[tied] Re: Mi- and hi-conjugation in Germanic

--- In cybalist@yahoogroups.com, Miguel Carrasquer <mcv@...> wrote:

> I don't think there any reason to doubt that "apple" was
> mobile in PBS. It's a consonant stem, originally
> proterodynamic (*h2ábo:l, *h2ábolm., oblique *h2(a)bélos ~
> *h2(a)búlos), but reshaped in Balto-Slavic to the
> generalized mobile type *h2abó:l, *h2ábolim etc.
> It was thematized in Latvian and Lithuanian, remaining
> mobile. In Slavic, the suffix *-ko can either be dominant or
> recessive. If it was recessive, the word should have
> remained mobile, which isn't the case. If the suffix was
> dominant (like the cognate -IcI/-Ice, -ikU, -akU,
> -UkU/-Uko), the word should have become immobile (which it
> is) and stress should have been fixed on the thematic vowel.
> The latter is not the case, which means that the stress was
> retracted to the acute root syllable. This cannot be due to
> Hirt's law, because the acute is not of laryngeal origin,
> and because the retraction skips a syllable.

But is there retraction? The suffix -uka- forms class-2 derivatives
in Lithuanian, pointing to *-úko-. Would that not give *a:blúka >
*abl'Uko, whence with Stang *j'ablUko? Where is the double
retraction?

>
> >> >Was the infinitive morpheme ever accented in the iterative -

ati

> >> >verbs?
> >>
> >> It was everywhere else, so yes. And it's bê"gajoN, vê"dajoN
> >> anyway (as opposed to, say, vita"joN, z^ela"joN), so it has
> >> nothing to do with the infinitive morpheme per se.
> >
> >Stang's Baltic grammar gives some Lith. and Latv. examples with
> >radical accent: klú:poja beside klú:po, ry´moju (sic) beside

ry´mo,

> >Latv. me~ta~ju 'werfe hin und her', te~ka~ju 'laufe hin und her',
> >ne~sa~ju 'trage hin und her' (p. 360). The Lith. infinitives are
> >klú:poti 'kneel' (I guess this really means 'remain kneeling'),
> >ry´moti 'lean, remain leaning'. This looks to me like an old
> >structure *té:k-a:-tei with initial accent already in Balto-

Slavic.

>
> I don't know what significance can be ascribed to this
> handful of forms. The vast majority of Lith. verbs in -oti
> are accented -óti

I know.

>(by Hirt's law),

We don't know that.

>no matter what the
> accentual characteristics of the root. In the index to LIV,
> the most convenient source I have at hand, we have:
>
> -ýti 30 '-yti 14
> -é:ti 73 '-e:ti 3
> -óti 10 '-oti 0

Doesn't mean a thing.

If the effects of a proposed specifically Slavic accent rule appear
in Latvian I get supicious.

>
> >> >If Winter + d + t can trigger the same retraction as
> >> >clusters
> >>
> >> d+t *is* a cluster. What do you mean? Which retraction by
> >> clusters?
> >
> >We were told many messages ago that the Slavic type Russ. tonú,
> >tónes^' (with omega) from *tópn-e- (via Dybo + Stang) had been
> >brought about by a general accent retraction caused by consonant
> >clusters of some types. Thomas Olander presented it as Slaaby-
> >Larsen's analysis, not dissimilar to a theory put forward by van
> >Wijk.
>
> If I rememeber correctly, Slaaby-Larsen's law was about
> *non*-retraction in the presence of clusters (però <
> *p(t)etróm).

Thomas Olander reported on this list, Dec. 9, 2004:

"> A tentative and, to some extent, theory-neutral formulation of
the law is:

>
> In Pre-Slavic, words with mobile accentuation containing a medial

cluster C1C2 (where C1

> = obstruent, C2 = any consonant, probably except j and w) get

fixed root-stress (yielding

> CS a.p. a or - via Dybo's law - a.p. b)." [End of quote]


> >I yet have to digest the full message, but it would go a long
> >way towards unifying the original structures. Now, if *some*
> >clusters caused retraction, why could sê"sti, ê"sti not have been
> >produced by the same process?
>
> Because clusters per se don't cause retraction in
> infinitives. If we have *legtí with *-ght- versus *strígti
> with *-gt- then there is no difference between what you are
> saying (some clusters, namely the ones with C1 = *b, *d, *g,
> cause retraction) and what I'm saying (Winter's acute causes
> retraction).

It gives the impression that Winter does not produce retraction
alone either, but some clusters cause retraction when helped by
Winter (while others can do it alone).

> >> >> Failure of Hirt's law due to laryngeal breaking (*i/uH2/3):
> >> >> byla` (but by"ti)
> >> >> vila` (but vi"ti)
> >> >> gnila` (but gni"ti)
> >> >> pila` (but pi"ti)
> >> >> z^ila` (but z^i"ti)
> >> >
> >> >I do not see a significant correlation in this.
> >>
> >> It's clear that Hirt's law didn't work in these verbs, or
> >> the l-ptc. wouldn't be mobile, so something else caused the
> >> accent retraction in the infinitive. I think it's my turn
> >> to throw in some Latvian: bût, vît, dzît, pît.
> >
> >The agreement between Slavic and Latvian only shows that the
> >repartition was Balto-Slavic, which we would assume anyway.
>
> Of course: Francis/Normier is at least dialectal IE, Hirt's
> law is at least Balto-Slavic.
>
> What's your explanation?

That some aorists retained their regular circumflex, while others
gave in to the pressure from the acute of the other forms.

Jens