Re: Questions about Hungarian

On Tue, 4 May 2004 14:16:43 +0100, Racsko Tamas <tracsko@...> wrote:

>  Your're probably right but the term "partitive" still makes me
>trouble speaking about PFU.
From I have read, PFU did not have a partitive.  The partitive case is a
feature of the Balto-Finnic languages which derived it from the earlier
ablative in -tA.  Semantically, this makes sense: ablative means "from X,"
which can be construed to mean "from the whole of X" (singular) or "from
the set of X" (plural).  The use of the partitive as an indefinite
accusative happened even later.  Again, the semantic shift is easily seen:
Tahtoisin kahvia 'I would like (one) from (whole of) coffee' > 'I would
like some coffee.'

>  My questions: Do you know other inharmonic morphemes that became
>harmonic in PFV? And not just in PFV but in Proto-Ugric and in Proto-
>Permic because the latter two branch have continuants of a PU/PFU
>harmonic ablative *-tA (e.g. Erza Mordvin ablative -To/Te, elative
>-sto/ste [T = assimilated as t/t'/d/d']? Etc.
Is there a sound change in Erzya Mordvin such that final /a/ became /o/ and
final /ä/ became /e/?

>  I think ther're much less questions if we reconstruct PU *-tA. E.g.
>there're examples of the change -A > -i in Finnic branch.
What are the examples?  And what would cause such a change?

>> What are, by the way, PU's vowel harmony rules?
>
>  According to the my sources, the morpheme chain was harmonic. The PFU
>had a different set of vowels in the first syllable and in the non-
>first ones. (Similarly -- but not identically -- like in Lappish
>dialects). The first syllable may contain "velar" (back and central)
>vowels (/a/, /o/, /u/, /3/) or "palatal" (front) ones (/a/, /E/, /i/,
>or maybe /y/). If the first syllable was velar, the following syllables
>contained velar /a/ or neutral /e/; if the "head" was palatal, the rest
>were palatal /E/ or neutral /e/.
Sammallahti reconstructs the following vowel inventory for PU: back
(/a/, /o/, /u/, /ï/ [high back unrounded]), front (/ä/, /e/, /i/, /y/).

>  The case suffix *-ti breakes the above rule in two points. (1) A case
>suffix can't be the first syllable in a morpheme chain, therefore its
>vowel should have been /a/, /E/ or /e/. (2) It shows palatal
>homomorphism, i.e. it doesn't have a velar variant for velar noun
>stems.
According to Sammallahti, PU could have the following vowels in non-initial
root syllables: /a/, /ä/, /i/, /ï/.  With palatal harmony in place, it
seems obvious that non-initial vowels also collapsed to a simple low-high
distinction.

The PU ablative in -tA is related to the PIE ablative -(e)d and the Turkic
locative -da/de.

- Rob

Reply