From: Mate Kapovic
Message: 36292
Date: 2005-02-14
----- Original Message -----
From: "willemvermeer" <wrvermeer@...>
To: <cybalist@yahoogroups.com>
Sent: Monday, February 14, 2005 7:07 AM
Subject: [tied] Re: Various loose thoughts
>
>
> --- In cybalist@yahoogroups.com, "Mate Kapovic" <mkapovic@...> wrote:
> >...
>
>
> > As I understand it, Kortlandt assumes that the long neo-acute in
> gen. pl.
> > *gor'U > *go~rU in a. p. c is a regular development (this is also
> the case
> > in the a. p. c present tense like *neset'I > *nese~tI as you state
> in your
> > 1984 article). Thus, he assumes that in *st'olU there was no Dybo's
> law
> > because *stolU is short not long (and the a. p. b gen. pl. is long
> > analogically to a. p. c gen. pl.). But I really don't see what is
> gained by
> > supposing this kind of development. Analogical development of gen.
> pl. in a.
> > p. b is required anyway and the lengthening in present tense
> (*neset'I) is
> > also not very conviencing.
>
>
> Very briefly and superficially.
>
>
> A. For the GPl what is needed is a source for the length alternation
> you find in SCr, Sln and Slk, and in traces in Cz and P. In
> Kortlandt's system, the loss of the stressability of final jers is
> that source. Once length is present in (c), its spread to (a) and (b)
> is a trivial analogy, as is its virtually total elimination in modern
> Czech. But it has to start somewhere and that's a point where most
> systems are acutely vulnerable.
Well, I frankly do not see what is gained by this. You assume that the
neo-acute is long regularly but you still need analogy for a. p. b. If you
take starting *gor'U and *zen'U (after Dybo's Law) you can easily get *go~rU
and *ze~nU by assuming again a simple analogy to long-vowel stems like
*gol~lvU or *tra~vU. The long neo-acute has just become a marker of the gen.
pl. of a. p. b and c, even if the root vowel is elsewhere short.
> B. For the present tense, the mechanism explains a number of
> distributions that are actually attested, e.g. the link between
> length of the thematic vowel and type (c) which Boutelje (and much
> later Stang) established for Central Slovak, and also the type of
> distribution of length and brevity you have in such C^ak dialects as
> Susak and Vrgada. Living systems all over the Slavic world have
> intricate distributions of length and brevity and it just does not do
> to state, as has been done so often, that these are due to analogical
> carry-over of length found elsewhere because there were no similar
> distributions elsewhere. Most accentological theories don't even
> begin to explain those distributions.
I have a different opinion of course. I do not agree with your analysis.
Vrgada forms with the ~ in 3. sg. can be explained by a simple analogy to 3.
pl. Also, you adduce in your 1984 article examples from Jurkovo Selo
(Zumberak): pli:je^s, pli:je^, but pli:je``mo, pli:je``te which has the
original length distribution according to you. This is unlikely since, if
the length is original in 2. and 3. sg., we would not expect a long root
vowel (it would have to be shortened). And if you say that 2. and 3. sg.
were originally long everywhere you cannot explain the length in cases like
tre:se``s, tre:se`` etc.
As goes for the other theories not explaining the distributions, it's not
really true. The other theories just have a different perspective. They tend
to see in these forms different kinds of levelling which is hardly
unexpected. All the cases of length on the thematic -e- can be easily
explained separately.
> > Also, I cannot find an explanation in Kortlandt 1975 why does the
> nom. sg.
> > of a. p. b have a neo-acute (*poN~tI) if there was no Dybo and no
> retraction
> > from the jer.
>
>
> Yet if you follow his reasoning in the "konj" chapter (13-19) closely
> you'll find out easily that it is the regular reflex and that it is
> *not* a neo-acute in the sense in which that term is used by
> Kortlandt (notably p. 17, which, by the way, is still awaiting
> elaboration).
I must be blind, I cannot find the explanation for the neo-acute in *poN~tI
(which is not really a neo-acute according to Kortlandt and you as it seems,
I have no idea why).
> A propos the latter: the terminological system comprising such words
> as "acute", "neoacute", "circumflex", "neocircumflex", etc. is more
> of a liability than anything else because it conflates synchronic and
> diachronic considerations, prevents vowel length from being taken
> seriously and effectively obstructs our view of the synchronic
> properties of the various systems that must have succeeded each other
> in the final phase of shared Slavic innovations (and which produced
> those innovations). Different accentologists use those words with
> different meanings. In extreme cases one guy's neoacute can be
> another's circumflex. It is much more fruitful (and flexible) to talk
> in terms of tone, vowel length etcetera.
I don't think it's such a problem in non-Kortlandt accentology. For
instance, if you have two rising accents you have to name them - so one is
called "acute" and the other "neo-acute" because it's a newer rising accent.
And as concerns vowel length, I think that Kortlandt's theory has many
shortcomings in that area. For instance, assuming the Croatian/Serbian nom.
rúka must be analogical to the acc. sg. ru^ku totally unnecessary.
Mate