Hunspell-marking and j Surfacing for short passives (

[albbas]

This issue was created automatically with bugzilla2github

Bugzilla Bug 699

Date: 2008-06-05T11:55:59+02:00 From: Thomas Omma <> To: Trond Trosterud <> CC: lene.antonsen, sjur.n.moshagen

Last updated: 2015-03-16T14:16:01+01:00

[albbas]

Comment 2758

Date: 2008-06-05 11:55:59 +0200 From: Thomas Omma <>

dekčoj dekčoj deakčut+V+TV+Der/PassS+V+Ind+Prt+Sg3 should be dekcui

fámuhuhttuj fámuhuhttit+V+TV+Der/PassS+V+Ind+Prt+Sg3 should be fámuhuhttui

I am working on it, but the rule "Postvocalic j Surfacing" (where the cause prabably is) is really messy and hard to work with

[albbas]

Comment 2768

Date: 2008-06-05 15:56:56 +0200 From: Thomas Omma <>

cvl 1.194: "Postvocalic j Surfacing" j:i <=> ( %>: ) ( »: ) Vow: ( %>: ) ( :0 ) ( %>: ) ( »: ) ( Dummy: ) _ [ Cns | ': | ( %>: ) ( »: ) ( :0 - Y5: - Y6:)+ (Hyph) [# | #:] ] ; !more and more dummies...

changed to :

"Postvocalic j Surfacing" j:i <=> ( %>: ) ( »: ) [Vow:|:0] ( %>: ) ( :0 ) ( %>: ) ( »: ) ( Dummy: ) _ [ Cns | ': | ( %>: ) ( »: ) ( :0 - Y5: - Y6:)+ (Hyph) [# | #:] ] ; !more and more dummies...

to get j:i in short passives fámuhuhttui, dekčui etc

fámuhuhttui is okey now, but we still have dekčoi (the j surfacing works, but not the vowel-rising u:o

[albbas]

Comment 2772

Date: 2008-06-09 15:05:50 +0200 From: Thomas Omma <>

"Stem Vowel in Imperative Plural 1st Person and Passive Stem" Vx:o <=> Cns _ ( %>: ) ( »:) [ Y4: | Y7: ] ( j u v v o | j u n | j u p m ) ; where Vx in (a i u) ;

changed to:

"Stem Vowel in Imperative Plural 1st Person and Passive Stem" Vx:o <=> Cns _ ( %>: ) ( »: ) [ Y4: | Y7: ] ( %>: ) ( »: ) (Y5:) [Cns: - j:i|#:] ; where Vx in (a i u) ;

to prevent vowel-rising in front of j:i cvl 1.195

bug fixed

[albbas]

Comment 2783

Date: 2008-06-23 14:26:40 +0200 From: Trond Trosterud <>

Unfortunately, removing the requirement for vowel before the post-vowel change had some negative consequences:

jna. did not work, but ina. did:

ina. jna+ABBR+Adv

jna. jna. jna. +?

So, reverting the twol rule back again gives us the opposite:

jna. jna. jna+ABBR+Adv

fámuhuhttui fámuhuhttui fámuhuhttui +?

Now, we should look for how to get both right.

[albbas]

Comment 2784

Date: 2008-06-23 14:27:08 +0200 From: Trond Trosterud <>

Note to Lene as well.

[albbas]

Comment 2785

Date: 2008-06-23 15:43:35 +0200 From: Trond Trosterud <>

Checked in 1.197, with the following comment:

j:i <=> Vow: ( %>: ) ( :0 ) ( Dummy: )* _ [ Cns | ': | ( %>: ) ( »: ) ( :0 - Y5: - Y6:)+ (Hyph) [# | #:] ] ; ! Kept Vow:, but accepted Dummies at length The problem with the old one was that it did not have a compulsatory underlying Vow in the left context, i.e., it opened for jnv. as inv. etc. By allowing any number of Dummies we are ab le to keep the Vow restriction. The present version accepts both jnv. and fámuhuhttui.

I keep it open to see whether it behaves, but tests looked good: I ran the new testament through it, and got no strange behaviours.

[albbas]

Comment 2796

Date: 2008-06-24 22:38:06 +0200 From: Trond Trosterud <>

So, it had fatal consequences. What I then did was I reintroduced a j in a pass contlex. cf. svn log.

[albbas]

Comment 2812

Date: 2008-08-10 09:29:19 +0200 From: Trond Trosterud <>

I would like to see a report on how this issue is doing.

[albbas]

Comment 2813

Date: 2008-08-10 09:31:33 +0200 From: Trond Trosterud <>

I now has gone through the comments, and the j rule behaves. Bug closed.