Re: Getting Married

[cowan (John Cowan)]

la djim. kartr. cusku di'e:
> Some predicates are inherently 
> commutative, some are non-commutative, and some can be seen either way
> (like cinba "kiss").  Now these are semantically the same:
> 
> 	la xeris. speni la selis.	la selis. speni la xeris.
> 	Harry is married to Sally	Sally is married to Harry
> 
> However, binxo "become" is not commutative if you define it as "become ...
> under conditions ..." or in my style, "x1 changes so event x2 becomes true".

I do not understand how events can "become true".  An event happens or it
doesn't happen, whereas a predication is true or not true.
The place structure of "binxo" is

	x1 becomes x2 under conditions x3

which makes no reference to an event.  "nu binxo" is the event of something
becoming something else.

> Specifically x1 changes.  Written out in full:
> 
> 	la xeris. binxo le nu la xeris. speni la selis.
> 	Harry becomes (married to Sally)

What this says is "Harry becomes the event of Harry marrying Sally", and
is false.  A person cannot become an event (under any normal epistemology,
at least).  A better translation would be:

	la xeris. binxo lo speni be la selis.
	Harry becomes the-one-who-is married to Sally.

> This doesn't say that Sally becomes anything.  In lojban-today, you
> could conceivably say
> 
> 	se binxo le nu la xeris. speni la selis.
> 	It comes to pass that Harry is married to Sally (commutatively)

Nope, that has to mean "Harry marrying Sally becomes something-or-other."
What an event can become is not clear to me: perhaps completed or interrupted
or some other predicate such as those.

> However, under my proposal this would need an explicit "da" placeholder 
> in binxo x1 because "Harry" would be retro-replicated into a vacant binxo 
> x1.

In Lojban, sumti places can >always< be elided.  There is no such thing as
a required placeholder, except to keep the places straight (which can be
done with place markers of selma'o FA).

> What you really need is the explicit "and vice versa" operator :

It is "soi", which comes in several flavors.  The most basic kind is 
"soi <sumti> <sumti>", where the sumti are typically vo'a-series pro-sumti.

	la xeris. speni la selis. soivo'avo'e
	Harry is married to Sally reciprocally-between-1st-and-2nd-places.

If the 2nd sumti is omitted, it is presumed to be a reference to the
immediately preceding expressed sumti:

	la xeris. speni la selis. soivo'a
	Harry is married to Sally reciprocally-with-the-1st-place.

This syntax allows other kinds of reciprocity involving other places:

	mi klama le zdani le zarci soivo'evo'i
	I go to-the home from-the office reciprocally-between-2nd-and-3rd
	I go to the home from the office and to the office from the home.

> (It might be seen as a blemish that sometimes you need to plug some place
> (case) with "da" to block retro-replication; however, it's much more 
> common that you have to drag an argument of an event out into the main 
> phrase -- retro-replication is justified on Zipfean grounds.  And I just
> think more naturally with retro-replication available.)

If your replication rules can be made useful in Lojban at all, it is only
in the context of lujvo place structures, which are currently not defined.
(-gua!spi doesn't have a distinction like that between tanru and lujvo,
for the benefit of everyone else.)

-- 
cowan@snark.thyrsus.com		...!uunet!cbmvax!snark!cowan
		e'osai ko sarji la lojban