NAI

[Logical Language Group <lojbab@ACCESS.DIGEX.NET>]

>From: ucleaar <ucleaar@UCL.AC.UK>
>Subject:      Re: NAI
>
>> Since NAI is a permissible standalone word in an utterance, it is
>> a mistake to consider it solely a suffix.
>
>What would a context be for an utterance containing only NAI?

Umm.  Elliptical denial of someone else's statement or expression, where
the ellipsis can clearly be inferred to apply to a word that takes NAI

a:  mi'o klama le ckule
b:  .ui
a:  nai

Note that NAI can go on other things besides UI, but those are the
easiest to use in examples.

>Is {du'u nai kei} grammatical? What does it mean?

I doubt it - du'u requires a bridi, and there is none.  Standalone NAI
is a sentence fragment, just like bare sumti, bare attitudinals, bare
tenses, etc.

>> It is of course a word because it meets the Lojban definition of a word.
>
>I don't know what that is.  But at any rate, I meant "word" in the vague
>but general way it is understood in linguistics.

I think there are more than one such definition.  Even Lojban in a sense
has more than 1 definition.  To me "lenu" is psychologically 1 word,
even though it is grammatically 2 words.  To call it 1 word DOES require
some kind of idiom status, since a parse of a construct using lenu
places the "le" separate from a construct "nu bridi kei".

>> IN short, I do not see what the point of your claim is - it sounds
>> like you wish to choose another definition of "word", one which
>> complicates the morphology and the grammar of the language.
>
>There isn't an agreed definition of "word" in linguistics (and it is my
>personal belief that grammar contains no entity with characteristics
>corresponding to the meaning of the English word _word_), but linguists
>would generally agree that a question like "does this utterance contain
>20 words or 1?" has some empirical content and is worth debating.

My claim was that the definition of "word" linguistically is determined
by the characteristics of each language.  In the case of Lojban, we KNOW
the characteristics of the language fully, and decided in advance what
set of characteristics corresponded most closely to linguistic "word" as
we/JCB understood it.

>I am not advocating complication of Lojban morphology.  I am tentatively
>claiming that it already is more complicated than had hitherto been
>thought.

If it can be understood as being simple using a modified paradigm with
*slightly* non-standard definitions, especially of terms that have
rather broad definitions in the first place, then I think that the
morphology itself thereby proves the definition is appropriate.  Just as
the definition of "word" has to adapt in an heavily agglutinative
language, so might other linguistic terms.

>At minimum, a word must occupy its own node in syntactic structure, and
>I was suggesting that NAI doesn't, and is therefore not a word.

It does, since it is defined to do so.

>I gave two reasons.  The first is semantic:  Lojban in general has no
>idioms - the sense of a phrase is fully predictable from the meaning of
>its parts, whereas the sense of a word is not fully predictable from the
>meaning of its parts.

I think it is not true that Lojban has no idioms.  It may have few
of them at this point.

But given this definition, "nai" must be a separate word, since if
"uinai" were a single word, then by the above its sense could not be
fully predictable from the meaning of its parts.

>Second, and more interestingly, UI are in general invisible to other
>words, but they appear to be visible to NAI.

They are also visible to SI, CAI, ZO, ZOI, and other UI.  A variety of
other words are all "recognized" by NAI, but not all of them.

>How so?  This is accounted for if the bond between UI and following NAI
>is morphological.

But the straight explanation works too.  Lojban has two sublanguages,
that of predications, and that of attitudinal discursive metalinguistic
expression, which may be expressed simultaneously, but interact only
minimally.  NAI has a role in each sublanguage.  Few other words do.

>> Just as wishing that we called the apostrophe (or h) a consonant
>> because it happens to be one in most of linguistics, but would make
>> the Lojban design less clear, is less than productive.
>
>I have always approved of what Lojban nowadays does in most cases where
>it had used technical terms nonstandardly:  it uses Lojban.  So we've
>dropped "lexeme" and use "selmao" instead, etc. etc.  In this case a
>Lojbab term to replace "consonant" would be appropriate.

Generally we have done so when a) a JCB usage of a standard term has
major misunderstandings because people read the "message" entirely
wrong; and b) when we have a concept for which there is no corresponding
standard English term.  The Lojban sense of consonant(morphological) is
close enough to the definition of consonant(phonological) that I think
it should be acceptable.  If I were a wise and knowledgeable linguist, I
would find and point out parallel examples from other languages, where
perhaps using a standard definition of a term leads to an incorrect
generalization about the language with lots of apparently unexplainable
exceptions, whereas a slightly modified definition eliminates the
exceptions.

>But in my point about NAI I was not quibbling about terminology.

You were, more or less, quibbling with our term "word"

lojbab