selma'o PU and BAI

[cbmvax!uunet!math.ucla.edu!jimc]

Lojbab suggested that some of this would be of general interest on the
list, so here is an edited version.  > = jimc, >> and unmarked = lojbab.

>Back in the old days, I (jimc) also noted the irregularity that some
>modal operators (Lojban <BAI>) were effectively converted with (the
>equivalent of) "se" and some were not. 
>A count of the then-prevailing modal operators showed that more
>of them had the tagged sumti in x1 than in x2.  This corresponds to
>Lojban's without-se pattern.  The concept of putting a conversion on a
>modal gave some interesting sentences, but was not accepted by the
>Institute....

>It's true that the Institute grammar failed to split up the compound
>tenses into their separate words.  However, in 1980 pc published a 
>definition of how to interpret
>the compound tenses as nested offsets.  In the same essay he defined the
>tense reference point; the main PU is an offset from the reference.  I
>thought it was neat, and still do.  It is my impression that this
>definition for tense offsets is still in use by Lojban.  However, the
>interval specifications have been expanded quite a bit.

A good summary of what happened, yes.  The intervals expanded to
dominate while intensional and extensional tenses were added to make the
language more natural.  Then the perfective set (ZAhO) was added to
match our now heavy emphasis on events - but that set is really part of
the 'interval' grammar'.  Finally we added directions in space to allow
PU and VI compounds to be truly in parallel, even though we don;t really
in any natural language treat space the same as time.

>> Actual usage has of course moved the lexeme BAI members into near
>> equivalence with associated gismu,
>
>A natural result of linguistic drift by people trying to learn the
>language.

An excellent point, and one that needs to be made to everyone (and one
you won't like where I lead it to, jim).  Right now, people are learning
the language, and so the drift you are talking about takes place towards
the famiuliar.  It could also have drfited towards the malglico.  John
C. has reminded me that "ci'a", "cu'u", "ja'o" "du'o" and "ca'i" are all
incredibly malglico - they were added because we have English uses to
parallel them that I think could/should be expressed by tagged sumti in
Lojban.  ("cu'u" by the way was your old 'freemod' marker for
identifying the speaker of a quote.  It originally meant "said" in the
sense "said John".  I added your freemods to BAI if I saw that they
could take a sumti.  Then added a few more.

But tied to "cusku", "cu'u" means any form of 'expressor', not just
verbal, and your limited but clear use starts to get foggy.

The BAI words were not, and I still think ARE not intended to be exact
abbreviations for their corresponding "fi'o gismu" constructs.  They are
related, in some cases closely, but I think there is a metaphorical
component in them, possibly a malglico one (especially in the case of
"ja'o" and "ci'a" if unchanged).  If you want to exactly express a
subordinate selbri, use fi'o, not BAI, which is so highly ellipsized that
it becomes something else from its corresponding fi'o.

Indeed by linguistic drift, with both features in the language, BAI will
inherently come to mean something different from the fi'o 'equivalent'.
They will move apart because they are both there TO move apart.  But this
will only occur, if at all, after people have learned the language.  Until
then, BAI will drfit towards fi'o gismu.

		-- >jimc >>lojbab