[lojban] Re: semantic space

["Craig" <ragnarok@pobox.com>]

>>>>>>>>>Lojban doesn't really classify relations at all:
>>>>>>>>
>>>>>>>>the gismu list is a large and disorderly bag whose merit is that it
>>>>>>>>blankets semantic space,
>>>>>>>
>>>>>>>has this really been demonstrated?
>>>>>>
>>>>>>
>>>>>>How does one demonstrate that? (Or, for starters, measure semantic
>>>>
>>>>space?)
>>>>
>>>>
>>>>>I would have thought one thing Lojban does very well is classifying
>>>>>relations. "Semantic space" is just a (sometimes) convenient metaphor.
>>>>>There is no real space out there to be carved up into concepts.
>>>>
>>>>
>>>>Well, in a sense there is. If we only had thirty-five gismu (the lowest
>>>>minimum I've heard claimed as necessary) then we would need them all to
>>>>encompass a great deal of stuff.
>>
>>
>>>No, we could just pick thirty-five things that we thought were worth
>>>talking about.
>>
>>
>> Seeing as the thirty-five figure is claimed for a natural language, which
>> can talk about everything...

>Why should a natural language be able to talk about everything?  Can you
>talk about astrophysics in Hittite?

Hittite, an extinct language, has never had speakers who could discuss
astrophysics in *any* language. But I defy you to list thirty-five things
that aren't so vague as to make the claim meaningless such that I can't come
up with a number thirty-six that you'd probably be able to discuss in
Nahuatl.