Why biologists still argue about what a species is

Ask a working biologist to define a species and, often enough, you will get a pause before the answer. This surprises people. Species seems like the most settled word in science — children learn it before they learn long division — yet the professionals have never agreed on a single definition. That pause is not neglect or evasion. It is the residue of more than a century of serious attempts, each one good, each one leaking somewhere.

The definition that keeps slipping

The most famous attempt says: if two populations interbreed and produce fertile offspring, they belong to one species; if not, they are separate. The idea, associated above all with the ornithologist Ernst Mayr, is elegant because it points at something real — the flow of genes. It explains the mule neatly: horses and donkeys stay distinct because their offspring are almost always sterile.

Then the edges appear. Many ducks hybridize readily yet remain recognizably distinct species. Coyotes, wolves, and dogs interbreed where their ranges meet. Microbes trade genes sideways across lineages that have nothing to do with parentage. Entire kinds of organisms reproduce without sex at all, so the interbreeding test has nothing to test. And fossils — the deep history of life — can never be asked about their mating habits.

So biologists kept adding tools. Shape and structure, which is what Linnaeus had to work with. Shared ancestry, traced as the smallest branch on the tree of life that can be told apart from its neighbors. Ecological role. Lists of proposed definitions run long enough that the situation has earned its own name: the species problem.

Different questions, one word

It is tempting to read all this as science failing at its job. I read it as the opposite: a record of people discovering that their word was doing several jobs at once. A field guide needs a bird you can recognize in bad light, which means shape and voice. A museum curator needs specimens that can be reliably told apart, which means diagnosability. An evolutionary biologist wants lineages — who descends from whom. These are not rival answers to one question. They are answers to different questions, given by people with different work in front of them.

Most stubborn definitional arguments I have sat through, in and out of science, have this structure. The parties keep offering better and better answers to different questions, and each experiences the other's answer as an evasion. Naming the question each definition serves is usually more productive than shopping for a definition that ends the argument — which, at the edges of the living world, may not exist.

What the disagreement is good for

A caution, though, against the cheap conclusion. "Definitions are just tools" can slide into "so taxonomy is arbitrary," and it is not. The disagreement is disciplined. Claims are checked against specimens, gene flow, and ancestry, and classifications that stop being useful or communicable get revised. Some proposed criteria have effectively lost. The fuzziness lives at the edges, not the middle: nobody doubts what a horse is because horses and donkeys can produce a mule. Vague boundaries are compatible with real things — noon and dusk are both real times of day even though daytime has no sharp edge, and height is real even if tall is contestable.

There are habits worth carrying out of this. When an argument refuses to die, ask what question each side is actually answering. Expect edges to be soft; the world was never obligated to sort itself into our nouns. And resist the demand for one true definition when several good ones, clearly labeled, will do — a toolbox with marked drawers beats a single universal tool, and it is the labels that let people cooperate.

What strikes me most, in the end, is the quiet of it. Biology's most basic word has stayed open for a century without stopping anyone from working. Taxonomists kept counting, comparing, describing, and conserving. The argument stayed alive because the underlying reality stayed complicated — and because, as far as I can tell, nobody wanted to win by flattening the world to do it. That, more than any definition, is what arguing well looks like.