Salmoneus wrote:zompist wrote:Unfortunately, "this really ought to be a language family because of stuff completely outside linguistics" is not convincing.
Why on earth not? If you know that something is a language family, that should be convincing, and clearly you can know that something is a language family for non-linguistic reasons [just put people in a big box for ten thousand years, and other than the possibility of spontaneous generation (which is generally discounted) you know that the resulting languages will be a language family - no linguistics needed]
Why on earth not? Because we tried it. Early attempts at classification mixed linguistic criteria with region, race, technological development, or Biblical taxonomy. This turned out not to be a generator of truth, once people went back and reclassified based on linguistics alone.
Your box analogy begs the question. You can't prove that Amerindian is a family by assuming that the ancestors of the Amerindians all spoke languages in one family.
There are just too many assumptions in your theory. A small population of hunter-gatherers need not speak the same language; indeed, they're notorious for having tiny languages. Genetic descent does not correlate with linguistic families. Family tree development (which is clearest in the language of former agricultural empires) is not the best model for hunter-gatherer languages. We really know nothing about what language looked like 15,000 (or more) years ago.
(On spontaneous generation of spoken language, I can't advocate that either, for this period. But we do have an example of the phenomena: the humans of Earth have spoken languages, so it happened at least once. We don't know when. We don't know what happened next. And we don't know when, if ever, it stopped being a possible process. There are interesting things like mother-in-law languages, twin languages, and sign language generation, that suggest that it's not simply impossible. In a world of agriculturalists there is always a dominant spoken language to swamp these efforts; in a hunter-gatherer situation in an unsettled continent, I'm not so sure.)
The diversity of the Americas is a scientific puzzle. Scientific puzzles are great; they're opportunities to learn something!
Why are there as many as a hundred families there, when Eurasia makes do with a dozen or so?
And if we are interested in propositions purely in terms of assigning praise or blame to the person who first proposed them, that would be relevant. But from time to time, we may be interested in whether a proposition is actually true, at which point it doesn't matter whether the formulator of the theory believed he was being given the idea by a colloquy of tangerine pixies.
Fine, if we ever know by actual science, then you may go back and affix a gold star to Greenberg's grave. In the meantime, science doesn't work by assessing whether the pixies have spoken or not. Greenberg's work is criticized by linguists not because people think he's wrong, but because his methodology is suspect.