coirodo
I've been thinking for a long time this idea, and I wanted to write it
down in order to discuss it (and not to forgive it). What if we
stablish some:
*Levels* (beginner, intermediate, expert...)
I know, it's really obvious this was discussed before, but I think I
have something larger in mind.
(this is the obvious part)
A level shall be a series of contents, or concepts of lojban grammar,
along with a quantity of minimum vocabulary. There would be 3 or more
levels (the more, the better), in ascending order of difficulty.
(this may be not)
A level shall be very well defined. With this I mean to enumerate the
possible grammar structures that can be in a lojban text readable for
a student of the level. The vocabulary for that level should be
divided in the existing categories (gismu, cmavo, lujvo...), and
establish a subset of those words.
For (a toy) example, a beginner should know:
- gismu, and the place structure.
- how to say his name.
And the subset of words:
gismu = {klama, blanu, zdani}
cmavo = {mi'e, i, coi}
This way, it's easy to say whether a text would be readable for a
beginner or not. It would be *so* easy, that a program could do that.
So, we could make a program that classifies the entire lojban corpus.
(this occurred to me when I was looking for a text to read, so I could
practice a little, but I always had a difficult concept (actually,
vocabulary is not a problem, since I could search in a dictionary))
The advantages of this:
- Having a classified corpus would be beneficial for people writing a textbook.
- Also for them, the levels should be a good place to start.
- Homogenization. If many are attracted by this, and start writing
(or modifying) many textbooks, the topics between books could be
interchangeable for students.
The enumeration of vocabulary for each level should be easy. For each
group of words (gismu, cmavo, lujvo...), we could order them by
frequency (I think I saw something like that with gismu), and
establish limits there.
The enumeration of grammar is the tricky one. If you like the idea, we
should discuss the number of levels, and the contents of them.
I think it could be easy for me, if someday this levels grow up and
become usable, to modify some of the existing parsers into a
classifier.
If I remember something else, I'll post it. In the meantime, I'd like
to know what you think about this.
mu'o mi'e .leos.