2015-10-16 14:26 GMT+03:00 Alex Fink <[log in to unmask]>:

> On Fri, 16 Oct 2015 14:04:54 +0300, Gleki Arxokuna <
> [log in to unmask]> wrote:
>
> >2015-10-16 13:44 GMT+03:00 Alex Fink <[log in to unmask]>:
> >
> >> I dare say that I
> >> find a metaphorical case system obviously more ergonomic than a numbered
> >> one.  A numbered system demands that each predicate's complementation be
> >> learned separately, and is fundamentally hostile to transferring
> knowledge
> >> of one predicate that one knows well to another unfamiliar one.
> >
> >I certainly disagree with this one. Where do you learn this metaphorical
> >case system from?  From your mother tongue, supposedly a SAE one, isn't
> it?
> >Then certainly from auxlanish purposes such metaphorical system can be
> >learnable faster ("donor" is marked with Dative-like case tag etc.)
>
> SAE is not the only source of metaphoric case systems.  If I were
> designing a metaphoric case system for an ergonomic log-leaning language in
> fully serious mode, I would start instead from an analysis like that done
> in the paper of Hartmann, Haspelmath, Cysouw
>
> http://cysouw.de/home/articles_files/cysouwhartmannhaspelmathCOEXPRESSION.pdf
> recently mentioned here, and do something like run a clustering algorithm
> on the space of microroles.
>

It's certainly possible to make even a numbered case system as more optimal
by making each of core predicates to belong to one of frames. But I have no
clue why metaphorical system has any advantages over a non-metaphorical one.


> If your argument for numbered cases is "using any case system in the
> well-populated region of natlang space will give speakers of some natlangs
> an unfair advantage, so we have to do something entirely different"...
> whatever sort of argument that is, it's not an argument from ergonomics.
>

Only auxlangs aim at making a language faster to master (while leading to
all sorts of problems from that).


>
> Somewhat impertinent question: is there any feature of Lojban which you
> think is actually suboptimal?  Your attitude towards Lojban on this list is
> one I might characterise as boosterism (if clear of the taboo sort of
> advocacy).
>

All languages are suboptimal.
I may seem to boost Lojban only because I don't create any new conlangs
(unless you consider practicing existing conlangs "creation" of them).


> Alex
>