Jan 08 2024

Categorization and What’s In a Name

Categorization is critical in science, but it is also very tricky, often deceptively so. We need to categorize things to help us organize our knowledge, to understand how things work and relate to each other, and to communicate efficiently and precisely. But categorization can also be a hindrance – if we get it wrong, it can bias or constrain our thinking. The problem is that nature rarely cleaves in straight clean lines. Nature is messy and complicated, almost as if it is trying to defy our arrogant attempts at labeling it. Let’s talk a bit about how we categorize things, how it can go wrong, and why it matters.

We can start with an example that might seem like a simple category – what is a planet? Of course, any science nerd knows how contentious the definition of a planet can be, which is why it is a good example. Astronomers first defined them as wandering stars – the points of light that were not fixed but seemed to wonder throughout the sky. There was something different about them. This is often how categories begin – we observe a phenomenon we cannot explain and so the phenomenon is the category. This is very common in medicine. We observe a set of signs and symptoms that seem to cluster together, and we give it a label. But once we had a more evolved idea about the structure of the universe, and we knew that there are stars and stars have lots of stuff orbiting around them, we needed a clean way to divide all that stuff into different categories. One of those categories is “planet”. But how do we define planet in an objective, intuitive, and scientifically useful way?

This is where the concept of “defining characteristic” comes in. A defining characteristic is, “A property held by all members of a class of object that is so distinctive that it is sufficient to determine membership in that class. A property that defines that which possesses it.” But not all categories have a clear defining characteristic, and for many categories a single characteristic will never suffice. Scientists can and do argue about which characteristics to include as defining, which are more important, and how to police the boundaries of that characteristic.

Returning to planets, they clearly orbit the sun, but since many objects do that is a necessary but insufficient criterion for the category. We don’t want every asteroid to be a planet, so we can add that planets have to be big enough that gravity pulls them into a sphere. But a sphere to what tolerance? Also, there are some bodies that would be a sphere if they were not rotating so fast, but their spin distorts them into an oval. Do they still count? Also many moons are spherical, so we have to exclude objects that are revolving about another object other than the sun. Are we there yet – spherical objects orbiting a parent star but not other objects?

For a while, that was it. But it became apparent that there are potentially many hundreds of objects orbiting our sun that fit this definition, diluting the utility of the category of planet. Ceres, for example, which is the largest asteroid in the asteroid belt between Mars and Jupiter, fits this definition (and for a time was categorized as a planet). Is it a spherical asteroid, or is it a planet orbiting among the asteroids? But what triggered this controversy was the discovery of planetary objects in the Kuiper belt, and the realization there may be hundreds of them. So astronomers added another defining characteristic – a planet must also dominate and clear out its orbit around its parent star. That will nicely exclude Ceres and all or at least most Kuiper belt objects. But – it also excluded Pluto, because its moon Charon was deemed too large to be just a moon. It is more accurate to say that Pluto and Charon orbit each other, and share an orbit around the sun. Astronomers famously created a new category, Dwarf Planet, and placed Ceres and Pluto into this category, along with several Kuiper belt objects.

And we’re just talking about clumps of rock and gas orbiting stars. Imagine how messy nature gets when we talk about something like biology. Linnaeus was the first scientist to attempt to categorize all life into a single system, in his Systema Naturae published in 1735. Linnaeus used a nestled hierarchical classification system, which actually works well in biology since the evolutionary branches of life are nestled hierarchies. The challenge for Linnaeus is that he was operating at a time before evolutionary theory, before molecular biology and genetics. What he had at his disposal was gross morphology – how living things looked.

Most famously he divided plants mainly by their number of sexual organs, stamen and pistils. This made for easy categorization, but was largely arbitrary and was controversial even at the time. As we have learned more and more about biology, the classification system has evolved and become increasingly complex. But it has also become more accurate, reflecting actual relationships rather than just superficial characteristics. It is a great example of the challenges of categorization. Is the duck-billed platypus a mammal? Well, what are the defining characteristics of mammals – they are vertebrates, warm-blooded, have hair or fur, give birth to live young, and nurse their young with mammary glands. A platypus looks like a mammal, but it lays eggs. It’s clearly not a reptile. So – do some mammals lay eggs? Is the platypus in a category all its own (with other monotremes, like the echidnas).

One solution is called cladistics – where all the messiness of biology is reduced to a single defining characteristic – evolutionary branching order. That’s it. This does allow for an unambiguous system and it does reflect an important underlying reality. But some biologists are not entirely happy with this system, because it does not consider things like morphological distinctiveness. In a purely cladistic system, all birds are just one tiny branch of dinosaurs. While this is evolutionarily true, it does not reflect their diversity, their disparity from other dinosaurs, and their importance as a category of animals. It’s the Pluto thing all over again.

What this reflects is that there is often no objectively correct or more scientific choice when it comes to categorization. There are just tradeoffs. We have to decide whether to prioritize precision vs utility, for example. What are we using the categories for? How do they guide science? How much fuzziness are we willing to accept, how much to we lump vs split, and which characteristics are truly defining? What do we do with the inevitable “exceptions”?

The true controversies, however, come into play once we try to categorize humans. These categories can have real implications for people’s lives. They are no mere abstract scientific exercise. That is one reason it is so important to recognize what categories truly are – they are ultimately choices we make that reflect biases and value judgements. They do not automatically reflect objective underlying reality. Some are better than others, but we have to define what “better” means.

No responses yet