Linguist Joseph Greenberg devised a system of classifying Native American languages which divides them in three categories: Eskimo-Aleut, Na-Dene & AmerInd. Another name for Na-Dene would be Athabascan, I suppose. All others that don't fit into the first two categories are lumped together as AmerInd. Needless to say, there has been plenty of disagreement with this system. Lumping the Tarascan language with AmerInd would be a mistake, since it's not related to other languages. The system raises alot of questions in my mind. Is it relying mainly on vocabulary while neglecting to consider syntax structure? Does anyone have a better system? BTW, has anyone ever heard of a language called Eskimantiq or Eskimantic? I have not found any info on this possible language. Apparently, Native American languages are difficult to categorize.

Ruhlen, following Greenberg, listed Athabaskan as a subfamily of Na-Dene. He also reported that one of Greenberg's primary arguments was the prevalence of t-initial second person singular.

Athabascan, Eyak, and Tlingit make up Na-Dene. Greenberg put Haida in there too, though everybody else considers Haida an isolate.

But you know what I don't get is how Greenberg could just create a hugely diverse macro-family like Amerind, and not have jack to back it up. I mean, it's ridiculously internally diverse, if it is a family at all, which I highly doubt.

Greenburg was a lumper. With regard to the Amerind hypothesis, I think this tendency got the better of him.