Make assumptions, humans are pretty fucking good at that. If that isn't good enough then design studies, run experiments, attempt to extrapolate unobtainable data. If you find that to be lacking, just define the words more succinctly to incorporate the obersved traits of other species.

I think, however, that the words are already succinct. We can classify many species according to the definitions of the two words. Of course, jamming 1 million species into a dichotomy is mcfucking stupid. We can use more words, create new words, or eliminate the species that don't fit (which is what we do, historically) (I took a break from writing this for reasons and now I don't know where it was going. This was not intended to be rude or anything; I'm not sure if it could even be interpreted as such. I just thought it would be fun to write)