While lojban tries to give a convenient treatment of anaphor and
anaphoric resolution (that is, allowing the speaker to be thoroughly
precise in anaphoric linking, but also to remain as much fuzzy as he
likes), lojban tries also to establish the 'best' semantic space which
can be used in order to communicate efficiently, and this implies that
meaning is correctly carried along utterances...
but...
the anaphor problem and the semantic-space problem are two really
different ones, and, whereas the former is clearly a lojbau
problematic, the latter does not appear to be so evidently connected
to lojbau-ness than its companion.
Indeed, a consequence of thorough semantisation is that no metaphor
can be derived from usage, and this does not seem to be a real
interesting property of the language, nor is it even characteristic of
the 'logical-ness' of it. That is : a logical language, stating with
no fuzziness what is to be said *could* use metaphors, after all.
Another problem is with metonymy/synechdoch : rightly, I already know
that rafsi such as the one for "nose", for instance, include the maybe
importante case where the nose would be a noselike appendix of some
very strange alien species, or even a noselike protruding piece of a
thing, which we would like to call 'a nose', to keep things simple.
But, keeping open such *interpretations* is the way to let the wolf in
the sheepyard, for they are indeed a way of encoding numerous
synechdochs, metonymies, and perhaps... metaphors.
I recently took several minutes to read extensively a paper on frogs,
and they speak of the shoulders and the toes of these frogs. But in
what sense 'a toe' or 'a shoulder' of a frog still has a relation with
the similar terms for a human being ? Then, we admit to have shifted
a little form human-being-ness to frog-ness, but then, why not going
further, and perhaps one day we find that a bactery toe is something
we should account of.
So, what is Lojban position concerning metaphors and the like ?