations.

Perhaps your dino-mad daughter has been reviewing scientific papers and been
rejecting all non-cladistic papers submitted for publication.

>> > All of the above will be *explicitly*
>> >discussed so "the community of biologists as a whole" may draw their own
>> >conclusions as to the validity of my (and previous author's) choices.
>>
>> But nevertheless these are YOUR choices, based on YOUR subjective judgment.
>>
>MY knowledge. MY work. My understanding.

Yes! Since they are yours, they are also subjective. There is no objective
means for determining synapomorphies in cladistic analysis. That is what
evolutionary systematists have been saying all along. Now we finally have
cladists admitting that their method is subjective, but only after my
criticism of the unreliability of their procedures for ascertaining
synapomorphies.

>> Or perhaps your rejection of their data is colored by YOUR desired outcome.

>Perhaps. Perhaps we are about to learn something new about development.

Strong hope imposes upon weak evidence, as Gould once said. The hope that the
bird manus is homologous with the theropod manus is much stronger than the
evidence.

[Tom DiBenedetto]

Cal King wrote:

>According to R. Hinchliffe (1997, Science 278:597), the
>use of embryological evidence "...represents a different methodology in
>ascertaining homology from that adopted by many paleontologists, who use
>multiple synapomorphies."

This is a foolish statement. It lays out a false dichotomy between
synapomorphies and embryology. Embryological studies are hardly
something new, they have been used by systematists (including
cladists) for a very long time.

>The bird manus evidence shows that cladists, because of the constraints of
>their methodology have a great deal of difficulty distinguishing
>synapomorphies from homoplasies.

Once again, there is nothing in cladistics which compels one to
identify digits in a certain manner, nor is there anything in
darwinian systematics that compels one to identify digits in a
differet manner.
This is getting really boring.

>>No one has claimed that you need to be a cladist in order to be a
>>comparative anatomist.

>But cladism discourages the comparative anatomist from interjecting judgment
>into his analysis.

False. Anatomical judgement is used all the time,,,for anatomical
issues, not for systematic issues.

> Problematic characters, for example, cannot be removed
>from the analysis without introducing subjectivity.

False. Characters which apear in a cladistic matrix are tested
hypotheses of homology. They would not appear there if they were so
problematical that the systematists did not have confidence in
proposing the hyptothesis of homology.

>> Homology is hypothesized after long and extensive anatomical study.

>Show me evidence of such studies.

Read the damn literature. You are the fool who is making baseless
charges. Go to the library and read any cladistic study.

>If the assumption of homology can be falsified, then the tree that is
>constructed from these characters would then be falsified.

It depends on all the evidence.

>Not all consensus cladograms show polytomies.

huh? Do you even know what a consensus tree is?

> Some cladists reject polytomies altogether.

huh? Do you even kinow what the different types of polytoomies are?

> Hence the test of congruence for homologies do not exist.

Brilliant logic. I think "Cal King" doesnt exist.

> They
>don't exist because conflicting cladograms are melted into consensus
>cladograms, instead of being used to falsify homology.

Huh? I think the time has come for me to do somehting a bit more down
to earth. I will leave you argue with the little green men who run the
United Nations. :)

>>And once again, on what grounds can you claim that the cladists didnt
>>study the manus as much as anyone else?
>
>What evidence do you have that they did? Feduccia said Gauthier didn't.
>That is my evidence. So what is your evidence?

And there you have it,,,,"cal" the scientist. "My proof is that
someone said that someone else did......".

[Cal King]

In article <34ee6fce....@news.itd.umich.edu>, td...@umich.edu (Tom
DiBenedetto) wrote...
>

>Cal King wrote:
>
>>. Your argument is like saying that it
>>>would be circular to assert (for the purpose of testing) that an
>>>element is gold, then to test that assertion, and to conclude on the
>>>basis of the test that it really is gold.
>
>>It would be a circular argument if the presumed gold element being tested is
>>tested using the presumed gold.
>
>huh? I dont follow,
>A homology is a similarity due to descent. If I claim that two
>characters are homologous, and then I deduce, from the principle of
>descent, a qualtiy which these characters should have if they truly
>are homologous (congruence), and then I test to see if that quality is
>present, and then draw a conclusion from the result,,,there is no
>circularity.

"I do not expect to convince many cladists."--A. Cronquist (1987 A botanical
critique of cladism. Bot. Rev. 53: 1-52)

[Zen Faulkes]

Hello,

Cal King wrote:

> Cracraft's analyses is reason why many systematists are
> turning to more reliable developmental evidence to test hypotheses of
> homology.

There is no strong reason to assume that any particular class of data
is more reliable for testing hypotheses of homology than any other
dataset.

George V. Lauder has written the best critique of this argument,
calling it the search for the "locus of homology." I *think* the
reference is this one: Lauder, G. V. 1986. Homology, analogy, and the
evolution of behavior. Chapter 1, pp. 9-40 In The Evolution of Behavior,
M. Nitecki and J. Kitchell, Eds. Oxford University Press. If this isn't
the right book chapter, let me know and I'll dig out the right one.

Zen Faulkes
Biology, McGill University

[Cal King]

In article <6ci0i3$c...@sifon.cc.mcgill.ca>,
Zen Faulkes <zfaulkes@*NOSPAM*bio1.lan.mcgill.ca> wrote:

He reached his conclusion prematurely. A decade later, as Hinchliffe
(1997) stated, systematists are indeed turning increasingly to embryology for
evidence of homology. It sure beats the cladistic way of using a tree to test
for homology, since the tree is itself constructed with the assumption of
homology, using the same characters that are supposed to be tested.

[Cal King]

In article <34ed7453....@news.itd.umich.edu>, Tom DiBenedetto
(td...@umich.edu) says...

>
>Cal King wrote:
>
>>According to R. Hinchliffe (1997, Science 278:597), the
>>use of embryological evidence "...represents a different methodology in
>>ascertaining homology from that adopted by many paleontologists, who use
>>multiple synapomorphies."
>

>This is a foolish statement. It lays out a false dichotomy between
>synapomorphies and embryology. Embryological studies are hardly
>something new, they have been used by systematists (including
>cladists) for a very long time.

Yes, even Hennig thinks that it may provide clues concerning character
polarity. His followers, however, have instead chosen to use outgroup
comparisons to ascertain polarity almost exclusively.

>>The bird manus evidence shows that cladists, because of the constraints of
>>their methodology have a great deal of difficulty distinguishing
>>synapomorphies from homoplasies.

>Once again, there is nothing in cladistics which compels one to

>identify digits in a certain manner, nor is there anything in
>darwinian systematics that compels one to identify digits in a
>differet manner.

There isn't. In fact, it was Ostrom (who is not a cladist) who, according to
Hecht, used overall similarity to group brids and theropods. The cladists,
because they are uncritical about homologies, and who in fact rely on the
results of their analysis to assess homology, therefore accepted Ostrom's
opinion as evidence of homology in the similarities between birds and
theropods. The paper I cited by Channing also shows that cladists can take
phenetic data and assume uncritically that these characters are in fact
homologies without further investigation.

>This is getting really boring.

You can break the monotony by giving us evidence that Gauthier actually
analyzed his characters. So far you have argued by purely using unsupported
assertions.

>>>No one has claimed that you need to be a cladist in order to be a
>>>comparative anatomist.
>
>>But cladism discourages the comparative anatomist from interjecting judgment
>>into his analysis.
>

>False. Anatomical judgement is used all the time,,,for anatomical
>issues, not for systematic issues.

Again, more unsupported assertions.

>> Problematic characters, for example, cannot be removed
>>from the analysis without introducing subjectivity.
>

>False. Characters which apear in a cladistic matrix are tested
>hypotheses of homology.

More unsupported assertions. If they are tested, how come many studies end up
with an average C.I. of 0.5. Some even had a dismal C.I. of below 0.3.

>They would not appear there if they were so
>problematical that the systematists did not have confidence in
>proposing the hyptothesis of homology.

We have you admitting here that problematic characters are rejected. The
criterion for such rejection is of course subjective. Hence the evolutionary
systematists were right that cladistics is very much a subjective method.

>
>>> Homology is hypothesized after long and extensive anatomical study.
>
>>Show me evidence of such studies.
>

>Read the damn literature. You are the fool who is making baseless
>charges. Go to the library and read any cladistic study.

The library has a million books. Which one should I start with? War and
Peace or Pride and Prejudice?

>>If the assumption of homology can be falsified, then the tree that is
>>constructed from these characters would then be falsified.

>It depends on all the evidence.

Yes, the total evidence approach, in which homoplasies can outweigh
synapomorphies.

>>Not all consensus cladograms show polytomies.

>huh? Do you even know what a consensus tree is?

Yes. Are you still claiming that all consensus cladograms show polytomies?

>> Some cladists reject polytomies altogether.

>huh? Do you even kinow what the different types of polytoomies are?

What are those types?

>Huh? I think the time has come for me to do somehting a bit more down
>to earth.

Like making a lot of unsupported assertions without giving any reference and
then disappear. Yes, you have done that before.

[Tom DiBenedetto]

Cal King wrote:

> A decade later, as Hinchliffe
>(1997) stated, systematists are indeed turning increasingly to embryology for
>evidence of homology. It sure beats the cladistic way of using a tree to test
>for homology, since the tree is itself constructed with the assumption of
>homology, using the same characters that are supposed to be tested.

yo "cal", you are spiraling off into the ether again. Lets face it,
your charge of circularity has fallen flat, and your reference to
embryology is equally absurd. Cladists use embryological characters
just like adult characters. They are coded as homology hypotheses and
subject to cladistic analysis. They can be synapomorphies,
plesiomorphies, or convergences like any other character. There is an
extensive literature, written by cladists, which deals with these
types of characters.
Once again, why dont you get off this childish game of trying to
disparage people whose work you dont understand, and be serious for a
change. Try to find out what really goes on in the scientific world,
so that these discussions can be a bit more intellegent and
interesting.

[Tom DiBenedetto]

Cal King wrote:

>>A homology is a similarity due to descent. If I claim that two
>>characters are homologous, and then I deduce, from the principle of
>>descent, a qualtiy which these characters should have if they truly
>>are homologous (congruence), and then I test to see if that quality is
>>present, and then draw a conclusion from the result,,,there is no
>>circularity.

>"I do not expect to convince many cladists."--A. Cronquist (1987 A botanical
>critique of cladism. Bot. Rev. 53: 1-52)

well thanks "cal",,I realize that this is about as close as you are
probably emotionally capable of coming to an admission that your
arguements are groundless.

[Cal King]

In article <34EC6779.3DDC@no_spam.mcgill.ca>, Jason Anderson
<jasona@no_spam.mcgill.ca> wrote:
>Cal King wrote:

[Another try...]

>> As I recall, you already indoctrinated your 5 year old daughter into saying
>> that "birds are dinosaurs." :)
>

>Um, not me. No kids, the dog is more than I can handle.

My mistake, I apologize to Jason. I misread your earlier post. :(

>> Perhaps the cladists have managed to become editors of many scientific
>> journals and are thus in a position to reject non-cladistic papers.
Whatever
>> the reason, if a systematist cannot get his work published unless he uses
>> cladistics, then sooner or later, the reality of needing to publish will
> force
>> most to become cladists. This is a rather unfortunate trend because
fashion,
>> not scientific rigor, becomes the driving force in scientific research.
>>
>PERHAPS little green men from mars run the United Nations.

That is not my claim. You are free to think that if you want.

> PERHAPS you
>are really Arnold Kluge, screening potential submitters to Cladistics
>for idealogical purity.

Does he do that? I have no idea.

> PERHAPS I'll accidentally eat poorly prepared
>Fugu at the Sushi bar for lunch-making this my last post.

I hope not. I wish you well.

BTW, since the peer review process is anonymous in many journals, the
rejection of articles for publication on ideological, or even purely personal,
grounds is not inconceivable.

>> But nevertheless these are YOUR choices, based on YOUR subjective judgment.
>>
>MY knowledge. MY work. My understanding.

Your subjective knowledge and your subjective understanding. No problem with
that. It also proves what cladistics is not objective, despite the cladists'
claim.

>> Or perhaps your rejection of their data is colored by YOUR desired outcome.
>
>Perhaps. Perhaps we are about to learn something new about development.

Actually that is what the cladists HOPE, that there is no serial homology
between bird hand and bird feet. They are welcomed to perform their own
studies to try to find evidence that this is indeed the case. Right now, they
don't have evidence, just strong hope, that bird digits are homologous with
theropod digits.

[Cal King]

In article <34ec9980....@news.itd.umich.edu>, Tom DiBenedetto
(td...@umich.edu) says...
>

>Cal King wrote:
>
>> A decade later, as Hinchliffe
>>(1997) stated, systematists are indeed turning increasingly to embryology
for
>>evidence of homology. It sure beats the cladistic way of using a tree to
test
>>for homology, since the tree is itself constructed with the assumption of
>>homology, using the same characters that are supposed to be tested.
>
>yo "cal", you are spiraling off into the ether again. Lets face it,
>your charge of circularity has fallen flat,

Again, more unsupported assertions. You are entitled to your opinion, but at
least try to give us some reasons why.

>and your reference to
>embryology is equally absurd. Cladists use embryological characters
>just like adult characters.

That is irrelevant. We were discussing the use of embryological evidence to
assess character homology. Their use in phylogenetic analysis is not the
issue. So cladists use embryological characters, so what? Hinchliffe, Burke
and Feduccia are using embryological evidence to assess bird and theropod
digit homology. Please don't confuse character analysis with phylogenetic
analysis.

[Cal King]

In article <34ed9aea....@news.itd.umich.edu>, Tom DiBenedetto
(td...@umich.edu) says...

>>"I do not expect to convince many cladists."--A. Cronquist (1987 A botanical

>>critique of cladism. Bot. Rev. 53: 1-52)
>
>well thanks "cal",,I realize that this is about as close as you are
>probably emotionally capable of coming to an admission that your
>arguements are groundless.

No, Cronquist means that many cladists do not have open minds.

[Tom DiBenedetto]

Cal King wrote:

>> Embryological studies are hardly
>>something new, they have been used by systematists (including
>>cladists) for a very long time.

>Yes, even Hennig thinks that it may provide clues concerning character
>polarity. His followers, however, have instead chosen to use outgroup
>comparisons to ascertain polarity almost exclusively.

That is another issue (and, btw, although many cladists use outgroups,
others have continued the investigatio of ontogentic polarity
criterea). The issue is using embryological characters as characters.
Cladists have *never* resisted doing that, and have, in fact advocated
it.

>>Once again, there is nothing in cladistics which compels one to
>>identify digits in a certain manner, nor is there anything in
>>darwinian systematics that compels one to identify digits in a
>>differet manner.

>There isn't.

well good, so lets put this to sleep finally. The manus dispute is a
dispute about anatomy.

>You can break the monotony by giving us evidence that Gauthier actually
>analyzed his characters. So far you have argued by purely using unsupported
>assertions.

Read Gauthier's work "cal". I am not going to dig it out myself and
type it into my newreader for you. You are the one making the charge.
What kind of nonsense is this that you think you can make such charges
and expect everyone else to run around proving you wrong?

>> Characters which apear in a cladistic matrix are tested
>>hypotheses of homology.

>More unsupported assertions. If they are tested, how come many studies end up
>with an average C.I. of 0.5. Some even had a dismal C.I. of below 0.3.

huh? A C.I. which is low is explicit proof that there has been
testing. The value represents the number of times homology hypotheses
have been falsified. sheeesh

>>They would not appear there if they were so
>>problematical that the systematists did not have confidence in
>>proposing the hyptothesis of homology.

>We have you admitting here that problematic characters are rejected.

They never make it to the matrix. Is this a revelation to you? Are you
so unfamiliar with how systematics is done?

>The criterion for such rejection is of course subjective. Hence the evolutionary
>systematists were right that cladistics is very much a subjective method.

"cal", you constantly confuse systmetics with anatomy, and you use the
word "subjective" in a very loose sense. Judgements are made in
anatomical issues. THat does not mean they are subjective, since they
revolve around explicit standards in the field of anatomy, and are
open to argument with other anatomists. And all this is completely
separate from systematic issues.

>>>> Homology is hypothesized after long and extensive anatomical study.

>>>Show me evidence of such studies.
>>

>>Read the damn literature. You are the fool who is making baseless
>>charges. Go to the library and read any cladistic study.
>
>The library has a million books. Which one should I start with? War and
>Peace or Pride and Prejudice?

No, cal, go to the biology section. Do a search on (e.g.) Gauthier.
Then read all his work. Yourself. Then you might have an impression of
how much he has studied anatomy. Or even better, give him a call. I
think he is still at the calacad,,,just up the road. Ask him how many
critters he has looked at. Y'know,,,,we scientist-types do things like
that,.,,,,talk to eachother,,,find out what is really going on.

>> Do you even know what a consensus tree is?
>
>Yes. Are you still claiming that all consensus cladograms show polytomies?

yes. The ones I and most cladists use.. So what do you think a
consensus is?

>>> Some cladists reject polytomies altogether.

>>huh? Do you even kinow what the different types of polytoomies are?

>What are those types?

very good "cal". An admission that you dont know what you are talking
about. I am impressed with your ethical progress. But instead of
asking me (since it is clear that you will argue with anything I say),
why dont you read a little and find out yourself?

[Tom DiBenedetto]

Cal King wrote:

> We were discussing the use of embryological evidence to
>assess character homology.

So what? What makes you think that this is anything new in
systematics, or amongst cladists?

[Cal King]

In article <34ec9ebf....@news.itd.umich.edu>, td...@umich.edu (Tom
DiBenedetto) wrote...

>The issue is using embryological characters as characters.

An emphatic no. The issue is the use of embryological evidence to ascertain
homology.

>Cladists have *never* resisted doing that, and have, in fact advocated
>it.

Cladists are definitely resisting Burke and Feduccia's use of embryological
evidence to ascertain bird-theropod manual digit homology. The cladists don't
like the findings because it shows that those cladists who had assumed
homology in this character were PAUPably wrong.

>well good, so lets put this to sleep finally. The manus dispute is a
>dispute about anatomy.

It is a dispute about character homology and thus the soundness of
phylogenetic hypotheses based on homoplastic characters. The anatommical
evidence provides evidence that the cladist's synapomorphies were in fact not
synapomorphs.

>>You can break the monotony by giving us evidence that Gauthier actually
>>analyzed his characters. So far you have argued by purely using unsupported
>>assertions.
>
>Read Gauthier's work "cal". I am not going to dig it out myself

Then on what basis did you claim that he "obviously" analyzed the characters?
Feduccia (1996) said he didn't. You argue that Gauthier did but you have no
evidence to back up your claim.

>What kind of nonsense is this that you think you can make such charges

I didn't make the charges. Feduccia did.

>and expect everyone else to run around proving you wrong?

You asserted that Feduccia was wrong, but provided no evidence to support your
assertion.

>huh? A C.I. which is low is explicit proof that there has been
>testing.

It is an explicit proof that characters were chosen indiscriminantly and that
there was little, if any, a priori character analysis.

> The value represents the number of times homology hypotheses
>have been falsified. sheeesh

Thus a low value shows that there are a large number of homoplasies, which
imply that the a priori character analysis, if there was any, was shoddy at
best.

>>We have you admitting here that problematic characters are rejected.

>They never make it to the matrix. Is this a revelation to you?

Not to me. Evolutionary systematists have always pointed out that in any
analysis, the number of characters is only a subset of those present, and
hence there must be choice of characters. Such choice is of course
subjective. In fact, they argue that cladists choose characters
subjectively and assign polarity subjectivelly. Cladists also choose
characters that are easy to dichotomize.

>"cal", you constantly confuse systmetics with anatomy,

Actually you are trying to confuse systematics with anatomy.

>>>huh? Do you even kinow what the different types of polytoomies are?
>
>>What are those types?
>
>very good "cal". An admission that you dont know

I am just asking you to list those types since you said there are different
types.

>But instead of
>asking me (since it is clear that you will argue with anything I say),
>why dont you read a little and find out yourself?

Read what, your mind? Since you again gave me no reference concerning the
various types of polytomies, only the assertion that there are different
types.

[Peter Nyikos]

td...@umich.edu (Tom DiBenedetto) writes:

>Peter Nyikos wrote:

>> A matrix is a set of homology hypotheses which emerge from
>>>detailed study of the organism. It is only if a percieved similarity
>>>survives anatomical study (as the similarity between octopus and
>>>vertebrate eyes do NOT)

>>In other words, you rely on the subjective judgment of anatomists.
>>Funny, I thought you and Mike Noren were the ones who claimed
>>cladistics was fully objective.

>huh?

Or was it just Mike Noren? Anyway, I even said at one point
that Noren's view of "objectivity" seemed to be such that
eventually all his ideal systematists could be replaced by computers.
Either you or Mike came back and said that is exactly what is
good about his ideal of systematics.

>>Can ANY amount of anatomical insight falsify the most parsimonious
>>tree in your eyes?

>After all these years I still cant figure you out Peter. Is this
>serious?

Of course. Don't you recall our exchange about the pollex
disappearing and re-appearing several times in a sequence
of successive nodes?

Don't you realize that you did NOT explicitly answer my
question below?

I guess all I can do is ignore my suspicions and forge ahead.

>To put it simply. Morphological cladists must be expert anatomists. We

>study the anatomy according to the standards of the field of anatomy.
>Cladistics per se is a sytematic method. It does not tell me how to do
>anatomy. Cladistic principles tell me how to deal with characters once
>they are defined. They are defined by Tom the anatomist, not Tom the
>systematist.

But they are subject to revision in the light of cladistic
analysis, and the bottom line is, the last word belongs to
the most parsimonious tree. There is no court of higher
appeal if something as bizarre as a pollex or even
the vertebrate eye disappearing and reappearing is to be
found in the tree, is there? The best you can do is keep
running matrices with more and more organisms and more and
more characters, in the hope that a detailed enough cladistic
analysis will vindicate your anatomical instincts.

The "objectivity" of cladistics has to do with systematic

>procedures. Character defintions are presented, they are made clear.

>Any weighting which is done is done so explicitly, and must be
>defended.

Weighting has nothing to do with considering some apparent characters
to be more important than others, does it?

[I say "apparent characters" because "weighting" has just to
do with hypothesizing that two features in two different animals
are either "the same character" or "not the same character", doesn't it?]

The matrix is presented, as well as the algorithm
>identified. In a properly done analysis, any reader of the paper
>should be able to retrace the authors steps and arrive at the same
>conclusion, or have every arguable issue laid out, and the authors
>reasons for taking certain positions identifiable. This is in contrast
>to the "evolutioanry systematists" who use subjective "judgement",
>somethiong which does not allow for an "objective" resolution.

Yes, your anatomical insights are "objectively" either
confirmed or falsified by the most parsimonious tree, eh?

[Peter Nyikos]

michae...@nrm.se (Mike Noren) writes:

>Thusly nyi...@math.scarolina.edu (Peter Nyikos) spake unto all:

>(snip)
>: In other words, you rely on the subjective judgment of anatomists.

>: Funny, I thought you and Mike Noren were the ones who claimed
>: cladistics was fully objective.

>You wouldn't mind supplying a quote to back your claim that I've

>stated that "cladistics" was "fully objective", would you? You see, as
>is too often the case when you paraphrase my opinions or claims, I
>don't recognise them as mine.

See my reply to Tom. Did it refresh your memory?

Peter Nyikos -- standard disclaimer --
Professor, Dept. of Mathematics
University of South Carolina
Columbia, SC 29208

[Zen Faulkes]

Hello,

Before I get to what Cal King wrote, I should point out that the
chapter I mentioned in the first post on this thread wasn't the one I
was thinking of, although it deals with some of the same issues. The
chapter I was thinking of is in _Homology_ (Ed. B.K. Hall, Academic
Press, 1994).

> We were discussing the use of embryological evidence to
> assess character homology.

Certainly, embryological evidence is very good and interesting. But
it does not have any _a priori_ status as a better data set for
evaluating hypotheses of homology.

> Their use in phylogenetic analysis is not
> the issue. So cladists use embryological characters, so what?
> Hinchliffe, Burke and Feduccia are using embryological evidence to
> assess bird and theropod digit homology. Please don't confuse
> character analysis with phylogenetic analysis.

So the claim is that an hypothesis of homology can be assessed
independently of a phylogenetic hypothesis? As the modal definition of
homology is some variation on "a character inherited from a common
ancestor," this would seem very difficult indeed. Identifying a
structure as a homologue in two taxa implies something about their
phylogeny; i.e., they shared an ancestor. Unless, that is, one uses
the pre-Darwinian definition of homology (a small minority do).

[Peter Nyikos]

Jason Anderson <jasona@no_spam.mcgill.ca> writes:

>Peter Nyikos wrote:
>>
>> [posted and mailed, with request for a CC of any reply]

>(BTW, replace spam trap with bio1.lan to reply)

Thanks, I merely deleted it and the message bounced.

I'm doing an encore because I allude to two upcoming changes of
thread title.

>> Jason Anderson <jasona@no_spam.mcgill.ca> writes:
>>
>> >Cal King wrote:
>>
>> [Tom DiB?]

>(Naw, me)

>> >> >Homoplasies are not identified *except* in light of a tree.
>>
>> >> That is a common claim among cladists, but it is demonstrably false.
><snip>
>> >See Tom DiBennidito's reply.

><snip>

>> >> Another
>> >> example is the giant panda's opposable "thumb." It is not a real digit at
>> >> all, but is formed instead from the radial sesamoid. Hence it is not a
>> >> homolog of the vertebrate thumb at all.
>>
>> >Correct. Nobody would score it as "thumb (+)" either, except your
>> >strawman cladist.
>>
>> That's only because a careful anatomist could tell it is NOT a pollex.
>> In the example I gave Tom about a year ago, I very carefully
>> used the word "pollex" to remove all ambiguity about something
>> an expert anatomist might identify as such.
>>
>> I said that I would take a cladogram that has the pollex
>> disappearing and re-appearing at successive nodes several times
>> as wrong even if it was the most parsimonious tree; and that
>> I would regard that feature as falsifying not only the tree
>> itself but also the very principle of maximum parsimony.
>>
>You are free to do so if you wish. Why would this character be so
>crucial as to "falsify...the tree [and] the principle of maximum
>parsimony" anyway?

If it isn't crucial enough, try "the vertebrate eye" in place
of "the pollex". And if that isn't crucial enough, is there
ANY anatomical or paleontological evidence that could falsify
these things in your eyes?

Podial elements are notorious for disappearing and
>reappearing evolutionarily.

Including the pollex? I have yet to see any example of it
reappearing after disappearing.

Many secondarily aquatic tetrapods cease
>ossifying podials, or ossify them at increasingly delayed times of
>onset.

I was careful to say "the pollex", which assumes that any
other digits would be ossified. So your statement is relevant
only if the solitary
digit is the pollex (do you know of any tetrapod for which
this is true?

Others, notably ichthyosaurs, mosasaurs and cetaceans *increase*
>ossification to lock the elements into rigid paddles (while adding
>phalanges at a startling rate).

Phalanges, yes, but it has been claimed that one can make out
the digits even in ichthyosaur paddles which at first look
like a jumbled mass of phalanges in reproductions. Only in
ichthyosaurs is it true that actual *digits* are added.

If you can wait a bit I can check to
>see if any lineage first delays podial ossification (say _Hupesuchus_)
>while later increasing ossification.

I certainly can wait, but if the digits are treated in the
same way then you are missing my point.

>Mostly I get the feeling that you strongly subscribe to Dollo's
>principle. Is this correct?

Sorry, I've forgotten what Dollo's principle is.

[deletia to be replied to on "The Decline and Fall (?) of Comparative Anatomy"]

[deletia to be replied to on "Homology and Synapomorphy"]

>> >> One of the ways to determine homology a
>> >> priori is comparative anatomy.

>> >See above, and earlier posts. Comparative anatomy is what most time is
>> >spent on-before a matrix is built. One establishes *synapomorphies* in
>> >this manner-using a rigorous series of tests.

>> In other words, one looks through the literature to see what
>> other cladists have done, and takes their *conclusions* about
>> homology as starting points for what should be synapomorphic, no?

>No, see above.

I did, but your opening line read,

"Homoplasies are not identified *except* in light of a tree."

I right now am taking "what other cladists have done"
>and am working, character by character, to assure myself that they are
>correct. I disagree with the authors on several major points (so far).
>At the same time I am reexamining as many specimens as possible to fing
>*new* characters and assure myself that previous descriptive work was
>done properly. The result will be a different matrix with different
>characters (some new, others removed or redefined), different character
>coding and different operating assumptions, because my views of
>evolution are different.

Thanks for the clarification.

All of the above will be *explicitly*
>discussed so "the community of biologists as a whole" may draw their own
>conclusions as to the validity of my (and previous author's) choices.

>Mind you, all of this stems from a revision, with *detailed* anatomic
>description and alpha taxonomy, of a previously poorly known group.
>Cladistic analysis is only one facet of a much larger study. Other
>areas interest me more (generating life-history data from fossils, for
>one).

If I had decided to go on majoring in zoology, I think I'd be
doing this kind of work too.

>> The final test, though,
>> >is the tree-it tells *homology* from false homology-in light of all
>> >other homologies. Things like the cephalopod eye would not be included
>> >in the analysis because it is not a synapomorphy with the vertebrate
>> >eye-it fails the test (a couple, truth be told) of similarity.
>>
>> Similarity is a relative concept. You aren't seriously claiming
>> that cladists will score any two things that show the slightest
>> dissimilarity as different *a priori* do you? That wouldn't even
>> allow for individual variation within a litter of puppies.
>>
>Octopod eyes are structurally and developmentally different.
>Intraspecific variation is accounted for in the alpha taxonomy.

Ah, but my point is, where does one draw the line?

[item on bird and theropod digits deleted. Will pursue it with
others who are interested in pursuing it further]

>
>> [...]
>>
>> >> Hence it is better to perform some character analysis to
>> >> determine character goodness and to discard characters that are suspect prior
>> >> to systematic analysis than to assume a priori that all characters chosen are
>> >> synapomorphies, then plunge into an analysis that assumes all characters are
>> >> of equal importance in the evolution of a group of animals. Better character
>> >> analysis makes for better systematic analysis. In other words, quality is
>> >> more important than quantity.
>>
>> >I agree :o. But what you seem to want to refuse to acknowledge is that
>> >characters ARE analysed before included in a matrix.
>>
>> +++++++++++++++++++++++ dogmatic cladistic posting mode on
>>
>> Either the analysis took the form of earlier cladistic use
>> of trees or else it was subjective.
>>
>> +++++++++++++++++++++++ dogmatic cladistic posting mode off
>>
>????????????
>You are posting "dogmatic cladistic" nonsense, I'm afraid.

I was just describing how Mike Noren came across to me
in past posts, although he never put it that way. See my
reply to Tom DiB on why Mike came across to me that way.

[Peter Nyikos]

This is my second followup to a post that appeared on
"Methods and characters (Re: Quetzalcoatlus northropi!!!)"

Jason Anderson <jasona@no_spam.mcgill.ca> writes:

>Peter Nyikos wrote:

>> Jason Anderson <jasona@no_spam.mcgill.ca> writes:
>>
>> >Cal King wrote:

[about the fact that the panda's "thumb" is not really a pollex:]

>> >> This fact was revealed again by
>> >> comparative anatomy, long before systematists started arguing whether the
>> >> panda is a bear or a raccoon.

>> >FYI, cladists are among the few remaining comparative anatomists left in
>> >biology.

>> How do you know?

>Well, comparative anatomy as a taught subject, believe it or not, is
>actually in many institutions being weeded out of the undergraduate
>cirriculum. Nobody wants (or can) teach it because the focus (and
>money) has moved to molecular biology.

*sigh* I'm reminded of something a sage old professor of mine
once wrote, "Things will get worse before they get better,
because we have sold ourselves to the devil of external funding."

>Comparative anatomy is still being used by functional morphologists,
>systematists and paleontologists.

Sad, really. I remember in my undergraduate days that the
very next course a zoology major was to take after General Zoology
was Comparative Vertebrate Anatomy. I thought that was a very
wise choice and still do.

Now, with molecular biology ruling the roost, I suppose the
very concept of a "zoologist" will become radically shrunk;
what molecular biologist wants to be pegged as someone who
isn't interested in the molecular biology of plants and fungi
and monerans?

Of these groups, systematists and
>paleontologists are most concerned with finding evolutionary
>relationships-and cladistics is the dominant paradigm for doing so.
>Functional morphologists are also using phylogenies to reconstruct the
>evolution of structures and functional units-these again are based upon
>cladistic work predominantly. Hold outs for evolutionary systematics or
>other techniques are few and becoming fewer every year.

Does the "techniques" part apply to the use of ML or NJ to
molecular biological reconstruction of phylogeny?

Evolutionary systematics is now identified not by whether
it uses cladistic techniques, but whether it allows paraphyletic
taxa into its bookkeeping system. And before long, it will
be probably be identified with those who want to hang on to the Linnean
system in whatever form.

[Peter Nyikos]

This is a followup to a post that appeared on

"Methods and characters (Re: Quetzalcoatlus northropi!!!)"

Jason Anderson <jasona@no_spam.mcgill.ca> writes:

>Peter Nyikos wrote:

>> Jason Anderson <jasona@no_spam.mcgill.ca> writes:
>>
>> >Cal King wrote:

[Deletia of things already replied to on the original thread and on
"The Decline and Fall (?) of Comparative Anatomy".

>> >> Using presumed synapomorphies to construct a tree and then using the tree to
>> >> identify synapomorphies amounts to a circular argument.

>> >Again, one uses *synapomorphies* to construct the tree. The tree
>> >interprets *homology* from *homoplaisy*.

>> Synapomorphy *is* homology, isn't it?

>No. Synapomorphy is a "shared derived trait". Homology is "same organ,
>regardless of form or function (Owen)".

Well, that seems to clash with the apparent consensus on
sci.bio.evolution last year on what homology is, because
it was agreed, for example, that bat wings are homologous
to bird wings *as forelimbs* but not *as wings*, and it
would seem that what distinguishes wings from mere forelimbs
has partly to do with function. Yes, ostrich wings don't
function that way, but they ARE synapomorphic with wings
that are used for flight, hence they are wings and homologous
*as wings* to the wings of eagles.

>One examines the animals under study. One notices a character, which
>seems to be present in one group but different in another group of study
>animals (or plants, etc...). One must now apply a series of tests to
>establish homology. The first set is collectively referred to as "the
>test of similarity". Is it in a similar place? Does it relate with
>similar other organs/bones/etc. Does it scale geonetrically? Does it
>form from similar embryologic tissues?

What is the function of that word "similar"? Besides the
bit about bird and theropod digits that you don't want to
get further into, there is the long-hot topic on talk.origins
whether an amniotic membrane that forms by delamination from
the cytotrophoblast is homologous to one that forms by
spreading from the epiblast. I have concocted the "Peter Principle"
which says that to say they are homologous is to promote
the word "homologous" beyond its level of competence. [Recall
the formulation of the original Peter Principle? Peter was
the person's last name, by the way.]

Is it similar to other
>structures, but appear at an earlier stage developmentally?

>If it passes similarity, it must next pass the "test of consilience" Is
>the structure under study and the structure with which one wishes to
>establish homology present in the same organism? (Say wings and arms in
>angels) If so, the two structures cannot be homologous.

Good point, sometimes overlooked. However, if genes can be duplicated,
why not whole limbs?

>If the structure passes all of these tests, one has an apomorphy-a
>shared trait. Assuming the outgroup has a primitive state it will be a
>synapomorphy-shared *derived* trait. Is it a homology? No, because
>there is one further test to go-the test of congruence-or, how does it
>stand up compared with all of the *other* shared derived traits? How to
>tell? Run the matrix of all synapomorphies through the computer (or do
>it by hand if you like). If the character is congruent with the tree,
>it is a homology.

And a synapomorphy.

>If it requires additional ad hoc statements to
>explain away an incongruent distribution (statements of reversal or
>convergence) then it is a homoplaisy-a false homology.

And a `false synapomorphy', in the light of your earlier words:
"If the structure passes all of these tests, one has an apomorphy-a
shared trait. Assuming the outgroup has a primitive state it will be a
synapomorphy-shared *derived* trait."

>They are not interchangable terms.

I interchanged them just now.

[Peter Nyikos]

Interesting--for once Tom and I are in agreement. See my
own reply to Anderson's statement on "Homology and Synapomorphy".

[Cal King]

In article <6ck4nj$i...@sifon.cc.mcgill.ca>,
Zen Faulkes <zfaulkes@*SPAMBLOCK*bio1.lan.mcgill.ca> wrote:

>> We were discussing the use of embryological evidence to
>> assess character homology.
>
> Certainly, embryological evidence is very good and interesting. But
>it does not have any _a priori_ status as a better data set for
>evaluating hypotheses of homology.

It is not infallible. But it has proven to be a very informative source of
homology information. It is better than, say, claiming that 2 characters are
synapomorphs because they merely look alike. Functional, adaptive
considerations can certainly help weed out some of the problematic characters.
Systematists dating back to Darwin have learned not to trust adaptive
characters in phylogenetic analysis because they are prone to being
convergences. Indeed, it is the cladist's uncritical use of convergent
similarities which have led to such PAUPably wrong hypotheses as a loon-grebe
clade and an owl-hawk clade.

> So the claim is that an hypothesis of homology can be assessed
>independently of a phylogenetic hypothesis?

Yes, definitely. Using Dollo's Principle, many convergent characters can be
identified.

>As the modal definition of
>homology is some variation on "a character inherited from a common
>ancestor," this would seem very difficult indeed.

If their similarities are superficial but differ in their details, then they
are unlikely to be homologs. Or if the similarities are in some trivial
characters, such as the presence or absence of claws in mammals, but their
differences are in such significant characters as mode of reproduction, then
we can be sure that the similarities are convergent even if we don't have a
detailed tree.

>Identifying a
>structure as a homologue in two taxa implies something about their
>phylogeny; i.e., they shared an ancestor. Unless, that is, one uses
>the pre-Darwinian definition of homology (a small minority do).

Since life on earth in monophyletic, all organisms shared an ancestor. The
question is not whether, but how long ago?

[Peter Nyikos]

Jason Anderson <jasona@no_spam.mcgill.ca> writes:

>Cal King wrote:
>>
>> In article <34EAF811.45FC@no_spam.mcgill.ca>, Jason Anderson
>> <jasona@no_spam.mcgill.ca> wrote:
>>
><snip>

>> Perhaps the cladists have managed to become editors of many scientific
>> journals and are thus in a position to reject non-cladistic papers. Whatever
>> the reason, if a systematist cannot get his work published unless he uses
>> cladistics, then sooner or later, the reality of needing to publish will force
>> most to become cladists. This is a rather unfortunate trend because fashion,
>> not scientific rigor, becomes the driving force in scientific research.

>PERHAPS little green men from mars run the United Nations. PERHAPS you

>are really Arnold Kluge, screening potential submitters to Cladistics

>for idealogical purity. PERHAPS I'll accidentally eat poorly prepared

>Fugu at the Sushi bar for lunch-making this my last post.

Perhaps comparative anatomy is becoming more and more the domain
of cladists. That's a much fairer analogy--to something you
yourself wrote a few days ago, discussed further on the thread,

"The Decline and Fall (?) of Comparative Anatomy".

[Jason Anderson]

Peter Nyikos wrote:
(BTW Peter, a response to your pers email is forthcoming-I just need to
homologize the mammalian podial terminology with the lower tetrapods
with which I am accustomed)

>
> This is a followup to a post that appeared on
> "Methods and characters (Re: Quetzalcoatlus northropi!!!)"

> >> Synapomorphy *is* homology, isn't it?
>
> >No. Synapomorphy is a "shared derived trait". Homology is "same organ,
> >regardless of form or function (Owen)".
>
> Well, that seems to clash with the apparent consensus on
> sci.bio.evolution last year on what homology is, because
> it was agreed, for example, that bat wings are homologous
> to bird wings *as forelimbs* but not *as wings*, and it
> would seem that what distinguishes wings from mere forelimbs
> has partly to do with function. Yes, ostrich wings don't
> function that way, but they ARE synapomorphic with wings
> that are used for flight, hence they are wings and homologous
> *as wings* to the wings of eagles.
>

Limbs are a synapomorphy of Tetrapoda-so all tetrapods will have them,
unless secondarily lost. At the level of birds and bats, forelimbs
would be a symplesiomorphy-a shared, primitive trait. The secondary
alteration into wings highlights interesting issues of how movement
through a fluid medium constrains the direction which evolution may
follow-don't you think? However, the wings of birds and bats are
*analogous*, not *homologous*, structures. So, the tree shows that the
forelimb *as forelimb* is shared due to ancestry between birds and bats
*at the level of Tetrapoda*-a homology; it is a *plesiomorphy* at the
level of *Amniota*; and the secondary adaptations to powered flight are
*analogous* structures between the two groups-all at the same time. Who
said that a cladogram does not contain much information? :)

> >One examines the animals under study. One notices a character, which
> >seems to be present in one group but different in another group of study
> >animals (or plants, etc...). One must now apply a series of tests to
> >establish homology. The first set is collectively referred to as "the
> >test of similarity". Is it in a similar place? Does it relate with
> >similar other organs/bones/etc. Does it scale geonetrically? Does it
> >form from similar embryologic tissues?
>
> What is the function of that word "similar"? Besides the
> bit about bird and theropod digits that you don't want to
> get further into, there is the long-hot topic on talk.origins
> whether an amniotic membrane that forms by delamination from
> the cytotrophoblast is homologous to one that forms by
> spreading from the epiblast. I have concocted the "Peter Principle"
> which says that to say they are homologous is to promote
> the word "homologous" beyond its level of competence. [Recall
> the formulation of the original Peter Principle? Peter was
> the person's last name, by the way.]
>

Similar in this case would be "the same"-with the proviso that up-stream
changes may change how we identify the precursor tissue.

As to what goes on in talk.origins and the "Peter Principle"-I've never
heard of it, have never peaked into talk.origins, and have not really
participated here for to terribly long-so references to past threads are
lost on me.

> Is it similar to other
> >structures, but appear at an earlier stage developmentally?
>
> >If it passes similarity, it must next pass the "test of consilience" Is
> >the structure under study and the structure with which one wishes to
> >establish homology present in the same organism? (Say wings and arms in
> >angels) If so, the two structures cannot be homologous.
>
> Good point, sometimes overlooked. However, if genes can be duplicated,
> why not whole limbs?
>

They do, at least presumably, between arthropod taxa ("orders"?)-not
just the limb but full segments- but never in tetrapods. Perhaps its a
constraint placed upon development due to the complex interplay of the
regulator genes involved in signalling morphogenesis along the axis and
limb bud of the animal concerned. Again presumably there would be a
"Shiva" stage- say a human with 4 identical working arms, before natural
selection would be able to further modify one set of arms into a flight
apparatus while preserving the origional function in the other. One
would expect to find fossils showing this transitional morphology, thus
preserving our ability to identify homology.

> >If the structure passes all of these tests, one has an apomorphy-a
> >shared trait. Assuming the outgroup has a primitive state it will be a
> >synapomorphy-shared *derived* trait. Is it a homology? No, because
> >there is one further test to go-the test of congruence-or, how does it
> >stand up compared with all of the *other* shared derived traits? How to
> >tell? Run the matrix of all synapomorphies through the computer (or do
> >it by hand if you like). If the character is congruent with the tree,
> >it is a homology.
>
> And a synapomorphy.
>
> >If it requires additional ad hoc statements to
> >explain away an incongruent distribution (statements of reversal or
> >convergence) then it is a homoplaisy-a false homology.
>
> And a `false synapomorphy', in the light of your earlier words:
> "If the structure passes all of these tests, one has an apomorphy-a
> shared trait. Assuming the outgroup has a primitive state it will be a
> synapomorphy-shared *derived* trait."
>
> >They are not interchangable terms.
>
> I interchanged them just now.
>

As Tom pointed out, my use of these terms are not followed by him (and
countless others, I'm sure). Let me rephrase what I said above in Tom's
terms-One identifies potential homologies and enters them into a
matrix. A tree is produced which is minimizes appeals to ad hoc. The
characters which survive this test of congruence are homologies-and
synapomorphies. Despite following this nomenclature "Cal" would still
be wrong in his assertion that "one uses a tree built with
synapomorphies to identify synapomorphies"-unless the meaning of
synapomorphy sensu me ("Cal's" prepositional clause)= synapomorphy sensu
Tom ("Cal's" direct object) , which it does not due to my adhearing to a
piddly little semantic point. If I did accept Tom's definition, then
one would use not use a tree built of synapomorphies-because the
synapomorphies are read off the tree. The analysis is the test of the
hypothesized homologies-the identification of true homologies is the
result of the analysis.

[Tom DiBenedetto]

Cal King wrote:

>>The manus dispute is a dispute about anatomy.

>It is a dispute about character homology and thus the soundness of
>phylogenetic hypotheses based on homoplastic characters. The anatommical
>evidence provides evidence that the cladist's synapomorphies were in fact not
>synapomorphs.

No, the embryological evidence, were it to be valid, would be evidence
that one of the synapomorphies was false.

>Then on what basis did you claim that he "obviously" analyzed the characters?

The fact that I have read works of his which entailed masses of
anatomical work,,,as I told you,

>I didn't make the charges. Feduccia did.

You repeated it with complete confidence. Why do you not feel
compelled to verify Feduccia's charges before you repeat them?

>You asserted that Feduccia was wrong, but provided no evidence to support your
>assertion.

But you assert that he is right with no evidence to back it up except
his assertion itself. Since you raised the issue, and made a serious
charge, you have the responsibility to justify it.

>>huh? A C.I. which is low is explicit proof that there has been
>>testing.

>It is an explicit proof that characters were chosen indiscriminantly and that
>there was little, if any, a priori character analysis.

Clearly the words of one who has never done any systematic work!!!!

>Thus a low value shows that there are a large number of homoplasies, which
>imply that the a priori character analysis, if there was any, was shoddy at
>best.

I challange you to get out from behind your computer and do some
systematic work, and to come back to us with a matrix and tree with a
suitably high CI which is not the result of hiding inconvenient data.

> Since you again gave me no reference concerning the
>various types of polytomies, only the assertion that there are different
>types.

Sorry buddy, I think it far more helpful for the long term tenor of
these discussions if you are forced to assume responsibility to read
widely and get a considerably broader view of the field. You will
probably be exposed to a lot of interesting things on your search for
the truth about polytomies. Good luck.

[Tom DiBenedetto]

Peter Nyikos wrote:

>> Cladistic principles tell me how to deal with characters once
>>they are defined. They are defined by Tom the anatomist, not Tom the
>>systematist.

>But they are subject to revision in the light of cladistic
>analysis, and the bottom line is, the last word belongs to
>the most parsimonious tree. There is no court of higher
>appeal if something as bizarre as a pollex or even
>the vertebrate eye disappearing and reappearing is to be
>found in the tree, is there?

More characters.

> The best you can do is keep
>running matrices with more and more organisms and more and
>more characters, in the hope that a detailed enough cladistic
>analysis will vindicate your anatomical instincts.

Or refute them, hence allowing you the opportunity to learn something.
The point of systematic analysis is not to vindicate your instincts,
Peter. You can use your ego and assertiveness in their raw form if
that is your only goal.

>Weighting has nothing to do with considering some apparent characters
>to be more important than others, does it?

Of course it does. Why else would you weight?

>[I say "apparent characters" because "weighting" has just to
>do with hypothesizing that two features in two different animals
>are either "the same character" or "not the same character", doesn't it?]

No. Why would you need to weight in order to make a simple homology
statement?

>, your anatomical insights are "objectively" either
>confirmed or falsified by the most parsimonious tree, eh?

yup, they either are or are not congruent; something which is expected
of true homoogies.

[Mike Noren]

Thusly nyi...@math.scarolina.edu (Peter Nyikos) spake unto all:

: >>In other words, you rely on the subjective judgment of anatomists.

: >>Funny, I thought you and Mike Noren were the ones who claimed
: >>cladistics was fully objective.

:
: >huh?

:
: Or was it just Mike Noren?

Or was it none at all.

: Anyway, I even said at one point

: that Noren's view of "objectivity" seemed to be such that
: eventually all his ideal systematists could be replaced by computers.
: Either you or Mike came back and said that is exactly what is
: good about his ideal of systematics.

I don't know what to say of the above. I have never ever claimed that
"cladistics", the whole shebang, is fully objective, and you will not
be able to show that I have. What I _HAVE_ said is that identifying
the groups is objective, and naming them is not.

Please feel free to post any quote where I say that cladism is fully
objective. Or retract.

Michael Norén, Doctoral student, Tel: Int +46 (0)8 6664236
Swedish Museum of Natural History, Fax: Int +46 (0)8 6664125
Dept. of Invertebrate Zoology
P.O.B. 50007
S-104 05 Stockholm, Sweden

[Mike Noren]

Thusly nyi...@math.scarolina.edu (Peter Nyikos) spake unto all:

: michae...@nrm.se (Mike Noren) writes:
:
: >Thusly nyi...@math.scarolina.edu (Peter Nyikos) spake unto all:
:
: >(snip)
: >: In other words, you rely on the subjective judgment of anatomists.

: >: Funny, I thought you and Mike Noren were the ones who claimed
: >: cladistics was fully objective.
:

: >You wouldn't mind supplying a quote to back your claim that I've

: >stated that "cladistics" was "fully objective", would you? You see, as
: >is too often the case when you paraphrase my opinions or claims, I
: >don't recognise them as mine.
:
: See my reply to Tom. Did it refresh your memory?

No, it did not. I have no recollection whatsoever of ever having
stated that cladism was fully objective. Identifying the groups is
objective (quite different from any system which allows paraphyly) but
naming the groups is not.

Document or retract, please.

: Peter Nyikos -- standard disclaimer --

[John Harshman]

Just a possibly illuminating question for this thread: Are shark mandibles
in any way homologous to adult mammalian mandibles? (Let's ignore
embryonic states for this purpose.) I can think of arguments for and
against. My personal take-home message: homology is a somewhat slippery
concept when you get above the nucleotide or nucleotide site level. (Still
useful, though.)

[Mike Noren]

Thusly nyi...@math.scarolina.edu (Peter Nyikos) spake unto all:

: >> +++++++++++++++++++++++ dogmatic cladistic posting mode on

: >>
: >> Either the analysis took the form of earlier cladistic use
: >> of trees or else it was subjective.
: >>
: >> +++++++++++++++++++++++ dogmatic cladistic posting mode off
: >>
: >????????????
: >You are posting "dogmatic cladistic" nonsense, I'm afraid.
:
: I was just describing how Mike Noren came across to me
: in past posts, although he never put it that way. See my
: reply to Tom DiB on why Mike came across to me that way.

Although I've never actually claimed that cladism is fully objective,
have I. You're unable to document that I ever have, aren't you.

But I'm not likely to get a retraction from you, am I.

Please, feel free to clarify what I'm supposed to have said.
Preferably backed up by an actual quotation.

I mean, for F's sake, I must've exchanged several hundred mails and
post with you over the last two years - I _MUST_ have said something
stupid sometime, somewhere, so you don't have to make up stuff I'm
supposed to have said!

[Cal King]

In article <34eddbb4....@news.itd.umich.edu>, td...@umich.edu wrote:

>Cal King wrote:

>>It is a dispute about character homology and thus the soundness of
>>phylogenetic hypotheses based on homoplastic characters. The anatommical
>>evidence provides evidence that the cladist's synapomorphies were in fact not
>>synapomorphs.
>
>No, the embryological evidence, were it to be valid, would be evidence
>that one of the synapomorphies was false.

If it was false, then the cladogram constructed using this character to show a
close bird-theropod relationship is also false, because the common ancestor of
birds and theropods would be found among the pre-dinosaurian thecodonts.
Birds would then be no more closely related to theropods than they are to
sauropods or even the ornithischian dinosaurs. If the hand evidence is
correct, then their other similarities could only have arisen through
convergence as well. Hence Gauthier's (1986) hypothesis that birds are a
subgroup within the theropods would also be false. It is thus a systematic
dispute, not just an anatomical dispute.

>>Then on what basis did you claim that he "obviously" analyzed the characters?
>
>The fact that I have read works of his which entailed masses of
>anatomical work,,,as I told you,

The study in dispute is Gauthier's 1986 paper linking birds with theropods.
What evidence do you have that he actually analyzed the characters in that
paper, or in any other paper, since you still speak in generalities, without
any specific references.

>>Thus a low value shows that there are a large number of homoplasies, which
>>imply that the a priori character analysis, if there was any, was shoddy at
>>best.

>I challange you to get out from behind your computer and do some
>systematic work, and to come back to us with a matrix and tree with a
>suitably high CI which is not the result of hiding inconvenient data.

"Hiding inconvenient data" is something that cladists apparently don't do.
They simply throw all the questionable characters into the analysis and keep
their fingers crossed. Too bad the cladistic gods don't always answer their
prayers. If there is a priori character analysis but the questionable
characters are still included in the phylogenetic analysis, then the result is
unreliable. One cannot claim to be "objective" in character choice and also
claim that the characters one chooses are good characters. Bad characters
make for a bad phylogenetic analysis. Homoplasies in, garbage out.

>> Since you again gave me no reference concerning the
>>various types of polytomies, only the assertion that there are different
>>types.
>
>Sorry buddy, I think it far more helpful for the long term tenor of
>these discussions if you are forced to assume responsibility to read
>widely and get a considerably broader view of the field. You will
>probably be exposed to a lot of interesting things on your search for
>the truth about polytomies. Good luck.
>
>Tom DiBenedetto td...@umich.edu

I don't know how widely you have read, but you definitely haven't read
Gauthier's (1986) analysis. If you did, you could have provided us with
evidence from his paper to falsify Feduccia's (1996) claim that Gauthier never
analyzed the characters linking birds and theropods.

[Cal King]

In article <6ck4da$o...@redwood.cs.sc.edu>, nyi...@math.scarolina.edu (Peter Nyikos) wrote:
>Path:

>Evolutionary systematics is now identified not by whether
>it uses cladistic techniques, but whether it allows paraphyletic
>taxa into its bookkeeping system.

Doubtful. There is no shortage of methodological disputes. Cladists claim
that they don't choose characters, and thus their method is objective.
Evolutionary systematists still believe that the choice of characters is the
most important part of a phylogenetic analysis and will continue to challenge
cladistic results using character analysis. As the cladist Kurt Schwenk says,
character analysis is the soft underbelly of phylogenetic analysis and is
always vulnerable to attack. Zhou recently analyzed the characters used to
identify Mononykus as a bird and showed that most of those postulate
synapomorphs Mononykus purportedly shares with birds are actually digging
adaptations that are found in many unrelated lineages of extant digging
mammals, for example moles. The uncritical choice of "synapomorphs" by
cladists will always be an area that the evolutionary systematist will choose
to criticize cladistic results.

> And before long, it will
>be probably be identified with those who want to hang on to the Linnean
>system in whatever form.
>

>Peter Nyikos -- standard disclaimer --

Again, there are many cladists who will gladly use the Linnaean system.
Remember that example of the lizard families? A cladist split 3 families of
iguanian lizards into 9 families. There is no consensus on classifications
even among cladists. Carroll (1997), for example, says that cladistic
classification is in a state of flux.

[Cal King]

In article <6ck5f0$o...@redwood.cs.sc.edu>, Peter Nyikos
(nyi...@math.scarolina.edu) says...

>>> Synapomorphy *is* homology, isn't it?

>>No. Synapomorphy is a "shared derived trait". Homology is "same organ,
>>regardless of form or function (Owen)".

Synapomorphs, as used by modern cladists, need not be a homology. In fact,
cladists do not care if they are synapomorphs. Many cladists even claim that
homoplasies can be informative of relationships and trees can be constructed
using homoplasies.

Here is a short excerpt from an exchange between Holtz and I:

====================
Message-ID: <346CDC5A...@umail.umd.edu>
Newsgroups: sci.bio.paleontology

[Cal King]:
> Homoplasy is rampant in evolutionary history you know, even by the cladist's
> own measure.

[Holtz]:
Indeed, I made a slide that says "Homoplasy, thy name is "Theropod"!"
Nevertheless, there is signal in them there noise...
====================

And then there is this exchange between Brochu and I in talk.origins:
==========================
[Cal King]:
>> Yeah, but polarity assessment is pointless if homology has not been
>> demonstrated.
>
>Incorrect.

>chris

I would leave this last comment for other systematists (cladists and
noncladists alike) to judge. BTW, Hennig is spinning in his grave faster than
any croc ever did. ;-)
=========================

>Well, that seems to clash with the apparent consensus on
>sci.bio.evolution last year on what homology is, because
>it was agreed, for example, that bat wings are homologous
>to bird wings *as forelimbs* but not *as wings*, and it
>would seem that what distinguishes wings from mere forelimbs
>has partly to do with function.

What distinguishes them isn't necessarily function, but often morphology.

>Yes, ostrich wings don't
>function that way, but they ARE synapomorphic with wings
>that are used for flight,

Ostrich wings are still bird wings. They are degenerate bird wings, just as
ostrich feathers are degenerate feathers.

>hence they are wings and homologous
>*as wings* to the wings of eagles.

Correct, even though ostrich wings do not have the same function as eagle or
albatross wings.

>>One examines the animals under study. One notices a character, which
>>seems to be present in one group but different in another group of study
>>animals (or plants, etc...). One must now apply a series of tests to
>>establish homology. The first set is collectively referred to as "the
>>test of similarity". Is it in a similar place? Does it relate with
>>similar other organs/bones/etc. Does it scale geonetrically? Does it
>>form from similar embryologic tissues?

Well, I doubt that cladists have time to do all of these tests, or even most
of them before they plunge into an analysis. Mostly they just rely on
available information. About the only test cladists have time to perform
with their scores of characters is the test of similarity. If they look
similar, then they are assumed to be synapomorphic.

>What is the function of that word "similar"? Besides the
>bit about bird and theropod digits that you don't want to
>get further into,

Similar looks, that was exactly how the cladists claim that bird wings and
theropod hands are homologous. They point out that Archaeopteryx has the same
phalangial formula as some theropods, but alas, phalangeal formula is
individually variable in Archaeopteryx, as Feduccia (1996) pointed out.
Apparently, the cladists did not perform the embryological test on the bird
manus even though in the passage above they claim that it is one way to test
for homology. In fact, the cladistic party line is that Burke and Feduccia's
data is wrong. It is wrong because they contradict previous cladistic
assumptions of bird-theropod digital homology.

> Is it similar to other
>>structures, but appear at an earlier stage developmentally?
>
>>If it passes similarity, it must next pass the "test of consilience" Is
>>the structure under study and the structure with which one wishes to
>>establish homology present in the same organism? (Say wings and arms in
>>angels) If so, the two structures cannot be homologous.
>
>Good point, sometimes overlooked. However, if genes can be duplicated,
>why not whole limbs?

If the wing arises from a duplicated gene, then it is by definition not a
homolog of the arms.

In fact, Mayr and Ashlock have an even stricter definition of homology. If a
common ancestor possesses the genotype but not the phenotype, for example for
stalked eyes, and that the phenotype is independently developed in 2 daughter
species, then it is not a homolog but a homoplasy. Cronquist argues that this
sort of parallelism is rampant in the angiosperms, which are difficult to
classify as a result.

>>If the structure passes all of these tests, one has an apomorphy-a
>>shared trait. Assuming the outgroup has a primitive state it will be a
>>synapomorphy-shared *derived* trait. Is it a homology? No, because
>>there is one further test to go-the test of congruence-or, how does it
>>stand up compared with all of the *other* shared derived traits? How to
>>tell? Run the matrix of all synapomorphies through the computer (or do
>>it by hand if you like). If the character is congruent with the tree,
>>it is a homology.

Hmm, that is how Cracraft concluded that the similarities which he used to
lump the owls and hawks are synapomorphs. DNA-DNA hybridization and long
standing theories about owl-hawk convergences of course contradict his claim.
This last test is no test, but is actually a misleading circular argument on
which the cladists have overly, often exclusively, relied to their own
detriment.

[Tom DiBenedetto]

Cal King wrote:

>>No, the embryological evidence, were it to be valid, would be evidence
>>that one of the synapomorphies was false.

>If it was false, then the cladogram constructed using this character to show a
>close bird-theropod relationship is also false, because the common ancestor of
>birds and theropods would be found among the pre-dinosaurian thecodonts.

Maybe, but maybe not. If the character is not homologous, it must be
removed from the matrix, or recoded to reflect the new understanding.
Then you rerun the analysis.

> If the hand evidence is
>correct, then their other similarities could only have arisen through
>convergence as well.

Not necessarily. The hand evidence could be the convergence. I am not
an expert on these critters. Apparently you know something about the
nature of the evidence here. So tell us, if the taxa are really
related the way you claim, then how many characters which the
"cladists" consider synapomorphies would have to be reinterpreted as
convergences? How many, and what are they?

> Hence Gauthier's (1986) hypothesis that birds are a
>subgroup within the theropods would also be false. It is thus a systematic
>dispute, not just an anatomical dispute.

No "cal". try to think with a little precision. It is an anatomcial
dispute which might have consequence for relationships.

>The study in dispute is Gauthier's 1986 paper linking birds with theropods.
>What evidence do you have that he actually analyzed the characters in that
>paper, or in any other paper, since you still speak in generalities, without
>any specific references.

Once again, "cal", stop trying to avoid responsibility for shooting
off your mouth on something which YOU have no knowledge of. You are
the one who is broadcasting Feduccia's charge as if you were
completely convinced that he is right. Do you accept on faith anything
that an anti-cladist says? You are the one who is so quick to charge
others with religous zealotry, or dogmatism, yet you are also the one
who seems to have no sense of responsibility for the things you say.
Do you think it makes you are a serious, or intetresting thinker to
merely advocate the ideas or accusations of others? Why dont you think
for yourself?

>>I challange you to get out from behind your computer and do some
>>systematic work, and to come back to us with a matrix and tree with a
>>suitably high CI which is not the result of hiding inconvenient data.

>"Hiding inconvenient data" is something that cladists apparently don't do.

correct.

>They simply throw all the questionable characters into the analysis and keep
>their fingers crossed.

No, there is a difference between questionable and inconvenient. If
you were to get off your jihad trip and think clearly for a moment you
would be able to see this. Questionable means you cannot formulate a
homology hypothesis with confidence. No scientist will put forth a
hypothesis unless they have reason to believe it is true. Inconvenient
data is data which contradicts the outcome you prefer. Cladists do not
hide inconvenient data, even though it subjects them sometimes to
clueless charges such as yours.