jeff_l
(Jeff_L)
#1
488 8 Verbal derivation
(505) LoImpfP Definite Participles for Light V-Final 'butcher' and 'drink'
MaSg FeSg MaPl
a. 'butcher' (LoImpfP Ι-Χ-ύι-Γ)
w-α i-t-iss-n t-ά 0-t-isa-t w-i t-iss-nen
b. 'drink' (LoImpfP /-sdssA-/)
w-α i-ssessas-n t-ά 0-saessae-t w-i sasssae-nen
Disregarding the accents for the moment (see below), we observe schwa
(rather than ae) as the result of VV-Contraction from /i-ae/ in the singular
forms. There is also a schwa representing III before PI suffix -nen. These are
the normal manifestations of hi and call for no special phonology here.
The accentual alternations in (505) are remarkable. The PI participle has an
accent shift onto the stem-final a or ae. The fact that the FeSg participles
0-t-is9-t and 0-sasssae-t are entirely unaccented, forcing phrasal accent on the
preceding demonstrative, indicates that Default Accentuation applies at a late
point. As shown below, FeSg participles of heavier verbs show the same
accent shift (Rightward Accent Shift (132)) that applies to the PI participle.
As shown by paradigms of (non-participial) LoImpfP verbs in non-subject
definite relatives (§12.1.2-4), the basic principle is that Rightward Accent Shift
applies when the ablaut accent χ-pel targets the first stem syllable, and when
there is no (potentially) syllabic subject prefix (like 3MaSg i-). Given that
FeSg LoImpfP definite participles undergo Rightward Accent Shift (since the
3FeSg t- prefix is zeroed and is therefore not syllabic), the conclusion is that
e.g. 0-t-isa-t in (505) starts out as /t-t-isi-aet/, becomes /0-t-is3-ast/ by
Rightward Accent Shift (after the t- prefix is deleted), and then abruptly loses
its accent as the /a-ae/ combination contracts to 3. Since there is no reason for a
marked accent to be deleted in the course of VV-Contraction, this requires a
special morphophonemic rule, χ-Erasure (136).
It is important to make sure that Default Accentuation does not apply at
the unaccented /0-t-iss-aet/ stage, since this would lead to surface penultimate
accent, e.g. the incorrect #0-t-is3-t. In the MaSg participles, e.g. i-t-Isa-n,
there is no reason not to allow precisely this ordering, though we would get the
same surface penultimate accent anyway since the marked ablaut accent χ-pcl
occurs precisely on this syllable.
Consider now (506), where the underlying LoImpfP stem is bisyllabic.
These forms bring out the parallelism in accent between FeSg and PI LoImpfP
definite participles, in opposition to MaSg counterparts.
