[lojban-beginners] Re: Distinguishing Type 3 and 4 fu'ivla

["H. Felton" <fagricipni@gmail.com>]

On Mon, Aug 24, 2009 at 6:21 PM, Minimiscience<minimiscience@gmail.com> wrote:
> de'i li 24 pi'e 08 pi'e 2009 la'o fy. H. Felton .fy. cusku zoi skamyxatra.
>> Is there a way to always distinguish between Type 3 and Type 4
>> fu'ivla? Â"cidjrspageti" is a Type 3 fu'ivla based on Lojbanizing
>> "spaghetti" to "spageti", and then attaching "cidj-" from the rafsi
>> for "cidja" for "food" and an r-hyphen. ÂHowever, upon first
>> encountering that word in text could it not also be a Type 3 fu'ivla
>> based on a longer natural language word?
> .skamyxatra
>
> Assuming you meant "could it not also be a Type 4 fu'ivla," yes, it could be
> either, and there is no foolproof way to tell the two types apart. ÂConsidering
> that external knowledge is required in order to understand any type of
> {fu'ivla} in the first place, I don't think this is much of a drawback.
>
> mu'omi'e .kamymecraijun.

I'm still working on my learning program to help me with memorization,
but I also thinking about a program that could _morphologically_
identify "words" in Lojban text.  If I see "lenubrivla", the program
could identify it as a multiple word, and split it in to "le nu
brivla" -- _this_ program would not attempt to identify meaning.
Given what you just said it seems that the catagories will have to be:
"multiple word" (needs to be split), "cmavo", "gismu", "lujvo", "Type 4
fu'ivla", "possible Type 3 fu'ivla".  Note at the level I am talking
about the program will identify "kobra" as a gismu and "pa'ai" as a
cmavo even though such valsi aren't defined in the vocabulary lists
-- If I implement "look up", it will be after morphological
identification; this also means that the program will identify
"becrbai" as a lujvo (assuming I haven't make a mistake in generating
a lujvo from those two rafsi); it could not be identified as illegal
due to there being no rafsi "bec" until after "look up".