If we look at standard a.p. a, b and c adjectives in Slavic,
we get the following picture:
a.p. a:
sta"rU, sta"ro, sta"ra
a.p. b:
bê'lU, bêló, bêlá
a.p. c:
'drugU, 'drugo, drugá
The definite (long) forms are:
a.p. a:
sta"rUjI, sta"roje, sta"raja
a.p. b:
bê'lUjI, bê'loje, bê'laja
a.p. c:
drugÚjI, drugóje, drugája
These long forms are peculiar in several ways:
- In a.p. b we have retraction of the stress: bê'lUjI,
bê'loje, bê'laja for expected *bêlÚjI, *bêlóje, *bêlája.
- In a.p. c we might have expected something like *drugUjÍ >
drugÚjI, drugóje or *drugojé, drugája or *drugajá (cf.
Zaliznjak p. 125: "1. Odnosloz^naja slovoforma s minusovoj
markirovkoj, stojas^c^aja v konce taktovoj gruppy posle
nedodnosloz^nogo ènklinomena, menjaet svoju minusovuju
markirovku na +. Primery: <slovo sé>, <zemlju tú>, <varjú
vy> 'predvarjaju vas', <mirÚ sI> (iz <mirU sÍ> po p. 3
bazisnogo pravila). 2. Esli taktovaja gruppa sostoit iz
neodnosloz^nogo ènklinomena, okanc^ivajus^c^egosja ne na
<U>, <I>, i ènklitiki (ili peremarkirovannogo p. 1
odnosloz^nogo ènklinomena), to takoj neodnosloz^nyj
ènklinomenon fakul'tativno menjaet svoju minusovuju
markirovku (poslednjuju, esli ona ne odna) na +. Primery:
<styz^jú sja>, <varjú vy>, <zemljú tu>, <obá ta> -- narjadu
s <styz^ju sjá>, <varju vý>, <zemlju tú>, <oba tá>."
[1. A monosyllabic wordform with minus marking, standing at
the end of the prosodic group after a polysyllabic
enclinomenon, changes its minus marking into a +. Examples:
<slovo sé>, <zemlju tú>, <varjú vy>, <mirÚ sI> (from <mirU
sÍ> by step 3 of the Basic Rule). 2. If the prosodic group
consists of a polysyllabic enclinomenon not ending in <U>,
<I>, plus enclitics (or a monosyllabic enclinomenon
re-marked by rule 1 above), then such a polysyllabic
enclinomenon optionally changes its minus marking (the last
one, if there are more than one) into a +. Examples:
<styz^jú sja>, <varjú vy>, <zemljú tu>, <obá ta> -- besides
<styz^ju sjá>, <varju vý>, <zemlju tú>, <oba tá>."]
The more or less traditional explanation for the a.p. b long
forms is that by "Stang's law" the accent was retracted from
a long circumflex vowel to the preceding syllable. After
Dybo's law, we would have had: *bêlU'jI, *bêlo'je, *bêla'ja,
which contracted to *bêly~, *bêle~ (or *bêlo~), *bêla~, and
then retracted (with neo-acute on the first syllable) to
attested bê'ly:, bê'le:/bê'lo:, bê'la:.
There are two reasons why this cannot be correct: (1) the
retraction takes place where there never was contraction, as
in Russian bélyj, bélo[j]e, bélaja, bély[j]e; (2) there is
no such retraction in a.p. c (there should have been no
difference between contracted drugy~, druge~, druga~ and
*bêly~, *bêle~, *bêla~).
Zaliznjak's MAS-approach is to assign the marker +Re to the
"thematic vowel" of long adjectival forms. This means that
the vowel is capable of being stressed, but that a preceding
-> turns into a +. We have:
star-U-jI, star-o-je, star-a-ja = stárUjI, stároje, stáraja
+ +Re + +Re + +Re
bêl-U-jI, bêl-o-je, bêl-a-ja = bê'lUjI, bê'loje, bê'laja
-> +Re -> +Re -> +Re
drug-U-jI, drug-o-je, drug-a-ja = drugÚjI, drugóje, drugája
- +Re - +Re - +Re
which is correct, until we take the polysyllabic-root
adjectives into account.
We have (based on the Russian data) four types:
a.p. a:
gla"dUkU, gla"dUko, gla"dUka
gla"dUkUjI, gla"dUkoje, gla"dUkaja
a.p. a/b:
gotóvU, gotóvo, gotóva
gotóvUjI, gotóvoje, gotóvaja
a.p. b:
teNz^élU, teNz^eló, teNz^elá
teNz^élUjI, teNz^éloje, teNz^élaja
a.p. c:
ve``selU, ve``selo, veselá
vesélUjI, veséloje, vesélaja
Assigning +Re to the a.p. c forms here would have resulted
in *veselÚjI, *veselóje, *veselája, which is incorrect. The
explanation through Stang's law works here, which of course
doesn't make up for the fact that it *doesn't* for the
"standard" 2-syllabic mobile adjectives, and that there's
still no contraction in Russian.
So what is going here? I have no good explanation for
Russian vesëlyj, vesëlaja either. I was a bit afraid that
the polysyllabic a.p. b adjectives would bring down my whole
theory about Proto-Balto-Slavic a.p. II (theme-stressed
paradigms), because a pre-Dybo form *teNz^élo, teNz^éla (=>
teNz^eló, teNz^elá by Dybo's law) is simply unacceptable in
that scheme (a medial non-acute stress would have been
retracted long before Dybo's law). It has to be *teNz^eló,
*teNz^elá, but how do I explain teNz^éloje etc.? But as it
turns out, the "medial stress retraction law" (what I
usually, but perhaps confusingly, refer to as "Stang's
law"), retracts the stress only by one syllable, as is clear
from the trisyllabic o-stem neuters: R. veretenó, pl.
veretëna; res^etó, res^ëta; tenetó, tenëta. A compound
*teNz^elóje would have been retracted to teNz^éloje, which
is exactly what we find. But that explanation does not work
for a.p. c (well, it works for veséloje, but not for
drugóje).
I still feel that the fixation of the stress in the long
adjectival forms must be related somehow to Dolobko's law
and with the fixation of the stress in prefixed nouns
(Zaliznjak p. 153ff.) of the potópU-type, where the base
word gets plus-marking (more precisely acute marking:
bezboródU), even if it was originally minus (a.p. c) or ->
(a.p. b). But my earlier view that the stress was simply
fixed on the initial syllable in a.p. a/b and on the final
syllable in a.p. c is wrong: it works for sta"roje,
gla"dUkoje, bê'loje and drugóje, but it does not for
gotóvoje, teNz^éloje and veséloje.
=======================
Miguel Carrasquer Vidal
miguelc@...