Re: Some things about lojban

["Arthur W. Protin Jr." (GC-ACCURATE) <wetblu!uunet!PICA.ARMY.MIL!protin>]

Hello folks,
    I just read the comments by Yary Richard Phillip Hluchan
and I want to add to it.  I have not read much of anything on
the new grammar, but my comments are based on content of that
letter, my knowledge of computer science, and common sense.

    I have heard from many sources that "most people have a
short-term memory of about seven items," which seems to fit
my observations.  But it was pointed out to me that various
context maintenance and background processing chew up some of
the seven, and it is best for understanding if a paper is
limited to only three ideas in discussion at any point.  If
the three discussion ideas are independent of the parsing
of the English, then there are only four left and using them
all for routine parsing would be cutting it close.  If prefix
notation will free up the variables through chunking (or any
other ploy), then that has to be superior.

    On the subject of machine use of the language, I have two
points.  First, computers are there to serve people, to off
load us of work.  We now write programs in higher level languages,
not because it is any easier for the computer to understand,
but inspite of the extra work the computer has to do, because
it is more productive for us.  Machines are getting faster
and more powerful, and have been for longer than Loglan has been
around.  The only time it makes sense to bias the decisions for
the computer's ease is when the savings is in orders of magnitude.
A 40% reduction in effort for computer manipulation translates into
only a year or less earlier delivery.  A thousand fold (1000x)
difference causes about a 21 year delay which is slightly longer
than I want to wait, but 7 years for a ten fold difference can
be acceptable.

    The second point on machine use of Lojban is simply: I
want Lojban for use by humans!!!  I don't want anything to 
jeopardize that!  

    (If researchers think that they can make computers understand
English, with all its nonsense, why should we have to cut corners
to have them understand Lojban.)


    On a side note (a very side note): I read the photocopy
of the newpaper article about the Lojban-only dinner party
and want to comment on some translations.  Of course there
was a problem with translating "hamburger", because it is really
"Hamburger".  For hints about its translation look to the
"Frankfurter", the "Manhattan" and the cheese "Danish".  The
popular American fast food entree is a sandwich in the style
of Hamburg as presented at the Hamburg Worlds Fair.  For a 
descriptive translation: it is a patty in a bun; where a patty
is an individual portion sized meatloaf, and a bun is a variety
of roll which is an individual portion sized bread loaf.  The
meatloaf is made from shaping a serving of mediumly ground/chopped
meat, usually beef or beef plus others but also reasonably
venison, lamb, or even turkey.  The patty/meatloaf may also
contain: egg; breadcrumbs; chopped onion, garlic, and/or pepper;
and spices.  (I think at this point that the analysis is more
than detailed enough to come up with a better translation, and
I there will suspend it,  unless, of course, someone out there
wants more of this abuse.)


From phri!sci.ccny.cuny.edu!rpi!zaphod.mps.ohio-state.edu!think!snorkelwacker!bloom-beacon!eru!luth!sunic!mcsun!hp4nl!ruuinf!praxis!jagversm Thu May 24 11:29:38 EDT 1990
Article 4102 of sci.lang:
Path: marob!phri!sci.ccny.cuny.edu!rpi!zaphod.mps.ohio-state.edu!think!snorkelwacker!bloom-beacon!eru!luth!sunic!mcsun!hp4nl!ruuinf!praxis!jagversm
>From: jagversm@praxis.cs.ruu.nl (Koen Versmissen)
Newsgroups: sci.lang
Subject: possible readings of "John seeks a bike or a fish"
Keywords: natural language, quantification, intensionality
Message-ID: <3320@ruuinf.cs.ruu.nl>
Date: 22 May 90 12:29:36 GMT
Sender: news@ruuinf.cs.ruu.nl
Lines: 40
Status: RO


What are the possible readings of the sentence
>John seeks a bike or a fish< in natural language?

Technically there are, I think, eight readings:
"to seek" being an intensional verb, both "bike" and "fish"
can be either intensional or extensional. Furthermore, the
disjunction can apply to either the speaker or John.
Let me paraphrase:

1. There are a bike and a fish, and John is trying to find
   one of these (he doesn't care which one).

2. There are a bike and a fish, and John is trying to find
   one of these (but I don't know which of the two he's
   actually looking for).

3. There is a bike, and John is trying to find either this
   bike or a (possibly non-existent) fish (he doesn't care...)

4. There is a bike, and John is trying to find either this
   bike or some fish (but I don't know...).

5. & 6. Similar to 3. & 4., but with the roles of "bike" and
	"fish" interchanged.

7. John is looking for a (possibly non-existent) bike or a
   (possibly non-existent) fish, and will be satisfied when
   he has found either one.

8. John is looking for a bike or a fish, both possibly non-
   existent, but I don't know which one he's looking for.

I hope this is clear.
This came up during a talk on quantification in natural
language. A theory thereof should account exactly for those
readings acceptable in natural language. But which ones are
in fact acceptable?

Koen.


From phri!sci.ccny.cuny.edu!rpi!zaphod.mps.ohio-state.edu!samsung!cs.utexas.edu!uunet!munnari.oz.au!uhccux!lee Thu May 24 11:30:21 EDT 1990
Article 4105 of sci.lang:
Path: marob!phri!sci.ccny.cuny.edu!rpi!zaphod.mps.ohio-state.edu!samsung!cs.utexas.edu!uunet!munnari.oz.au!uhccux!lee
>From: lee@uhccux.uhcc.hawaii.edu (Greg Lee)
Newsgroups: sci.lang
Subject: Re: possible readings of "John seeks a bike or a fish"
Message-ID: <7809@uhccux.uhcc.hawaii.edu>
Date: 22 May 90 19:01:19 GMT
References: <3320@ruuinf.cs.ruu.nl>
Organization: University of Hawaii
Lines: 62
Status: RO

>From article <3320@ruuinf.cs.ruu.nl>, by jagversm@praxis.cs.ruu.nl (Koen Versmissen):

>What are the possible readings of the sentence
>>John seeks a bike or a fish< in natural language?
>
>Technically there are, I think, eight readings: ...

As to what senses are actually possible in English, I couldn't say.
I recall that Montague argued in "A Proper Theory of Quantification
in English" that `seek' is not the same as `try to find', but I
can't remember how that went.

Approaching the matter transformationally, one would make somewhat
different predictions from yours.  I think some of your senses would
violate movement constraints.  Here are the possible forms I've
come up with so far, using conjunction-reduction or a kind of
pseudo-combinatory notation:

(a)= 2	a bike John-tries John-finds or a fish John-tries John-finds
(b)= 7	John-tries( (a bike John-finds) or (a fish John-finds) )
(c)= 8	John-tries(a bike John-finds) or John-tries(a fish John-finds)
(d)	John-tries(a bike John-finds) or a fish John-tries John-finds
(e)	a bike John-tries John-finds or John-tries(a fish John-finds)
(f)	John-tries(a (bike or fish) John-finds)
(g)	a (bike or fish) John-tries John-finds

Only the first 3 correspond to senses you enumerated.  The others
are:
 (d) John is trying to find a possibly non-existent bike, or there is
     a fish that John is trying to find.
 (e) There is a bike that John is trying to find, or John is trying
     to find a possibly non-existent fish.
 (f) John is trying to find a possibly non-existent thing which, to
     him is either a bike or a fish, he's not sure.
 (g) There is a bike or a fish, I'm not sure which it is, that John
     is trying to find.

There follows a key to the reductions from more standard logical
forms to the above, and a couple of sample reductions:

John-finds(x), bike(x), fish(x)
John-tries(f(x)) -> (John-tries(f))(x)
(Ex)f(x) -> some f
some(f and g) -> a (f g)
f(x) & g(x) -> (f and g)(x)
f(x) v g(x) -> (f or g)(x)

"There is a bike John finds or there is a fish John finds."
(Ex)(bike(x) & John-finds(x)) v (Ey)(fish(y) & John-finds(y))
-> (Ex)(bike and John-finds)(x) v (Ey)(fish and John-finds)(y)
-> (some and bike John-finds) v (some and fish John-finds)
-> (a bike John-finds) v (a fish John-finds)
-> (a bike John-finds) or (a fish John-finds)

"There is either a bike or fish that John finds."
(Ex)((bike(x) v fish(x)) & John-finds(x))
-> (Ex)((bike or fish)(x) & John-finds(x))
-> (Ex)((bike or fish) and John-finds)(x))
-> some ((bike or fish) and John-finds)
-> a (bike or fish) John-finds

				Greg, lee@uhccux.uhcc.hawaii.edu