L2/18-053

Title: New Indic Syllabic Category Consonant_Initial_Postfixed
Author: Roozbeh Pournader
Date: 2018-01-24

Proposal
========
Create a new property value for Indic Syllabic Category,
"Consonant_Initial_Postfixed" and populate it with U+1A5A TAI THAM CONSONANT
SIGN LOW PA.

Background
==========
UTC Action Item 151-A148 asked the author to propose a new property value of
"Consonant_Suffixed" for Unicode 11.0.  After dicussions with Liang Hai and Ken
Whistler, the author is proposing it as "Consonant_Initial_Postfixed".

This category was expected to be a further division of the existing category
Consonant_Succeeding_Repha which temporarily holds that character.  Here is
what the comments for Consonant_Succeeding_Repha say for Unicode 10.0:

  # Repha Form of RA (reanalyzed in some scripts), when succeeding the main
  # consonant. Additionally, U+1A5A TAI THAM CONSONANT SIGN LOW PA has been
  # temporarily classified in this class due to its behavior, although it is
  # not a RA.

Characters in the Consonant_Succeeding_Rapha class indicate a /r/ phoneme that
precedes a consonant in pronunciation (either in modern use or historically),
but follows the consonant in a Unicode string.

The problem with the existing classification is that U+1A5A TAI THAM CONSONANT
SIGN LOW PA is not a repha.  It's a /p/ phoneme that precedes the consonant
it's paired with in pronunciation but follows it in encoding.  The example
given in L2/17-120 is <letter low ka, letter low pha, consonant sign low pa>,
which is pronounced /kap pʰaʔ/.

Thus, the category Consonant_Succeeding_Repha needs to be split in two, one to
keep the existing rephas, and another to hold others (just one character at the
moment).  This is similar to the existing pair of categories,
Consonant_Preceding_Repha and Consonant_Prefixed, the first of which holds the
rephas and the other the rest.