Should a Lisu name divide at its tone letters? ꓡꓳ, ꓟꓬ ꓢꓴ reads family ꓡꓳ because Lisu writes tones with ASCII , and ; #1269

Lisu (the Fraser alphabet) marks tone with letters that look like Latin punctuation. Legacy and much current Lisu text writes those tones with the ASCII characters ., ,, ; and : rather than the Unicode tone letters at U+A4F8–U+A4FD. Per the Unicode Core Specification (chapter 18), , in Lisu text is a tone letter, apart from its use as a thousands separator. The Lisu comma is a separate character, U+A4FE ꓾, which #1266 added to COMMA_CHARS.

The parser reads the ASCII tone letter as a family comma:

>>> parse("ꓡꓳ ꓟꓬ ꓢꓴ")    # given 'ꓡꓳ', middle 'ꓟꓬ', family 'ꓢꓴ'
>>> parse("ꓡꓳ, ꓟꓬ ꓢꓴ")   # family 'ꓡꓳ', given 'ꓟꓬ', middle 'ꓢꓴ'
>>> parse("ꓡꓳꓹ ꓟꓬ ꓢꓴ")   # U+A4F9 tone letter: given 'ꓡꓳꓹ', family 'ꓢꓴ'
>>> parse("ꓡꓳ,; ꓟꓬ ꓢꓴ")  # family 'ꓡꓳ', given ';'

Evidence for priority: none of the corpora hold a Lisu name, and Wikidata has 0 humans with a lis label (checked 2026-11-08). The misreading is certain where it happens, but nobody has been seen writing Lisu names to a parser.

Possible shapes, if this is worth doing:

• Treat a , or ; that follows a Lisu letter as part of that word. This is a script-conditional rule of the "certain for this one behavior" kind that AGENTS.md permits, because the script alone settles what the character is.
• Record this as an accepted limit under rules/C's Background and close the issue.