In my last lecture of Historical Linguistics, I brought in a guest lecturer, a fellow PhD student, who was an ardent Nostraticist. I hadn’t discussed Nostratic with him for years. To my astonishment, I watched him recant Nostratic right before my eyes. And the way he did it was by making fun of Starostin et al., grasping for cognates.
What do *I* think about Nostratic? It’s plausible, and it uses the comparative method, which the long range Greenbergian macrofamilies do not. It’s not generally accepted, and the scepticism is warranted given the time depth and the likelihood of noise in the data. Unproven, but wouldn’t be horrified if it turned out true. But hard to see, given current attitudes and the tenuousness of the relations, what it would take for to be proven true…