About Miguel's *ei ~ *i ablaut etc., Piotr states:
>>Actually, this "Great Vowel Ship" scenario has been considered (e.g. >>by
>>Schmitt-Brandt, if I remember correctly), but of course there's a >>long
>>way to go from a suggestion to a fully developed and cogently >>argued
>>theory. There are lotsa messy details to be handled.
My view doesn't require "long vowels" in any earlier stages beyond
IndoTyrrhenian. It would have started in ProtoSteppe (c.9000 BCE) where only
three distinct vowels existed [*a, *i, *u]. Keep with me, here. Here's the
exact situation of the shift:
Steppe IndoTyr Early/Mid IE LateIE
*i *ei *ei *ei/*oi/*i
*e:i *e:i *e:i/*o:i/*ei/*ou
*u *eu *eu *eu/*ou/*u
*e:u *e:u *e:u/*o:u/*eu/*ou
This first shift in Early IndoTyrrhenian certainly must have gone something
like *i to *ei (that is, */@y/) and *u to *eu (*/@w/). No long vowels yet.
The lengthening only would have come after this initial shift. The question
is "What caused the lengthening?".
Currently, I believe that it was the eventual dislike for monosyllabic
nouns, somewhat like in Modern Mandarin or Early Proto-Uralic (but I digress
on this last one because it's an essay in itself).
Point is, monosyllabic nouns like IndoTyrrhenian *rek: "chief" (taken as
noun directly from the verb *rek:- "to rule") became *re:k: to appease some
sort of subconscious dislike for nouns with "short duration", we might say.
In Mandarin, we have the same phenom. Monosyllabic nouns are icky, replaced
in this case by disyllabic words via suffixation (hai > haizi "child") or
reduplication (xing > xingxing "star"). In IndoTyrrhenian, this syllabic
constraint is simply solved a little differently via syllabic lengthening.
The rest of the quantitative changes are caused by compensatory processes
via the loss of *s, *x or whatever phoneme, hence some differences between
nominative and vocative (an unmarked case that lacks nominative *-s) and the
lengthening of IE *wedo:r and *pxwo:r via the original existence of final
*-x, a collective suffix (earlier: *wedorx, *pxworx). The lengthening in
*k^e:r is also due to the loss of the final dental stop (hence earlier
*k^ert without lengthening).
As for Miguel's insistance of voicing in certain instances of *s, I implore
him to see reason. There is no IE *z and it need never be. All can be
explained better without it, as you will see further below.
Miguel quips some absurd and non-Occam-compatible explanations:
>[...] The first question is harder to answer. I don't know. One >could
>imagine something like *wokw + *-z (1)> *wo:gwz (2)> *wo:gws >(3) *wo:kws,
>with (1) voicing and lengthening caused by /-z/, (2) >/z/ > /s/
>unconditionally, (3) devoicing before /-s/, but that's >hardly
>"parsimonious".
Yes, it's very low on parsimony, isn't it? Piotr's first question is
actually _easy_ to answer when you understand the IndoTyrrhenian constraint
on monosyllabic nouns. The word *wo:kWs is simply derived from Early IE
*we:kW, a straight-forward nominal derivative of the verb *wekW- seen
elsewhere. Note that the Early IE nominative was unmarked. The lengthening
was caused by the very syllabic constraint against monosyllabic nouns that
I've mentioned ad supra. No need for *z at all. No need for messy
voicing/re-devoicing. No need to rewrite nominatives or accusatives to suit
your theory. It's all very straight-forward.
Note also that Miguel's obsession for voiced phonemes to explain
compensatory lengthening is ridiculous unless we also claim laryngeals and
every phoneme involved in the process are voiced too (thereby, entangling
oneself further within a deficient and extravagant theory).
Occam, eat your *k^e:r out! :)
Miguel on IE "water":
>To connect *wo(:)dr with *wedo:r, Rasmussen needs to assume both >ablaut
>and a suffix; the latter assumption seems gratuitous to me.
Poor Rasmussen may not be aware of the underlying penultimate nature of the
IE accent. The reason for the change of accent is obvious in Mid IE where we
have *weder versus *weder-xe. The first has accent on first syllable due to
the penultimate accent, the second item has accent on the second syllable.
When final vowels disappear and Mid IE *e splits into Late IE *e & *o, we
obtain *wodr versus *wedorx (later: *wedo:r).
On a side note, I believe I mentioned a supposed Early IE *kWetWe-x "eight"
a while back, with final collective suffix. I should have stated *kWetwe-xe.
With the permanent placement of the collective suffix (needed to distinguish
it from the singular "four"), the accent would in turn be permanently taken
away from the first syllable by penultimate accent.
Come Late IE, a stage to be known for its syllabic reductions, the numeral
would be reduced to *kWtwox-. The subsequent loss of connection between
*kWetwo-r-es "four" and *kWtwo-x-u "eight", together with the need to fight
the ugly initial consonant cluster in "eight" led to the irregular prothetic
*o-, hence *oktox- (*oktoH2-). There, that's better. Carry on.
Miguel:
>*e ~ *o Ablaut as originally quantitative **a ~ *a: Ablaut).
No way, Jos�. Qualitative ablaut cannot be very ancient at all (as opposed
to quantitative ablaut, which is). This is another extravagant theory of
yours that only produces more absurd questions that could have been avoided
in the first place. If we accept that *e/*o ablaut is simply derived from a
recent earlier stage where only *e existed, we don't have to go into long,
boring and trivial explanation about why a supposed Early IE verb like *sed-
"to sit" alternated with *sa:d-. Instead, we just have *sed-, period,
without verbal alternation at all and without need to ramble on needlessly.
The qualitative alternation is quite clearly determined for the most part by
accentual alternations in the Tonal stage of Late IE.
Piotr gives insight:
>A general remark: some historical linguists seem to be anxious to >make
>laryngeals so tangible that no critic could possibly deny their >presence
>-- isn't it some atavistic angst inherited from the times >when the
>laryngeal theory was less securely established?
Interesting. Of course, as I've stated very early ago, the direction of
science itself is always affected in some way by history, politics and
social preferences of the times. If this weren't so, marijuana would be as
legal as coffee in North America. Ooh, that reminds me. I think it's
caf�-time again, see yo'll! :)
- gLeN
_____________________________________________________________________________________
Get more from the Web. FREE MSN Explorer download :
http://explorer.msn.com