Re: [edict-jmdict] Sorting out adj* PoS tags

["Jim Breen" <jimbreen@*********>]

On 01/04/2008, Jean-Luc Leger <reiga@dspnet.fr.eu.org> wrote:

>  On Tue, Apr 01, 2008 at 11:38:12AM +1100, Jim Breen wrote:
>  >
>  > (a) the current "adj-na", "adj-no" and "adj-t"markers be left unchanged;
>
>  maybe "adj-no" should be reviewd too.
>  "n" AND "adj-no" seems redundant to me ..

Well, I don't know about that.

AFAICT not all nouns can or do take a の, so I think it's useful to
mark those that commonly do.

Also, I keep encountering words which seem only to be XXXのYYY. There
are 300 or so entries at present where the POS is only "adj-no".
Some are gairaigo like ホームレス, but others are like 育ち盛り, which
you can't really imagine being in "育ち盛りが..." situations. (Bad
example, as I see about 0.5% of uses *are* 育ち盛りが...)

>  > (d) "adj-pn" be restricted to terms recognized as 連体詞 (有らゆる, etc.)
>
>  meaning : invariable word _always_ used in a prenominal position ?

Dunno. The strict ones like その, この, etc. are, but I'd hat to be dogmatic.

>  > (e) everything else either use the catch-all "adj" tag, or maybe replace
>  > it with "adj-f" (adjectival function/form) to stop people who are not used
>  > to the funny adjectival practices in Japanese from thinking they are "real"
>  > adjectives.
>
>  I'd rather have "adj-f" than a mere "adj".

I think I would too.

Jim

-- 
Jim Breen
Honorary Senior Research Fellow
Clayton School of Information Technology,
Monash University, VIC 3800, Australia
http://www.csse.monash.edu.au/~jwb/