A Rhapsody on Why It's Hard to Define Concepts that Seem Obvious, and Why It Matters
Why are some things so hard to define? Why do we struggle to identify what a “catch” is in the National Football League? Why did Supreme Court Justice Potter Stewart once say he didn’t know if he could ever define hardcore porn, but that “I know it when I see it”? How can it be that some people are 100% sure about the definition of a woman and some people say it requires biological expertise?
How we go about defining and classifying things does not just describe the world, but contributes to our understanding of the world. It follows that an understanding of how we define and classify things would help to understand the human experience. It also follows that much of our understanding of the world is in fact constructed by other humans who have done the defining and classifying. So much of what we know and assume about the world is taught to us before we ever observe it ourselves—if we ever observe it ourselves. I want to explore the idea that two opposing classification systems can both complement each other and reinforce our biases, and therefore have wide-ranging societal consequences.
The first system is more familiar and comes to us from Aristotle. This is where an object or concept is defined by properties it has. A chair is a structure consisting of a seat, four legs and a back on which people sit. This sort of thing. This type of set theory was the basis of Carolus Linnaeus’ system for classifying the natural world. It is an exercise in data entry: Observe the world, note the features and characteristics of various organisms and then classify and define them by shared characteristics.
In the 1970s, psychologist Eleanor Rosch developed what has come to be known as Prototype Theory, which suggests that every concept in language has a real-world equivalent that acts as a “prototype” of that word. This prototype is the most pure form of a concept, the one that nails it on the head. I say “chair” and you think of a solid object with four legs and a back. I say “folding chair.” Still a chair, but not the first thing that came to mind. I say “bean bag chair.” You sit on it, but is it really a chair? It has many, but not all, of the properties of the prototype chair. You could add swivel chairs, office chairs, etc. As Lakoff and Johnson write, “We understand bean bag chairs, barber chairs, and contour chairs as being chairs not because they share some fixed set of definite properties with the prototype but rather because they bear sufficient family resemblance to the prototype.”1
Rosch’s system is more intuitive than those in the tradition of Aristotle. It also speaks, I think, to common online debates like “Is a hot dog a sandwich?” Yes, you can list properties that imply a hot dog is a sandwich—it is bread with a filling inside—but it just doesn’t feel like a sandwich. Going in reverse, it helps to explain why it seems easy to identify porn but is hard to define it.
The Aristotelian method is subject to errors and revision. Although most of you probably learned about the ranking system of the animal kingdom, biologists now generally agree that the hierarchical system is of limited value and may someday be discarded entirely. As we’ve learned more about evolution and the way in which animals are related, the system breaks down. Some birds are more closely related to crocodiles than they are to snakes, even though both crocodiles and snakes are reptiles. Linnaeus, who died in 1778, would have never even heard the word “dinosaur” let alone the theory that birds are related to them.
This isn’t that surprising when you consider that nature doesn’t care about our definitions. The natural world deals primarily in spectrum and gradation. It is human beings who decide when a shade of red is so different it becomes a shade of orange. It’s humans who decide when the temperature increases by “1 degree” (and we can’t even agree on one universal system for that). So, of course definitions and classifications will change—we’re always learning new things. But people have a tendency to assume science is the study of settled facts, rather than our best understanding of them at the time subject to constant revision and clarification.
Humans have an especially sordid history when it comes to defining and classifying other humans. From slavery to eugenics to the holocaust there are innumerable historical examples.
But few stories drive home just how bad we still are at this than the story of Caster Semenya2, a two-time Olympic gold winning middle-distance runner and intersex woman. For over a decade, Semenya has been on the defensive as people claim she is not really a woman and shouldn’t be allowed to race other women. In 2009, the International Association of Athletics Federation ordered she take a test to prove she was a woman. In 2019, the IAAF, now known as World Athletics, proposed rules that athletes like Semenya would have to medically reduce testosterone levels to compete with women. As recently as last month, they put forth new guidance for trans women and “women with differences in sexual development.” Although there was once a leaked news report claiming to represent Semenya’s chromosomal makeup, she has never publicly released such information3. World Athletics has also referred to her as a “biological male” in legal proceedings despite her birth certificate, which clearly states she is female.
Many South Africans have rallied around Semenya. To them, it smacks of discrimination and outsiders trying to impose their definitions on them. This is not new to South Africans. The Population Registration Act of 1950 classified South Africans by race, which determined legal status. These decisions were made by white people. A census in 1985 determined that over a thousand people had suddenly changed race. As Ariel Levy writes in the New Yorker (Warning: some of the descriptions that follow may be unsettling):
“Taxonomy is an acutely sensitive subject, and its history is probably one of the reasons that South Africans—particularly black South Africans—have rallied behind their runner with such fervor. The government has decreed that Semenya can continue running with women in her own country, regardless of what the I.A.A.F. decides.
“South Africans have compared the worldwide fascination with Semenya’s gender to the dubious fame of another South African woman whose body captivated Europeans: Saartjie Baartman, the Hottentot Venus. Baartman, an orphan born on the rural Eastern Cape, was the servant of Dutch farmers near Cape Town. In 1810, they sent her to Europe to be exhibited in front of painters, naturalists, and oglers, who were fascinated by her unusually large buttocks and had heard rumors of her long labia. She supposedly became a prostitute and an alcoholic, and she died in France in her mid-twenties. Until 1974, her skeleton and preserved genitals were displayed at the Musée de l’Homme, in Paris. Many South Africans feel that white foreigners are yet again scrutinizing a black female body as though it did not contain a human being.”
This passage brings into focus one of the issues that I think is best explained by Prototype Theory: It’s not that Baartman did not satisfy the properties of a woman according to an Aristotelian style classification system, she merely did not sufficiently resemble what Europeans were used to seeing. Their worldviews did not contain someone like her. Upon encountering her, they treated her like she wasn’t human at all.
Sometimes, these methods can reinforce each other. For example, life itself is hard to define. I’ve written before that some think the distinguishing characteristic of life is any organism that tends toward growth and increased organization rather than decay and increasing chaos. But that isn’t necessarily the same as an organism being alive. A dead body still has living cells within it. So when is something alive? When does a fetus become a human being? When it has a heartbeat? What is a heartbeat?
For many the answers to these questions are self-evident. Some of this of course has to do with religious beliefs. But setting questions of faith aside, for many, it seems obvious that since humans require a heartbeat to be alive, the presence of a heartbeat signals life. This line of reasoning moves from Prototype Theory to Aristotelian—trying to add definitions to what seems obvious and then attempting to legislate based on the Aristotelian definition. Things with heartbeats are the purest idea of life, so we should define life as a thing with a heartbeat. There’s no real reason this has to be the case though, of course. Many plants and animals don’t even have hearts, and they are alive.
In my opinion, we need to acknowledge both kinds of systems. We need some method of accounting and defining things by characteristics—but we must be aware that this is not the source of ultimate truth. It’s just a necessity so that we can communicate and refer to things with a mutual understanding of what is being referred to. However, we also should acknowledge this isn’t how we experience things in real life. Just look at children. Without ever being taught a thing, they can intuitively grasp relations between things based on resemblance.
A kind of rallying cry in some circles is that “facts don’t care about your feelings.” But many things we consider facts are very much shaped by feelings, feelings about how to classify and define things. It’s impossible to separate what we call “facts” from our learned experiences, and it’s both humbling and helpful to keep that in mind.
This quote comes from their book “Metaphors We Live By”
I cannot recommend reading this New Yorker article in strong enough terms, especially if you are unfamiliar with Semenya. When I first read the article back in 2009, it had significant impact on my understanding of sex and gender, and humanity generally.
And even if she had, chromosomal makeup is not totally determinative of biological sex. Individuals with Swyer syndrome have an XY chromosomes but external female genitalia but non-functioning gonads. And is is merely one example. There are others.