People like to label things.
One of the best-known examples of a classification scheme is our biological taxonomy, which is meant to show the relationships of different organisms to each other. But no taxonomy is set in stone; even our biological taxonomy has undergone many changes in the last century.
This got me thinking: how does our method for classifying organisms work? And what’s the point of a taxonomy, anyway?
The Long Reach of Aristotle
Probably the most well-known of the early schemes for classifying species was developed by Aristotle, during his stay on the island of Lesbos (Leroi, 2014). Based on careful observations and in some cases, dissections, he named about 500 birds, mammals, and fish. Any taxonomical system needs a set of organizing principles, and in Aristotle’s case, he classified organisms according to five major processes: metabolism, temperature regulation, information processing (e.g. sensation), heredity, and embryonic development.
Some of Aristotle’s predictions, as well as those of his pupil, Theophrastus (the so-called father of botany), have been borne out; others have not. But what’s more interesting is the belief of many early philosophers that these taxonomies described natural kinds – in other words, that a “species” refers to an actual category of being, “out there” in the world, as opposed to a concept described by language. One implication of this view of taxonomies was that they were fixed, e.g. a rigid hierarchy. This eventually led to an idea called the Great Chain of Being – a hierarchical structure of everything on earth, as ordained by God.
Over the years, many debates were waged around this issue. Some philosophers argued that a taxonomy represented an absolute truth about the world, while others argued it simply reflected how our minds conceptualized the world.
Many centuries later, Carl Linnaeus implemented a more standardized model for naming organisms (binomial nomenclature), which is still used today. Soon after, the advent of Darwin’s model of evolution through natural selection gave rise to a new way of thinking about organisms, and the relationships between them. Theories of evolution showed that the supposedly discrete “species boundaries” was actually much more fluid – that, over many years, and under certain conditions, the forces of natural selection could act upon a population in such a way as to produce divergent species. This led to visualizations such as the “Tree of Life”.
A Species by any Other Name
The above section suggests that, come Darwin, taxonomists had essentially “figured it out” regarding the classification of species. But this isn’t the case. The problem of dividing and classifying species is still an issue today, and has undergone many transformations.
One way of classifying species is by their morphology, their structure or form. This can include external features like their shape, size, and color patterns, as well as internal features like their anatomy. Form and function are often related in biology, so this principle often works pretty well. But in many cases – such as the convergent evolution of particular features or structures – a classification based on organism morphology is misleading.
Another principle is reproductive viability, which argues that two organisms are of the same species if they can produce fertile offspring. This definition of a species also works pretty well, and seems to capture something about the forces of natural selection that morphology does not (or does only indirectly). But while this works well for species that reproduce via sexual reproduction, it is difficult to apply to asexual species.
The advent of DNA sequencing provided a third method for classifying species: genetic similarity. Genetic comparisons between species have allowed scientists to find species differentiations where morphological or reproductive models have failed; in other cases, genetic comparisons have revealed similarities missed by other models.
A fourth principle is ecological niche: how an organism (or population of organisms) fits into an ecosystem. One can ask how that organism produces and consumes resources in the ecosystem, and which other organisms it competes with. For example, the dung beetle uses feces as a food source, and thus fulfills a particular ecological niche.
In any case, all of these frameworks for classifying organisms have been very helpful in taxonomizing the world around us into a hierarchy of species. But this raises another question…
Why Taxonomize, Anyway?
What is the purpose of a taxonomy?
In my view, the purpose behind a taxonomy of some individuals, such as a taxonomy of biological organisms, is to facilitate a more rigorous exploration of the relationship among those individuals. By carving up the world in a certain way, we can uncover patterns that we might not otherwise.
Crucially, however, any good taxonomy should have some sort of organizing principle. One way to think of an organizing principle is an assumption: assuming that principle P is a good way to describe the relationships between individuals, let’s consider how all these individuals relate as a function of P.
Some examples of organizing principles for classifying species were described in the section above. When considering the structure of a taxonomy, it is very important to keep in mind what organizing principle was used in creating that structure. The relationship between any two individuals in a taxonomy is a function of this principle. For example, if our organizing principle of species classification is genetic similarity, the distance between any two species is a function of their genetic similarity.
That might sound obvious, but it’s actually surprisingly easy to mistake a taxonomy for an absolute truth about the world around us. A taxonomy is a representation of the world around us from a particular perspective (the organizing principle), but it is not identical with the world around us. Conflating the two is sometimes called mistaking the map for the territory. It’s a common error, and one that arises partly because of our reliance on language. Language allows us to flexibly categorize the world around us with ad hoc categories (e.g. “things that fly” vs. “things that don’t fly”), but it’s easy to mistake these linguistic labels as somehow representative of the thing itself. This is why it’s so important to have a consistent organizing principle in the construction of a taxonomy.
A Mildly Immodest Proposal
You might be thinking: this is all sort of vaguely interesting, but where exactly is it going?
I have a particular proposal in mind. Specifically, I’d like to propose a new way of taxonomizing organisms – a new organizing principle. Before, we discussed a few different organizing principles for classifying species: morphology, reproductive viability, DNA, and ecological niche. We also discussed the purpose of a taxonomy; one of the main motivations to build a taxonomy, as I see it, is to visualize and discover relationships between individuals, and perhaps find a larger pattern governing these relationships. A taxonomy can also help guide researchers in their work, providing a basis for which relationships to explore.
This proposal has two components. One is theoretical, and the other is more practical.
Component 1: A New Organizing Principle for Taxonomizing Species
We should add a new dimension to our classification scheme: the cognitive and interactional features of an organism. These are broad categories, but my intent is to capture a dimension that current frameworks do not; the closest existing framework is ecological niche, but this perspective is more about an organism’s role in some environment. Here, I’m interested in exploring similarities and differences among organisms in terms of:
- Some characterization of that organism’s cognition, e.g. the way in which it perceives and processes information;
- Some characterization of that organism’s interactional behavior, e.g. the methods by which it interacts with other organisms (either the same kind of organism, or other kinds).
I consider both these dimensions to be intricately linked. Many cognitive features constrain the ways in which an organism interacts with others; and the ways in which an organism interacts with others constrains the kinds of cognitive processing that must go on.
Here are some examples of cognitive/interactional features that I think would be interesting to explore:
- Communication: how does an organism send and receive signals? What kinds of signals can be sent and received?
- Spatial navigation: how does an organism move through space? What kinds of perceptual information does it use to guide this navigation (if it’s guided at all)?
- Cooperation: to what extent does an organism cooperate with conspecifics (or other species)? What is the nature of this cooperative behavior (e.g. altruistic, collectivistic, etc.)?
- Social dynamics: what is the size of this organism’s social groups? Are these groups hierarchically structured, and if so, what is the nature of this hierarchy?
Importantly, I’m not trying to imply that no one has already explored these features. In contrast, each of the dimensions listed above is a rather “hot topic” in one – or potentially more than one – field. “Communication systems” are studied by linguists, biologists, semioticians, and more, usually with slightly different perspectives, but with similar underlying goals. In some cases, there already are taxonomies within these fields.
What I’m suggesting is that we do some more meta-analytical work to more rigorously apply these existing taxonomies across this board. For example, Hockett’s “design features” of language (Hockett, 1968) are often discussed in terms of what makes human language unique, but an interesting project would be to enumerate, using research from all the requisite subfields, which design features apply to which communication systems among which organisms. This would allow us to group together organisms on the basis of which “design features” they share.
Component 2: A More Fine-Grained, Multi-Dimensional Taxonomy (and a visualization tool)
Once we build such a taxonomy, we should also build an interactive visualization tool. My goal here is to facilitate visualization based on:
- Different organizing principles (DNA, social dynamics, etc.)
- Different levels of granularity (particular genetic components, particular social behaviors, etc.)
We don’t need to unify these separate taxonomies into a single “taxonomy of everything”. Again, any given taxonomy is just one way of carving up the world. Equally interesting is the ability to visualize how each taxonomy carves up the world in different ways.
This would let us compare relationships between organisms based on different organizing principles. Let’s say we’re interesting in the relationship between humans and bees. We could visualize this relationship through their genetic similarity, their evolutionary history, their body morphology, or similarities in how they communicate with conspecifics.
The more fine-grained analysis would also let us make comparisons at differing levels of granularity. For example, we could try to predict which particular genetic components predict similarities along particular cognitive abilities. This could give us insight into the evolution of particular cognitive or interactional behaviors; some might share a common ancestor, and some might be the result of convergent evolution – e.g., two organisms independently evolved a similar communication system because it happened to be adaptive in a similar environment.
Hopefully it’s clear that while taxonomies are useful, they by no means express an “absolute truth” about the world. They do, however, express a “conditional truth”: assuming some organizing principle, how do things in the world relate to each other?
More practically, I think a visualization suite such as the one above would be a great boon to scientists and laypeople alike. It’s by no means a trivial project, but the advantage is that any progress would be helpful. Any of the examples I listed above (communication systems, etc.) could be an interesting starting point for compiling a taxonomy based on the research literature.
Leroi, A. M. (2014). The lagoon: How Aristotle invented science. Bloomsbury Publishing.
Hockett, C. F., & Altmann, S. A. (1968). A note on design features. Animal communication: Techniques of study and results of research, 61-72.
 Note that in this post, I’ll be focusing in particular on our taxonomy for biological species, but many of the points I make (particularly in the section entitled “So why taxonomize, anyway?”) apply to any classification scheme, whether it’s a web ontology or a hierarchy of our favorite foods.
 Other taxonomical schemes actually predate Aristotle and the development of taxonomies in the West. Namely, Shen Nung, Emperor of China (~3000 BC) created a taxonomy of herbs according to their medical value. About 1500 years later, Egyptians began illustrating the relationships between plants (again, for medicinal uses), suggesting an understanding of and emphasis on taxonomical principles. A more comprehensive post (written by someone more educated than me in the history of taxonomies) would delve into more detail comparing the developmental trajectory of these schemes, but for the sake of brevity, I’m focusing on Aristotle as the origin point for the current conceptualization of biological taxonomy in the West.
 Where “soon after” actually means a difference of about 100 years.
 The tree of life is still hierarchical in nature, of course, but the hierarchy is organized by evolutionary history and relationship.
 This problem persisted for a long time in philosophy of language regarding the meaning of words. My personal take on the issue is that the meaning of a word is a mental concept, e.g. some sort of representation in the mind/brain.
 This idea is explored to comedic ends in The Analytical Language of John Wilkins by Jorge Luis Borges.
 A brief caveat is in order. I don’t necessarily subscribe to the view that “information-processing” is the best metaphor or paradigm for thinking about something like cognition or intelligent behavior. Alternative (or potentially complementary) perspectives, such as enactive cognition, show a lot of promise as well, but information-processing is probably the dominant view.