From: Willem Vermeer
Message: 39486
Date: 2005-08-03
On 2 Aug 2005 at 13:37, elmeras2000 wrote:
[On the Moscow revision of de Saussure's law:]
> It's not ran~komis, ran~kose again, is it?
Bingo! Plus ça change ....
[On Winter's law:]
> I do not see what is surprising about glìnda (1), Latv. gni~da, Sl.
> *gni"da (a). If there was a sequence *-ind- in it at some point,
> that should produce *-i:nd- regardless of the position of the accent.
But that way you don't get the Slavic form.
[On Illich-Svitych's use of dialect descriptions to prove the non-occurrence of the transition of (b)
to (c) in masculine o-stems.]
> How much of it is wrong? I had got the impression that there were
> indeed some errors of detail in the dialect descriptions, but that
> the general message remained sound. Could you illustrate how wrong
> this impression is?
It is emphatically not a matter of errors of detail. Unfortunately this can't be done briefly.
The central point is the following. SCr dialects are very complex. As a consequence, quite a few
dialect descriptions are not reliable enough for Slavic accentology to operate with. This is an
elementary truth known to everybody who deals with that kind of material. (And the same holds
for Slovene and Lithuanian.)
A. "Istria".
The allegedly Istrian material (from Nemanic/ 1883-1885) is from a description that mixes
information from many different but unspecified places (also outside Istria) and normalizes the
results. That is enough to make it quite worthless as a source. (Dialectologically speaking the
area covered by Nemanic/ is known to be extremely heterogeneous.) When Ill.-Sv. dredged up
Nemanic/'s work, it had not been used by serious accentologists for more than half a century.
B. Susak.
The Susak material is from a description written by three authors: Josip Hamm, Mate Hraste and
Petar Guberina (Hrvatski dialektolos^ki zbornik 1, 1956). Most of it is by Hamm (including the
lengthy glossary), the chapter on inflection is by Hraste; the very brief section on phonetics is by
Guberina. Much about this dialect grammar is worrisome.
The authors have not managed to determine whether or not the dialect has a tonal contrast and in
which positions it has contrastive vowel quantity. To make up for that deficiency they present
their material in a phonetic transcription in which tone and length are noted impressionistically
(rather than contrastively, as is traditional). The result is an immense amount of doublets and a
complete absence of clarity about the prosodic system, i.e. about the very subject accentology
happens to be about. This point alone is a good reason to be extremely cautious.
There are systematic discrepancies between Hraste's and Hamm's contributions. Since they
cannot both be correct, at least one of them has to be unreliable in important respects. The
material Illich-Svitych operates with is limited to Hraste's chapter, completely lacking parallels in
the sections written by Hamm.
Both authors have a problematic track record as dialectologists. Hamm, who was a specialist on
medieval texts, had a quite limited dialectological experience. Hraste is well known for his work
on the dialects of his native central Dalmatia (Hvar and Brac^), but his work on dialects further
away from his native area has a very poor reputation. In 1951 Stjepan Ivs^ic/ took the
unprecendented step of warning the scholarly public about the unreliability of Hraste's later work.
Considering the track record of the two principal authors there is reason for caution, particularly
in the case of Hraste.
What I find most disturbing is that Ill-Sv. appears to have missed the warning signs that are so
abundant both in Nemanic/ and in the Susak description. In both cases you really don't need a
background in SCr dialectology to realize that something is terribly wrong.
[For further discussion see my article in Folia Linguistica Historica 5/2, 1984 (pp. 358-361) and
my appendix to the second edition of Werner Lehfeldt's book (2001, pp. 134-138).]
All the best,
Willem