On Thu, Oct 17, 2013 at 03:20:04PM -0400, qiihoskeh wrote:
> On Tue, 15 Oct 2013 12:52:59 -0300, Hugo Cesar de Castro Carneiro <hcesarcastro@...> wrote:
> >I thought of something close to this.
> >But the main difference is that I used two distinct genitives for the
> >agent and the patient. A possessive-gentive was used for the agent as
> >he/she would be the "owner" of the action, and a standard genitive
> >was used for the patient as he would simply be a complement of the
> >action verbnoun.
> >Nowadays I started to think about dative sentences. One of my ideas
> >was changing a sentence like "I send a letter to you" to something
> >like "The sending of the letter from me to you", or simply "Sending
> >letter-GEN I-ABL you-ALL" (respectively genitive, ablative and
> >allative cases). Another possibility was "Letter I-ABL you-ALL
> >sending-PERL" (PERL = perlative case) - "The letter from me to you
> >through sending".
> That's beginning to look like one of H.S. Teoh's languages.
It seems that when one abstractly considers the most straightforward
representation of trivalent verbs like 'to give', one inevitably arrives
at some kind of directional system similar to this. It makes one wonder
why such a system isn't attested in natlangs, since once arrived upon,
it would seem desirable to retain due to its intuitiveness. :)
Or so I thought, when I first invented the Ebisédian case system, which
also influenced Tatari Faran.
Recently, though, I've been thinking that perhaps an even more
fundamental system would look like Henrik Theiling's Trukva
, in which clauses are constructed not
from verbs and nouns, but adverbials in a serial-verb-like system. This
appears to have interesting parallels in a recently-mentioned natlang
(sorry I forgot the name) where transitive clauses are constructed from
monovalent serial verbs, e.g. "to take something outside" is expressed
as "(take something) (go outside)". A trivalent clause like "I give a
book to the woman" would thus be expressed as something like "(I go)
(pick up book) (give woman)".
Now, on a more conlangy note, if we take such a serial verb system and
let the verbs eventually affix onto the nouns they act on, then
eventually verbs frequently used in serial constructions (take, give,
go, etc.) could be reanalysed as noun case affixes, and so one would
arrive at a system where "I give x to y" would be expressed as "I-AGT
book-ACC woman-DAT", which is pretty close to the system suggested
Computers aren't intelligent; they only think they are.