[lojban] polysemy of {nai}

On Friday, December 7, 2012 12:28:51 AM UTC+4, lojbab wrote:

I think this is more appropriate for the main list.

la gleki wrote:
> Even the current grammar has two meanings of {nai}.
> Such "polysemy" (although lacking ambiguity in any case) might lead to
> inconvenience for newbies.
> Why {nai} actually means
> 1. to'e (UInai)
> 2. na (NU NAI = NU NA KU ZOhU, the same with connectives and BAI)?
>
> The proposal http://www.lojban.org/tiki/Move+NAI+to+CAI adds the third
> meaning (na'e).

There is one "meaning" - a syntactically appropriate afterthought
negation of a single word. The semantics of that negation are specific
to what is being negated, but generally it is a scalar/contrary negation
(cf. na'e) of the specific word being marked. Sometimes the nature of
the construct means that a scalar negation is effectively equivalent to
a contradictory negation (cf. na) (this is especially the case for
logical connectives, by intent).

I understand that on boolean scale {na'e=to'e} but what is {na} then?

As a scalar negation, it is NOT the equivalent of to'e when attached to
a UI, but rather na'e (generalized rather than extreme contrary
negation).

na'e is {cu'i ja to'e} (grammar ignored), isn't it?

naicai would be the afterthought "nai"-like equivalent of
to'e when attached to UI. That said, sometimes a scalar situation
degenerates to the point where to'e and na'e are equivalent in meaning.

This is not the case with some UI that have {cu'i} as an appropriate point on the scale.

The separate words exist for those situations when the scale is NOT
degenerate.

> Next question is why {nai} should move to CAI and then to UI when UI
> have no truth value?

It shouldn't, and I have no idea why such a thing would be proposed (I
haven't read the cited proposal, and personally don't consider any
proposals until/unless formally brought before byfy - not that I know
what the procedure for doing so would be these days).

One more vite that it shouldn't be done. Therefore, the poll is closed.

moving to CAI - may be.

moving to UI - no.

We specifically considered that when solving the negation problem. Most
languages have oversimplified and degenerate forms of negation (probably
because logical complexity is "inconvenient for newbies"). TLI Loglan
does so. Lojban specifically tried to improve on that situation.

> If so why having {to'e}, {no'e} and {na'e} and if they can be always
> optionally replaced with {nai}, {cu'i} and some experimental cmavo (e.g.
> {ne'e}) correspondingly?

They can't be so replaced, unless some proposal screws up the language
in an attempt to oversimplify the negation problem. Having multiple
words allows the semantics of each situation to resolve over time with
usage evolving the way each word is interpreted.

That's what I'm proposing. Separate words for different meanings.

Note also that nai is afterthought (like UI) while the NAhE family of
words are forethought and can be used with larger constructs than a
single word.

UI/CAI can be used with larger constructions, don't they?

lojbab