a phonetic mystery

[Koontz John E]

On Thu, 22 Mar 2001, Pamela Munro wrote:
> All of David's impressionistic judgments sound correct to me (thus,
> John, it's not the case that there's no nasalization after nasals!).

I think I should say that probably this is my (incorrectly) simplified
version of something David told me (or a class I was in), and so should
not be flatly attributed to David in the terms I used as I so blythely
did.  Or to put it frankly, I clearly misquoted David on this one.

I remember that Pat Shaw presented a paper at, I think, the second Siouan
& Caddoan Conference which pointed out among other things that there were
a number of cross-dialect correspondences in Dakotan that belied the
simple idea that Santee has d/_V[or] : n/_V[nas], Teton ditto with l vs.
n, and Assiniboine n before both.

I don't recall the details at the moment, though one problematic case is
the diminutive.  I've also noted that some Dakotan dialects nasalize final
i in some enclitics (like =xti(n)) and some don't.  This variability
occurs across Siouan generally, for that matter, when the relevant
morphemes are shared.