Richard:
> The first observation then is that Glen's construction is not
> impossible, although admittedly fairly infrequent. (Now who's
> proposing new peculiarities for pre-PIE? :)
I think one pecularity is forgivable in a theory of a regular
Pre-IE rather than a sea of idiosyncrata from O-fix to double-
long vowels as per Jens' views. They are all equally rare if
not rarer when amassed together in an unlikely theory of a
single language.
Further, one can't compare Jens' idea which predicts a huge list
of non-existent case forms with my view that only predicts that
*ya was the original endingless locative of *ya-s > *yo-s, supported
by a necessary endingless locative *i that Jens in fact agrees
to to explain the later indicative and locative endings. No comparison.
Another added insight is that while I call this endingless locative
form a "locative", it was in fact used for other cases that were
undeclined by the older case system. So it is also a dative since
the creation of *-syo predates the innovation of dative *-ei.
> In what follows I will call the affix OWN.
Exactly the example I thought of but in *NOUN-s-ya, the "own"
of "John's own book" is expressed by *ya, showing that *ya
MUST be in the locative (or dative, take your pick) since it
is that morpheme that necessarily must carry the genitive
nuance that the noun in *-s cannot unambiguously transmit.
> From tatpurushas and Semitic constructs, one might wildly
> suggest zero! I have no idea whether the connection, if any,
> would be areal or genetic.
Neither, the locative was in fact zero, hence *ya, not nom. *ya-s
here. Locative *-i was ironically based on an endingless locative
*i of another pronoun stem *i-, and came later so we can't
expect the locative to be anything other than endingless at this
stage.
> For animate thematic nouns, the phonetic forms in the singular,
> assuming nominative in [z], are:
>
> 1. *-osyo (final /o/ rather than /e/ unexplained)
> 2. *-osyos (only governing a noun in the nominative)
> 3. *-esyos (governing a noun in any case)
Ugh, I already explained why *-e isn't found here. It's because
*-syo < *-sya, not from **-sy&. Both Jens and I agree that e/o
alternation is caused by a single vowel that changed differently
before voiced and voiceless segments, however at the end
position, it became *-e in order to explain vocative forms of
thematic stems which were in effect endingless. Since we have
both *tesyo (e/o alternating) and *tosyo (non-alternating), we see
that one is from a paradigm based on *t&- and the other based
on *ta-. The former is an enclitic form, the latter "full".
Parallels with English "an", pronounced [n] when unstressed
and [An] when stressed show how two paradigms can exist.
So *-syo is based on the non-enclitic form *ya which no doubt
needed to be clearly pronounced to convey that it is a
genitive, not a nominative noun.
= gLeN