On Fri, 10 Dec 2004 13:01:57 +0100, Miguel Carrasquer
<
mcv@...> wrote:
>
>On Fri, 10 Dec 2004 04:23:52 +0000, "Anders R. Jørgensen"
><ollga_loudec@...> wrote:
>
>>--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:
>>
>>>
>>> >I think the comparison
>>> >between Slav. (pre-Dybo) *-íko- : Lith. -ìka-, Slav.
>>*-ímo-
>>> >: Lith. -ìma- can hardly be ignored.
>>>
>>> Neither can the comparison between Slav. *-ikó- : PIE
>>> *-ikó-, Slav. *-imó- : PIE *-imó-.
>>
>>Just bear in mind that we anyway have to explain why Baltic decided
>>to put the stress on the -i- in *-íko- and *-ímo- and on the
>>thematic vowel in *-inó- and *-iskó-. There is no sound law for
>>that either (as far as I know).
>
>There should be.
There is.
I had more or less skipped the parts on suffixes up to now,
because I first wanted to understand roots and desinences.
Now I had no choice but to start to read them.
OSA, p. 85-108 lays down some of the rules quite clearly.
In Lithuanian, the dominant (stressed) suffixes -ta-, -kla-,
-sta- (and probably -va-) cause metatony of acute mobile
roots, which acquire circumflex: dz^iau~tas, dõtas,
garu~z^tas, -mau~tas, etc.; klõstas, mõstas, au~ks^tas,
etc.; -bu~:klas, -de~:klas etc. Other dominant suffixes
also turn mobile roots immobile, but do not cause metatony:
-tuva-, -inja-, -is^kja-, etc. It is clear that the
difference is that these last suffixes are bisyllabic: the
stressed part (-vá-, -já-) loses the accent and causes
matatony of the first, unstressed, part (here vacuously, as
-tu-, -in- and -is^- are not acute: a proper example is
-õkas, Slav. -akÚ).
Dybo, Zamjatina and Nikolaev suggest that the same thing
happend in Slavic. Words in -dlo have the following
distribution:
immobile acute: AP(a) bí'dlo, my''dlo, etc.
immobile circumflex: AP(b) c^eNdló (kridló, sidló are given
as circumflex, but I think they're acute)
mobile acute: AP(b): dêdló, peNdló, stadló, verdló, z^eNdló,
z^erdló, z^Irdló, *perdló, vidló, bydló, z^idló (exceptions:
gÚ'rdlo, ó'rdlo).
The development would have been: PIE *de:tló- > de~tla-
(accent retraction & metatony = Lith. de~klas) > de:tlá-
(Dybo's law = Slav. dêdló). I think the retraction is
Lithuanian only (perhaps Latvian dèsts, tìkls suggest East
Baltic in general [*]), because PSlav. *dé~:dlam would have
given masculine dêdlÚ after Dybo's/Illich-Svitych's law
(like dváram -> dvorÚ), and it's more parsimonious to assume
that in Slavic nothing happened (PIE *de:tlóm -> Slav.
dêdló).
The retraction-with-conditional-metatony rule explains why
Lith. has -ìka-, -ìja-, -ìnja- for the oxytone suffixes
*-ikó-, *-ijó-, etc. (Slav. -IkÚ, -IjÍ), and how it lost the
category of oxytonic neuters (de:klá- > de~kla-) and
oxytonic (mesotonic) verbs (-jé-, -éje-, verbs etc. become
barytonic [and circumflex]).
[*] On the other hand, the infinitive suffix *-téi retains
the stress in z^aût, plaûst, duôt, graûzt, maût, bût, etc.
Also, it is stated in OSA, p. 87: "dazhe esli [...]
vtorichnymi mogut okazat'sja preryvistaja intonacija v
latyshskom i nepodvizhnaja a.p. slavjanskix prezensov na
*-je-...". I don't have any references on Latvian accents
(all my Latvian grammars are accentless), but if that is
indeed the case, Latvian would have retained at least in
those cases oxytonic/mesotonic accentuation on non-mobiles,
like Slavic.
=======================
Miguel Carrasquer Vidal
mcv@...