Re: [edict-jmdict] JMdict changes - RFC

["Melissa Boiko boiko.melissa@********* [edict-jmdict]" <edict-jmdict@***************>]

Hi,

> An example for entry 1584660 (明日/あした) might be "<re_pa
pa_type="hm">3</re_pa>" with the あした reading indicating a higher pitch on
the 3rd mora.

As a suggestion, I feel like it should be made clear whether the 'pitch
accent' element is describing pitch (phonetics) or accent (phonemics), as
these are very different things.  From the preceding text, and from normal
lexicographic practice of Tokyo Japanese, I would expect the example above
to be trying to say there's an accent on the 3rd mora (as is indeed the
case for *asitá*)*. *The position/presence of the accent (a single digit)
is enough information to deduce the pitches of all moræ in Tokyo J., e..g. 3
means LHH-L *àsítá-gà*.

However, the wording of the quote and the attribute name suggest instead
that it’s not indicating an accent on the 3rd, but merely a high pitch on
the 3rd, which does not tell you what’s the accent (an unaccented trimoraic
word would also have a high pitch on the 3rd mora).  A purely phonetic
description would need more information per entry, with (here) 4 (no. moræ
+ 1) re_pa elements (to distinguish cases like LHH-H* nàmárí-gá* ‘the lead’
from LHH-L* nàmárí-gà* ‘the accent’).

A phonetic description of all moræ  is redundant if the goal is to describe
Tokyo J. only. But doing it this way could simplify the job of app makers,
and could be useful if the goal is to eventually describe all dialects.
But even then, it would be perfectly possible to use dialect-specific
phonemic models (register + accent position for Kinki, word class for
Kagoshima, accent position + particle accent for Noto etc.) and let
implementers infer the pitch of each mora themselves if that’s wanted
(that’s how the Kokugo Daijiten does it). The important thing is to be
clear (in attribute names etc.) about what's being described, pitch or
accent.

Warm regards,
Melissa

2020年2月18日(火) 5:37 Jim Breen jimbreen@gmail..com [edict-jmdict] <
edict-jmdict@yahoogroups.com>:

> Some changes are proposed to the JMdict structure, both
> to add some new information fields to entries, and to change
> the ways some of the fields are structured. The proposed changes
> are outlined at:
> https://www.edrdg.org/wiki/index.php/JMdict:_Next_Generation
>
> The changes have been discussed within the editorial group and now
> we are seeking comments from the wider community. Feedback from
> people using JMdict in apps and websites will be especially welcome.
>
> Jim
>
> --
> Jim Breen
> Adjunct Snr Research Fellow, Japanese Studies Centre, Monash University
> http://www.jimbreen.org/
> http://nihongo.monash.edu/
>
>
> ------------------------------------
> Posted by: Jim Breen <jimbreen@gmail.com>
> ------------------------------------
>
>
> ------------------------------------
>
> Yahoo Groups Links
>
>
>
>