Re: [edict-jmdict] adj-i/adj-ix revisited

I read through the entirety of the previous discussion as well. I'm happy with Jim's proposed solution because

1) Most importantly, it allows us to merge ～いい and ～よい entries back together. I made a couple of arguments in the original thread why having いい and よい together is preferable. One big reason is that we (editors and other contributors alike) have been completely unable to make sure the separated entries are kept uniform.

2) "よい" inflects differently from normal adj-i in -そう as よ_さ_そう -- based on that alone it should arguably have its own tag. (The issue that adj-i-tagged 無い and ～無い entries also inflect as ～なさそう remains though.)

3) We don't have to cast it as "non-native speakers making up their own PoS" as it will be up to the actual dictionaries how to present this to their users. If they just want to present adj-ix as adj-i to the end user and just use adj-ix for conjugation tables, etc., this will allow for that.

I understand Rene's gripe though and the addition of a restr_infl/no_infl reading tag (I don't think a kanji tag would be needed) in addition to the merge would be even better. That would also allow for 無い and ~無い to be tagged adj-ix instead of adj-i. I'm supportive of Jim's suggestion with or without such a tag, though.

Marcus

On Wed, Sep 18, 2019 at 11:58 AM René Malenfant rene_malenfant@*********** [edict-jmdict] <edict-jmdict@***************> wrote:

I'm not entirely sure what the problem with the two entries is other than it is a little cumbersome (at least if you want to ensure that the ~よい and ~いい entries are consistent with one another, which of course we do). As a regular editor, I don't ever recall that being too much of an issue. The only "issue" really has been that no one really agrees on how the いい/よい situation should be handled, so it keeps getting raised.

Is there some other benefit to having いい/よい merged that I am not seeing? From an editor standpoint I don't think it's much of a problem to have them separate, and I think whether they are separate or combined is mostly invisible/irrelevant to the average end user. But I don't use the dictionary for any automated text processing etc. Is there some benefit there?

Rene

On Sep 17, 2019, at 11:40 PM, Francis Bond bond@******** [edict-jmdict] <edict-jmdict@***************> wrote:

I think the advantage of giving them the same tag is that you capture the fact that any _expression_ that takes いい almost always also allows よい --- we don't need two entries for e.g. 気持ち (が) いい, we need one entry with a note that says いい can be replaced by よい (or if you prefer, that いい is simply an alternate pronunciation for よい that applies only when it is specifically in its よい form). But we should decide on what we want the index form to be 気持ちがいい or 持ちがよい or 気持ちがいい/よい or ...

Currently we have two entries:

(I) 気持ちいい【きもちいい】 (exp,adj-ix) (ant: 気持ち悪い) good feeling; feeling good; (P)

(Y) 気持ち良い; 気持ちよい; 気持ち好い; きもち良い【きもちよい】 (exp,adj-i) (See 気持ちいい・きもちいい) (ant: 気持ち悪い) good feeling; feeling good

Note that (Y) links to (I) but not visa-versa. I would prefer these to be a single entry or at the least mutually linked, ...

On Wed, Sep 18, 2019 at 10:22 AM René Malenfant rene_malenfant@*********** [edict-jmdict] <edict-jmdict@***************> wrote:

Hi.

It's been a long time, so I forget some of the details of this... But if I recall, my argument was that we should follow the kokugos in acknowledging that--contrary to what various Introduction to Japanese books for foreigners say about よくない being the negative form of いい--いい really only has 終止 and 連体形 forms, and therefore いい itself never actually changes form (hence it gets a different PoS to indicate its irregular lack of inflection: adj-ix). It's only よい that inflects (as a regular adj-i). いい is simply an alternate pronunciation for よい that applies only when it is specifically in its よい form. So:

よい = adj-i (regular i-adjective)
いい = adj-ix ("special" i-adjective)

I still think that is the technically correct stance (and AFAIK all the Japanese dictionaries back that up, although I haven't gone and checked again).

If we want to compromise and put them all in an [adj-ix] tag because the "technically correct" solution is too cumbersome or unsightly, I guess that's okay by me as long as:

The description of the [adj-ix] tag describes the technically correct stance as described by the kokugos.
The conjugation table for WWWJDIC does not incorrectly suggest that よくない is a conjugation of いい.

Rene

On Sep 17, 2019, at 10:59 PM, Jim Breen jimbreen@********* [edict-jmdict] <edict-jmdict@***************> wrote:

Back in 2014 I split the 〜良い and 〜いい adjectives into two
groups with adj-i and adj-ix POS tags respectively to reflect their
different inflection patterns. It actually made rather a mess and in
early 2018 we had a long discussion about how to resolve it. The
discussion fizzled out, and I left the バランスがいい entry sitting in
the edit queue as a reminder. I'd like to try and get it sorted out.

In our discussion we identified three options: (a) stay with the
present split, (b) make then all "adj-ix", (c) make them all "adj-i"
with a note and/or a "no-conj" tag for the 〜いい forms. On 6 March
2018 René Malenfant summarized and scored the options, then on 8
March Francis Bond made some extensive observations and rescored
them.

I've been re-reading of all the discussions and the back-and-forth proposals,
and I think the cleanest solution is to put them back together, with the
special "adj-ix" POS tag to indicate that collectively they behave differently
from the usual 形容詞. As Francis wrote: "I think it is better to acknowledge
this (situation) with a single special tag that encompasses both of them,
and we can add a complete description in one place, ..."

I just checked the entity tag labels in the DTD and they are:
<!ENTITY adj-i "adjective (keiyoushi)">
<!ENTITY adj-ix "adjective (keiyoushi) - yoi/ii class">
so if we take this approach nothing has to be changed there.

I know this approach is not everyone's preferred solution, but
I'd really like to see this one resolved. The main work will be to
put the ~100 entries back together. I'm prepared to do the legwork.

Reactions? Comments?

Jim

--
Jim Breen
Adjunct Snr Research Fellow, Japanese Studies Centre, Monash University
http://www.jimbreen.org/
http://nihongo.monash.edu/

--
Jim Breen
Adjunct Snr Research Fellow, Japanese Studies Centre, Monash University
http://www.jimbreen.org/
http://nihongo.monash.edu/

--

Francis Bond <http://www3.ntu.edu.sg/home/fcbond/>
Division of Linguistics and Multilingual Studies
Nanyang Technological University

Re: [edict-jmdict] adj-i/adj-ix revisited - closure?