Decline a title word a trailing rule reads, in the bound join (#637, #640)

[derek73]

Summary

P5's bound given-name join takes the word after abdul. It already declined a maiden marker and a suffix piece, but it took a title word even where a trailing rule would have read that word as a title. The joined piece was then no title word to read:

Input	Before	Now
Berg, abdul Prof. (#637)	given 'abdul Prof.'	title 'Prof.', given 'abdul'
John Smith, Jr., abdul Dr. Smith (#640)	suffix 'Jr., abdul Dr. Smith'	title 'Dr.', suffix 'Jr., abdul Smith'
Berg, abdul Prof. Ali	given 'abdul Prof.'	unchanged: the title stands inside the name
abdul Sir Smith Berg	given 'abdul Sir'	unchanged

The join now declines a title word that a trailing rule reads as a title:

• After a family comma: a word H5's trailing run takes from the given part. The LENIENT join reads no peel (fix(group): the bound-given reserve and join ask what assign reads (#401, #421) #423), so tail_reading is asked here, and only when a title word follows the bound word.
• In a part past the second comma: any title word, since C2 reads them all as titles.

Decided (Derek): this narrower clause (a′), not declining every title word (a). P5 states that a title word standing in the name is a name word and joins like one (H3), and (a) undid that:

• six tests failed against master's suite, including a P5 rules example and the two tests named for that clause;
• 4,044 bound-given grid parses moved, 3,000 of them joins split with no title read ((Bob), Abu Bishop → given 'Abu', middle 'Bishop').

Measured (against 4455962, py3.11, import asserted on each side; recipes in decisions.md#P5)

• Bound-given grid (5,184 texts under three orders, recipe in the entry): 1,044 parses move. Every one moves a title word into the title field, and none moves only a report.
  • 660 are in a part past the second comma, e.g. John Smith, Jr., Abu Bishop Ali Smith → title 'Bishop', as C2 already reads John Smith, Jr., Bishop.
  • The rest come after a family comma or an empty head.
• Should Smith, John van der Berg Prof. read title 'Prof.' as John van der Berg Prof. does? #623's given-part grid: 280 moves. 260 are the same class. The other 20 split the join with no title read (Smith, abdul Prof. V, Jr.), the limit recorded below.
• Corpus names and case texts: only Berg, abdul Prof., the row that had pinned the gap.
• No moves on the S2 stress grid, Where does Freiherr St John Smith MA start: should a title that is also a particle open the name? #624's title-particle grid or Merge the trailing run's particle runs in group, so S2's read needs no BOUND stop or report gathering #625's particle grid.
• Five differential gates: all exit 0.
  • Each 2.x ledger gets fix(#637) and fix(#640) rules. Should Smith, abdul Dr. read title 'Dr.' as Smith, abdul Ali Dr. and abdul Smith Dr. do? #637's fields differ by baseline: 2.0.0 and 2.1.0 read middle 'Prof.', while 2.2.0 and 2.3.0 read given 'abdul Prof.'.
  • At 1.4.0 the broad comma rule already explains both new examples, and its claims are re-recorded with dated notes.
• Frames: the reference rows don't move.
  • A comma-less name with a bound given word pays nothing extra, because the site is checked before the title test.
  • A family-comma bound name with no title behind it pays 1 (Berg, abdul salam 208 → 209).
  • One with a title after it, after a family comma, pays about 10 to 20 more for the H5 read (Berg, abdul Bishop 208 → 218).

Tests

• title_word_trailing_behind_a_bound_pair_after_a_comma (Berg, abdul Prof.) used to pin the gap as a non-normative row. It now pins the title, as fix(#637).
• New row a_bound_join_declines_a_title_past_the_second_comma, fix(#640).
• rules.md#P5 examples: three new ones, including the Berg, abdul Prof. Ali boundary. rules.md#H5's gap paragraph becomes an example.
• New rows from the review:
  • John Smith, Jr., abdul Sir Smith (fix(#640)): the bare-title half.
  • Berg, abdul Prof. Dr. (fix(#637)): the whole trailing run rather than the last word.
• Negative controls:
  • With no title declined, 11 tests fail.
  • With the past-the-second-comma half alone removed, 5 tests fail.

Review

Recorded, not narrowed. After a family comma, the decline asks how the comma-less name would read H5's run, and the given part's own reading differs on some shapes. There the join is declined and no title is read:

• a numeral VI to X (Smith, abdul Prof. VI → middle 'Prof. VI')
• V before a third comma part
• a dotted title-particle, which P6 attaches (Smith, abdul St. → family 'St. Smith')

A bound word that is also a particle, before a dotted title-particle, goes the other way: the title is read where an ordinary given word doesn't read it (Smith, Abu St. → title 'St.'; Smith, Zed St. → family 'St. Smith').

All of these are garbage input, recorded the same way as #623's disagreement, and rules.md#P5 states them in an Accepted clause. The sweep found nothing reading worse than master on a real name, and no comma-less name moved.

Fix commit:

• adds the two rows above;
• corrects option (a)'s count, the controls, a row note's history and a ledger note;
• moves the decisions entry into P5's dated entries and adds notes to the Should Smith, John van der Berg Prof. read title 'Prof.' as John van der Berg Prof. does? #623 entry and to abdul Smith Jr Ma loses the family name — P5's reserve does not mirror S2's bare-acronym fork #425's paragraph on the post-comma path (which renders under Rethink the credential-run rule (S2): read a name's trailing run once, over name units, and bind it before any join #614's bullet);
• adds P5 and C2 to each other's interacts;
• fixes two release sentences;
• checks the site before the title test.

Second fix commit: H5's Accepted paragraph now excepts the shapes P5 records, and the frames sentence gains the family-comma case.

Docs

• rules.md#P5: a new decline clause.
• rules.md#H5: the Accepted paragraph that recorded this as a gap (tracked under Should a title written after the name still be a title? ("John Smith Prof." → family Prof.) #316) now states the reading.
• decisions.md#P5: a new entry. The 2026-09-09 entry that recorded the gap is marked fixed.
• Release log: two bullets extended, with every released version's reading checked.

Closes #637
Closes #640

🤖 Generated with Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 99.02%. Comparing base (4455962) to head (baa310c).
⚠️ Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master     #643   +/-   ##
=======================================
  Coverage   99.02%   99.02%           
=======================================
  Files          46       46           
  Lines        4512     4513    +1     
=======================================
+ Hits         4468     4469    +1     
  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.