Thanks, Ivan. I'll have to re-read that paper, which is definitely an
important one.
I would say that the main issues I discuss are much simpler and much
more basic. The problem with the within a configuration tests is a
special case of a well known (since the early days of GMM) problem in
Procrustes data: they are great when one analyses all the superimposed
points together, but one cannot interpret variation at specific
landmarks. Nobody I know argues about this. Then, however and in the
context of modularity/integration, one might argue on a practical ground
whether this makes or does not make a difference. We're still learning
on this, but the problem is there and probably the answer is
data-specific. This is a secondary issue, however, compared to the
really basic ones:
- Regardless of how one treats the landmarks (and semilandmarks) having
more points does not guarantee accuracy and a large set of points cannot
be used as the meter of comparison to justify why one is using many
points (invalid tautology).
- Generalizing from a few specific cases is unlikely to be accurate.
- Semilandmarks may be very useful but, regardless of the clever maths
used to manipulate them, they are not the same as landmarks: for
instance, sliding using the minimum bending energy does not 'create'
biological equivalence. Sliding, for me, adds a layer of complexity and
further alters the covariance structure using pure maths with (at least
in the methods I know) no biological model behind it.
- Yet, sliding may help in other ways. I am (almost) agnostic on this.
Maybe in some cases it can produce geometric equivalence, but one cannot
cherry pick ad hoc examples and geometric equivalence does not guarantee
biological equivalence (which will be, in any case, specific to the
study question, as beautifully said many times by Oxnard & O'Higgins).
Indeed, a cucumber and a holoturia may have an almost identical geometry
but hardly any homologous feature.
This is why I don't see using high density morphometrics as the default
option, but I am not against it, as long as it is needed and one is
honest about the cons (not just the pros). Although my criticisms are
inspired mainly by a few specific papers, those same papers have also
interesting and useful ideas. The misleading overstatements, however,
should not be there ... but I am sure I made similar slips in some of my
articles, with the blessing that, not being a leading morphometrician,
nobody reads and bothers about those!
Cheers
Andrea
On 23/09/2020 16:35, ivan perez wrote:
Dear all
I remember that the problem discussed by Andrea is described in a very
interesting (and beautiful) way by Richtsmeier et al. (2002). A classic
paper, it is important to read it again
All the best from La Plata
ivan.
Richtsmeier, J. T., Burke Deleon, V., & Lele, S. R. (2002). The promise of
geometric morphometrics. *American Journal of Physical Anthropology*
*119*(S35),
63-91.
El mié., 23 sept. 2020 a las 5:36, andrea cardini (<[email protected]>)
escribió:
Dear All,
this is out now (first link allowing to get a pdf but only for a day;
second link for those who have subscriptions - of course, send me an
email if you don't make it to get a reprint either way).
https://rdcu.be/b7zXr
https://link.springer.com/article/10.1007/s00435-020-00499-w
This paper is about questions such as: does "high density morphometrics"
really equate to accuracy? Does sliding make semilandmarks into
landmarks? Is geometric correspondence the same as biological
equivalence? Is there no problem with the 'within a configuration' tests
of modularity/integration?
I believe the simple answer to all these questions is: NO. However, the
article is neither against semilandmarks nor against testing for
modularity/integration. It's against unsupported misleading claims and
overstatement. Many methods can be useful (if important for the specific
study question) even with limitations, but the limitations should be
stated and not be overlooked.
I take the chance to thank again the editors, reviewers and all those
with whom I have discussed these issues over the years (regardless of
how much we agree or disagree).
I hope you're all well, despite the difficult times we're in.
All the best
Andrea
--
Dr. Andrea Cardini
Researcher, Dipartimento di Scienze Chimiche e Geologiche, Università di
Modena e Reggio Emilia, Via Campi, 103 - 41125 Modena - Italy
tel. 0039 059 4223140
Adjunct Associate Professor, Centre for Forensic Anthropology, The
University of Western Australia, 35 Stirling Highway, Crawley WA 6009,
Australia
E-mail address: [email protected], [email protected]
WEBPAGE: https://sites.google.com/site/alcardini/home/main
FREE Yellow BOOK on Geometric Morphometrics:
https://tinyurl.com/2013-Yellow-Book
ESTIMATE YOUR GLOBAL FOOTPRINT:
http://www.footprintnetwork.org/en/index.php/GFN/page/calculators/
SUPPORT: secondwarning.org
--
You received this message because you are subscribed to the Google Groups
"Morphmet" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/morphmet2/dcac1c85-8872-0d60-5cb8-97d7ad956ab6%40gmail.com
.
--
Dr. Andrea Cardini
Researcher, Dipartimento di Scienze Chimiche e Geologiche, Università di
Modena e Reggio Emilia, Via Campi, 103 - 41125 Modena - Italy
tel. 0039 059 4223140
Adjunct Associate Professor, Centre for Forensic Anthropology, The
University of Western Australia, 35 Stirling Highway, Crawley WA 6009,
Australia
E-mail address: [email protected], [email protected]
WEBPAGE: https://sites.google.com/site/alcardini/home/main
FREE Yellow BOOK on Geometric Morphometrics:
https://tinyurl.com/2013-Yellow-Book
ESTIMATE YOUR GLOBAL FOOTPRINT:
http://www.footprintnetwork.org/en/index.php/GFN/page/calculators/
SUPPORT: secondwarning.org
--
You received this message because you are subscribed to the Google Groups
"Morphmet" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/morphmet2/eef8205a-56f9-e361-759a-68a6ca40081b%40gmail.com.
