In an unsearchable and potentially ephemeral comment to the original posting, Professor Lawler kindly presented the following answer:
It’s possible, but we’re not there yet. One can study the evolution of languages, and a great deal is known about them, and they can be grouped into clades of various characteristics.
The problem is that, while biological evolution (for practice, count the biological semantic metaphors in the OQ) is Mendelian, language evolution is LaMarckian, in that acquired characteristics are passed on, and gene analogs are widely shared.
Language cladistics is much closer to bacterial cladistics, and that’s a mess too, for the same reasons — bacteria share genes.
I’ve marked this posting Community Wiki because it is John’s answer not my own, and so I deserve no reputation from it.
