Thursday, April 18, 2013

Typological species (?) and autapomorphic species

This post continuing my series on species will perhaps not do its topic full justice, but unfortunately I currently cannot seem to find the time to write as much.

The term "Typological Species Concept" (TSC) is one that you can hear rather often from taxonomists, generally accompanied by a sneer, but it appears surprisingly ill-defined. Googleing around a bit one can, for example, encounter these teaching materials from Southern Illinois University which describe it as follows: "typological species are defined by similarity to a type specimen or ideal type." This is then dismissed as biologically unjustified and followed by several alternative species concepts.

Defining the TSC like this is in my opinion rather strange because it confuses systematics and nomenclature. No matter how a systematist circumscribes species and no matter what species concept they use to do that, afterwards a name has to be assigned to each of these species, and that is always done with reference to a type specimen. A species that contains the type specimen of, say, Bellis perennis will be called Bellis perennis regardless of whether it is a biological species, a phenotypic cluster species or a phylogenetic species*. Those are simply the rules of nomenclature and not another distinct species concept.

So, be the TSC formally defined as such somewhere or not, I will now describe how we in the business appear to use the term. As mentioned above, it is generally used as an accusation, as in "yeah, that guy used a typological species concept, his treatment is completely useless". What the detractors mean in that case is that the taxonomist in question has used a very unbiological and schematic approach to circumscribing species. Typical criticism would include:
  • Excessive splitting, especially "single character taxonomy" in which a specimen showing any morphological deviation whatsoever is immediately given taxonomic recognition as a variety, subspecies or even species
  • Failure to take the plasticity of the organisms into account, basing taxonomic decisions on characters that are variable under varying environmental conditions; this is often due to the taxonomist not having seen the study group in the field
  • Failure to reflect the hypothesized relatedness of the samples in the classification, in extreme cases assigning varieties that are admitted to be most closely related to each other to separate species because they differ in a character that was arbitrarily considered "important enough" to define the species level
As an example, we might consider a breeding group of daisies in which some populations have a pappus and others don't. Now although there are such cases this happens rarely - usually an entire species of daisies either has a pappus or it hasn't. A competent taxonomist using the Phenotypic Cluster Concept would look at overall variation and conclude that the character "pappus presence" does not correlate with anything else and that the overlap in morphological variability between the populations with and those without a pappus is otherwise so huge that they should be one species. A colleague using the Biological Species Concept would find that the two forms interbreed and make them one species. In contrast, a taxonomist with a typological approach (as commonly understood) might well consider them two different species. Their gut feeling would be that presence or absence of a pappus has always been a "taxonomically important character" in daisies, and that is that.

In other words, where a taxonomist using the Phenotypic Cluster Concept would use morphological data as evidence to find gaps in overall variation under the assumption that that gap tells us something about the underlying biological reality of reproductive isolation and/or selection, a taxonomist using the TSC (whether consciously or not) is really only interested in the morphological characters as such, to find convenient boxes to which to assign the specimens.

It should not escape our notice that there is an interesting parallel here between the question of biological or clustering approaches versus typological approaches to species delimitation on the one hand and the question of cladism versus "evolutionary" systematics on the other. The user of the more sophisticated approaches cares about the relatedness of the specimens, whereas the user of the typological approach cares about characters considered so important (because they are striking or because they were useful in other cases) that they justify circumscribing species to contain populations that are more closely related to populations of a different species than they are to any other population of their own species. Likewise, the cladist cares about the relatedness of species, whereas the proponent of paraphyletic taxa cares about characters considered so important ("evolutionary divergence"!) that they justify circumscribing supraspecific taxa to contain species that are more closely related to species in another supraspecific taxon than they are to any other species in their own supraspecific taxon. But that just as an aside.

Finally, I find it interesting in this context to consider what is called the Autapomorphic Species Concept in the list I linked to earlier. Please note that I haven't read Nelson & Platnick (1981), and it has been some time since I read most of the book that contains the Wheeler & Platnick (2000) article, so I have to rely on the definitions being reported accurately, but the concept is described as being about the "smallest aggregation of (sexual) populations or (asexual) lineages diagnosable by a unique combination of character traits". That also sounds awfully typological to me - the question is not primarily about interbreeding, relatedness or gaps in what is understood to be a complex, multidimensional field of morphological variability, but it is about a unique combination of characters. I fear that this approach would also yield very "splittery", mechanistic and artificial species circumscriptions, unless I misunderstand the meaning of that description. I could see more sense in using these smallest diagnosable units as operational taxonomic units in a cladistic analysis and uniting several of them into one species based on their relationships on the resulting cladogram, but a group of specimens defined by a unique character combination is likely to be only a small part of a breeding group or cluster species.

Summary (Typological Species as commonly understood)

Grouping criterion

Species are groups of individuals sharing a combination of morphological characters. What characters are considered important enough for taxonomic recognition depends on the personal opinion of the taxonomist and on the study group in question.

Ranking criterion

Completely unclear. Again, what characters are considered important enough to say "species" instead of "variety" or "subspecies" depends on the taxonomist and the study group in question.

What is it good for?

The only justification to use it would be when there are too few specimens available for even very intuitive and qualitative phenotypic clustering. Otherwise the typological approach is today synonymous with incompetent taxonomy. One might wonder why somebody would use it then but yes, there are taxonomic treatments that were obviously written under a typological approach, presumably unconsciously because the author simply could not help but think in terms of neat but artificial "one-character-taxonomy".

Are species real?

While the philosophical origins of the typological approach are said to lie in the belief in the existence of an "ideal type" for every species (this was before the theory of evolution, of course) it is clear that the approach will recover biologically meaningful units only by accident. It is more concerned with describing convenient boxes than with reflecting biological reality, and as such typological species are unlikely to be real entities "out there in nature".

Are species a special rank?

Potentially yes. If species are seen as the smallest diagnosable units, i.e. the smallest groups of specimens with a unique character combination, then obviously there is nothing diagnosable below them. As such the species level is then special as the lowest that can be recognized at all. However, colleagues using a typological approach often recognize subspecies, varieties, and sometimes even subvarieties and forms under the species level. Unless they claim that their classification reflects biological species (without having demonstrated that), there is no reason to assume that the species level is anything species in those cases.


Nelson GJ, Platnick NI, 1981. Systematics and biogeography: cladistics and vicariance. Columbia University Press, New York.
Wheeler QD, Platnick NI, 2000. The phylogenetic species concept (sensu Wheeler and Platnick), pp. 55-69 in Wheeler QD, Meier R (eds.), Species concepts and phylogenetic theory: A debate. Columbia University Press, New York.


*) Conditions apply, as they say in advertisements. If there are several types in one species, the oldest name counts etc etc etc, not really important for present purposes.

No comments:

Post a Comment