Classificatory Compounds and Theme Form (RE: argument ....)

[Koontz John E]

On Tue, 5 Apr 2005 lcumberl at indiana.edu wrote:
> I agree with the point about *interpretation* of data in B&D. I don't
> include wo- 'food' in my discussion of "noun classifiers" (which I call
> "truncated nouns") because, in fact, it is not a truncated noun, like
> the others in the list;

Without the list in front of me, what I noticed in Regina's two examples
is that in at least one case, wi-, it's less truncation than loss of the
thematic affix that forms the independent stem.  In short, its a combining
form like s^uNk- vs. s^uNka 'dog' or siNl- vs. siNte' 'tail'.  There is
denasalization in many cases with wiN-, there is nasal spread in the full
stem wiNyaN (underlying |[wiNya]|), and there is an epenthetic glide -y-,
but I'd explain wiNyaN < wiN + a 'woman' as otherwise exactly parallel
with suNka < suNk + a.  There are relatively few cases of CV + a, but it
occurs.  Another is heya < he + a 'louse'.  I think there's also iNyaN <
iN + a 'stone'.

For that matter, parallel with cases like c^haNl- ~ c^haNte' 'heart' we
have c^huN ~ c^huNwe' ('w.'s older sister', I think!), though the parallel
older sibling terms are the only case of CV + e that I know of.

I'm not sure what the Dakotan extended form is for ho- 'camp circle'.  In
OP the independent for is hu'dhuga, in which the -dhuga seems to be a
locative element *roka 'inside', though the evidence is comparative, not
internal to OP.