Philosophy of Systematics

Review of Between System and Poetics, William Desmond and Philosophy After Dialectic

by Brendan Sammon

Unable to embed this for immediate viewing, so if you would like to read it you must download it.

How do fossils influence estimates of phylogeny?

by Ross Mounce

[Conference abstract]
Presented at the 3rd International Palaeontological Congress (IPC3)
Co-authored with my supervisor Matthew A Wills

Phylogeny is the grounding for most macroevolutionary and comparative studies, but conclusions are only asgood as the... more Phylogeny is the grounding for most macroevolutionary and comparative studies, but conclusions are only asgood as the accuracy of the cladograms on which they are based. Neontologists often discount the importance of fossils, claiming that their incompleteness makes them obfuscate or positively mislead cladisticinference. Given the importance of paleontological data in macroevolutionary work, therefore, do we need tobe especially cautious? Initial studies of 45 predominantly vertebrate data sets demonstrate that adding fossils is no more likely to change the reconstructed tree than adding living taxa. Moreover, there is no evidence that the presence of fossils decreases phylogenetic resolution. Do preliminary findings for vertebrates hold true for other higher taxa? And can we predict when fossils are likely to have their biggest impact? In general, we predict this will be most marked in volatile trees with uneven taxon sampling, and with large extinct branches. A more pressing question now is whether the fossil record biases the preserved sample of characters in such a way that the trees inferred from them are likely to differ from those that would be derived from more entire data sets. For example, all but the most exceptionally preserved arthropod fossils lack details of the appendages (antennae, mouthparts and legs, etc.) – precisely those characters thatare usually cited as most diagnostic or pivotal in phylogenetic analyses. Preliminary results show no systematic biases. However, we also investigate the putative phenomenon of stemward slippage resulting from the early taphonomic loss of the most derived characters.

Congruence
 between
 cranial
 and postcranial
 characters
 in 
vertebrate
 systematics

by Ross Mounce

[Conference abstract]
Co-authored with my supervisor Matthew A Wills.
Presented at Hennig XXIX.
Full research article to be submitted asap

Tests 
of 
partition
 homogeneity 
are 
routinely
 applied
 to 
matrices
 amalgamating
 data 
from 
more 
than 
one... more Tests 
of 
partition
 homogeneity 
are 
routinely
 applied
 to 
matrices
 amalgamating
 data 
from 
more 
than 
one 
molecular 
marker, 
or 
combining 
molecular
 and
 morphological
 evidence.
 Incongruent
 phylogenetic
 signals
 from
 different
 partitions
 are
 variously
 explicable
 in
 terms
 of
 different
 evolutionary
 histories
 (in
 the
 case
 of
 genes),
 different
 selective
 pressures,
 and/or
 different
 signal
 strengths
 and
 levels
 of
 noise.
 However,
 this
 approach
 has
 rarely
 been
 applied
 to
 partitions
 of
 morphological
 data
 sets.
 Here
 we
 investigate 
the
 homogeneity 
of
 signal 
from 
cranial
 and
 postcranial
 characters 
in
 over 
60
 published 
morphological
vertebrate 
data 
matrices.
Cranial
 and 
postcranial
 characters 
are
 often 
claimed
 to 
be
 under 
differing 
magnitudes 
and
 directions
 of 
selection.
 We 
test 
this
 assertion
 here, 
not
 least 
because
palaeontologists
 often 
have access 
to 
strongly 
biased
 samples
 of
 characters
 (or 
lack
 data
 from 
one 
partition 
altogether).
 Firstly, 
we
 compared
 levels
 of
 homoplasy
 and
 synapomorphy
 in
 cranial
 and
 post‐cranial
 partitions.
 Secondly
 we 
applied 
two 
randomization 
tests
 of
 congruence 
based 
on
 numbers
 of
 steps,
namely
 the
 incongruence
 length
 difference
 (ILD)
 test,
 and
 topological
 incongruence
 length
 difference
 (TILD)
 test. 
Finally, 
we 
explored
 a 
new
 randomization
 approach
 using 
tree‐to‐ tree
 distance
 metrics 
rather
 than 
tree
 length.
 We
 observed that 
the 
TILD 
test 
suffers 
an
 especially
 high
 type
 I
 error
 rate
 and
 appears
 to
 be
 insensitive,
 whilst
 the
 ILD
 and
 tree
 distance
 metric
 tests
 have
 a
 low
 correlation
 because
 they
 quantify
 valid
 but
 different
 aspects
 of
 phylogenetic
 signal.
Our 
results
 show
 that
 it
 is
 not
 uncommon 
for 
cranial
 and
 postcranial 
characters
 to
 contain
 significantly
 incongruent
 signals.
 Also, 
because 
missing
 entries
 are
 often
 preferentially
 distributed
 in
 one
 or
 other
 partition,
 this
 may
 have
 implications
 for
 the
 way 
in
 which 
incomplete 
fossils 
are 
resolved.