polysemy of {nai}

[la gleki]

Let me see if I understand negators correctly (scheme attached in a file to this post).

{na'e} says that we are somewere at another point but on the same scale.
{no'e} says we are in the middle of the same scale.

{to'e} says that we are at the opposite point of the same scale.

{na'i} says that we are outside this scale (i.e. this predicate relationship)

{na}. Here I have a problem. According to what I draw {na} means that we are not at this point of this scale and may be even outside this scale.

So for me {na} is (warning! bad grammar follows) {na'i ja na'e}.

But may be you prove me wrong (I'm not sure myself).

Anyway, I want all types of negation to fit on the same scheme.

Last time when I draw a similar scheme I could completely solve (at least for myself) the problem of subjunctives in lojban.

Now it's time for negation.

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 

[la gleki]

The scheme attached.

[Ian Johnson]

{na'e} is definitely not "in the middle or on the opposite end", it is just somewhere on the scale that isn't the point being tagged. Note that the scale needn't even be 1 dimensional, which is one reason why {to'e}, {no'e} etc. needn't make sense even when {na'e} does.

{na}, unlike everything else, is taken straightforwardly from logic; there are no questions about what it means, and it is orthogonal to both {na'i} and {na'e}. The only possible issue with {na} is the dispute as to how its scope works. CLL considers selbri-na to be equivalent to bridi-initial naku; standard idiom seems to consider selbri-na to be equivalent to naku in the exact position of the na. In situations where any confusion might arise I prefer to avoid selbri-na for this reason, using explicit naku or {na'ei} instead.

mu'o mi'e la latro'a

--
You received this message because you are subscribed to the Google Groups "lojban" group.
To view this discussion on the web visit https://groups.google.com/d/msg/lojban/-/3mwxUYZC6TUJ.
To post to this group, send email to loj...@googlegroups.com.
To unsubscribe from this group, send email to lojban+un...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/lojban?hl=en.

[Ian Johnson]

(I should add that there is no dispute as to what naku means.)

mu'o mi'e la latro'a

[Bob LeChevalier, President and Founder - LLG]

la gleki wrote:
> Let me see if I understand negators correctly (scheme attached in a file
> to this post).
>
> {na'e} says that we are somewere at another point but on the same scale.

On the negative side of the same scale, but not necessarily the opposite

> {no'e} says we are in the middle of the same scale.
> {to'e} says that we are at the opposite point of the same scale.
> {na'i} says that we are outside this scale (i.e. this predicate
> relationship)


> {na}. Here I have a problem. According to what I draw {na} means that we
> are not at this point of this scale and may be even outside this scale.
> So for me {na} is (warning! bad grammar follows) {na'i ja na'e}.

na has nothing to do with scales, but rather with truth tables. It is
contradictory negation, and in general says that the predication without
the na is false

> But may be you prove me wrong (I'm not sure myself).
>
> Anyway, I want all types of negation to fit on the same scheme.

They can't because negation is NOT a single scheme, conceptually. There
is contradictory negation and contrary/scalar negation.

The negation chapter of CLL goes into this at length.

> Last time when I draw a similar scheme I could completely solve (at
> least for myself) the problem of subjunctives in lojban.
> Now it's time for negation.

You have to know the problem in order to solve the problem. At the time
we solved it, the most comprehensive book we could find on the
linguistics of negation was _The Natural History of Negation_ by Horn

http://cslipublications.stanford.edu/site/1575863367.shtml

> 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}

na'e and to'e have nothing to do with booleans. Boolean negation is
contradictory negation, which uses na.


Sometimes the na'e and even the to'e of a predication has the same truth
value as the one without the scalar. That is not the case for
contradictory negation ("not the case" is a contradictory "na" negation
expressed in English)

>but what is {na} then?

contradictory (Boolean), not scalar

> 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 ingnored), isn't it?

No.

Let me concoct an example with a well-defined scale. Perhaps I'll
manage this with no mistakes, being rusty.

Let us arbitrarily define any natural number (positive integers) larger
than 100 as "large" and any number smaller than 100 as "small". Then

li 200 cu (je'a) barda ke rarna namcu
li 100 cu no'e barda ke rarna namcu
li 1 cu to'e barda ke rarna namcu
li 1 bi'i li 99 cu na'e barda ke rarna namcu
li za'u 1 cu na'eto'e barda ke rarna namcu
li ci'i cu to'eto'e barda ke rarna namcu

li 1 bi'i l00 na barda ke rarna namcu = naku zo'u li 1 bi'i l00 cu barda
ke rarna namcu
lo mlatu na barda ke rarna namcu (lo mlatu na namcu)

le cinfo cu barda ke rarna na'e namcu (emphasizing that na'e can apply
to the next word, and that it can be used when the "scale" is unclear to
mean "other-than")

"lo mlatu cu namcu" could also be marked with na'i on the entire
sentence, or on the words mlatu or namcu because it is
metalinguistically inappropriate to talk of cats as numbers.

If I had not strictly defined what "large" meant, then both of the
following could be true

li 200 cu barda
li 200 cu na'e barda

> 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.

correct. I said "sometimes"

> 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.
> :)

moving at all - no

changing the language, unless things are so truly broken that the byfy
(i.e. Robin) can't write it up - no

(and at this point, unless Robin says so, NOTHING is subject to a vote)

I am opposed *in principle* to language change by decree at this stage.
We aren't designing the language any more.

> 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.

"meaning" is itself an ambiguous term

> 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?

only by using it in a way that the word that it marks is a delimiter for
a larger construction (generally one of the construct-terminators or one
of the start-construct words - e.g. fu'e bu'o ke to tu'e) and at the
beginning of the sentence/utterance (where it is afterthought of
nothing). In those cases, the longer scope is inferred from what is marked

NAhE's and NA's scope are defined by the syntax rules

--
Bob LeChevalier loj...@lojban.org www.lojban.org
President and Founder, The Logical Language Group, Inc.

[la gleki]

Well, that ":)" was supposed to mean {zo'o}.

[John E Clifford]

Lordy, where is that solution for subjunctives.  I'd love to see that!

---

From: la gleki <gleki.is...@gmail.com>
To: loj...@googlegroups.com
Sent: Friday, December 7, 2012 3:37 AM
Subject: [lojban] polysemy of {nai}

--