Cc: "Ari Heljakka" <heljakka@iki.fi>,
        "Izabela Lyon Freire Goertzel" <izabela@goertzel.org>
Date: Mon, 15 Aug 2005 15:45:05 -0400
Message-ID: <JNEIJCJJHIEAILJBFHILKELPFFAA.ben@goertzel.org>
MIME-Version: 1.0
Content-Type: text/plain; charset="iso-8859-1"
In-Reply-To: <20050815192714.98630.qmail@web81306.mail.yahoo.com>
Importance: Normal
Errors-to: lojban-list-bounce@lojban.org
To: lojban@yahoogroups.com
From: "Ben Goertzel" <lojban-out@lojban.org>
Reply-To: ben@goertzel.org
Subject: [lojban] Re: Loglish: A Modest Proposal
Content-Length: 3386
Lines: 77


> Let me see if I understand what would be the
> point of all this.

The main point is pragmatic....

It seems that, after a pretty modest amount of work specifying Loglish
grammar and writing a Loglish parser (utilizing WordNet, FrameNet, and the
existing Lojban grammar), one would have a situation where

1) Loglish is much easier to train users in than Lojban

2) Loglish queries can automatically be used to query databases of knowledge
built up using English-language information-extraction tools (whereas to do
this using Lojban would require building a real Lojban-English dictionary,
including a translation of Lojban words into WordNet senses and Lojban
argument-positions into FrameNet argument labels, or something similar)

3) Loglish can immediately be used to discuss complex topics in any area of
discourse without needing to spend time continually inventing new vocabulary
words

The potential Achilles heel is, it might turn out that Loglish is a much
bigger pain to speak than Lojban, because of a psychological difficulty with
using English vocabulary within non-English syntax.  I really don't think
this will be a problem, but I can recognize it as the major risk with the
Loglish idea.

I emphasize that this is not entirely a theoretical discussion.  In
2003-2004 I managed a project building an NLP information extraction system
for a government customer -- and in late 2004 I actually tried to sell them
on making a Lojban query front end to the knowledge repository, but the idea
was just too weird for them.  We did build for them a system that lets the
user enter knowledge and queries in English, and then has a parser that
presents the user with a menu of possible parses of the sentence -- the user
then has to choose the correct parse (based on their semantic insight),
which is a pain.  I have a feeling that Loglish -- though still very weird
by ordinary societal standards -- might potentially be non-weird enough to
get funding from this customer or another one, in order to fund the building
of knowledge bases using Loglish and the building of fully-featured
professional Loglish software tools.

-- Ben G

>We can train a machine up in
> Loglish and then we can talk to it -- feed it
> information orask it question or program it or...
> .  Clearly we can do the same in Lojban and
> rather more easily, since Lojban is already open
> to unique parsing and decomposition, whiles
> Loglish needs to have it grammar specified to
> reach that point (and then people need to be
> trained to stick to that grammar).  Now one of
> the thngs we want the machine in question to do
> is process masses of linguistic data in English
> (I gather).  But these processes -- other than
> perhaps actual translation -- have nothing to do
> with Loglish, so, assuming the processing can be
> done, asking for it, directing it, and the like
> can be done as well in Lojban as in Loglish.  Ah,
> but Lojban has a limited vocabulary whereas
> Loglish has all of English (with some
> restrictions imposed by decomposition processes
> and parsing and perhaps other things (which
> restrictions people have to learn and abide by). >


To unsubscribe from this list, send mail to lojban-list-request@lojban.org
with the subject unsubscribe, or go to http://www.lojban.org/lsg2/, or if
you're really stuck, send mail to secretary@lojban.org for help.