Message-Id: <199310081948.AA27355@eli.CS.YALE.EDU>
Date:         Fri, 8 Oct 1993 15:45:00 EDT
Reply-To: protin@USL.COM
Sender: Lojban list <LOJBAN%CUVMB.bitnet@YaleVM.YCC.YALE.EDU>
From: Art Protin <protin@USL.COM>
Subject:      Re: deleting places
To: Erik Rauch <rauch>
Status: RO

While I am tempted to digress into the connotation of cliva being
leaving with or without travelling (Is this leaving your
girlfriend), I think it is much more productive to pursue
the native learning issue.

Jorge quotes me and comments:

>>         I believe that a property of teaching language by example
>>         is that the place structure will have to be somewhat looser
>>         and people will learn gismu initially as having minimal
>>         place structure.
>
> Well, in theory you don't understand the correct meaning of the gismu
> if you don't know all the places. That's in theory, of course.
>
>>         The less commonly used places will be
>>         learned later as enhancements of the base concept.
>
> This shouldn't be the case. That's why I dislike gismu with more than
> three places, even though some seem necessary. But in practice, I
> agree.

I don't know all the different, documented ways that children
learn language, but I paid very close attention to how my own
two children first learned to speak.  Words seemed to represent
whole events.  Examples are "up" meant "you, the listener, pick
up me, the baby"; "food" meant a meal and if the tone of voice
expressed desire it meant "please repeat the event of me, the
baby, eating a meal".  First, word were modified only by tone
of voice, as "food" meant either "this is a meal" or "I want
a meal".  Then, they were modified by gesture, like baby in
my arms points to doll or bottle and says "up".  The first
multiword sentences were the simple juxtapositioning of these
concept words, seemingly to use one concept to modify another.
For example, "eat you" was the same as "you eat" and it meant
that "you, the listener" replaced "me, the baby" in the eat-a-meal
scenario.  It was a major transition when they recognized the
beginnings of grammar - that word order actually meant something.

    Given the experience of these two children learning to speak
a language, English, entirely from example, I expect, if lojban
is ever successful with native speakers, that a lot of people
will violate the theory and learn gismu with only the place structure
they need to express their experiences to date.
    I strongly suggest that our design of this language should
safely accomodate most common failures and I expect eventually
that most speakers of lojban will not be familiar with the full
place structure of most gismu, that similar concepts will be
corrupted with erroneous places, and that new places will be
added as native speakers think about things differently because
of the language that they start from.

    (Side note: my (old?) gismu list shows "manku" as a one place
gismu for dark.  My loglan/lojban training leads me to expect it
to have at least two places as "x is darker than y".  I gather
from looking at blue and red that that property of loglan was
dropped.  Is this part of the reason I am seem so out of step
is that I still remember from the '70s "da blanu de" and
"da blanu" with the explanation that the second of these was
"x is bluer than some unspecified reference, ergo x is blue"?
(YES, I have been following (albeit irregularly) this language
for more than 20 of its more than 30 year history.))

    thank you,
    Art Protin


Arthur Protin <protin@usl.com>
STANDARD DISCLAIMER: The views expressed are strictly those of the author and
are in no way indictative of his employer, customers, or this installation.