Suppose a scholar of toys wanted to count the number of kinds of toys in the world. Since this information is not generally recorded, the scholar might look at the ways toys are used, the sales figures, or the designs recorded in the patents and trademarks offices in order to estimate the tips and numbers of toys. What we would not expect the scholar to do is to count the number of kinds that the scholar and his colleagues had written about, because all that would tell you is what kinds of toys are described by and found interesting to toy scholars.
Now replace “toy scholar” with “taxonomist” and “toy kinds” with “species”. No taxonomist would ever try to estimate the number of species in the world based on what and how taxonomists have gathered them together, because that would only tell us about the activities and predilections of taxonomists, right? Wrong.
Taxonomists suffer from a congenital disease: it’s very common, although not so widespread you can’t find taxonomists who avoid it. This disease is the fallacy of reification. Broadly stated it is this: if there’s a name for something, there’s a thing the name names.
The reification fallacy is applied in spades in a new paper that has gotten a lot of press: “How Many Species Are There on Earth and in the Ocean?” The authors apply the following technique: look at the average number of species included in larger taxonomic groupings in the Linnaean system of naming where we know a lot about the species, and extrapolate from that to the unknown species. There are names for these Linnaean ranks, so they must be real things, right?
However, when we do know the actual species for a group (say, birds, since they are roughly the same size as us and are hard to miss), we may very well sort the numbers of species into roughly the same sized groups because it suits us to do so. One simply cannot extrapolate from that to unknown groups. In short, as one critic said, we are measuring human activity, not the biological world.
How might it suit us to have similar numbers? There are several ways: one is Linnaeus’ own justification: it makes teaching and learning easier. Another is that we discriminate roughly the same amount of “similarity” based upon how deeply we have studied a group and so we will tend to make salient groupings that match our educated dispositions. A third is that some species have more economic value, and so commensurate groupings serve that purpose. And so on…
The Linnaean ranking system is marvellously fluid and adaptable. When Linnaeus proposed his five ranks (species, genus, class, order and kingdom – family and phylum in zoology or division in botany were added much later) he thought that might cover all the known and as yet unknown diversity of life easily. Within a century, the scheme had blown out to 25 ranks, and today there are many more.
What is worse, these ranks, besides being conventional, are arbitrary. You cannot compare the rank of “family” among cats with “family” among roses, for example. They have quite different meanings and extensions. Of course, if you take one group (say, eukaryotes, which are but one branch of a much larger tree) you may get similar and commensurate rankings simply because the same people are doing the taxonomies, as well as the fact that the degrees of difference used will be similarly evolved in each subbranch. But that’s about it. Terms like “phylum”, so widely appealed to as a measure of “body plan” diversity, are essentially meaningless except as reports of what the investigator thinks is important. In short, the ranks tell us about us, not the species.
I once wrote that a major criticism of the “biological” (better, reproductive isolation) species conception is that if it is adopted, by far the bulk of life will be excluded from being in species. I was, of course, referring to microbial, mostly non-eukaryotic, life. Microbes, it turns out, are largely excluded from this study. The 8.7 million species figure fails to deal with what must surely be the bulk of biodiversity in the world. I suspect this has to do with the authors’ macrobial bias (“kingdoms of life on earth” seems not to include microbes, which they note are too hard to identify. This seems to be due to the fact that most microbes are not isolatable in the laboratory, and so cannot be sequenced easily).
The final figures are largely meaningless except to tell us what we can expect if taxonomists continue to do what they do. I think this is interesting, but hardly an objective fact about the world. We may as well be cataloguing toys.
On a more constructive note, read this paper for a summary of how microbes, particularly bacteria, but also archaea, share genes, both as plasmids (small loops of DNA) and between chromosomes. Despite this ability to introduce novelty (the author notes that as much horizontal transfer occurs as mutation among most microbes), microbes still form recognisable groups and even phylogenies. The existence of sex is not a precondition for being a species, since they lack what we would call sex (or gender) and yet they share genes across wide evolutionary distances.
Also, a very good book on the philosophical issues regarding biodiversity is this one, by Jim Mcclaurin and Kim Sterelny: What is biodiversity?