On 07/09/2017 14:36, PETER BLEACKLEY wrote:
> I'll raise a wider issue, which many people will have heard me 
> expound upon at LCC6.
> The idea of "case" is based on the way that familiar IE languages 
> such as Latin mark the grammatical role of nouns. Applying it to 
> other languages might obscure interesting features of those 
> languages. For example, the question "How many cases does Tsez
> have?" is not a useful question, because it ignores the interesting
> way that affixes combine in Daghestanian langauges.

I seem to recall a thread discussing this no so long ago.  Earlier this

I agree that slavishly applying standard IE case (as we find, e.g. in
Sanskrit, Ancient Greek, Classical Latin etc) can often distort one's
view of a language.  Daghestanian language have forms that mark the role
of the NP within sentences so there is some sort of case system; but it
needs to be analyzed in terms of the languages themselves.  I agree that
labelling the all the locative forms as different cases (e.g.
contessive, superlative [NOT the usual meaning of the word!],
suballative, apudablative etc., etc.) is not helpful.

> Applying that to conlangs, if you're too hung up on IE-style case
> systems, there are all sorts of things you could be doing with your
> nouns that you'll never even think of. So rather than ask, "What sort
> of case system does my language have?", ask yourself "How does my
> language mark the role of nouns?" and you might get a more
> interesting answer.

Yes, indeed.   :)