Re: PIE meaning of the Germanic dental preterit

On Sat, 1 Mar 2008 11:00:17 -0600, "Patrick Ryan"
<proto-language@...> wrote:

>----- Original Message -----
>From: "Miguel Carrasquer Vidal" <miguelc@...>
>To: <cybalist@yahoogroups.com>
>Sent: Saturday, March 01, 2008 10:17 AM
>Subject: Re: [tied] PIE meaning of the Germanic dental preterit
>
>
>> On Fri, 29 Feb 2008 18:19:49 -0600, "Patrick Ryan"
>> <proto-language@...> wrote:
>>
>> >Thanks, Piotr.
>> >
>> >Locative looks like a real possibility to me.
>> >
>> >Perhaps you can answer something.
>> >
>> >Would IE *bh&i- result in Lithuanian <bai->?
>>
>> Some would say yes (*h2 > *&(2) > *a), but personally I
>> don't believe that "schwa indogermanicum" gives /a/ in
>> Balto-Slavic (it disappears).
>>
>> The root which Pokorny gives as *bho:i- ~ *bh&i- ~ *bhi:-
>> "sich fürchten", corresponds to LIV (p. 72) *bheih2- (*h2
>> because of CLuw. pi:ha-, Lyc. piXe- "fear" < *bhéih2-os).
>>
>> I'm not sure what Pokorny means by *bho:i-, maybe it's just
>> his way of notating *bhoih2- (with an acute diphthong in
>> Balto-Slavic terms). However, given the Sanskrit words
>> bhá:ma- "anger" and bha:mitás "angry", if from an original
>> C-stem *bhéh2-m, there may be grounds for positing an
>> original root *bhVh2i- ~ *bhVih2- ~ *bhVh2- with "Rasmussen"
>> laryngeal metathesis (*bhVh2i- before V, *bhVih2- before CV,
>> *bhVh2- before CC or C#). Most of the forms attested outside
>> of Indo-Iranian reflect the perfect (with present meaning)
>> with o-grade *bhóih2- ~ *bhóh2i- and zero-grade *bhh2i- or
>> *bhih2- (as assuredly in Latv. bîstuos < *bhi:-sk^é-).
>>
>> In Baltic we have Lith. baidýti vs. Latv. bai~dy^t, of which
>> the first appears to reflect *bai~d- with circumflex
>> diphthong (perhaps from original *bhoh2i-dh(h1)- > *bo:id- >
>> bai~d-), and the second *baíd- (< *bhoih2-dh(h1)-) with
>> acute diphthong (if Pokorny's Latvian data is correct).
>>
>> =======================
>> Miguel Carrasquer Vidal
>> miguelc@...
>
>***
>
>Thank you, Miguel, for commenting in your usual knowledgeable way.
>
>Your information certainly puts a crimp in my proposed reconstruction of
>*bho:(H)i-.
>
>So, if I can summarize what I understand from your data, the earliest PIE
>form could be *bheH(2)i- or *bheiH(2)-, with a considerable amount of
>apparently meaningless variation between the two. Is that correct?

Not meaningless. The variation is conditioned. Under the
stress we have ViH-V, VHi-CV and VH-CCV/VH-C#, in unstressed
position we have -Hi-V and Hi-C (where C is voiced) but
-&t-. Of course a perfect paradigm like:

*bhe-bhóih2-h2a
*bhe-bhóih2-th2a
*bhe-bhóh2i-e

*bhe-bhh2i-mé
*bhe-bhh2.-té (?)
*bhe-bhh2i-ér(s) [*bhé-bhh2y-r.s?]

would be prone to regularization, with singular either
*bhoih2- (as in Slavic bojati) or *bhoh2i- (as in Skt.
bi-bhá:y-a) throughout, and plural *bhh2i- or *bhih2-
throughout.

>And the forms pointing to *o(:) should be regarded as perfects being used as
>presents.
>
>In the case of Lithuanian <baidý-ti>, the main form under discussion, it
>appears the perfect is evidenced in the first of the two alternating forms:
>*bho:(H)i-dhe:(H)-, which, in reduced grade, would yield *bhoi-dhé:-, the
>form I suggested albeit gotten by a different route.
>
>I guess the big question I have, Miguel, do you believe we have *bhoi- (from
>**bho:í-) because it is the postulated endingless locative, or should we not
>have expected *bh&i- (zero-grade) if it is not a locative?

Baltic baidýti ~ bai~dît "to scare (away)" is a standard
causative-iterative in -i:ti. These can be made from a verb
or from a noun. In this case, at least in Lithuanian, there
is no verb <baid->, I think, but there is a noun <bai~das>
"scarecrow, bugaboo". Fraenkel at www.indoeuropean.nl is
down, so I can't check there.

The endingless locative applies only to old consonant stems
like *k^é:rd.


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