[Taxacom] Panbiogeography and molecular systematics critique

[John Grehan]

Hi Jason,

Some interesting points. I decided for simplicity to separate this into a
separate posting.

“has left you holding an untenable arcane view of biogeography whilst
ignoring the last 60 years of advances in phylogenetics.”

Ignore what advances?

“ attempt to neutralize any possibility of criticizing Croizat or
panbiogeography by labelling as suspect ... molecular phylogenetics”

While I might regard molecular phylogenies as phenetic, they are used
extensively in panbiogeography studies, including my studies. This
criticism is therefore invalid.

“fossil calibration (using minimum age as a the fine edge of the wedge to
push arguments such as confusing point calibration with log distributions),”

Actually we don't confuse them. The minimum age issue is one – and it is
absolutely clear that a minimum cannot be made into a maximum. The log
distributions are another – and it is absolutely clear that these are just
guesses. This criticism is therefore invalid.

“molecular clocks (your definition is "strict clock" from the 1960´s but
you don´t care),”

As far as I am concerned,molecular clocks are problematic when treated as
absolute.

“generally weaponizing the widely acknowledged limitations of all of these
and more fields.”

Don't understand that one. Many molecular studies admit that their findings
are only minimums and then go on in their discussions to treat them as
actual or maximal dates.

“Firstly because Croizat quite simply rejected cladistics, and if you
accept cladistics you are a vicariance cladist by definition and arguing
otherwise is rejecting reality”

Panbiogeography uses cladistic phylogenetics and so in this respect accepts
cladistics. You would have to explain how the methodology of
panbiogeography is the same as 'vicariance cladistics' for me to respond to
that one.

“bcause CP rejects dispersal”

No it does not. Repeating this false claim (without evidence) is a bit like
Trump repeating election fraud without evidence. Science should be better
than that.

“Trying to relabel cases of dispersal as "range expansion" is sophistry
aimed at making panbiogeography unfalsifiable.”

Ecological dispersal is an observable empirical phenomenon that does indeed
cause range expansion. Attributing the origin of allopatry to one off
(usually) chance dispersal is another matter – purely conjectural. Not
empirical.

“Biogeographic and fossil calibrations are complementary, not competing,
approaches to constraining molecular clock analyses, providing alternative
constraints on the age of clades that are vital to avoiding circularity in
investigating the role of biogeographic mechanisms in shaping modern
biodiversity." (2016).

I couldn't agree more – fossil calibrations give minimum dates, tectonic
calibrations give approximations of actual dates. Those are definitely
complementary. If a tectonic date is older that is complementary to the
minimum date by providing additional information on an older origin which
is consistent with the fossil calibration not precluding an older origin.

I don't think any biogeographer could disagree with this statement,

As above, I for one do not.

“he unfalsifiability of CP in al its glory. If something disagrees with CP,
CP is not at fault because it is perfect.”

Jason, what you and probably others are looking for here is an umpire who
can stand in and call out 'true' or 'false'. In this situation we are
dealing with competing research programs. Arguments in the philosophy of
science have been made that research programs include a hard core of
unfalsifiable assumptions that are necessary to make their research
programs work. Simple falsification falls down in this situation. There is
a nice paper on this problem for biogeography if anyone is interested.

“ML or Bayesian models do not assume point dates but a probability
distribution into the past. Please stop confusing the terms.”

Actually it is molecular systematists who seem to do this when they use
their 'probability' estimates (arbitrary guesses') to present dates that
are then used to argue for chance dispersal because they are younger than
potential tectonic events (in other words the divergence estimates are
treated as actual ages – otherwise they could not be used to exclude
earlier origins and earlier events).

“A molecular clock model does not assume monotonic (constant) rates within
genes or within lineages nor are clocks assumed a priori.  This is a straw
man argument.”

Well then, this goes back to Smissen et al 2005 “If this [tectonic
calibration] was then extrapolated to the tribal-level divergences (around
0.4 substitutions per site) within the Caryophyllaceae that we used as a
calibration for our dating, then the Caryophyllaceae are  projected to be
in excess of 10 billion years old.” This is true only if there is a
monotonic rate throughout the Caryophyllaceae

I can´t believe that this is a mistake when you guys selectively quote dos
Reis et al 2016:

“Thus, while the clade ages generated in Bayesian studies are extremely
precise and have excellent statistical support, because of the problem of
he priors they can be ‘grossly inaccurate’ (dos Reis et al., 2016) and
could be tens or even hundreds of millions of years too young.”

What's the problem here? This is what they said. Please explain.

Cherry picked quote: “Errors in the time prior and in the rate prior can
lead to very precise but grossly inaccurate time estimates. Great care must
always be taken in the construction of fossil calibrations and in the
specification of priors on times and on rates in a dating analysis.”

Not quite the same meaning, in particular with the added sentence ending
“tens or hundreds of millions of years…”. That is a misquote if I ever saw
one.

You lost me a bit on what was misquoted. Does Reis et al do say “fossil
ages provide good minimum-age bounds on clade ages, but assuming that clade
ages are the same as that of their oldest fossil is unwarranted and
incorrect” I agree fully.

“... NZ Alpine fault dislocation...an approximation of age for the
correlated groups.”

You are assuming as axiomatic that the taxa on either side of the fault
have been unable to cross since the tectonic event.

Not sure about axiomatic, but its evidence they have not – or at least show
no evidence of that, and keep in mind that the fault is a transform and the
disjunction is not only across the fault, but offset by sometimes hundreds
of km. This is an example of a tectonic calibration. Perhaps it is an
artifact resulting from some other event other than transform fault
movement. If so, no one has demonstrated that to be the case.

“As Hauenschild et al. (2018) observed, clade ages that are too young may
be calculated ‘based on too narrowly set priors, and vicariance thereby may
be erroneously ruled out’.”

Again, cherry-picked, out of context quote.

But there is nothing out of context about this. And none of the additional
quotes change that.

Shocker, they find both mechanisms at play. I can´t wait to see how you
ignore this part of the paper whilst using the bits that you agree with
(usually the necessary acknowledgement of study limitations).

Nothing shocking at all. Many molecular clock papers conclude both
vicariance and chance dispersal. Never suggested otherwise. The issue, as
you imply, is how they reach their conclusions. As I do not have a copy of
the paper I cannot comment on that here.

This is a pattern. I have kept correspondence for the last 10 years on
these exchanges and it is always the same rigmarole. But I still bite
because it is important to warn newcomers.

In science some issues are long standing, and by their nature, often
repetitive. For the 'newcomers', some summary points:

   1.

   Panbiogeography accepts dispersal as an ecological process of range
   expansion or persistence.
   2.

   Chance dispersal explanations for allopatry are often based on
   assumptions about phylogenetic age that may be underestimated.
   3.

   Chance dispersal explanations of allopatry are always problematic as
   they involve mutually exclusive requirements of isolation and accessibility.
   4.

   Tectonic correlation represents evidence of a historical relationship
   and origin that may be contested by competing research programs.
   5.

   Fossil calibrated divergence ages are always minimum estimates
   6.

   Priors of how much older are always guesses
   7.

   A molecular divergence estimate could be seen as a falsifier of younger
   origins just as a tectonic calibration could be seen as a falsifier of
   younger origins.
   8.

   A fossil calibrated molecular age is consistent with an older tectonic
   correlation age since the former does not exclude the latter.
   9.

   Many molecular papers are problematic because their biogeographic
   reconstruction treats molecular ages as actual or maximal when they are not.

May 2021 continue to be a good year for systematic and biogeographic debate.

Cheers, John Grehan