[lojban-beginners] Spaced-repetition based lojban course

[Chris Capel <pdf23ds@gmail.com>]

I sent much of this message as a reply, but it got scrambled, so I
figured I'd expand on it a bit and send it again.

A while back I was thinking about various ways to optimize the speed
and creative mental effort (i.e. the problem-solving kind of thinking
you can't do when you're trying to finish that last math problem)
necessary to learn Lojban. I observed that doing spaced repetitions is
a particularly easy way to learn vocabulary, and indeed many other
items. So the idea came naturally to apply the spaced repetition idea
to *all* lojban knowledge, by formulating it all into items suitable
for use in Supermemo or clones. Then I got a bit distracted. It would
work a lot like the Spanish Pimsleur series that Christopher
mentioned, except that while the Spanish series manually introduces
the spaced repetition effect, by having material repeated in
succeeding lessons, Supermemo schedules each item's next interval
based on its model of your memory.

To have a supermemo-based course be effective, it seems like you'd
have to include some basic concept items, as a general overview of
this or that aspect of the language--a gentle introduction. Then you'd
introduce some more specific points of the grammar, and have them
repeated a few days to firmly settle into the student's memory. Then
you'd introduce example items which use the grammar. Lots of example
sentences, both to and from Lojban, along with all of the vocabulary.
It'd be good to find a middle balance between explicitly explaining
the grammar and usage and having the student learn by example. Besides
explaining place structure, of course, gismu learning could commence
immediately, and would probably take up 40-50% the students effort.

I believe something like this could fully replace the LFB material,
and more. I think it'd easily be the preferred method for learning
Lojban, especially because all needed information is presented to the
user in the correct order--thus no need for looking things up--every
day, without the user having to make decisions about what to review.
There's no creative effort required to decide what chapters to do
next, or what material needs reviewing, or whether one has covered
this or that point of grammar thoroughly enough. There's no saying
"Oh, I need to go write this program or create this or that tool or do
this or that before learning lojban will be easy enough." Of course,
one could do those things on the side, but one's progress in learning
Lojban would not suffer for it.

For instance, I become interested in lojban again a few weeks back,
and started entering items into supermemo. In the past week I've
gotten very distracted from it, but because the items are in Suprmemo,
I'm still reviewing a few of them every day, so I'm not losing any of
the knowledge I'd built up, and when I want to, I can add more
vocabulary starting from where I left off. The dropout rate with a
course like this seems like it'd be a lot smaller than with other
methods. If the the whole grammar, plus tons of examples, were
included in this process, I'd have very little work outside of
supermemo needed to actually use the language.

So I'm wondering what everyone thinks of this idea. Is this something
that you'd find useful? Is the method obviously flawed, and if so, is
there any supplement that could be added to compensate? And, I
suppose, most importantly, does anyone have any interest in taking an
active role in formulating such a curriculum?

I'm still not very sure about the best way to formulate grammatical
knowledge and examples into good items. For reference, you can look
here: http://supermemo.com/articles/20rules.htm
But this doesn't give me a lot of hints.

Chris Capel

On 9/17/05, John Leuner <jewel@pixie.co.za> wrote:
> I read about the Pimsleur method a couple of weeks ago. The main point
> of Pimsleur material is that is carefully constructed using "Graduated
> Interval Recall". I searched the web for any papers or research about
> this model of memory retention but I only found hundreds of sites
> flogging Pimsleur language sets.
>
> Does anyone know more about this theory or have a link to some research?
>
> John Leuner
>
> On Fri, 2005-09-16 at 21:30, Christopher Zervic wrote:
> > I actually posess the entire Spanish Pimsleur series lessons 1-90. You
> > can probably find this or one of the other ones at Your Local Public
> > Library™.
> >
> > Pimsleur uses three voices, a native man, a native woman and a
> > narrator. Having the man and woman are somewhat necessary to
> > straighten out gender issues. The principles are anticipation and
> > repitition, each lesson introduces circa 10 new concepts, and hammers
> > them home. There is also a reading section (which for a language
> > posessing perfect orthographic mapping would be hardly necessary).
> >
> > A typical script would be something in he beginning would be
> >
> > Narrator: Let's say you meet a woman on the street. Say "Excuse me."
> > Listen and repeat.
> > Man: Perdón. (Pause) dón. (Pause) dón. (Pause) dón. (Pause) Perdón.
> > (Pause) Perdón.
> > Narrator: Say "Excuse me"
> > Man: Perdón. (Pause) Perdón.
> > Narrator: Ask her if she speaks English. Listen and repeat.
> > Man: ¿Habla Inglés? (Pause) ¿Habla Inglés?
> > Narrator: Do you speak English?
> > Man: ¿Habla Inglés? (Pause) glés (Pause) glés (Pause) Inglés (Pause)
> > Inglés.
> > Narrator: Do you speak English?
> > Man: ¿Habla Inglés?
> >
> > Little by little they introduce more phrases and each time  Later on,
> > it might sound like this:
> > Narrator: Have you seen Maria?
> > Man: Escucha y repita: ¿He visto María?  (Pause) visto (Pause) He
> > visto
> > Narrator: Have you seen Maria?
> > Man: ¿He visto María?
> >
> > and on and on...
> >
> > I don't know how well this would work with jbo, because Lojban as a
> > spoken language appears to be still a shade or two underripe.