aspirated and unaspirated caga

[ROOD DAVID S]

First, I resist writing the "gh" for the gamma; the letter g unambiguously
spells gamma before a vowel and the voiced velar stop in other positions;
the only reason for ever using a diacritic is to help you remember
that rule.

Second, neither Buechel nor Riggs ever, to my knowledge, acknowledged that
there is an aspiration contrast for "c".  You can't rely on either one of
them for that distinction.  Buechel's three-way distinction for ptk that
John describes is a consequence of evolution.  His early notes leave
aspirated forms unmarked and put the dots on the unaspirated ones.  His
later notes put the aspiration mark on the aspirated ones and leave the
plain ones unmarked.  But when they typed his notes to produce the
dictionary, they didn't tell us which ones were new and which ones were
old, so we can't tell whether an unmarked letter is an old aspiration or a
new plain recording.  Riggs never marked aspiration regularly at all.


Third, while "ice" is always aspirated, as John points out, you'll have
to check with the native speakers for the "icaga" pronunciatiion -- and it
may not help.  Unfortunately, all the palatalized forms of "kaga" that I'm
familiar with EXCEPTIONALLY aspirate that initial when it changes from "k"
to "c"; the best examples are the dative/benefactives, which go from kaga
to kichaga.  My little note in the volume of IJAL that was dedicated to
Eric Hamp speculates, on the basis of comparative evidence, that the "ch"
in this case is not from the "k" at all, but from PSI *y.  There is a
tendency that I haven't been able to find rules for for "k" to disappear
between vowels in derived forms.  I speculate that that's what happened to
kicaga.  Since PSi *y regularly gives aspirated "ch" in Dakotan (cf.
heart, wood, etc.), the sequence is: ***ki-kaga > **ki-aga > *ki -yaga
(epenthetic y; note that this is "edh" in Dhegiha) > ki-chaga.
	The problem is that I don't know whether "icaga" is old enough to
follow this sequence, or whether it's a newer form from i-kaga which, if
productively derived by currently used rules, would give unaspirated "c".
	The point is that even if the "c" of the 'stones' form is
aspirated, we still don't know whether it's from 'ice' or 'make'.

Frankly, I forgot about 'ice' when this discussion first came up, but
semantically that's a much more likely source than 'make'.  Chaga can be a
verb meaning 'for ice to form' (amachaga 'little bits of ice formed on
me'), but I'm not coming up with any good ideas to explain the use of "i-"
here.  Aren't there any native speakers out there reading this who can
help us?


David S. Rood
Dept. of Linguistics
Univ. of Colorado
295 UCB
Boulder, CO 80309-0295
USA
rood at colorado.edu

On Sat, 10 Aug 2002, Koontz John E wrote:

> On Fri, 9 Aug 2002, Patricia Albers wrote:
> > ... In this regard, I have been trying to track down the term icage,
> > which Walker (1983, pp. 220-221, 222-223, 227-228) gives to describe
> > 'white fruits' growing under the earth, suggesting the crystalline
> > formations in caves. ... The Cheyenne sometimes refer to gypsum as
> > "frost."  Could icage come from the word caga (Riggs 1968, p. 84;
> > Buechel 1970, p. 113).
>
> It may help to realize that c^ha'gha 'ice' (compounding form is c^hax) has
> an initial aspirated c^ (usually written just c in Buechel) as opposed to
> the unaspirated c^ of -c^aghe in ic^a'ghe, usually written with c-overdot
> in Buechel.
>
> Note:  Actually ptck with no marking would be in principle ambiguous in
> Buechel.  A ptck with an opening apostrophe (or half circle) adjacent
> upper right would be unambiguously aspirated, and a ptck with an overdot
> would be unambiguously unaspirated.  It's a two-way contrast encoded in
> three series of symbols, representing Buechel's attempt to improve upon
> the Riggs system which uses only a single ambiguous series.  However,
> hearing aspiration with c^ is harder (and I believe c^ never receives the
> velarized aspiration characteristic of Teton Lakota), so c essentially
> never receives the aspiration mark.  Usually, however, an undotted c is
> c^h.
>
> Note:  Also, gh is not aspirated.  It's just a fumbling attempt to
> represent gamma (voiced velar fricative).  Buechel uses g-overdot or, in
> the English index, just g.  We can't do that easily on this list because
> too many Siouan languages have g as a voiced stop.  I guess it depends
> what type of confusion you like.  The list is somewhat dominated by
> comparativists trying to use a single scheme across all the languages.
> But you will find lots of examples of use of popular orthographies that
> fit ASCII better, too!
>
> Returning to the subject, icage written just like that (assuming no
> diacritics are missing) could be either ic^ha'ghe or ic^a'ghe.  Only a
> native speaker familiar with Walker's usage could say which.  If the first
> form exists, but is specialized usage, not widely known, it's possible a
> native speaker might assume the latter incorrectly.  If the first form
> exists, it's possible it has something to do with c^ha'gha 'ice'.
>
> Note:  However, 'frost' is given in Ingham and Buechel (for Teton) as
> xeyuN'ka.  The Santee form is xewaN'ka or wo{']tasaka in Williamson, while
> Riggs gives xewaN'ke and says Teton has xeyuN'ke.  That's as much
> difference among the sources as I've noticed in some time.  The yuNk- ~
> waN'k- alternation reflects dialect variants of the 'sitting' positional
> auxiliary.
>
> Although the Riggs vs. Buechel and the others' testimony on the final
> vowel reminds us that there are contexts in which final a alternates with
> final e, there are no attested -e variants with the c^hax- root that I
> know of, and xe-SIT-e vs. xe-SIT-a variation is the only suggestion that
> such a variation is possible in something like this context.  (Of course,
> OP nughe, Osage naNghe, IO nuxe ~ noNxe 'ice' show that there is an
> e-final version of this somewhat irregular stem set in Mississippi Valley
> Siouan general.)  In Dakotan specifically, usually you find final a ~ e in
> connection with non-possessed vs. possessed, e.g.  s^uN'ka vs. thas^uN'ke
> 'his particular horse' (said to be an obsolete form).  Supporting this,
> many body part terms take final -e', but have a compounding stem that is
> consonant final.  There are some other a ~ e alternation contexts but I
> don't feel I understand the details.  To some extent e makes sense as
> marker of 'specificity'.
>
> To summarize, ic^haghe having something to do with c^ha'gha 'ice' is not
> impossible, but probably would require more support.  There's also the
> issue of explaining the i- in this case.  The best I could come up with is
> an irregularly denasalized iN- , the compounding form if 'stone', i.e.,
> perhaps something like 'frozen stone'.
>
> The possibility ic^a'gha from ka'gha (see below) makes more sense.
>
> > Icage means "something to make with" (Riggs (Riggs 1968, p.171;
> > Buechel 1970, p. 199).  Another and probably related word, icago,
> > refers to a mark or line that is drawn or sketched on something
> > (Buechel, p. 199),and Buechel also gives wakicaga as a name for a
> > sacred ceremony (Buechel 1970, p. 835).
>
> Here it's ic^a'ghe from ka'gha 'to make', as David Rood points out.  The
> shift of k to c^ or kh to c^h after i is essentually regular.  This form
> is essentially a nominalization of ic^agha 'to make something with
> soething'.  (This is one of those a ~ e contexts I'm not sure of.)  If you
> look in the surrounding articles, you can see that ic^a'gha has other less
> transparent meanings:  'to spring up (as grass, a child, etc.); to
> become'. Also 'to skim off', perhaps only compounded with ic^hu', i.e.,
> ic^a'ghe ic^hu'.  It also participates in some constructions meaning
> 'together', e.g., ic^a'gheya 'together'.  I don't think any of these are
> relevant, though the specific example of 'to become' given was iN'yaN
> ic^agha 'to become a stone'.
>
> Note:  Ic^a'gha is a stative verb inflectionally - imac^agha - though
> plainly it takes two arguments in syntactic terms, so this is one of the
> somewhat overlooked Siouan experiencer verbs.
>
> I think ic^a'gho 'to make a mark, draw a line, sketch' (or the thing so
> produced, as a noun) is simply a different stem, probably the root is gho,
> with prefixed ka 'by force' (which becomes c^a after i). It does occur to
> me that the 'make marks' version of *kax that appears in Ioway-Otoe and
> Winnebago might owe something of its semantics to interactions with a
> Proto-Siouan version of this stem.  There's no "change to o" or "add o
> suffix" morphological rule in Dakotan.  If one added ic^ax to a verb stem
> o, one would get ic^axo, not ic^agho.
>
> Waki'c^agha < wa - ki - kagha is 'to make something for someone; what is
> made for someone', the detransitivized form waki'c^agha from the dative
> ki'c^agha from ka'gha.
>
> In checking into this form I notice that Buechel gives kic^a'gha (with
> c-overdot, so explicit lack of aspiration) as 'to become ice again'. Is
> this correct?
>