Dakotan T-words and there equivalents in Siouan

[Koontz John E]

On Tue, 25 Apr 2006 Rgraczyk at aol.com wrote:
> Hidatsa /ts/ corresponds to Crow /t/ before /a/, and to /ch/ (that's the
> practical orthography) before a non-low vowel.  So from a historical
> standpoint, I guess we can say that PSi *t becomes s and *s becomes t.

Also per Randy:
> T-words in Crow are S(H)-words.  T > sh before all non-low vowels:
> i,e,u, and o.

I think PS *t only becomes Cr s before a, as in the possessed forms where
PMV *hta- 'alienable' comes out sa-, or the question words.  But I think
that Randy was confirming that when he said *T-words (question/indefinite
words) were S(H)-words in Crow.

I had a vague recollection that *t and *s merged, but clearly not, from
Randy's summary above!  In any event the history of *t, *s, and *s^ is
quite involved in Crow-Hidatsa and Mandan.  With various cases of
inversion. Apart from Randy's comment, Matthews, famously, has pointed out
that PS *s > Ma s^ and PS *s^ > Ma s.  I think this was the basis for
Wolff's earlier suggestion that the PS segments were *s and *c^.

Note that PS *t is being used here as a sort of shorthand for *ht, *th,
*t, and *t?.

===

Returning to Bruce's original question, I gather, Bruce, that you might be
more interested in the merged interrogative and indefinite behavior of the
Tpwords and perhaps in their syntax generally, than with their
phonological forms per se, or perhaps the T-categories distingusihed,
e.g., is 'why' distinguished form 'how', etc.

I have a slight tendency to tackle an expression like T-word in
phonological terms, but I assume you are using the phrase by analogy with
the English expression WH-words, or Latin Qu-words, etc.

As Bob says, in Dhegiha these are mostly ha-words, though there are three
or four other elements that figure in Dhegiha, e.g., *ta in 'what' forms,
*pe in 'who' forms, and (wi)awa(N)- in 'which of two' forms.  The last
probably has something to do with PS forms for 'one' and 'other'.  And, of
course, in Omaha-Ponca *ha- comes out a-, as it 'appens:  one of a small
set of forms where OP and Da have a vowel initial, but the rest of MVS has
h.  There are a few cases of *(h)a- as a question root elsewhere in
Mississippi Valley, mainly in Winnebago and IO, but otherwise *tV (V =
some vowel) seems to be the usual one, except for *pe which is common for
'who'.

The other bits attached to these roots tend to occur also with
demonstatives and some nouns as "tight postpositions" or other enclitic
modifiers.  I say "other enclitic modifiers" because I have trouble
thinking of things like *na(N) 'quantity' and *htaN 'extent' as
postpositions.  These things have always interested me, perhaps because I
was bitten as a teenager by a table of Latin "correlatives."