W3C home > Mailing lists > Public > public-i18n-mongolian@w3.org > April to June 2016

RE: Discussion on Manchu rules

From: Greg Eck <greck@postone.net>
Date: Thu, 16 Jun 2016 16:11:58 +0000
To: "ishida@w3.org" <ishida@w3.org>, "public-i18n-mongolian@w3.org" <public-i18n-mongolian@w3.org>
Message-ID: <SN1PR10MB097361E431168EFA3A20981BAF560@SN1PR10MB0973.namprd10.prod.outlook.com>
Hi Weizhe,

Welcome to the forum. Thank you for the detailed discussion points as listed.

As I am not so much up on Manchu grammar, I will ask others to step in and lead this discussion. One thing that I can do though is to give a listing of our current OT rulings in Baiti as they relate to Manchu. This will be out within the next few days.

Thank you, Weizhe, for the reference to the new GB draft on Manchu. That is new information to the most of us and is very helpful. If you know of other drafts like this for the other scripts, please pass them on to the forum. We have been searching for this type of information to no avail. So, references like this are very valuable. In particular, I would like to the see the 2015 GB draft proposal for Mongolian (U+1820 - U+1842). Todo and Sibe will be helpful as well.

And thanks to Richard for helping Weizhe to get on the forum.

Greg

>>>>>
Sent: Thursday, June 16, 2016 1:32 AM
Subject: Discussion on Manchu rules

Greg suggested that I join the discussion on Manchu encoding on this forum. I learned a lot from Martin Heijdra's excellent document on Manchu rules. Here are a few points on default behavior and Ali Gali that I would like to discuss.

(1) Syllable-final k (final or between vowels and consonants)

It is well-known that g/k/h takes the back form before a/o/uu and the front form before e/i/u. Rule 1g adopts the same grouping for syllable-final k: back in ak/ok/uuk, and front in ek/ik/uk. This makes sense, but does not correspond to actual usage. I think the rule should be back in ak/ik/ok/uk (exception: front in kuk/guk/huk), and front in ek/uuk. This would significantly reduce the frequency of FVS needed to encode Manchu texts.

(2) Final o/u

The current default for final o/u is the long form, and one needs FVS for the short form. I think the rule in Manchu is short form if the final o/u is the only vowel, and long form otherwise. The current default means that even the most basic syllabary "a uju" requires very frequent use of FVS. Unfortunately at least for o the correct behavior may be difficult to implement, because the code-point U+1823 is also used for Mongolian o, which has a different default behavior.

(3) zi and AG.i

In Mongolian Baiti, z+i gives si with vertical stroke (let me call this the longer form), while z+AG.i gives the shorter form: i is aligned with z in initial/medial position, and becomes a hook in final/isolate position. This assignment is baffling to me. In non-AG and AG words alike final/isolate zi is usually in the shorter form. And in Tong Wen Yun Tong (source for AG, available at
http://www.babelstone.co.uk/Mongolian/Resources.html) initial/medial zi is in the longer form. There seems to be no difference between AG and non-AG. In all positions the more common form should be represented by 
z+i, and the rarer form, if attested, represented using FVS. AG.i does
not seem necessary here. Is there any use of AG.i in Manchu at all?

(4) Initial AG.ng

In rule H it is remarked that the e in ngeng is "left out" and this is 
compared to the situation in Mongolian where e in eng is left out in 
certain words such as tngri. Here is another explanation. When followed 
by a vowel, AG.ng in initial position functions as a diacritic: it adds 
a circle to the initial form of the vowel. So ngeng is eng with circle, 
nge is e with circle, ngi is i with circle, ngei is ei with circle, ngen 
is en with circle, nga is "a" with circle. Seen in this way, the 
behavior is completely regular.

(5) Initial AG.a

Tibetan letter 'a (U+0F60), when used alone, is transcribed in Manchu 
with the isolate form of AG.a, as mentioned in rule K. TWYT lists 14 
more Tibetan syllables beginning with the letter 'a:
         'i   'u   'e   'o
    'an  'in  'un  'en  'on
    'ang 'ing 'ung 'eng 'ong
In Manchu these are transcribed by i, u, e, o, an, etc., with the head 
modified (compare lists 2-4 of the Tibetan section of TWYT with lists 1, 
11, 12 of the Sanskrit section). Parallel to the case of AG.ng, we can 
introduce an initial AG.a, which functions as a diacritic: it modifies 
the head of the initial form of the next character. The glyph of isolate 
AG.a also needs modification.

(6) AG gi, ge, gei

In Manchu AG, g is always transcribed using the back form of g, even if 
followed by i/e. The form of AG ge is peculiar: it has a forward tail, 
and the dot is placed to the right. The unusual positioning of the dot 
makes gei distinct from gi. The same holds for ghi, ghe, ghei. In 
Mongolian Baiti, g+FVS1+i gives AG gi, but AG ge, gei, ghe, ghei are not 
correctly rendered. One way to render them is to introduce ligatures for 
g+FVS1+e and gh+e. Another solution using FVS is proposed in a draft 
Chinese standard 
(http://www.nits.org.cn/UploadFile/xbw/File/%E4%BF%A1%E6%81%AF%E6%8A%80%E6%9C%AF%20%E6%BB%A1%E6%96%87%E5%90%8D%E4%B9%89%E5%AD%97%E7%AC%A6%E3%80%81%E5%8F%98%E5%BD%A2%E6%98%BE%E7%8E%B0%E5%AD%97%E7%AC%A6%E5%92%8C%E6%8E%A7%E5%88%B6%E5%AD%97%E7%AC%A6%E4%BD%BF%E7%94%A8%E8%A7%84%E5%88%99%E3%80%8B%EF%BC%88%E5%BE%81%E6%B1%82%E6%84%8F%E8%A7%81%E7%A8%BF%EF%BC%89.pdf), 
but it is not clear how gi and gei are differentiated in the draft.

Weizhe
>>>>>






Received on Thursday, 16 June 2016 16:12:34 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 16:07:52 UTC