>
>
> Also, one could do as follows:
> Take a root
> *ber-
> add irreal -s-
> *ber-s-
> inflect semi-thematically
> bher-s-óm
> bhér-s-s
> bhér-s-t
> (I think I'll leave out the plural)
>
> Voilà, sigmatic aorist.
Perhaps Jens' rule should be extended to:
stressed -ó- before voiced sound
unstressed -e- or nothing before unvoiced sound _and in open
syllable_ (ie. before syllable boundary)
In Danish we have
Amérika /ämé?Rikä/
and
amerikansk /ameRikä´nsk/ (where ä is a fronted /a/, not an /æ/)
That's because the syllables before the stress have the structure
-VC(C)-, under and after the stress -C(C)V-.
So it's
/ä-mé?-Ri-kä/
and
/am-eR-i-kánsk/ > /am-O-i-kánsk/
With Jens' rule, such a syllable behaviour would create havoc with
the ablaut vowels.
pre-PIE
nom.,gen.sg. padáz
acc.,gen.pl. padám
diversification for syntactic reasons
nom. pádz, gen.sg. padáz
acc. pádm, gen.pl. padám
>
nom. pá: , gen.sg. padáz
acc. pádm, gen.pl. padám
Jens' rule:
nom. pe:, gen.sg. pedóz
acc. podm, gen.pl. pedóm
Sanskrit-speakers try to fix it:
nom. pe:t, gen.sg. pedóz
acc. podm, gen.pl. pedóm
(because -d- is still not allowed after -e-)
Latin-speakers try to fix it:
nom. pe:s, gen.sg. pedóz
acc. podm, gen.pl. pedóm
(because -d- is still not allowed after -e-)
So, now I got past the possible criticism that acc. -Vm is not
stressed.
Torsten