Re: Genitive relationships (WAS: Construct States)

John Cowan wrote:
> We really know almost nothing about grammar-sharing; there is
> that one village in India where everyone speaks both a dialect
> of Kannada (Dravidian) and Marathi (Indo-Aryan), but the
> two languages are separate only in lexicon: they match up
> morpheme for morpheme, constituent for constituent.
>
> Here there is no lexical borrowing, and the two languages are
> felt as distinct solely because of their distinct lexicons.
There is a group of languages in the Amazon like that.  Each village
only speaks a single language, but one must marry outside one's language
group, thus everyone is at least trilingual, knowing their mother
tongue, the local lingua franca, and one other.  Many know all 20.  The
languages are almost relexes of each other, the grammar, morphemes,
etc., lining up almost exactly.  There is no lexical borrowing at all,
due to the need to keep languages seperate, and the view that the
languages consist solely of vocabulary.  I think that they are not
genetically related to each other.  Presumably, they were once quite
distinct, but over millenia of polylinguism, and numerous L2 speakers in
each village, they grew closer in grammar.  A change in one, an extra
distinction or loss of some distinction, diffusing among the others, as
their speakers marry, and move into, another village.