Keeping up with creationist claims
MAB
listed on the TO archive and more recent and 'sophisticated' claims.
Not that all creationist claims need to be listed or answered but the
list is out of dated.
For example creationists have been working on a way for the flood to
account for all plate tectonic activity at least since the last super
continent Pangea. See:
http://www.answersingenesis.org/articles/nab/catastrophic-plate-tectonics
For a more technical approach:
http://www.icr.org/article/catastrophic-plate-tectonics-flood
This situation wouldn't matter so much if the TO archive wasn't so
well respected as an important source of evolution/creation
information. I recently read a book called 'Why Evolution Works (And
Creationism Fails)' by Matt Young and Paul K. Strode. The book is
excellent and completely up do date on the latest developments in
evolution. However, one section that was very disappointing started on
page 157 titled 'Creationist Quibbles' about the age of the earth. The
authors mention moon dust and the decay of the earth's magnetism as
common creationist arguments. Not that some creationist still don't
use these arguments but for the most part they have been abandoned. A
book that is up to date on everything else should be up to date on the
major creationist arguments.
Don't get me wrong - the newer claims don't have any more validity
than the abandoned ones. They are more sophisticated however and do
try to account for more data.
I don't have an answer for this situation or the resources to address
it but I think it is important.
Mark A Buchanan
bpuharic
<marklynn...@gmail.com> wrote:
>Don't get me wrong - the newer claims don't have any more validity
>than the abandoned ones. They are more sophisticated however and do
>try to account for more data.
>
>I don't have an answer for this situation or the resources to address
>it but I think it is important.
>
seems to me the creationist argument is:
-we had NO idea this happened
-we can explain it using god
creationists are theological parasites, feeding off science to bolster
their ghost-based view of reality
macaddicted
> Don't get me wrong - the newer claims don't have any more validity
> than the abandoned ones. They are more sophisticated however and do
> try to account for more data.
>
> I don't have an answer for this situation or the resources to address
> it but I think it is important.
I once went to a panel in which Kenneth Miller was a participant. He
commented that creationists (and IDists IIRC) would not be happy until
every department of evolutionary biology was closed because there was
nothing left to learn.
The good part about god of the gaps is that there are still a lot of
gaps (even if you have to bring your own crowbar to open them up). The
bad part is that the gaps keep closing.
--
macaddicted
Wisdom is radiant and unfading and she is easily discerned
by those who love her and is found by those who seek her.
Wisdom 6:12 (NRSV)
John Wilkins
<9181010a-696b-4b9c...@e7g2000vbi.googlegroups.com>, MAB
<marklynn...@gmail.com> wrote:
The traditional manner for updating the archive is for someone to put
up a draft here, which we all pick to pieces, demolishing your
self-image and hope of advancement, before we put the finished version
up on the archive. Alternatively you could put up the problem, and
collect the responses into a FAQ. So get to it! I can personally
guarantee the self-image thing.
John Wilkins
macaddicted <macad...@REMOVETHISca.rr.com> wrote:
> MAB <marklynn...@gmail.com> wrote:
>
> > Don't get me wrong - the newer claims don't have any more validity
> > than the abandoned ones. They are more sophisticated however and do
> > try to account for more data.
> >
> > I don't have an answer for this situation or the resources to address
> > it but I think it is important.
>
> I once went to a panel in which Kenneth Miller was a participant. He
> commented that creationists (and IDists IIRC) would not be happy until
> every department of evolutionary biology was closed because there was
> nothing left to learn.
I met Miller a couple of weeks ago. He treats creationists as the
village idiots of science. Francisco Ayala was even more scathing,
calling IDevotees "embarrassing" in a room full of scientists. Both
agreed that arguing with them from scientific grounds was like shooting
fish in a barrel (another member of the panel, William Carroll, said
that he was in favour of animal rights, and so would not argue with
creationists for that very reason).
Here's a list of the people who were at that conference (apart from me
and the other 500 uninvited guests):
http://www.nd.edu/~reilly/darwinconference.html
I had dinner with Conway Morris, and Bowler, and Minelli, among others.
John Harshman
>
> I had dinner with Conway Morris, and Bowler, and Minelli, among others.
Did you happen to get into any arguments about the inevitability of Homo
sapiens?
John Wilkins
<jhar...@pacbell.net> wrote:
Yes I did. I also argued with Simon about the use of analogical
reasoning (i.e., assuming convergence).
macaddicted
> In article <1j9c7dv.1m2e4eh16oic34N%macad...@REMOVETHISca.rr.com>,
> macaddicted <macad...@REMOVETHISca.rr.com> wrote:
>
> > MAB <marklynn...@gmail.com> wrote:
> >
> > > Don't get me wrong - the newer claims don't have any more validity
> > > than the abandoned ones. They are more sophisticated however and do
> > > try to account for more data.
> > >
> > > I don't have an answer for this situation or the resources to address
> > > it but I think it is important.
> >
> > I once went to a panel in which Kenneth Miller was a participant. He
> > commented that creationists (and IDists IIRC) would not be happy until
> > every department of evolutionary biology was closed because there was
> > nothing left to learn.
>
> I met Miller a couple of weeks ago. He treats creationists as the
> village idiots of science. Francisco Ayala was even more scathing,
> calling IDevotees "embarrassing" in a room full of scientists. Both
> agreed that arguing with them from scientific grounds was like shooting
> fish in a barrel (another member of the panel, William Carroll, said
> that he was in favour of animal rights, and so would not argue with
> creationists for that very reason).
>
> Here's a list of the people who were at that conference (apart from me
> and the other 500 uninvited guests):
>
> http://www.nd.edu/~reilly/darwinconference.html
Do you know if they will be publishing the papers? Anything titled
"Teleonomic and Teleologic Causation in Phylogeny and Ontogeny" has got
to be a page turner.
I'm not sure which is worse, that I understand the title or that I am
anxious to read it.
>
> I had dinner with Conway Morris, and Bowler, and Minelli, among others.
> >
> > The good part about god of the gaps is that there are still a lot of
> > gaps (even if you have to bring your own crowbar to open them up). The
> > bad part is that the gaps keep closing.
John Wilkins
macaddicted <macad...@REMOVETHISca.rr.com> wrote:
I wish I could say that I understood it, but Genarro has a rather thick
Italian accent, and he talks too quickly.
From what I can recall it had something to do with treating metabolism
as teleological.
> I'm not sure which is worse, that I understand the title or that I am
> anxious to read it.
Yes, they will. Not my single author paper, but the one Paul and I put
up will be included.
Stuart
> There seems to be a growing gap between the list of creationist claims
> listed on the TO archive and more recent and 'sophisticated' claims.
> Not that all creationist claims need to be listed or answered but the
> list is out of dated.
>
> For example creationists have been working on a way for the flood to
> account for all plate tectonic activity at least since the last super
> continent Pangea. See:
>
>
This is complete bullshit. This used to be my field so I think I know
what
I'm doing here .
As a calculation, it is quite interesting but as a model for the earth
it
doesn't work. Baumgardner inappropriately uses a rheology referred to
as
the "stick-slip" rheology originally developed by Whitehead and Gans
to model
the dynamics of faults. Convection modelers such as Bercovici applied
it to mantle
convection, but only to the lithosphere (upper most) mantle which is
capable brittle
failure. Models like Bercovici's, Weinstein's (yours truly) and others
were an attempt
to understand how plate tectonics arises out of convecting mantle.
Baumgardner applies the stick-slip rheology to the whole mantle which
is inappropriate
as the mantle does not undergo brittle failure outside of subducted
lithospshere. Again,
interesting calculations perhaps, but not applicable to the Earth. In
my paper "Catastrophic Overturn
of the Earth's Mantle ...." which is referenced by this article
(eeeyeww) essetially gives a best case scenario for very rapid
subduction, and that is 5-10 millions years for a single
*catastrophic* event to occur. Never mind the rest of Earth's
history.
In short you can come up with a model to produce any timescale you
desire, however, it needs
to be relevant to the Earth, and this one makes for an interesting
planet, its just that it is not
Earth.
Stuart
John Harshman
distribution he doesn't like, or what?
John Wilkins
<jhar...@pacbell.net> wrote:
I gave a talk he attended in which I said that the most inductively
generalisable claims that could be made about biology, a historical
science, are those based on homologies, not on (adaptive) analogies.
This is because with analogies it is easy to construct them (Just-So
stories, for example) and almost impossible to test them, but also
because you get no inferential return out of classes constructed that
way that you didn't put into the construction itself. Phylogenies, on
the other hand, allow us a great deal of inductive generalisability.
They are, in other words, "projectible" (Nelson Goodman's term for
"straight rule" induction).
He likes adaptive analogies because he wants to claim that something
"like" humans was inevitable through evolution. I think any decent
phylogeny undercuts that presumption entirely, but he doesn't think
much of phylogenies. I think he's special pleading for humanity.
bobsyo...@yahoo.com
"MAB" <marklynn...@gmail.com> wrote in message
news:9181010a-696b-4b9c...@e7g2000vbi.googlegroups.com...
A possible solution/answer is to pay less attention to the morons who are
nothing more than (known) trolls of ignorance.
"I'll be back, "All seeing eye" and others constantly show that they are
here for nothing more than to play sick, childish games.
Ignore them, and they'll go away.
John Harshman
> In article <KLqdnUi3qMc...@giganews.com>, John Harshman
> <jhar...@pacbell.net> wrote:
>
>> John Wilkins wrote:
>>> In article <OvSdnfatv8c...@giganews.com>, John Harshman
>>> <jhar...@pacbell.net> wrote:
>>>
>>>> John Wilkins wrote:
>>>>
>>>>> I had dinner with Conway Morris, and Bowler, and Minelli, among others.
>>>> Did you happen to get into any arguments about the inevitability of Homo
>>>> sapiens?
>>>>
>>> Yes I did. I also argued with Simon about the use of analogical
>>> reasoning (i.e., assuming convergence).
>>>
>> He assumes convergence? You mean to explain putative homologies whose
>> distribution he doesn't like, or what?
>>
> I gave a talk he attended in which I said that the most inductively
> generalisable claims that could be made about biology, a historical
> science, are those based on homologies, not on (adaptive) analogies.
> This is because with analogies it is easy to construct them (Just-So
> stories, for example) and almost impossible to test them, but also
> because you get no inferential return out of classes constructed that
> way that you didn't put into the construction itself. Phylogenies, on
> the other hand, allow us a great deal of inductive generalisability.
> They are, in other words, "projectible" (Nelson Goodman's term for
> "straight rule" induction).
I thihk I disagree (if I understand what you're saying, often a big if).
Doesn't that deny the utility of comparative methods, e.g. Felsenstein's
independent contrasts, Maddison's concentrated changes test, etc.? If
so, I think that just-so stories can indeed be tested, as long as we
have multiple examples. But perhaps that isn't what yo mean.
> He likes adaptive analogies because he wants to claim that something
> "like" humans was inevitable through evolution. I think any decent
> phylogeny undercuts that presumption entirely, but he doesn't think
> much of phylogenies. I think he's special pleading for humanity.
That would be consistent with his books.
John Wilkins
<jhar...@pacbell.net> wrote:
Or perhaps it is. You *can* test a Just-So Story only to the extent you
have a phylogeny to identify how derived a trait is relative to the
clade in which it evolves. I tend to be rather skeptical of the
statistical methods, but I am incompetent to make a critique at a
technical level.
John Harshman
We agree that you can't test these things without a having a phylogeny
in hand, and can't derive the phylogeny using a priori statements of
non-homology. We appear to disagree on the validity of statistical
methods. For an intro, dated but still valid, to comparative methods you
might try Harvey & Pagel 1991. Here's a review (which I haven't read --
just the first hit to let you see the actual citation):
http://www.cscs.umich.edu/~crshalizi/reviews/harvey-pagel/
MAB
>
> The traditional manner for updating the archive is for someone to put
> up a draft here, which we all pick to pieces, demolishing your
> self-image and hope of advancement, before we put the finished version
> up on the archive. Alternatively you could put up the problem, and
> collect the responses into a FAQ. So get to it! I can personally
> guarantee the self-image thing.
Thanks for the instructions, advice, and warning. Hopefully I can find
the time to give it a good try.
Mark
Mark Isaak
I would be glad to keep adding updates if someone would hire me to do so
as a full-time job.
--
Mark Isaak eciton (at) earthlink (dot) net
"It is certain, from experience, that the smallest grain of natural
honesty and benevolence has more effect on men's conduct, than the most
pompous views suggested by theological theories and systems." - D. Hume
Mike Painter
<snip>
>
> I once went to a panel in which Kenneth Miller was a participant. He
> commented that creationists (and IDists IIRC) would not be happy until
> every department of evolutionary biology was closed because there was
> nothing left to learn.
I don't know if he even wanted to but I once listened to someone destroy his
chances with such a question.
Jean Michael Cousteau was at a dive convention in Anaheim CA in the early
70's.
I was on the outskirts listening to him field questions when somebody asked
if it was worth going into marine biology because there was almost nothing
left to learn.
I can't remember his polite answer but he managed to keep a straight face
until he saw me laughing.
That was the year we all thought we were going to see lava flowing
underwater off Hawaii, something few had ever seen.
The film started. It was in black and white, shot in a kelp forest.
About 30 seconds in a plaintive cry from the photographer told us he had
grabbed the wrong film when he left Hawaii.
It was a good film but it was years before I got to see what he had filmed.
R. Baldwin
c72435...@o31g2000vbi.googlegroups.com:
From a Google search, it looks like your paper has picked up quite a few
Creationist cites.
>
> In short you can come up with a model to produce any timescale you
> desire, however, it needs
> to be relevant to the Earth, and this one makes for an interesting
> planet, its just that it is not
> Earth.
It also needs to be relevant to physics. How could ANY planet made of
actual rock materials support "continental sprint"?
>
> Stuart
>
John Wilkins
<jhar...@pacbell.net> wrote:
Cosma always does great reviews, but it fails to convince me that
statistical methods are consonant with pure phylogenetics. On this
point I fear that I am being somewhat intransigent.
Here's my real problem: statistical methods give you phenetic groups,
groups formed by similarities. Similarities are, IMO, both component
and assay relative, and can be arrived at causally in several different
ways, and so while the properties uncovered are real, they are not
"natural" in the taxonomic sense. That is, they fail to identify
genealogical relations.
Now if statistical inferences are used in modern phylogenetics, and I'd
be a fool to insist they aren't, I have to ask what it is that they
present to us? Phylogenies are found on relations of identity, not
similarity, so is the argument that these methods uncover identity
relations in virtue of similarity relations? Isn't that just to put the
OTU before the taxon?
Or is there some better foundation? What I basically want to know is
what knowledge people think that statistical methods actually deliver.
What is it that we know when we have applied them that relies on them
and not, say, a conflation of homology with analogy?
rmacfarl
"John Wilkins" <jo...@wilkins.id.au> wrote in message
news:191120091741587117%jo...@wilkins.id.au...
[Struggling like Hell to follow this thread...]
What of cladistics? Isn't that the statistical basis for
inferring phylogenetic relationships? Do you deny / downplay the
power of this approach?
richardal...@googlemail.com
The factor which seems to be missing from all these calculations is
the energy needed to accelerate and decelerate the tectonic plates. If
tectonic plates were moving a million times faster than the observed
rate today - which is what is required to fit plate movements into
their vastly accelerated timescale - the kinetic energy of the plates
will be a million million times higher. That energy needs to come from
somewhere. Bearing in mind that the kinetic energy of tectonic plates
is dissipated as heat, it needs to go somewhere when the plates slow
down. The energy radiated by such decelerating plates would be enough
to vaporise the earth, and probably half the solar system.
RF
John Harshman
I'm afraid I understand little of what you're saying here. But you seem
to be objecting to statistical methods in phylogeny estimation. We were
talking about statistical methods in comparative biology, all of which
are subsequent to phylogeny. That's a whole new topic.
> Or is there some better foundation? What I basically want to know is
> what knowledge people think that statistical methods actually deliver.
> What is it that we know when we have applied them that relies on them
> and not, say, a conflation of homology with analogy?
At this point I'm unsure what statistical methods we're talking about.
Phylogenetic or comparative? Are we inferring a tree or are we using an
existing tree to learn about evolution? I had thought the latter, but
you seem to be talking about the former.
MAB
> On Tue, 17 Nov 2009 17:37:11 -0800, MAB wrote:
> > There seems to be a growing gap between the list of creationist claims
> > listed on the TO archive and more recent and 'sophisticated' claims. Not
> > that all creationist claims need to be listed or answered but the list is
> > out of dated.
>
> > For example creationists have been working on a way for the flood to
> > account for all plate tectonic activity at least since the last super
> > continent Pangea. See:
>
Disappointing that funding for a good resource like the TO archive is
a problem. Hopefully we can collectively contribute time to do some
updates.
Mark
John Stockwell
> On Tue, 17 Nov 2009 17:37:11 -0800 (PST), MAB
>
Creationism is an outgrowth of religious apologetics, which are
basically arguments that are intended to reinforce the beliefs of
believers, and give fence sitters reasons to believe. There are
two basic "sermons from science" that virtually all of creationist
arguments are modeled on. Both are now obsolete.
The first is "Have you every seen an atom?". At the time that this
was proposed, technology had not yet advanced to make imaging
atoms possible. So, the punchline is that most people believed in
atoms because some teacher told them to. So, if you accept
atoms on "faith" then you should be able to accept religious concepts
on faith too.
The second is "Bumblebees can't fly." At the time, the way that
appropriate wing size was calculated in aviation was the simple
wing loading computation, involving the ratio of the surface area
of a wing to the weight being carried by it. The wing loading of
a bumblebee suggests that it is out of the range of possible flight.
Technology had not advanced far enough at the time to calculate
the effect of vortices of air giving additional lift, so this was
supposed
to be something that "defied scientific explanation". Hence, science
doesn't know everything, so theism must fill the gap by allowing
for miraculous aspects of life.
The rest of creationism is nay-saying that is intended to put some
aspect of science into one or the other of these pigeon holes.
Stuart
wrote:
> Stuart <bigdak...@gmail.com> wrote in news:d3d7d96f-f705-4459-aa95-
> c72435f1f...@o31g2000vbi.googlegroups.com:
Which is why it will be the last paper on mantle convection with the
word
"Catastrophic" in the title, or anywhere else...
>
>
> > In short you can come up with a model to produce any timescale you
> > desire, however, it needs
> > to be relevant to the Earth, and this one makes for an interesting
> > planet, its just that it is not
> > Earth.
>
> It also needs to be relevant to physics. How could ANY planet made of
> actual rock materials support "continental sprint"?
>
Indeed.
Stuart
Stuart
That is basically correct except that in Baumgardner's computation,
the
stick-slip rheology is essentially self lubricating hence the energy
needed
to drive plate tectonics is much lower, still you simply can't pull
that much heat
out in 6000 years. If you did the surface heat flow over that time
period would be the
equivalent of an active volcano.
Stuart
John Wilkins
<jhar...@pacbell.net> wrote:
OK, in that case I have no deep problem. But I do wonder about the
recent paper identifying the Hobbit as a separate species on
statistical grounds, largely because I am entirely unsure what prior
work has been done on the statistical variance of primate species.
>
> > Or is there some better foundation? What I basically want to know is
> > what knowledge people think that statistical methods actually deliver.
> > What is it that we know when we have applied them that relies on them
> > and not, say, a conflation of homology with analogy?
>
> At this point I'm unsure what statistical methods we're talking about.
> Phylogenetic or comparative? Are we inferring a tree or are we using an
> existing tree to learn about evolution? I had thought the latter, but
> you seem to be talking about the former.
>
I am, and hence you can ignore me.
richardal...@googlemail.com
Even if the slip surface offers no resistance whatsoever, the problem
of kinetic energy still remains. If the plates were moving a million
times faster than they are now, they need to be slowed down by
something, and that means that the kinetic energy of the plate - which
is mass times the *square* of the velocity - has to be dissipated as
heat or converted into some other form of energy. The velocity may be
relatively small, but the mass is vast.
RF
John Harshman
A third topic! And back to species concepts. We can of course never be
sure we have identified fossil species, and I'm not even sure we can
quantify our uncertainty. But I know nothing about that paper.
>>> Or is there some better foundation? What I basically want to know is
>>> what knowledge people think that statistical methods actually deliver.
>>> What is it that we know when we have applied them that relies on them
>>> and not, say, a conflation of homology with analogy?
>> At this point I'm unsure what statistical methods we're talking about.
>> Phylogenetic or comparative? Are we inferring a tree or are we using an
>> existing tree to learn about evolution? I had thought the latter, but
>> you seem to be talking about the former.
>>
> I am, and hence you can ignore me.
I can at least register my confusion. Now if you want to discuss the
argument in favor of statistical methods in phylogenetics, we could do
that. As I mentioned at pizza, that particular battle has been more or
less over for a few years now, but we could revisit it.
John Wilkins
<jhar...@pacbell.net> wrote:
No, it's something I have to struggle through, myself. The battle may
be over among the cognoscenti, but I am a philosopher, and hence both
ignorant and confused.
John Harshman
thought of as a likelihood method. There are now at least two likelihood
models that are guaranteed to give identical results to parsimony. So if
not statistical methods, then what?
John Wilkins
<jhar...@pacbell.net> wrote:
Parsimony as a likelihood method? You'll have to explain that one, I
think, or point me at a paper that does.
Three item analysis is purported to be an alternative. I am unsure how
these various approaches can be tested, though. It's not like we
actually *have* a tree to test them against, independently. Constructed
trees don't do it for me, I'm afraid.
John Harshman
> In article <nKGdne5NoY-...@giganews.com>, John Harshman
> <jhar...@pacbell.net> wrote:
>
>> John Wilkins wrote:
>>> In article <qJ-dnUiO8cU...@giganews.com>, John Harshman
>>> <jhar...@pacbell.net> wrote:
>>>
>>>> Now if you want to discuss the
>>>> argument in favor of statistical methods in phylogenetics, we could do
>>>> that. As I mentioned at pizza, that particular battle has been more or
>>>> less over for a few years now, but we could revisit it.
>>> No, it's something I have to struggle through, myself. The battle may
>>> be over among the cognoscenti, but I am a philosopher, and hence both
>>> ignorant and confused.
>>>
>> Far be it from me to interfere with either. Of course, parsimony can be
>> thought of as a likelihood method. There are now at least two likelihood
>> models that are guaranteed to give identical results to parsimony. So if
>> not statistical methods, then what?
>>
> Parsimony as a likelihood method? You'll have to explain that one, I
> think, or point me at a paper that does.
Dang. Now I have to use my memory, and it's a problematic device. Oh,
wait, there's Google. The first such model was by Tuffley & Steel, Bull.
Math. Biol. 59:581–607, 1997. That should be enough to get you started.
Basics: parsimony can be considered as a likelhood method that assumes
each site on each branch is evolving under its own substitution
parameter. That is, it's a model in which the number of parameters
equals the number of sites times the number of branches (times the
number of possible transitions per site, if your characters aren't binary).
> Three item analysis is purported to be an alternative. I am unsure how
> these various approaches can be tested, though. It's not like we
> actually *have* a tree to test them against, independently. Constructed
> trees don't do it for me, I'm afraid.
Is it your contention that we don't know anything reliable about
phylogeny? That would make a lot of systematists very, very annoyed.
Frank J
> > There seems to be a growing gap between the list of creationist claims
> > listed on the TO archive and more recent and 'sophisticated' claims. Not
> > that all creationist claims need to be listed or answered but the list is
> > out of dated.
>
> > For example creationists have been working on a way for the flood to
> > account for all plate tectonic activity at least since the last super
> > continent Pangea. See:
>
>
> > For a more technical approach:
>
> >http://www.icr.org/article/catastrophic-plate-tectonics-flood
>
> > This situation wouldn't matter so much if the TO archive wasn't so well
> > respected as an important source of evolution/creation information. I
> > recently read a book called 'Why Evolution Works (And Creationism Fails)'
> > by Matt Young and Paul K. Strode. The book is excellent and completely up
> > do date on the latest developments in evolution. However, one section that
> > was very disappointing started on page 157 titled 'Creationist Quibbles'
> > about the age of the earth. The authors mention moon dust and the decay of
> > the earth's magnetism as common creationist arguments. Not that some
> > creationist still don't use these arguments but for the most part they
> > have been abandoned. A book that is up to date on everything else should
> > be up to date on the major creationist arguments.
>
> > Don't get me wrong - the newer claims don't have any more validity than
> > the abandoned ones. They are more sophisticated however and do try to
> > account for more data.
>
> > I don't have an answer for this situation or the resources to address it
> > but I think it is important.
>
> I would be glad to keep adding updates if someone would hire me to do so
> as a full-time job.
If everyone who criticizes ID/creationism would put their $ where
their mouths are and join NCSE, they could afford to hire you full
time.
In the meantime, it should be a relatively simple effort to remind
people that many (most?) anti-evolution activists these days do *not*
make most of those claims, particulary the YE ones. And even the ones
that do make those claims have been unable to, dispite decades of
finagling, get the evidence to converge on any particular age range.
>
> --
> Mark Isaak eciton (at) earthlink (dot) net
> "It is certain, from experience, that the smallest grain of natural
> honesty and benevolence has more effect on men's conduct, than the most
> pompous views suggested by theological theories and systems." - D. Hume- Hide quoted text -
>
> - Show quoted text -
John Wilkins
<jhar...@pacbell.net> wrote:
> John Wilkins wrote:
> > In article <nKGdne5NoY-...@giganews.com>, John Harshman
> > <jhar...@pacbell.net> wrote:
> >
> >> John Wilkins wrote:
> >>> In article <qJ-dnUiO8cU...@giganews.com>, John Harshman
> >>> <jhar...@pacbell.net> wrote:
> >>>
> >>>> Now if you want to discuss the
> >>>> argument in favor of statistical methods in phylogenetics, we could do
> >>>> that. As I mentioned at pizza, that particular battle has been more or
> >>>> less over for a few years now, but we could revisit it.
> >>> No, it's something I have to struggle through, myself. The battle may
> >>> be over among the cognoscenti, but I am a philosopher, and hence both
> >>> ignorant and confused.
> >>>
> >> Far be it from me to interfere with either. Of course, parsimony can be
> >> thought of as a likelihood method. There are now at least two likelihood
> >> models that are guaranteed to give identical results to parsimony. So if
> >> not statistical methods, then what?
> >>
> > Parsimony as a likelihood method? You'll have to explain that one, I
> > think, or point me at a paper that does.
>
> Dang. Now I have to use my memory, and it's a problematic device. Oh,
> wait, there's Google. The first such model was by Tuffley & Steel, Bull.
> Math. Biol. 59:581�607, 1997. That should be enough to get you started.
Thanks. I hope I have the math to deal with it.
>
> Basics: parsimony can be considered as a likelhood method that assumes
> each site on each branch is evolving under its own substitution
> parameter. That is, it's a model in which the number of parameters
> equals the number of sites times the number of branches (times the
> number of possible transitions per site, if your characters aren't binary).
>
> > Three item analysis is purported to be an alternative. I am unsure how
> > these various approaches can be tested, though. It's not like we
> > actually *have* a tree to test them against, independently. Constructed
> > trees don't do it for me, I'm afraid.
>
> Is it your contention that we don't know anything reliable about
> phylogeny? That would make a lot of systematists very, very annoyed.
>
My contention is that we had better know what it is that we think these
things are telling us. And isn't the default state of a systematist to
be very, very annoyed?
John Harshman
Hey, I don't have the math to deal with it. I just believe the parts
they put into words.
>> Basics: parsimony can be considered as a likelhood method that assumes
>> each site on each branch is evolving under its own substitution
>> parameter. That is, it's a model in which the number of parameters
>> equals the number of sites times the number of branches (times the
>> number of possible transitions per site, if your characters aren't binary).
>>
>>> Three item analysis is purported to be an alternative. I am unsure how
>>> these various approaches can be tested, though. It's not like we
>>> actually *have* a tree to test them against, independently. Constructed
>>> trees don't do it for me, I'm afraid.
>> Is it your contention that we don't know anything reliable about
>> phylogeny? That would make a lot of systematists very, very annoyed.
>>
> My contention is that we had better know what it is that we think these
> things are telling us. And isn't the default state of a systematist to
> be very, very annoyed?
You might well think so, but it may be a more local phenomenon, i.e. you
may have a biased sample. I do believe I know what I think these things
are telling us, by the way.
John Wilkins
<jhar...@pacbell.net> wrote:
Oy! I just got it! Talk about your opacity! -
"Definitions 1 (Phylogenetic trees, characters) A phylogenetic tree is
a tree T = (V (T ), E(T )) having no vertices of degree two and such
that each leaf (degree one vertex) is given a unique label from
{1,...,n}, where n is the number of leaves of T. We say that T is a
tree on n leaves, and write [n] for {1,...,n}. Where convenient, we
identify each leaf with its label. If every internal (non-leaf) vertex
of T has degree three, we say that T is binary. In the case of rooted
trees, we allow the root to have degree two.
A function ? : [n] ??? S, where S is a set of r states, is an (r-state)
character. When r = 2, ? is said to be binary. A function ?� : V (T)
??? S is called a state function for T; if ?� is such that ?�|[n] = ?
(that is ?� agrees with ? on the leaves of T) then ?� is called an
extension of ? (on T)."
From there it gets complicated...
I'll get back to you in about three years.
>
> >> Basics: parsimony can be considered as a likelhood method that assumes
> >> each site on each branch is evolving under its own substitution
> >> parameter. That is, it's a model in which the number of parameters
> >> equals the number of sites times the number of branches (times the
> >> number of possible transitions per site, if your characters aren't binary).
> >>
> >>> Three item analysis is purported to be an alternative. I am unsure how
> >>> these various approaches can be tested, though. It's not like we
> >>> actually *have* a tree to test them against, independently. Constructed
> >>> trees don't do it for me, I'm afraid.
> >> Is it your contention that we don't know anything reliable about
> >> phylogeny? That would make a lot of systematists very, very annoyed.
> >>
> > My contention is that we had better know what it is that we think these
> > things are telling us. And isn't the default state of a systematist to
> > be very, very annoyed?
>
> You might well think so, but it may be a more local phenomenon, i.e. you
> may have a biased sample. I do believe I know what I think these things
> are telling us, by the way.
>
Seriously, what are they telling you?
John Harshman
Right. All he's saying is that a tree has three branches meeting at each
internal node and that each terminal node is, well, terminal. And that
he has numbered the terminals.
> A function ? : [n] ??? S, where S is a set of r states, is an (r-state)
> character. When r = 2, ? is said to be binary. A function ?– : V (T)
> ??? S is called a state function for T; if ?– is such that ?–|[n] = ?
> (that is ?– agrees with ? on the leaves of T) then ?– is called an
> extension of ? (on T)."
Too many ???s for me to guess what that means.
> From there it gets complicated...
Do you know Steele, by the way? His papers on phylogenetics are
legendary for their opacity. Mathematicians may find it difficult to
communicate with biologists, much less philosophers.
> I'll get back to you in about three years.
Like I said, just read the comprehensible part.
>>>> Basics: parsimony can be considered as a likelhood method that assumes
>>>> each site on each branch is evolving under its own substitution
>>>> parameter. That is, it's a model in which the number of parameters
>>>> equals the number of sites times the number of branches (times the
>>>> number of possible transitions per site, if your characters aren't binary).
>>>>
>>>>> Three item analysis is purported to be an alternative. I am unsure how
>>>>> these various approaches can be tested, though. It's not like we
>>>>> actually *have* a tree to test them against, independently. Constructed
>>>>> trees don't do it for me, I'm afraid.
>>>> Is it your contention that we don't know anything reliable about
>>>> phylogeny? That would make a lot of systematists very, very annoyed.
>>>>
>>> My contention is that we had better know what it is that we think these
>>> things are telling us. And isn't the default state of a systematist to
>>> be very, very annoyed?
>> You might well think so, but it may be a more local phenomenon, i.e. you
>> may have a biased sample. I do believe I know what I think these things
>> are telling us, by the way.
>>
> Seriously, what are they telling you?
At this point I have to be told what you mean, specifically, by "these
things". But my claim is that we do have a tree to test methods against;
that is, there are certain nodes we are effectively certain of.