Positivism in evolutionary science

[P. Hovenkamp]

> James Francis Lyons-Weiler wrote: a lot
>> Tom DiBenedetto wrote: a lot

I write:

What seems to be the issue here is a problem of demarcation: how to distinguish
real patterns from accidental ones.

As an example, take your favorite dataset and use it to generate a cladogram,
using your favorite parsimony-program. All fine.
Now, drop the dataset on the floor (metaphorically speaking), reassemble it in
a hurry and again produce a cladogram. Does your favorite program say: sorry,
this data was not reassembled correctly? No, it just churns out another
cladogram, and the only indication that you'll see that your data were maybe
not in their original state is a lowered consistency index.

Now, Tom DiBenedetto seems to be convinced that data are never dropped on the
floor, while James Francis Lyons-Weiler is convinced that sometimes, data are
dropped and reassembled (still metaphorically speaking).  How do we tell the
difference. having only the reassembled data at hand? That is an interesting
and active field of investigation, and the only clear and undisputed result so
far is, in my opinion, that the outcome of a parsimony analyis on its own
(including values for CI, RI) is no great help (short of finding a CI of 1).
P. Hovenkamp
Rijksherbarium/Hortus Botanicus
The Netherlands
hovenkamp at rulrhb.leidenuniv.nl

This space intentionally left blank