Re: [tied] Some accentological thoughts...

From: Miguel Carrasquer Vidal
Message: 47663
Date: 2007-03-03

On Fri, 2 Mar 2007 22:36:13 +0100 (CET), Mate Kapoviæ
<mkapovic@...> wrote:

>On Èet, ožujak 1, 2007 10:27 pm, Miguel Carrasquer Vidal reèe:
>>
>> Yes, that works for a.p. c forms with final -tI. The
>> neo-acute length could have been analogically transferred to
>> a.p.'s a and b . But I don't believe that -tI just dropped
>> off. Novgorodian Russian uses forms with and without -tI in
>> the 3rd. person, with a distinction in meaning (the tI-less
>> carrying modal senses), which should go back to the PIE
>> present (*-e-ti) and the subjunctive or injunctive (*-e-t).
>> In my theory, the tI-less 3rd. person sg. -e: and pl. -oN:
>> also acquire length naturally (also through an analogical
>> development originating in a.p. c).
>
>Sure, but you have no proof that Polish was tI-less there...

That's a bit backwards. There is no evidence for thematic
*-e-tI in Polish at all.

>>>Besides, your theory is not really convincing. In Èakavian, there is only
>>>živete``, roni:te``, peèemo`` etc. There is no **-te:, *-mo: anywhere.
>>
>> Actually, according to my theory we wouldn't expect any
>> length there.
>
>Why, those are *accented a. p. c endings*, am I right?

I wasn't finished... :-)

>> It was the length of 3sg. -e: (and 3pl. -oN:)
>> that was transferred to the thematic vowel. So instead of
>> original
>>
>> *-oN
>> *-es^I:
>> *-e:
>> *-emU:
>> *-ete:
>> *-oN:
>>
>> we got "transposed":
>>
>> -oN
>> -e:s^I
>> -e:
>> -e:mU
>> -e:te
>> -oN:
>
>Very strange. You're just adding like 2 or 3 totally unnecessary
>sound-laws.

There's only one soundlaw (lengthening of a.p. c stressed
endings), which is needed anyway. The transfer of the length
from the desinence to the thematic vowel I wouldn't call a
soundlaw: it's an analogical process which is entirely
justifiable typologically: the paradigm adapts itself to the
shape of the third person (singular [*-e:] *and* plural
[*-oN:, *-oN:tI], in this case).

>The length of -e:- could simply be due to the length of -i:-,
>-a:- < -aje-.

That is of course a possibility, but I've never found it
convincing or appealing.


>> You said Slovene has neo-circumflex in the 1/2pl. and dual,
>> but I can only find it in the imperative, not in the
>> present. What am I missing?
>
>Nothing. Infinitive had the acute, present tense had the neo-acute, so no
>neo-circumflex there.

My mistake, I had completely overlooked the relevant
paradigm:

lé,sti "to climb" (< lê"zti)

lê.zem, lê.zes^, lê.ze,
lê.zeva, lê.zeta, lê.zeta,
lê.zemo, lê.zete, lê.zejo,

with neo-circumflex throughout.

>>>And
>>>there is no convincing analogy there since a. p. b does not have final
>>>accent in those forms. There is no *-té, *-mé in Czech as well.
>>
>> But there was -éme, -éte in Old Czech, and -ieme, -iete etc.
>> in modern Slovak.
>
>Yeah, but it's -éte, not -eté, that's my point. So your theory doesn't
>work.

On the contrary, my theory explains why we have no -eté.

>Why attribute -éte to -é (*and* you have to reconstruct *-e only,
>not *-etI), when you can get the length easily from i-verbs?
>
>> Since posttonic length is lost or unstable in West Slavic,
>> the distribution makes sense: Old Czech and Slovak have the
>> length only in a.p. c,
>
>I think that -é- occurs in a. p. a and a. p. b in Old Czech as well.
>
>> where it was tonic or pre-tonic, but
>> not in a.p. a/b, where the length was posttonic. The
>> (sporadic?) cases of length in Old Polish a.p. a verbs fit
>> into this pattern, although it's hard to explain why the
>> lengthened thematic vowel is not found in Old Polish a.p. c
>> verbs.
>
>Of course, the same thing is in Croatian. Typical Western S^tokavian has
>bu``de:s^ (A), me``lje:s^ (B) but pe`c^es^ (C). Thus, no length *only* in a.
>p. C (and in mo``z^es^, ho``c'es^ which preserve old a. p. B paradigm and in
>-je- a. p. a verbs like vje``rujem). Old this points to length on -e-
>being secondary since the one place you got length in Slovak is the one
>place you don't have in Croatian etc.

I wouldn't be so sure. Such a distribution can just as well
point to it being very old. For instance, Germanic and
Hittite have the perfect and no aorist, while Baltic and
Slavic have the aorist, but no perfect. That doesn't mean
that the perfect/aorist distinction is secondary.

=======================
Miguel Carrasquer Vidal
miguelc@...