> Dans un courrier dat=E9 du 27/05/99 14:06:35 , Josh a =E9crit :
> =20
> > Decided by its demiurge, yes. Arbitrarily decided by its demiurge, no.
> > You call me a "rare perfectionist" on your conlang page and then you g=
o=20
> on
> > to call my choice of objects arbitrary. I must say I find your lack o=
f=20
> > faith
> > disturbing (inside joke :).
> =20
> Perfection is "per-factio" not "per-natura". We say in French "le mieux=20
est=20
> l'ennemi du bien" ;-)
> =20
> My point is that i gave up considering the maps of roles of the verbs of=20
my=20
> conlang as universal ones. So i told Daniel to find his own maps rather=20
than=20
> picking them from a book. Transformative or lative maps are just easy=20
starts=20
> to figure out a whole system of roles (voices and cases).
> =20
> The semantic sememe of any word or any morpheme is decided either by=20
natlang=20
> usage or by a conlanger.
> In other words YOU can decide to make a verb like English "to enter" with =
"
> house" as a possible 1st object, or a verb like French "entrer dans" with=20
> same possible actor "maison" used with a prepositional 1st object, or else=20
a=20
> neologism like "to enter (through) 1st object (into) 2nd object (giving wa=
y=20
> to) 3rd object, etc.". Role-mapping depends on YOUR semantic definition of=20
> the verb including YOUR ranking of its various possible actors and YOU are=20
> the only one to decide both.
> I'm quite sure you've made the map and ranked the 10 first objects of eac=
h=20
> of your words. But (i) these maps are all yours and yours only and (ii)=20
some=20
> lazy conlangers are not willing to map anything beyond subject and 1st=20
object=20
> and decide to work with prepositions or verbs instead. And another lazier=20
one=20
> like i am decided to work with the subject only. The result is just the=20
same,=20
> except i remember better "addressing X" or "using Y's function" than "7th=20
and=20
> 8th object".
> =20
> > =20
> > No, this particular word, ve"g, belongs to a class of several roots=20
with=20
> the
> > same basic structure -- any complexity which can be taken out of the=20
> > individual
> > semantic word and used in the definition of the valency group=20
simplifies=20
> the
> > definitions of all words concerned.
> =20
> These are your own processive, transformative, applicative, lative, etc.=20
> role-mappings. You have "frozen" these schemes and ranked their actors. I=20
> have isolated each role of each actor of all schemes i could think of and=20
> given each of them a simplistic verbal definition like "to use", "to be=20
> located", etc. because you can remember these darn object-rankings but i'm=20
> zero at figures and abstraction : i need concrete words.
> =20
> The idea is to maximize this common
> > information, and this is not an arbitrary process.
> >=20
> =20
> this is for sure : common concepts imply common roles, usually with commo=
n=20
> ranking of these roles because the ranking of the roles is part of the=20
> semantic definition of the verb (ex : "to protect from" vs. "to prevent=20
from")
> .
> =20
> > You do not know the power of the Dark Side.
> >=20
> =20
> I am kind of colour-blind so never get down in a dark cellar lest i shoul=
d=20
> crunch my head on the opposite wall or never find my way out ;-)
> =20
> > Josh
> > =20
> =20
> Mathias
> =20