[Taxacom] [TAXACOM] Systematists as holists

[Bob Mesibov]

Hi, Gordon.

Sorry, I can't help with funding, and I'm not sure who can. A serious
obstacle to getting a global sampling going is that expert fieldwork is
competitive. There are only X people currently capable of doing it, and
Y dollars to pay them.

It isn't yet obvious to me how sampling can be freed from this
competitive arena. Maybe other TAXACOMers have examples? I keep my eyes
open for 'piggyback' opportunities. For example, here in Australia the
Royal Geographical Society of Queensland decided to get all sorts of
folks to go the major lat/long intersections in the State, and document
these 180-odd spots:

http://www.rgsq.gil.com.au/qldbydeg.htm

People, mainly non-expert locals, will be going to these spots and
taking notes and photos. I contacted RGSQ to ask that the spot-visitors
grab flora and fauna as well, and I offered to act as an intermediary
with Queensland Museum for collection protocols and sample curation.

Nothing's happened yet, but even if sampling is incomplete and
irregular, it would be a start in recruiting samplers from the world
outside Expert-dom.

Maybe a better example of distributing taxonomic tasks is
character/sequence analysis. Taxonomists prepare character matrices and
sequence comparisons. They use these to generate classifications and/or
phylogenies using a wide range of software programs. They do this
themselves, and they publish the results as their own work.

For heaven's sake, why? The goal is to produce a working
classification/phylogeny, a least-implausible tree given the inputs. Who
cares whether Fred Bloggs and assistants did it, or an online server
dedicated to the Bayesian analysis du jour, something like the
validation parsers set up by the W3C to check Web-page code?

To truly free up this process, we need a motley group of thinkers to
crystallise around building, testing and tinkering with the analytical
software. These people will argue just as loudly as the cladists and
pheneticists have in past years, with the difference that as part of a
tree-improvement network they will only be competing for results and
in-group recognition, like so many of today's tens of thousands of
open-source programmers, and not for academic prestige and citations.
These analysis enthusiasts will run and program the analytical software
servers.

You, the taxonomist, generate the raw data. You feed it raw online to
the tree-improvement network. You might get a message back saying 'Crap.
Get more data.' You might get a tree back built with a method you don't
like, but which is a less implausible tree than one built with a method
you like. The results from your data might intrigue network members for
technical reasons unrelated to biology, leading to new tweaks and new
directions in inference. The raw data, the programs playing with them,
and the results are all universally available online. The latter feed
into an automatically updated (and expert-moderated) online Tree of
Life.

You also publish your tree with the acknowledgement: 'Generated 11.05
GMT on 26 October 2009 by the Systematics Server'. Hours you would
otherwise have spent learning about and tinkering with mindbogglingly
complex, frequently updated inferential models can instead be spent
generating more data - or discovering new taxa.

For years I've been told that statistical phylogenetic inference is
non-subjective, reproducible, robust, etc. Sounds like machine work to
me. Why waste your time doing it? Why not give it to distributed amateur
specialists who think about nothing else in their spare time, and who
couldn't give a fig whether A is more closely related to B or C, or
whether a clade is monophyletic or paraphyletic?

I hear flame-thrower safeties being thumbed off...
-- 
Dr Robert Mesibov
Honorary Research Associate, Queen Victoria Museum and Art Gallery
and School of Zoology, University of Tasmania
Contact: PO Box 101, Penguin, Tasmania, Australia 7316
(03) 64371195; 61 3 64371195
http://www.qvmag.tas.gov.au/mesibov.html
---