The other definition is "a group of organisms capable of interbreeding and producing fertile offspring.". In this sense, it is not a "structure we humans impose on the world"" but a practical reality and has nothing to do with the human ability to distinguish between species.
Nothing is ever that simple of course.

I agree on all counts, especially the latter! As a definition, interbreeding is not too useful sorting out dinosaurs, obviously, but nor can we really use it to determine whether a fox-like creature is capable of cross-breeding with a wolf-like creature. Suppose only 0.1% of the offspring is viable to the extent of being able to have its own offspring. Should we count that as 'able to cross breed' or not? And if we did accept that, would we actually carry out the cross-breeding experiment that long in practice anyway? Probably not. And if you do accept it, how about 0.01, 0.001 etc etc.

Now, I've deliberately overstated my position to some extent. I'm quite happy to agree there are groups that definitely won't cross-breed however many candle-lit suppers we give them; it's more that I grow suspicious of a litmus test that says that creature is or is not a lion, for example. It's more accurate to say that there is some idealisation of what a lion is (shades of Plato!) but everything we call a lion is actually an almost-lion. Moreover, when an animal live in packs or prides, there can be genetic variation between them sufficient to identify an individual as belonging to a specific pride. That smacks of the very early stages of differenciation that could in time lead to new species, even though they are capable of interbreeding at the moment.

I am no expect on the matter, but it wouldn't suprise me if some of the Galapagos finches could interbreed, for example.