Nath Rao wrote:
> me wrote:
>
>> If it's possible to construct different PIEs, depending on which of its
>> daughter languages one wants to make it closest to, the arts of
>> reconstructing proto-languages allow for such wide variation that no
>> reconstruction can be expected to approach precision.
>
> The trouble is that more secure you want the reconstruction to be, less
> vocabulary you will have, making it harder to say a given thought in the
> proto-language.
We could augment its vocabulary with loanwords and use italics to indicate
which words are loanwords. In Schleicher's tale, there's vastram whose
appropriate cognate would seem to be vestment, vesture or vestiture in
English; the PIE can presumably be more satisfactorily reconstructed than
if all daughter languages have different words for it. To translate
something for which daughter languages don't have cognate words, we can
either borrow words from a sequential list of lending languages (the first
on the list being the primary lending languages) or coin words from PIE
roots in a way that mimics the coining of words from roots in the primary
lending language by break the word in the primary lending language into
morphemes, translating the morphemes into PIE morphemes and putting the
word back together and adding PIE inflections and/or affixes as
appropriate.
> In case of syntax, we have a real problem in that diachronic syntax is
> much less developed than phonology and morphology. I would agree that
> there we have no clue. [For example, why are narrative verbs in
> reduplicated perfect, and not root forms?]
>
> Nath Rao
|
