search

giellalt / lang-sme

Finite state and Constraint Grammar based analysers and proofing tools, and language resources for the Northern Sami language
https://giellalt.uit.no
GNU General Public License v3.0
6 stars 1 forks source link

Bahá� vuoinnat ( #191

Closed albbas closed 19 years ago

albbas commented 19 years ago

This issue was created automatically with bugzilla2github

Bugzilla Bug 62

Date: 2005-03-11T20:35:06+01:00 From: Marit Julien <> To: Thomas Omma <> CC: trond.trosterud

Last updated: 2005-05-12T02:57:55+02:00

albbas commented 19 years ago

Comment 203

Date: 2005-03-11 20:35:06 +0100 From: Marit Julien <>

According to our present lexicon, and also according to the dictionaries, the attribute form of bah

albbas commented 19 years ago

Comment 204

Date: 2005-03-14 09:40:55 +0100 From: Thomas Omma <>

I think bahás is the one and only attr. form, bahá isn't attr, but first part in compounds: Bahávuoiŋŋat. Not *bahá vuoiŋŋat. The same goes for for. ex. bahábeallji, bahádahkki etc. People are confused when it comes to attr. vs. first part in compounds, which isn't strange. Is it possible for you to compound?

albbas commented 19 years ago

Comment 205

Date: 2005-03-14 11:38:09 +0100 From: Marit Julien <>

But the thing is, the people who translated the New Testament DID use bahá as an attribute form, and now we cannot analyse phrases like 'bahá vuoinnat' correctly. We can't change the text, so we must do something else.

albbas commented 19 years ago

Comment 206

Date: 2005-03-14 12:53:45 +0100 From: Thomas Omma <>

Allow it as attr. form with sub-tag?

albbas commented 19 years ago

Comment 207

Date: 2005-03-14 12:56:22 +0100 From: Marit Julien <>

That seems to be the best we can do...

albbas commented 19 years ago

Comment 208

Date: 2005-03-15 10:46:43 +0100 From: Trond Trosterud <>

I'm not quite sure about the subtag. Now we have:

"<bahá>" "bahá" A Sg Nom @SUBJ "<vuoiŋŋat>" "vuoigŋa" N Pl Nom @SUBJ

With a subtag it would be:

"<bahá>" "bahá" A Attr Cmpnd @AN> "<vuoiŋŋat>" "vuoigŋa" N Pl Nom @SUBJ

or something like that. Strictly speaking, we do not operate with subtags, so we would eiher have to use the Cmpnd tag to signal this as a quasi-compound, or we should just give the tag A Attr, without Cmpnd. As a fix, I give bahá a lexicon on its own, with a separate attr form, and then a pointer to LAIKI. (hmm, alternatively I could hardcode an At entry in the lexicon, yes, let me do that instead.) I leave the bug open for a while in order to let us evaluate whether the new version is to permissive (to hard to disambiguate the two bahá), and whether we rather should have used the Cmpnd tag instead. If we choose to do so, all it takes is to hard code it into what is today "bahá At;" in the adj file.

albbas commented 19 years ago

Comment 266

Date: 2005-05-12 02:57:55 +0200 From: Trond Trosterud <>

This bug has accidently been stranded as unassigned. The bahá form now #has# an attr form, and someone thus has fixed it. The parser gives a good result for "Dat lea bahá vuoiŋŋat.", but not for "Bahá vuoiŋŋat leat boahtán.", so there still are things to do on the disambiguation. Those issues are irrelevant to this bug, though.