[lojban] Re: Loglish: A Modest Proposal

[John E Clifford <lojban-out@lojban.org>]

--- Ben Goertzel <ben@goertzel.org> wrote:

> 
> 
> Hi,
> 
> > So, you will just use Lojban grammar with
> English
> > words for the predicates -- well, English +
> "qui"
> > and "quu" (which latter is looking less
> useful
> > the more Lojban comes in -- {sepi'o} in place
> of
> > "quu weapon" for example (also simplifying
> the
> > matvch up to FrameNet categories).  This
> still
> > leaves all the Lojban cmavo to learn --
> without
> > the at leat occasional aid of matches with
> gismu.
> >  It might be better to use English for at
> least
> > BAI type cmavo to make the learning easier. 
> Of
> > course, then we again wonder about how much
> tis
> > is based on Lojban and whether the grammar
> can
> > take the shift (predicate quu would have to
> go
> > before, not after, probably).
> 
> Clearly, learning the cmavo is far easier than
> learning the cmavo + all the vocabulary.

Thgough not, I think, as much less as the numbers
suggest.  Notice, also, that all this is largely
predicated on continuing to learn Lojbna in the
worse known way: as isolated pieces of vocabulary
and syntax.  Hopefully, someone will soon (after
a mere 50 years) get around to an immersion
teaching system for Lojban (steal and modify any
of the dozens already out there for Uzbeki and
Tlon) and reduce the learning cost to the point
where it will largely (though never completely)
overcome this objection.  In theory, Lojban ought
to be completely learnable in a day and total
facility in a werek. The grammar part plus cmavo
can hardly be done in less.
 
> > > 1) Loglish is much easier to train users in
> > > than Lojban
> >
> > Yes, because of the vocabulary learning.  The
> > rest would be on a par, with the cmavo
> learning a
> > bit harder perhaps. I don't know how
> significant
> > the difference would be.
> 
> IMO, for an individual with some background in
> logic and a high IQ -- but without a
> photographic
> memory -- learning the vocabulary will be
> by far the most time-consuming part of learning
> Lojban.

Based, as noted, on a totally deplorable teaching
technique.  The fact that it is separated off
like this shows how bad things are.  There will
always be some cost, but it need not be separated
out.  How will total learning of Lojban compare
with total leaning of Loglish -- the cmavo will
be the same, the grammar will be about the same,
the vocab will slightly favor Loglish -- but
learning restrictions is often harder than
learning new words altogether (a similar remark
applies to grammar, so Loglish may have more
problems ther as well).
  
> For an individual with lower IQ and no
> background
> in logic, the formulation of thoughts in
> predicate
> logic may be difficult and unnatural, and may
> form
> the largest obstacle to learning Lojban.  I
> don't
> know if anyone has any experience with this --
> so
> far as I know, everyone who's bothered to learn
> a
> significant amount of Lojban is pretty clever
> and
> with an explicitly rationalist mindest.

Rationalist maybe, not always obviously rational
and in some cases demonstrably unacquainted with
and not competent in Logic.  This has not been a
hindreance in learning Lojban nor even of
reaching some of the higher levels of mastery.
Ithink we put too much stress on the relation
between the underlying syntax (vaguely predicate
logic) and some assumed thought processes.
 
> > > 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)
> >
> > I am not sure I understand this.  If the
> enquiry
> > is in Loglish than (grammar aside) the text
> can
> > be processed like ordinary English, using
> tools
> > devised for that ordinary language.  If the
> > inquiry say is in Lojban, either these tools
> have
> > to be replicated for Lojban or the Lojban
> query
> > has to be translated into English at least to
> the
> > extent of using English words in place of
> Lojban
> > ones.  Now, in Loglish this will have to be
> done
> > already with the cmavo, I suppose.  Anbd the
> > replacement in the case of predicates will
> > require  so fairly fancy processing toight
> > translation in the context (working through
> > WordNet and FrameNet and whatever else is
> > available.  On the other hand, presumably
> words
> > are ultimately to be represented as WordNet +
> > FrameNet entries for the purpose of
> processing in
> > the various ways.  If Lojban words were once
> so
> > represented, then -- without passing through
> > English -- they could be used directly,
> sparing
> > the need to diambiguate (or pick the right
> > meaning given all the contextual stuff). 
> That
> > is, the small extra effort (OK, not so small)
> > pays off bigtime in the end (and points the
> way
> > to a mass of databases that are less
> dependent on
> > English (or Lojban for that matter) than the
> > present plan.
> 
> Yes, if a Lojban dictionary were created that
> 
> -- included cognates of all the words in
> English

I would think that the ultimate aim was to get
away from English altogether.

> -- included mappings of each Lojban word into
> appropriate WordNet senses
> 
> -- included mappings of each Lojban
> argument-position
> into an appropriate FrameNet case-role

These seem like likely goals in any case (not
necessarily using WordNet and FrameNet, but some
such schemata).  Since the basic Lojban
vocabulary is small, unambiguous, and still
somewhat under its designers' control, the
fundamental part could be done fairly rapidly
(does WordNet have a set of basic notions in
terms of which all others are defined as FrameNet
seems to have a set of relations to cover all
cases?).  The task of accounting for the derived
values for lujvo -- and the imported values for
fuhivla -- will be more complex but still
relatively easy compared to dealing with the
whole of English -- or even a reasonable sample
(the classic 10,000 words, say).
To be sure, given the nature of the Lojban
community, some proposals will be disputed and
modified but this can be restricted, as has been
done already for cmavo (indeed, the process into
which this project might fit is already under
way).  We may not have what is needed at the
moment, but it is not so far off as you seem to
think (nor as expensive).

> then Lojban queries could be used to query
> English-
> language databases.
> 
> However, this kind of Lojban dictionary does
> NOT exist,
> and
> 
> 1) Creating it looks like a big job, far bigger
> than building a Loglish parser (since the
> latter can
> be easily done by combining the ideas of the
> existing Lojban
> grammar parser with my current software tools
> for handling
> WordNet and FrameNet)
> software tools
> 
> 2) So far as I can tell, no one seems to have
> the money
> or initiative to get this big job done.

See above.
 
> > > 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
> >
> > Word inventing doesn't take much time and
> only
> > need be done once in a given topic.
> 
> The need to do word-invention makes language
> usage an awful
> lot more difficult though.
> 
> Suppose I'm talking to a friend in Lojban
> about, say, polysemy
> 
> Then, if I don't know the word, I have to make
> up a word for
> polysemy, so we can use it in our conversation.
> 
> But if I'm not near a computer, I can't check
> whether someone
> else has already made up a word for this and
> uploaded it to
> jbovlaste....  So it may be that the word we
> made up is "wrong"
> in the sense of not being the first one
> invented for the intended
> meaning....
> 
> Even if I'm sitting at or near a computer,
> every time I encounter
> a word I don't know, I need to consult
> jbovlaste to see if it's been
> invented before, and otherwise I need to invent
> it.
> 
> This makes conversation in Lojban a lot slower
> than conversation in
> other languages -- NOT because of any intrinsic
> flaw 
=== message truncated ===

<<just because Lojban is young and not yet
complete.>>

Again, this is a flaw in present teaching
methods, where each word is learned separately. 
Properly traiined --  or indeed merely having
reached genuine competence -- a speaker would
have intenalized words in at least any area he
was involved with and would not have to keep
checking (any more than he does in English when
he strays from his area of knowledge).  And, of
course, there is not such thing as a wrong new
word in Lojban, except one tht replicates another
with another meaning.  But this is rather
unlikely, given the patterns of lujvo formation
(though nothing like impossible).  However, if a
word works in the present conversation, not being
confused with another meaning, then no harm is
done by this duplication.  It can be cleared up
later (as it at least occasionally has to be in
scientific jargon).

<<Of course, one can always use English words in
Lojban via the quotative
mechanism, but if one is going to do this
heavily, then one may as
well use Loglish ;-)>>

This is, of course, totally a bad idea of the
worst sort, though admittedly possible. I would
prefer (and all the teaching plans agree on this)

that the option not ever be mentioned in Lojban
learning (I remember the disasters it used to
crrate in Loglan, where whole conversation went
on in "Loglan" consisting of only Loglan
quotations -- and referential transfers.)

<<> And, of course, the lack of
> vocabulary is a temporary matter, not a long
term
> one -- and the sooner people start using Lojban
> to talk of many things, the sooner the problem
> will disappear.

Well, realistically, I am not that hopeful that a
lot of people
are going to start using Lojban to talk about a
lot of things
anytime soon.

I would like to see it happen -- I hope it
happens -- but it doesn't
seem terribly likely to me, frankly.>>

I agree, alas.  And it soes not even seem
feasible so long as people learn Lojban as they
are encouraged to do (a few have somehow managed
to get around or beyond that to more or less
correct learning techniques and they can do the
sort of thing we have in mind.  They are
presently too few to make a thoroughly
functioning community to generate the material
needed.)


<<I don't think a lot of people will start
learning and conversing in
Loglish any time soon either, even if a parser is
created and nice
educational materials are created.>>

Well, I think that if the learning could be
improved, a far larger segment of the present
community could be brought up to speed and that
would appraoch at least the critical mass needed.

<<However, I *do* think these languages may play
a valuable practical
role in mediating communication between humans
and AI's.

In this role, both Lojban and Loglish have their
own advantages.

Lojban requires less of the AI's, because it's
much less ambiguous
(the only significant ambiguity, so far as I can
tell, residing in tanru
and in reference resolution mechanisms).>>

Yeah, it has actually managed to have a reference
system that is less effective than English, a
fairly remarkable achievement in its own right
(to be sure, it has, in theory, a foolproof
reference system, but it has proven unworkable
even for written communication).

<<On the other hand, I believe Loglish requires
less of the humans, because
it requires much less vocabulary memorization and
invention.

I think that beefing up parsing software to
handle Loglish is going to be
easier than convincing a lot of humans to go
through the extra hassle of
learning Lojban in order to communicate with
AI's.

> As noted, Lojban grammar is close enough to
> English that the problem will likely be the
> restrictions, not new constructions. The
> experience with radically different grammars
from
> the home one would suggest that -- if Loglish
> really had such a different grammar -- it would
> be worse than learning Lojban vocabulary to get
> doing it right.

This is the main point on which our opinions
differ.>>

The difficulties of dealing with restricted or
different grammars while keeping vocabulary (like
the difficulties in sticking to restricted
meanings of familiar words) is pretty well
documented over a range of cases.  Using the
correct grammar when the vocabulary is novel has
also been tested out positively.

<<> > 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.
>
> Would Loglish seem (or be made to seem) less
> weird?

Yes, definitely.  From a marketing perspective,
one can pitch Loglish as a

"more logical version of English, modified to be
easier for computers to
understand"

which is a far better pitch than saying you want
knowledge-encoders to learn
a whole
new language.

It's not really quite true that Loglish is a
"more logical version of
English",
but this seems to me like a viable sort of
marketing pitch, and I have had
no
success in coming up with a workable pitch for
Lojban....

In a presentation context, you could show people
simple Loglish sentences
and they
would be able to basically understand them. 
OTOH, showing people Lojban
sentences
will just result in total bafflement...>>

Yes, I agree that seelling Loglish would be a lot
easier and that, therefore, if either of these
ideas is to get off the ground, Loglish is the
one that will make it. My point is only that this
is a sad state of affairs, since the adevantages
ultimately lie with Lojban (or some improved
version thereof).


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.