jeff_l
(Jeff_L)
#1
3.4 Ablaut 109
has been syncopated from underlying /t-a-fatar-t/ (leaving aside the question
of whether the first schwa is lexically accented). If so, the only unusual feature
of PI t-i-fattar and t-i-jazzal is the gemination. One might venture the guess
that the gemination is "designed" to protect the short V from the threat of
Syncope, but the forms are highly morphologized, and there are plenty of
similar cases among ablaut plurals where Syncope does apply, e.g. a-daris
'track', ablaut PI i-dras (but A-grm i-daras).
In any event, cases like t-i-fattar require a gemination formative Γ-ο2 as
an add-on to the usual nominal PI ablaut, to account for gemination of the
second stem C.
The same seems at first sight to be true of a specifically feminine type,
where the Sg is of the type t-a/e-CaePe and the PI is t-i-CaPP (§4.1.2.24).
Many but not all of these are VblN's of light V-final stems. Example: t-a-haela
'weeping' with PI t-i-hall. The PI vocalism is quite different from that of most
ablaut plurals, whose hallmark is a final-syllable a. One can partially
reintegrate t-i-hall into the mainstream by assuming that the L part of the PI
melody targets a deletable stem-final V, and is therefore inaudible on
the surface. The PI would therefore be something like /t-i-hallA/ with deletable
low vowel /A/. In this derivation, which I will question in a moment, the type
/t-i-hallA/ is roughly parallel in structure to the type t-i-fattar (preceding
paragraph). Since gemination affects the second stem C, in this case the 1 of Sg
t-a-haela, one could argue that T-c2 is at work in PI t-i-hall just as in t-i-fattar.
However, this won't work, and it will be necessary to dissociate the PI
types t-i-hall and t-i-fattar. To begin with, note that t-i-hall has a marked
accent on the surface final syllable (underlying penult), and this accent is not
carried over from the Sg (unaccented t-ahaela). In regular PI ablaut, an
accent in the Sg is carried over to the PI, with some exceptions (irrelevant
here) where a Sg accent is deleted in the PI. It may seem as though PI
t-i-fattar also shows an unusual accent shift vis-a-vis Sg t-o-ftaq-q, but if we
represent the stem here as /-fatar-/ we can account for the accent in both Sg
and PI forms. In Sg t-a-ftaq-q from /t-a-fatar-t/, the accented schwa is deleted
by Syncope, and the unattached accent then reattaches on the first syllable to
the left. No such phonological analysis will account for the accents in Sg
t-a-haela, PI t-i-hall.
Secondly, while PI t-i-fattar clearly shows T-c2, gemination of C 2 , in PI
t-i-hall it is best to recognize a gemination formative Γ-f, i.e. gemination of
the stem-final C. In t-i-hall itself, "C 2 " and "stem-final C" converge on the 1,
but there are at least three stems of the Sg shape t-a/e-CaePuQ-t that have a PI
t-i-CPaQQ, where C 2 (i.e. P) is unaffected and gemination applies specifically
to the third and final stem C (Q). An example is Sg t-e-faetel-t 'lamp', PI
t-i-ftall. Other plurals of this type are t-i-dbayy 'towns' and t-i-mdall
'sandstorms'(206.c) (§4.1.2.24.
