Message-Id: <m0pfYQ2-0000R2C@xiron.pc.helsinki.fi>
Date:         Sat, 12 Mar 1994 13:28:28 EST
Reply-To:     jorge@PHYAST.PITT.EDU
Sender:       Lojban list <LOJBAN%CUVMB.bitnet@FINHUTC.hut.fi>
From:         Jorge Llambias <jorge@PHYAST.PITT.EDU>
Subject:      Re: Even Madder (A no-go :-)
To:           Veijo Vilva <veion@XIRON.PC.HELSINKI.FI>
Content-Length: 3115
Lines: 85

Veijo joins in the fun:

> It occurred to me last night that -- had we the option -- it would
> be possible to construct a system of connectives which would be
> completely regular and fulfill all the requirements except perhaps
> that of maximal brevity. It would have no overloading and still
> utilize a minimal number of cmavo.

I don't agree that overloading, as you are defining it, is a problem.
The connectors retain their meaning when connecting different types
of objects. The result is different because the connected objects
are different, not because the connectors mean different things.

> The structure would be:
>
>     [gi operand-type connective] operand1 operand-type connective
>        operand2

The problem with this is that you could use, say {ja} in the prenex
and {je} in the middle, and what would that mean?

I think that the whole point of forethought vs. afterthought is to be
able to specify the connection first, or in the middle. (We could come
up with a proposal to allow "final connection" too, but better not... :)

If you wanted a different connective for each, you could use:

 gije ..... gi  .....
      ..... gije ....

 guje ..... gu  .....
      ..... guje ....

 gaje ..... ga  .....
      ..... gaje ....

etc. one for each type of connection (tanru, sumti, bridi, bridi-tail,
and paragraph if you want).

But that would not be more clear than "overloading", because you'd
be giving the vowels two different interpretations: which logical
connective, when used with "j", and which type of connection, when
used with "g". It would in fact be much more confusing.

As it is now, GA is used for most forethought connection, so that
the "overloading" is already there. {je} is also used for both
bridi and tanru connection. Yet nobody is confused by the overloading,
because the surrounding text makes it immediately clear what type of
connection is being used.

To me, bridi and bridi-tail connection are not really two different
concepts, so to have different words for them is an artificial
distinction. Sumti connection is different from bridi connection
(although in the case of the logicals, you can expand from one to the
other), and tanru connection is a third type. But the distinction
is given by the elements being connected, not by the connectors
themselves, which keep their meaning.

> I know this system doesn't stand a chance but I wanted to present
> it to give you some food for thought :-)

You did, thanks :)


> BTW. is there a zo'e type and/or vague connective?

Yes, {do'e} of selmaho BAI. Like all tags, it can be used as
a connective, although not everywhere. It's restricted to
forethought connection {do'egi....gi....}, and bridi
afterthought {.... .ido'ebo.....}.

My proposal would probably (I don't know if John YACC-checked this one)
allow {.... gido'ebo ....} as well.

Now it can't be used with bridi or tanru afterthought, but I'm
not sure why, since {mi do'ebo do} doesn't seem to conflict with anything,
and neither does {blanu do'ebo xunre}.

This might be material for a MAD PROPOSAL NUMBER 5.

>
>   co'o mi'e veion
>

co'o mi'e xorxes