Abstracting and Seeing
Natural Classifications are often discussed as separate issues to our own experiences and observations. Debate rarely touches on the issue of what a natural classification means to a taxonomist or systematist. If natural classifications, like homologies, are supposedly "abstract things" that taxonomists make up in order to make sense of the world, why do we see them? Can we see abstractions or are we doing something else?
Abstractions are hypotheses conjured up in order to formulate a system or artificial "classification" - that is, categorizing things based on a set of subjective rules or characteristics. We may state for instance that anything with six legs is an insect, and so forth. We may view the world through our chosen abstraction and thereby simplify what we are observing. We call this abstracting. It is not observing but simply glancing, looking for particular characteristics rather than observing the whole organism. Artificial classifications are abstractions. An invertebrate, an organism without vertebrae or a spinal column, are unobservable. Reptiles and unicorns are also unobservable but we know what they are and what characteristics define them. Abstracting is not seeing or observing, but simply employing a hypothesis to do the job of "seeing" for you.
Not being able to "see" something does not mean it does not exist. At times artificial classifications contain real (evolutionary) characteristics. Microscopic organisms or electrons too are unobservable to the naked eye, but can be discovered through tools such as microscopes or experimentation. Invertebrates, like reptiles are not only unobservable, they are also undiscoverable.
Seeing is what taxonomists do in order to know an organism. When we see, we do not go through a predefined list of characteristics. Seeing a dog does not make us recite a list of mammalian characteristics. We just see and recognize it. This is what taxonomists do.
Taxa define themselves rather than the other way around. A white blackbird for instance is still a blackbird even if an artificial classification lacks some characteristics or contains a few conflicting ones. An artificial classification for instance may place the American Robin into the Muscicapidae (Old World flycatchers) because they are similar to the European Robin. We can, however, see that the American Robin is the same as a Blackbird, even though they are similar to European Robins. That sameness is a relationship or homology. Natural classifications are defined by homologies rather than the similarities between arbitrary sets of characteristics as in artificial classifications or systems.
The difference between seeing and abstracting is one that divides the biological community. That division runs deep, especially in the case of total similarly versus relationship.
Homology (Relationship) and Phenetics (Similarity)
Homology is about relationship
When the same characteristic manifests itself in other taxa (e.g., a forearm appearing in the wing of a bat), we have found a homologue (sameness). Homology is about the relationships between two homologues when compared to a third (the Cladistic parameter). Homologies are discoverable and tell us that taxa are related.
Phenetics is about similarity.
When a similar characteristic manifests itself in other taxa (e.g., a leaf ratio of 2:3), we have found a similarity. Phenetics is about the overall similarity between two homologues only. Similarities are generated, that is proposed, and predict that taxa may be related based on a measure of confidence.
Homology, then, is evolutionary, as it discovers homology. Phenetics is non-evolutionary as it only predicts degrees of similarity. The differences are evident and should be construed as a negative criticism on our behalf.
Phenetics is simply an artificial classification system. Even though it endeavours to use homologies, it can never discover homologies only propose them. The tools used in phenetics are borrowed from mathematics (statistics) in order to replace what we do naturally, that is observe. In this sense a number of techniques are phenetic. These range from DNA barcoding to Bayesian analysis to cladistic analysis (optimisation etc.). They are based on the fundamental rule of similarity and hierarchy mixed with ad hoc proposition of evolutionary mechanisms, many of which as at best hypothetical.
Homology (monophyly) is a natural classification system. It is a discovery based on sameness and similarity. The tools we use in homology are unsophisticated and at times phenetic (i.e., relying on similarity). Since homology is discoverable and can be observed, taxonomists have been relying on their own intuition to find out what these natural groups are by studying homologies. The same taxonomic method is used today, but is not as thorough as one would hope. Phenetics, however, has helped develop a number of tools that imitate what we do naturally, in order to test whether our groups are truly groups (monophyly). Where our objections lie is how these tools are used. Rather than use phenetic tools to test our hypotheses, many are replacing the practise of testing with observing, without recourse to our own experience. In other words we are allowing algorithms (no matter how good) to do the seeing for us.
All numerical or phenetic methods are secondary to our own knowledge. This should not be interpreted as a Jacobian outburst, but rather as a cautionary statement. Already many systematists are not looking at their organisms and proposing hypotheses of their groups. Instead they are accepting these tests as hypotheses. In our view phenetics, as a system for testing potentially monophyletic taxa, are becoming primary in taxonomy and systematics. This has resulted in non-taxonomic practises to govern taxonomy (i.e., DNA Barcoding, molecular systematics etc.). Our grudge is not against phenetics, a field worthy in its own right, but at those who feel that it is all there is to taxonomy and systematics.
Given the non-evolutionary nature of phenetics, namely its inability to find or even recognize monophyletic groups, why do we need it in systematics and biogeography? Its simplistic aim of proposing similarity is convoluted by numerous algorithms, neither of which advance our field or our knowledge of the natural world. The Great Phenetic Revival returns us to an age old argument fought at the end of the 18th century, namely artificial classifications versus natural classifications. Nothing in our field today has advanced our understanding of classification (i.e., homology and monophyly) since 1858, a point that we will return to later.