'town'/'friend' (was Re: iron/ metal ...)

[Koontz John E]

A while back David Costa questioned my suggestion that PS *htuNwaN 'clan,
town' might underlie PA *ooteeweni 'town', or at least represent a loan
from the same unknown source, hypothetically a language associated with
"Mississippian" cultural phenomenon. He argued:

"This strikes me as rather implausible. The root is Proto-Algonquian
*/o:te:-/ (*o:te:weni is just a nominalization off that), & it seems to
have meant 'dwell together as a group'. It's very well integrated into the
Algonquian lexicon in that it appears in several different reconstructible
collocations, plus it's found throughout the family. So if it was a loan,
it really was all the way back at the PA level. If the Siouan cognate is
only present in one or two branches of Siouan, this would strongly suggest
to me that the borrowing had to have been Algonquian -> Siouan, which is
actually the usual direction."

I thought I owed to him to explain my reasoning a bit more fully, but it's
taken me a while to get around to this.

'clan, town'

Siouan

Te othuN'wahe 'cluster of houses, village, town; Washington, DC'
   -thuNwaN (~ -thuN ?) suffix on names of Dakota subtribes
   thuNwaN'=yec^a 'to dwell at a place'
Sa othuN'we 'cluster of houses or tents, village, town'
   thuNwaN'=yaN 'to make a village, to dwell somewhere'
   -thuNwaN ~ -thuN suffix on names of Dakota subtribes

Note that terms for 'clan' are along the lines of (thi)os^paye, oyate,
owe. I'm not sure if the -thuN variant of the suffix occurs in Teton
usage.  It's probably essentially a fast speech form in any event.

OP ttaN'waN 'town' (ttaN'aN with rearticulation in fast speech.)
   ttaN'waN=gdhaN 'tribe, clan'
Ks ttaN'maN, ttaN'maN=laN 'town, camp, village'
Os ttaN'waN 'town'
   ttaN'waN=laN 'gens'
Qu ttoN'waN, ttoN' 'town'
   o'ttaN=knaN 'tribe, nation'

The -gdhaN, -laN, -knaN are all probably positionals, based on the 'round
or compact' root *raN.  K- may be orientive in the context, cf.
itti=...gdhaN 'to place in the belt'.  The gdhaN is actually an inflected
element, i.e., this is ttaN'waN=...gdhaN, cf. OP ttaN'waN=iNgdhaN 'we
constitute a nation for him'.

Bi taNyaN', taaN', taN' 'town, village'

Biloxi probably had aspiration, and this t is not marked for
"non-aspiration," so it may be aspirated.  Dorsey specifically compares
the Biloxi form with Dakotan and Dhegiha forms.

The combination of Siouan attestations suggests something like PS
*htuN'waN, perhaps *htuNwaNh, but the final -he in Teton, the -y- of
Biloxi, the first syllable vowel in Biloxi, and the second syllable vowel
in Santee are all perhaps a bit problematic or irregular, though nothing a
Siouanist couldn't comfortably attribute to analogy or unknown suffixing
patterns.  The aN ~ e looks like nominal ablaut, with -a perhaps by
analogy from -e, since -a ~ -e is more common. The -he might be a
positional, and so on.  The root is also a bit unusual in form, but has
its parallels.  In short, this is a regular set for any but the
obsessively picky.  It is not universally attested, of course, but is is
presumably replaced by innovations in where it is not attested.  Winnebago
and Chiwere, for example, use for 'town' forms based on 'to dwell' + a
positional, i.e., Wi c^iinaN'k, IO c^hi'naN, both suggesting earlier
*hti-raNk-

My knowledge of the Algonquian forms is limited to what appears in George
Aubin's now somewhat dated summary, p. 118.  He gives PA *ooteeweni sg.,
*ooteewenali pl. 'town, village'.  This appears regularly in Fox
ooteeweni/ooteewenani and Shawnee hoteewe/hoteewena, but undergoes
metathesis of -we- and -na- in Ojibwe ooteenaw/ooteenawan, Cree ooteenaw/
ooteenawah, Penobscot o'tene/o'tenal, and Natick otan/???.  These are the
only forms he sites though there may well be others.  I believe here that
*-i is the singular inanimate suffix, and *-ali the associated plural.

Granted that apart from the metathesis there is nothing irregular here, I
know that loans can often look much like inherited material, if borrowed
between rather similar languages and we do have here an essentially
"central" distribution, apart from the Penobscot and Natick forms, and
these last share the metathesis that appears in all forms north and east
of Ojibwe. In short, one could see a loan here without too much
difficulty, the metathesis occurring in Ojibwe and being propagated beyond
that.  The oo- initial syllable is also interesting, given that o- is a
Siouan locative prefix, appearing in, for example, Dakotan and Quapaw
forms with this root.  We're comparing a Siouan model, not necessarily one
of the presently attested forms, along the lines of *o-htuN'waN with
Algonquian *ooteewen-.

In other directions we find that (Muskogean) Choctaw has "ta.maha" (from
Byington, with a, for underdotted a - short a?) 'town'.  This also occurs
in the Mobilian trade language in essentially identical form.  As far as I
know, the usual Muskogean terms for 'town' are something like okla, e.g.,
Chickasaw okla.  I don't know if Alabama oola is related, since it also
has okla 'friend', and the 'friend' term seems to be the basis of 'town'.
However, my data are restricted to Choctaw, Chickasaw, and Alabama.

The CSD also suggests investigating further (Caddoan) Arikara ituu'nu.  I
had noticed an element which I think was spelled tuh in various Arikara
clan names at some point, but my source for this is still apparently boxed
from my series of moves over the last few years.  I don't want to attach
any real emphasis to the Caddoan forms since I don't have any real
knowledge of how relevant forms work across Caddoan.

The possibly correlating friend terms:

Te kho'la 'friend (of male)'
   lakhota ~ lakhol= 'allies'
Sa kho'da 'friend (of male)'
   dakhota ~ dakhol= 'allies'

The difficulty of associating the friend and ally terms has been discussed
in the past.  It is possible that -la and -da in 'friend' are connected
with the diminutive Te =la and Sa =daN ~ =na.

OP iNdakkudha (perhaps a Dakota loan) 'friend' (archaic)
Ks kko'ya 'friend (of male?)'
Os ihko'dha ~ hko'dha ~ hko'wa (recent)'friend (of male?)'
Qu kko'da ~ kho'ta 'friend (of male?)'

The OP term in common use today is khage' 'friend (of male)', perhaps
related to similar terms khagesaNga 'younger brother' and khage' 'fourth
son'.

IO itha'ro (~ ithadu ~ ithara) 'friend' (*i-hta- Ps3-ALIENABLE-)
Wi hic^ako'ro 'friend'  (hi-c^a- < *i-hta- Ps3-ALIENABLE-)

The Dakotan forms suggest *hkoRa ~ *hkota, while the Dhegiha ones seem to
suggest *hkora.  The Winnebago might be *kro, but is probably *hkoro, and
the Ioway-Otoe version may be a worn down version of the same or *(hko)Ro
or *(hko)to.  In short, the Siouan forms, though more or less universal in
Mississippi Valley, do not correspond regularly between the lowest level
of subfamilies, though they are regular within each of these (unless we
count Winnebago-Chiwere as a lowest level subfamily).

The Algonquian comparison here, again from Aubin, p. 114, is *nii0kaan-a
'my fellow clansman (man speaking)' (0 for theta), for which Aubin
supplies Ojibwe niikkaaneen? 'my brother (man speaking)', niikkaaniss 'my
brother (man speaking), my friend (man speaking)', Potawatomi nikkane
("probably"), Menomini neehkaah 'my brother (man speaking)', neehkaan 'my
fellow participant in a rite', Fox niihkaana 'my friend (man speaking)'.
As I understand it, in these forms final -a is the animate singular, nii-
is the first person possessive, and the -ss or -iss in the one Ojibwe form
is the diminutive.  Again, it is possible that there are further
attestations beyond what is cited here.

This may be comparable to Choctaw "kana" (from Byington, perhaps
representing kaana, since the first a is not dotted) 'friend', and also
Chickasaw inkaNna? 'friend'.  This might be a good place to set out a
short list of Muskogean (and Mobilian) forms:

Gloss            town         nation, people     friend

Mobilian         tamaha       ???                mokula
Choctaw          ta.maha      okla               kana
Chickasaw        okla         imaaokla?          inkaNna?
                              okloshi
Alabama          oola         ???                okla

It appears that these concepts are related, etymologically, in Muskogean,
and that the tamaha and kana forms are somewhat intrusive in the domain.
The ? here is a final glottal, which I believe is a nominalizing suffix
here.