- From: Richard Wordingham <richard.wordingham@ntlworld.com>
- Date: Tue, 18 Aug 2015 02:33:39 +0100
- To: <public-i18n-mongolian@w3.org>
Richard Ishida's document has an error in the isolated forms for QA and GA. The Q and TR170 scheme uses a clever but extremely confusing trick for the presentation forms plus basic letters. For QA, the basis of the trick is as follows: (a) QA and GA are traditionally associated as a pair differing in phonation, and are only distinguished at the start of a syllable. At the end of a syllable, GA is used rather than QA. (b) Thus, QA is always followed by a vowel. (The rare exception does not invalidate the argument.) (c) Feminine QA with orkhitz E does not insert MVS. Therefore feminine final QA does not exist. (d) Feminine QA ligates with all the feminine vowels, and thus there is no distinction between initial and medial feminine QA. Final feminine QA does not exist. (e) There is no distinction between initial and isolated consonants. Therefore, for 'presentation forms', feminine QA may be represented by isolates; the presentation forms classified as initial, medial and final (i.e. pre-MVS) forms are used only for masculine QA. The undotted feminine QA is coded in isolation as <QA>; the dotted feminine QA is coded in isolation as <QA, FVS1> as recorded in Unicode Version 8.0.0. In Richard's table for QA, the Q column should be corrected as follows: 'isolate' changes from initial masculine QA to undotted feminine QA. 'is+fvs1' changes from undotted feminine QA to dotted feminine QA. 'is+fvs2' changes from dotted feminine QA to blank. The USVL column has one change: 'is+fvs1' changed from blank to dotted feminine QA. Note that in the symbol for the letter QA in code charts is the character string <QA, ZWJ>. The same trick is used for GA, but less successfully. Presentation forms for syllable final feminine GA are provided by <ZWJ, GA, FVS3, ZWJ> and <ZWJ, GA, FVS1> (UVSL, TR170) or <ZWJ, GA, FVS2> (Q). Note that in the system of TR170, FVS1 is a toggle on medial and final GA. For medial GA, it toggles the dotting, while for word-final GA it toggles the gender. For example, if I have interpreted the examples correctly, in Row 17 of the table straddling pages 5 and 6 of the main text of TR170, <<initial SA, medial I, final feminine GA>> is encoded <SA, I, GA>, while <<initial SA, medial I, final masculine GA>> is encoded <SA, I, GA, FVS1>. In Richard's table for GA, the Q column should be corrected as follows: 'isolate' changes from initial masculine dotted GA to initial feminine GA. 'is+fvs2' changes from initial feminine GA to blank. 'f+fvs1' changes from blank to undotted final masculine GA. Note that the symbol of the letter GA in code charts is the character string <GA, ZWJ>. (Other) Richard.
Received on Tuesday, 18 August 2015 01:34:20 UTC