'eight' some more

[Koontz John E]

On Wed, 28 Apr 2004, David Costa wrote:
> > I think this is more of a methodological objection than a linguistic one,
>
> Linguistics and its methodology are one and the same thing.

This is kind of a high level issue in methodology, however.

> I basically feel it's a severe violation of Occam's razor to claim that the
> M-I forms were borrowed from a hypothetical reconstructed form in one
> language when actual attested forms that match the M-I words better are
> known to exist in another language.

It's really only one form in Tutelo (or whatever).  The multiplicy of
forms is in the ears of the beholder and in MI.

> > The Tutelo variants are alternate perceptions of one thing, not a perceived
> > alternation between two things.
>
> I know. But that doesn't really matter for our purposes. It was presumably
> just allophonic variation in Tutelo, but M-I speakers, who did NOT have
> allophonic variation in their own language between liquids and /n/, would
> not have perceived it that way. M-I speakers weren't borrowing the
> underlying Siouan form, they were borrowing the phonetic Siouan forms.

Not allophonic.  Random subsignificant, unconditioned variation in
production or free variation in non-native perception of "identical"
productions.  Allophonic would be li (i oral) vs. niN (in nasal), at
least in principle.  Here we have to assume that the speaker was trying to
say niN in every case, while the hearer(s) heard sometimes li, sometimes
ni, or even thought from a single repetition that it might have been
either li or ni.  Granted, I wasn't there and the details of Tutelo
phonology will always have an element of uncertainty about them.

> > Nevertheless, the d vs. r problem is potentially more serious than the r vs. n
> > vs. l one.  As I recall, Bob Rankin opted for Tutelo over Ofo partly because
> > he thought Tutelo l more like MI r or n than Ofo t (which might have been
> > rather d-like).
>
> Right, if an older Chiwere or Ofo form would have been pronounced more like
> *[pataare] or *[pataani], that probably would have come out in M-I as
> /pataali/~/pataani/. That would be another fact tilting the argument towards
> a Tutelo borrowing.

What about phadaaniN?

> > But why do we need to assume that the Siouan 'eight' forms go as far back in
> > MI as we can push them?  We don't have any evidence of them, perforce, before
> > contact.  Could they have been borrowed after MI moved westward to the
> > vicinity of the Mississippi?
>
> Well, that would leave open the question of why the Siouan loan is attested
> through *all* known M-I dialects, including Miami. That is, if the word for
> 'eight' was borrowed by the Illinois from Chiwere speakers around the
> Mississippi River, why do Indiana Miami and Wea dialects have the word as
> well?

What is the date of the Indiana Miami and Wea?  My understanding is that
at present MI communities in Oklahoma crosscut the Miami vs. Illinois
distinction.  In other words, only the Indiana Miami and Wea communities
would be Miami without Illinois population infusions?  However, ... in her
atlas suggests ...

> > Tutelo /aka'aspee/ is definitely a clear Southeastern form.
>
> How so? Is /aka'aspee/ itself a loan from somewhere? A quick look at my
> Chickasaw and Creek dictionaries didn't reveal anything similar.

Sorry, the term Southeastern is ambiguous.  I meant Southeastern Siouan,
or Tutelo plus Biloxi-Ofo.  Tu akaa'spee, Bi akaxpe, Of akape.  =>
*akaas^pe.  I think Bob has pointed out that MVS *s^aakpe and SES
*(a)kaas^pe look like metatheses of each other.

> The Miami-Illinois form for 'six', /kaakaat(i)hswi/, isn't the form the word
> would be expected to have at all, given sister-language cognates like Ojibwe
> /ningodwaaswi/,  Potawatomi /ngodwatso/, Shawnee /nekotwah0wi/, and Fox
> /(ne)kotwaa$ika/ (from a probable PA form */nekwetwa:$i(ka)/). Given the
> sister language forms, the M-I form might be expected to be something like
> **/ninkotaat(i)hswi/. 'Influence' from Tutelo /aka'aspee/ isn't a terribly
> satisfying explanation, but I'm open to any other influences I might have
> missed.

In short, given that Spe vs. Swe is not a contast in Siouan,
kaakaat(i)hswi is fairlyy similar to akaaspee.  The leading k is missing.

JEK