[stefantalpalaru]

We make this classification, so we can adjust the rules until our model fits reality at an acceptable level. So let's define species as groups of individual of a similar genotype. We were not able to define and measure genetic distances in the past, but now with the ability to read whole genomes, we can do just that.

> sometimes hugely different species can interbreed and produce fertile offspring

Then they are obviously not hugely different. Were we just comparing phenotypes when we decided to classify them as different species? Let's improve on that.

[Amezarak]

> So let's define species as groups of individual of a similar genotype. We were not able to define and measure genetic distances in the past, but now with the ability to read whole genomes, we can do just that.

How similar?

Consider the following: imagine that you assembled a gathering of all your ancestors in the male line for say, the past two million years. Let's say, for illustrative purposes, that the gap between each generation is twenty years. So you'd have 5 people to represent a century - you, your father, your grandfather, your great-grandfather, and your great-great grandfather. That makes 50 people per millennium, 500 people per ten thousand years, 5,000 people per hundred thousand years, and 50,000 people per million years. So two million years of ancestry gets you 100,000 people, the size of a small city, or about enough people to fit in the Rose Bowl stadium.

So really, not that many people in the grand scheme of things. Now line all those people up in order going all the way back to the first guy, who we'll just call Adam. The difference between Adam N and Adam N+1 is very small. There's no point at which you can say "Adam N is this species and Adam N+1 is that other species." But there's a hell of difference between you and Adam.

The same logic applies to life in general, the tree is just a lot bigger and messier and we're talking about distant cousins rather than ancestors, and some lines are extinct.

But going back to your proposed definition - similar genotype - we can have all kinds of bizarre situations. Say we require the genotype to be 99% similar. By this definition, we conclude that Adam M and N are the same species. We also come to the conclusion that Adam N and Adam O are the same species. But wait! Adam M and O are only 98% similar! What do we do - assign multiple species labels to the same individuals?

Species has proven a necessary and useful abstraction, but it nevertheless doesn't hold up well under scrutiny. That's fine, as long as we recognize its limitations.

> Then they are obviously not hugely different. Were we just comparing phenotypes when we decided to classify them as different species? Let's improve on that.

Genotypes certainly provide useful information, but they can also muddy the waters.

Assume that we take the definition of "any two organisms that can interbreed and produce fertile offspring are the same species" as our preferred definition. (This has its own problems: for example, did you know mules sometimes are fertile, just not usually? Did you know it's conceivable that humans and chimpanzees can interbreed?) It is entirely possible for A and B to be distantly related genetically, yet able to interbreed, while A and C are closely related and unable to interbreed. "Can interbreed" is not necessarily a proxy for genetic distance.