[derek73]

Phase 2a of #752: the two vocabulary markers that record a clash with a name are now derived from name readings instead of written as flags. No compiled set moves: nameparser/config/_compiled.py rebuilds byte for byte.

What changes

• The derivation (tools/vocab/build.py, new build check 11):
  • SUFFIX_ACRONYMS_AMBIGUOUS is now the words with an acronym suffix reading and a given or family reading.
  • The particles outside NON_GIVEN_NAME_PARTICLES are the words with a particle reading and a given reading.
  • Each holds wherever in the files the two readings sit.
• The data. 44 words got 49 name readings, carrying out items 1 to 3 of decisions/vocabulary-model's 2026-10-23 Phase 2: derive ambiguity from clashing readings and extend the gate to every reading kind #752 entry (Record vocabulary phase 2's agreed design in the RFC and decisions (#752) #881), with counts from tools/name_usage.py taken 2026-10-25: the seven flagged credentials that lacked one (ba, do, ed, jd, lac, ma, meng) and the 38 flagged particles other than bon and dí (do is in both lists). The flags came off 97 readings.
• A name reading may stand without a count, but it needs a note saying why. Eight particles have no count in any source (aen, bat, binte, freiherr, freiherrin, thoe, ua, vande).
• ld stays unmarked. Its given count moved into its credential reading's note, so ld stays unmarked. Writing it as a given reading would mark ld: measured on this branch, that moves 3 ctgov and 5 FEC names (no PSC record writes it).
• The old flags are refused. The build refuses may_be_name (either value) and may_be_given = true, and the message names the reading to write instead. A particle whose word has no given reading must still state may_be_given = false, and writing false beside a given reading is refused. Check 9 now asks for a given reading.
• may_be_initial stays a property on the connective reading. It is not derived.

Neutrality, measured against 632685c

The branch is merged with master at b581482 (2.6.0 and #881). Master's code changes since 632685c are docstrings only, and _compiled.py still equals master's, so these hold on the merged tree; the suite and the committed snapshot were re-run on it.

• build.py --check: current, and _compiled.py is unchanged in the diff.
• snapshot.py: matches.
• cldr_agreement.py: all seven records match (CLDR, GND, Wikidata, conventions, Wikidata native, both locale views).
• snapshot.py --local, with the baselines recorded from a master checkout with --tree: ctgov-officials, fec-individuals and tse-donors all match.
• local_agreement.py psc-individuals: matches the record.
• Negative controls: the new mutation rows fail against master's build.py, and the rows the old check already refused still pass there.

Docs

• nameparser/vocab/README.md: a Derived markers section, may_be_given as a "stated" property, count-less name readings, and the adding-a-word step.
• decisions/vocabulary-model: a dated 2026-10-25 entry recording how 2a carried out the design entry, and its Standing lines. The Describe each word by its readings: per-word TOML vocabulary compiled to today's sets (phase 1) #737 and Classify the particles in und.toml by language and kind #749 entries are marked superseded in part.
• Lines made stale by the change: decisions/ma-do and decisions/vocabulary-exclusions (Standing and reference lines), a correction in decisions/C8 (putting sa's credential back now marks it), and the RFC's do example, phase 1 flags line and phase table.
• AGENTS.md's vocabulary lines.
• The config-module comments.
• No release-log bullet: nothing user-visible moves.

Checks

• uv run pytest, -n0 -m serial, --python 3.15
• mypy and ruff
• sphinx-build -b doctest
• A design-docs review (nine findings, fixed in the second commit), and a second review of that fix (four findings plus the C8 line, fixed in the third)

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.07%. Comparing base (b581482) to head (87733ee).

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #883   +/-   ##
=======================================
  Coverage   99.07%   99.07%           
=======================================
  Files          47       47           
  Lines        4781     4781           
=======================================
  Hits         4737     4737           
  Misses         44       44           

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.