Which abbreviations that open a name should the vocabulary list: Excm., Mohd., Mt.? #1115

The question

A period-marked word opening a name can be a title (Excm., Engr.), a given name written short (Mohd., Fco.), the start of a surname (Sta. Maria, Mt. Pleasant), or a typing slip (BEUMLER. WEAVER, CAROL). Nothing in the word's shape tells these apart. The proposal is to let the vocabulary decide: list the titles, the abbreviated given names and the abbreviated surname heads we can find, and let each slip read wherever the rules put it.

That is how rules/H6 (a title before a family comma) already works, so H6 keeps its vocabulary-only reading. What changes is its reason and its vocabulary:

Excm. Sr. Lloret Palol, Josep Antoni Carles Joan
  today:        family 'Excm. Lloret Palol', suffix 'Sr.'
  excm listed:  title 'Excm. Sr.', family 'Lloret Palol'

Why H6's reason needs rewriting

decisions/H6 declined H2's shape before a family comma twice. Both reasons have since gone stale:

• 2026-10-17, Dr. Smith, John reads family Dr. Smith: a title before a family comma isn't peeled #701: "every CLDR case is a listed title". Two CLDR samples have since joined the corpus with Read the field and the comma from what stands beside it: H1 behind a comma (#720), C12 (#656) #735 and break that: Excm. Sr. Lloret Palol II, … and Excma. Dra. Van Dalen Hoogeveen II, ….
• 2026-10-18, Xyz. née Jones, John keeps née Jones in the family where Smith née Jones, John reads maiden Jones: should a word before a family comma count as a surname for the maiden marker? #710: the cost was Sta. Maria and its siblings (1,057 US Census 2010 bearers), and "reopening it would need a stop for sta/sto". Should Juan Sta. Maria read family Sta. Maria? Today it reads middle Sta., family Maria #725 made both saint particles, so a shape peel that kept H6's particle stop would now leave them whole.

A scan for what a shape peel would reach (below) finds the reason that still holds. The slot before a family comma also holds given names written short (Md. Abu Shahin, FCPS, MD, Mohamad. S. AlSalhi, …), stray periods (CARL. G, EMILSON, Frances. M Cowan, …) and abbreviated surname heads that are not particles (MT. PLEASANT, JANE). So the proposed H6 Because reads: "a period-marked word there may be a title, a credential, a given name written short, a surname's head or a slip, and its shape tells none of these apart, so only the vocabulary claims it."

What the corpora hold (measured 2026-11-06)

Recipe: every distinct name in the 15 local corpora and the committed tools/differential/corpus*.jsonl whose first word matches [^\W\d_]{2,}\. followed by a space, keeping the words that Lexicon.default() does not list as a title, gated title, particle, given abbreviation or suffix. That gives 560 distinct words in 18,617 names, against dd2650ef.

Listed in a pack, off by default. A pack choice already reads these. They show up here only because the default lexicon lacks them:

• cs/sk degrees: MUDr. 2,323, JUDr. 846, PhDr. 732, RNDr. 570, MVDr. 393, MgA. 323, PaedDr. 255, PharmDr. 146, MDDr. 84
• es: Lic. 1,428, Mtro. 333, Mtra. 192, Lcda. 89, Arq. 72, Lcdo. 32
• vi: ThS. 240

Excluded on purpose: Bc. 8,724 (the US nursing BC), Mst. 104 and Sig. 4. Their recorded reasons should be re-read against these counts.

Unlisted titles (candidates):

• Engineering: Engr. 70, Eng. 29, Ir. 15, Er. 7
• Academic: Doc. 308, RSDr. 90, Drs. 15, ThLic./ThMgr./ThDr., Akad. mal./Ak. mal., Prim., Mgtr., Profr., Mtr.
• Spanish: Licda. 14, Dip. 15 (diputado), Sen. 28, Gral. 12, Cor., Tte., Biól., Act., LAE., LCP., MAP., MVZ.
• Czech ranks: mjr., plk., npor.
• Modifiers: Assist. Prof. 19, Ass. Prof. 7, Em. prof.
• Others: Ven. 19, Ku. 11 (Nepali Kumari), Adv., RR. (Javanese Raden Rara), med., Excm./Excma. (CLDR)
• Dipl. 19 as a word of its own (Dipl. Ing., Dipl. Arch.). The hyphenated Dipl.-Ing. is already a de academic word.

Given names written short (candidates for a given reading with same_as, the shape Chas. takes since #789):

• Muhammad: Mo. 44, Mohd. 37, Moh. 28, Muh. 26, Moch. 11, Mhd. 5, Ach. 5
• Others: Fco. 7 (Francisco), Sk. 18 / Sh. 5 (Sheikh), Kh., Th., Ph., Jan.

The Muhammad forms have no credential reading, so they don't run into what kept md from taking a given reading (decisions/C3, the md decline). Each of them reads as a title by H2's shape today, wherever it opens a name: Mohd. Ali Hassan → title 'Mohd.'. Md. itself stays an accepted cost (decisions/indic-honorifics, #853). Whether these new readings change that is a question for this issue.

Surname heads (candidates for particles): Mt. (MT. PLEASANT). The search for more is part of the work.

Slips (Mohammed., Abdullah., BEUMLER.): not listed. They read wherever the rules put them.

Proposed work

1. Amend rules/H6's Because and add a decisions/H6 entry recording that Should Juan Sta. Maria read family Sta. Maria? Today it reads middle Sta., family Maria #725 removed Xyz. née Jones, John keeps née Jones in the family where Smith née Jones, John reads maiden Jones: should a word before a family comma count as a surname for the maiden marker? #710's cost, and the stated reason that replaces it. Date or drop the CLDR line.
2. Batches by class, each with bearer counts (tools/name_usage.py --boxes, tools/vocab/word_report.py --rank), as Phase 2: derive ambiguity from clashing readings and extend the gate to every reading kind #752's phase 2 batches ran:
   • the Muhammad abbreviations and other given names written short
   • engineering and academic titles
   • Spanish titles not yet in es.toml
   • the remaining titles and the modifiers
   • Mt. and any other surname heads found
3. Decide, for each pack-only word, whether it should be on by default. For example, MUDr. is not borne as a name anywhere, so it may not need a Czech choice.

Open questions

• Which pack-only degrees should be on by default?
• Do Bc., Mst. and Sig. keep their exclusions in light of these counts?
• C1: many names this reaches are given-first names whose comma sets off a credential C1 doesn't know (Ing. Jitka Červenková, MScM, Ns. Andan Firmansyah., S.Kep., M.Kep.). Is listing those post-nominals part of this, or its own issue?