Predictivity vs Useful

Ken Kinman kinman2 at YAHOO.COM
Mon Feb 14 07:57:51 CST 2005


Michael (and other strict cladists),
     You think I confuse useful and predictive, apparently because the definition you gave for useful is very narrow.  To me usefulness is a broad term, and predictivity is just one of the many useful aspects of classifications (useful as meaning "full of uses", i.e. multipurpose).  If you concentrate too heavily on predictivity, the usefulness of classifications inevitably suffer on other important fronts (relative stability, historical continuity, balance, consistent "pigeonholing", not so complex that they are an enormous pain to learn and remember, not to mention having to relearn when a new classification arises due to a new "cladogram du jour").

     The goal should NOT be to simply maximize predictivity (at the expense of other uses), but to maximize the overall level of usefulness.  In the case of Reptilia, high homoplasy inherently weakens predictive power (as do "scrappy" fossil taxa), so much so that the short-sighted pursuit of more and more predictivity is largely an exercise in futility.  The predictivity you can achieve with Diptera (not heavily dependent on fossil evidence) will always be much higher than with most other taxa, so Hennig's strictly cladistic approach fares much worse (horribly so with amphibians and reptiles).  Furthermore, degree of homoplasy and scrappiness of fossils are just two factors that affect how much emphasis on predictivity is optimal for a given group.  And remember that predictivity is not only powerfully positive when it works, but powerfully negative when it doesn't.  It cuts both ways!!  The trash heap of discarded (or should-be discarded) clades among the dinosaurs alone has just started to become apparent.

     Going back to that "Stegocephalia" page on Tree of Life, to me that whole cladogram continues to be Class Amphibia.  The internal cladistic topology seems to change from year to year, or worker to worker, but I can simply modify the order of the taxa and recode them.  The contents remain relatively stable, even as new fossils are added.  The meaning and content of Tetrapoda are pretty constant, not shifting from year to year because Gauthier arbitrarily defined it as a crown group.  If you were a "tetrapod", but didn't leave any descendants, many don't make the cut any more (UNLESS they find a worker with a cladogram that has them once again cladistically splitting off between extant amphibians and the Amniota, then you are back in the Tetrapod Club for the time being, at least according to those who adhere to that particular cladogram).

      As one who believes in the benefits of cladistic analysis (when done well), it really pains me to have to argue so strenuously against strictly cladistic classification (which works well much of the time).  But I have no choice, because the excesses of strict cladism (especially the mindless assault on formal paraphyletic taxa) is ruining the classifications of so many taxa in so many different ways.  Ironically even their predictivity can be adversely affected when sloppy predictions are made in spite of the risks of high homoplasy, not to mention poor character choice, incorrect rooting, and other pitfalls to which strict cladism is particularly vulnerable.

     The main purpose of a classification should be information storage and retrieval (which goes way beyond just pigeonholing taxa).  For the specific purpose of increasing predictivity, you can and should use cladograms, and don't confuse the limited uses of your "cladificiations" as a proper replacement for a multipurpose classification.  And please stop redefining taxa like Tetrapoda and Amphibia to meet your narrow goals.  It's narrow-minded, destabilizing, and a great way to get your colleagues to tune you out with glazed over eyes.  Most of all, stop confusing the benefits of strict cladifications (which are limited) with the benefits of cladistic analysis.  Finally, those who seem to think I am either incapable of learning and understanding modern methods, or on the other hand, some kind of manipulative sophist who gets pleasure tricking people into substandard ways of reasoning, neither is the case.  If there is any sophistry going on here, it is those who have managed to convince so many that paraphyly is unnatural (or are perhaps themselves incapable or unwilling to learn and understand methodologies that might improve upon their own).
            ---- Ken Kinman




More information about the Taxacom mailing list