A couple of days ago, I referred to the point that the dividing line between the sciences and the humanities is not particularly clear, and there are some forms of research that do not fall clearly on one side of that divide. Taxonomy, the practice of classifying and characterising organisms, has been described as one such practice. So it is worth asking - is taxonomy a science?
First, of course, one needs to establish what exactly science is, a topic that has taken up entire volumes. One of the main characteristics of science, though, is the use of the scientific method. Hopefully, you would have had this explained to you in high school, probably as a scientist constructs a hypothesis about something, devises an experiment to test that hypothesis, then finds whether the results of the experiment support the initial hypothesis, refining the hypothesis as needs be (for instance, a hypothesis that hydrogen and oxygen combine to form water may be tested by burning hydrogen in oxygen and finding whether water is produced). Lather. Rinse. Repeat. For many branches of science, however, this is not really an option. In studies such as phylogeny, astronomy or geology, we generally can't directly experiment on our subjects as such. To use Stephen Jay Gould's metaphor, we can't rewind the tape of time, fiddle with the parameters, and see how things could have turned out. Nevertheless (contrary to what some have said), the scientific method is still applicable to these studies, but instead of directly experimenting on the object of study, the method goes somewhat more like this - the researcher will collect a body of observations, construct a hypothesis that explains the observations, then collect further observations and see whether the initial hypothesis still explains the available body of observations (for instance, observations that koalas are only ever seen eating eucalyptus leaves leads to the hypothesis that they have a diet solely composed of such leaves, which may then be tested by continued observations on koala diet). These two methods may be referred to as the experimental or "hard" scientific method and the observational or "soft" scientific method. Needless to say, the two methods blend into each other (is a developmental geneticist staining embryos to find where a certain gene product is expressed conducting experiments or observations?), and are not exclusive.
As a discipline, taxonomy can actually be divided into two distinct but interleaving components - systematics is the identification of relationships between organisms, while nomenclature is the process of determining what the various groupings of organisms identified by systematics should be called. Nomenclature, it should be stressed, is not a scientific process. The name Homo sapiens is not science, just as the equation "E = mc2" is not science. The process by which Einstein established that energy was equivalent to mass multiplied by the speed of light squared was science, but there was no inherent reason why the resulting equation had to be expressed in the format used. Had Einstein used a completely different set of symbols (say, "۞☺۩♫"), it could have still meant exactly the same thing. Similarly, the name given to a taxon is simply the linguistic tag used to identify that taxon, not the taxon itself. Once a particular tag has been established for a particular taxon, it is in every researcher's best interest to continue using that tag rather than inventing a completely new tag every time that taxon is referred to but this is a question of communication, not of science.
Systematics, on the other hand, is a scientific process - most of the time. Imagine I have a collection of unsorted arachnid specimens in front of me (which, as it happens, I pretty much do). I divide these specimens into smaller clusters whose components are morphologically more similar to each other than to specimens in other clusters. For the sake of argument, say I decide that these clusters represent separate species. These species are then able to be tested. I may look at further specimens to see if the boundaries between my various species remain constant, or whether there are specimens that blur the boundaries between clusters. I may use alternative data sources, such as genetics or biogeography, to see if my clusters remain consistent across methods. But while my initial division may be able to be tested scientifically, did it represent a scientific process itself? You might argue that it did not - that it involved a purely subjective judgement about similarities between specimens on my part. What about species that have been erected on the basis of single specimens, and so cannot be said to have been properly tested? I might reply that my own experience, and what I've learnt from the experience of others*, may have taught me a great deal about what kind of characters are likely to be reliable in distinguishing taxa. But is this a scientific progress, or an application of learning? What is the difference?
*In the past I've complained about the errors of past workers such as Carl-Friedrich Roewer complicating the taxonomy of harvestmen that I work on. At the same time, it cannot be stressed enough that I can only criticise the work of my predecessors because I have the published experience of later workers to draw on - Hickman, Martens, Staręga, the Goodnights, even Roewer himself. This is what Isaac Newton was referring to when he noted that "If I have seen farther it is by standing on the shoulders of giants".
In the end, I have to fall back on a quote from Bonde (1977) that I've used before: "An important aspect of any species definition whether in neontology or palaeontology is that any statement that particular individuals (or fragmentary specimens) belong to a certain species is an hypothesis (not a fact)". In my initial establishment of a species (especially if said species is only based on a single or very few specimens) I am essentially proposing a hypothesis that may be tested at a later date. The hypothesis itself may arguably not be science, but the scrutiny it will later be held up to almost certainly will be.
Bonde N. 1977. Cladistic classification as applied to vertebrates. In Major Patterns in Vertebrate Evolution (M. K. Hecht, P. C. Goody, & B. M. Hecht, eds.) pp. 741-804. Plenum Press: New York.
Drones, Silicon Valley and biology: The future isn't here yet
1 hour ago in The Curious Wavefunction