o/u

[Frances Karttunen]

On Sep 17, 2004, at 7:07 AM, Davius Sanctex wrote:

> [Frances] A number of people have compiled extensive lists of attested
> u for
> /o:, o/ and found no predictable pattern.  It doesn't correlate with
> length
> or with stress.
> ------
> [Davius] What about a statistical pattern, it seems that spelling < u
> > is
> more frequent following labial or labio-velar consonants /p, kW/.
> Another
> question, what about the alternances /i, e/ in Nahuatl loanwords to
> Spanish?
> Is this a problem of the same type?
>
The alternation of i/e is a different phenomenon.  For one thing,
within Nahuatl itself this process is definitely phonological, not
orthographic.

Nahuatl has distinct front unrounded vowels /i/ and /e/ (both long and
short).  All short front vowels (including /a/), however, have a
tendency to drift upwards toward /i/ and thence to nothing at all.
Lockhart discusses this at some length under the rubric of "weakening."

Nahuatl's characteristic /tl/ (a unitary consonant, not a consonant
cluster) derives historically from *t before the vowel *a.  That's why
tle- 'fire,' tloh- 'hawk,' tli:l- '(black) ink, soot,' and -tloc 'close
to' are exceptional and rather problematic for reconstruction.
Otherwise, you will notice that syllable-initial tl precedes a, except
in the absolutive suffix -tli/-tl.  In the case of the suffix, the
historical *a has drifted all the way to /i/ and thence has disappeared
entirely providing that there is a stem vowel to the left of the tl.

Stem-internally there's plenty of drift of e > i, providing variants
such as tletl/tlitl 'fire,' etc.  This drift is more pronounced in some
geographical dialects than in others.

Frances