What is striking about Sokal & Rohlf's "Intelligent Ignorumus" is that it totally undermines our in-built ability to classify, even groups we do not know. Any one from the southern hemisphere would know how to group "Song Birds" based on characters that have not been pointed out to them. It is what we do naturally. Consider the European Robin (Erithacus rubecula), the American Robin (Turdus migratorius) and the European blackbird (Turdus merula). Based on their common names we group them as Blackbird (E. Robin, A. Robin). But if we look at them, it becomes clear that the European Blackbird looks like an American Robin. See for yourself. You don't need a detailed list of pre-defined characters; it is what we do naturally. But Sokal & Rohlf don't think so.
The Intelligent Ignoramus is a simpleton. They are "unprogrammed" with little to no training in biology, meaning the are unable to identify the organisms before them as "bees" and therefore are presumed to lack the ability to classify bees in general. Sokal & Rohlf's test was to assess:
The two technicians managed to find a similar classification of bees. Anyone looking at bees for the 170 hours (as mentioned in the results) would obtain a general knowledge of bee morphology. Obviously this time would be remarkably reduced if a trained taxonomist had pointed out the relevant morphology and the characters that relate various bee groups. Instead of seeing the obvious, Sokal & Rohlf make an enormous blunder. What if the technicians had measured all the specimens, that is quantifying the qualities that helped them group the bees? Quantities can be standardized and therefore automated. In fact, one "... could hire teams of technicians to study the specimens, make the necessary measurements, and record the data and perhaps even select the characters themselves. One step beyond this would be to automate the entire process completely" (Sokal & Rohlf 1970: 318).
The future of taxonomy envisioned by Sokal & Rohlf - groups of "Intelligent Ignoramuses" coding taxa to be processed by numerical methods - is typical of phenetics and their attempts to remove taxonomists from doing what they do best. Now, all of this is starting to sound familiar. "Intelligent Ignoramuses" that can identify and classify taxa without the burden of taxonomic training have reappeared in the guise of DNA Barcoding (identification) and the Phylocode (classification).
Can we re-label DNA Barcoding as Phenetic? Usually phenetic methods or techniques are considered, but rarely do we ever identify phenetic ideas or intentions. The Great Phenetic Revival is the revival not only of phenetic methods but the ideas endorsed by early phenetists like Sokal & Rolf. Read phenetics as Numerical Taxonomy and one quickly realizes that it's about numerical data - quantities, not qualities - and about obtaining such data mechanically and processing it quickly. The taxonomist is not a machine. He or she does not seek to provide measurements. The aim is to discover homologies. The same is true for classifications. They are based on monophyly, not on some general rule of classification. Unfortunately, the Great Phenetic Revival is about the rise of Intelligent Ignoramuses, those that wish to supplant taxonomy and systematics with phenetics under the guise of helping taxonomy. What a frightening thought.
2 comments:
I personally do not see anything wrong in "coding taxa to be processed by numerical methods", and I would not say either that this is typical of phenetics, otherwise even cladistics analyses will be accused of being phenetics! :o
Regarding the DNA barcoding, Rob De salle has a very nice note discussing some of your concerns, basically the distinction between species identification and the discovery of species.
I am not sure if phenetics is a word that fits to the barcoding, I would say that is typological maybe but not phenetic.
Coding taxa or areas (in a matrix) is problematic in both systematics and biogeography (see Williams and Ebach 2004, and Ebach et al. 2003). You mentioned that "... otherwise even cladistics analyses will be accused of being phenetics!" this is a very interesting point you make and one that Dave and I will address soon.
"Phenetic" is a concept that we use to describe Soklal and Sneath's synthesis. Barcoding however is based on similarity and not relationship (two different things entirely) and therefore represents a phenetic world view.
Typology is a term that I should have discussed in the Terminology post. In fact, typology is an evolutionary concept that attempts to break away from any one particular evolutionary mechanism. The typologists wanted to find evidence for evolution based on the relationships between taxa and their homologues (homology) rather than proposing hypothetical ancestor-descendant relationships based on a doctrine or synthesis. Typology is structuralist rather than functionalist and has very little to do with barcoding.
The popular myth that typology is "bad" comes from Mayr who demonized it because did fit into his evolutionary synthesis. Most current definitions of typology are based on such "opinions" rather than on historical facts.
For more info on the history of typology see:
Amundson, R. 1998. Typology reconsidered: Two doctrines on the history of evolutionary biology. Biology and Philosophy 13: 153–177.
Winsor, M.P. 2003. Non-essentialist methods in pre-Darwinian taxonomy. Biology and Philosophy 18: 387–400.
Post a Comment