Phase 2: derive ambiguity from clashing readings and extend the gate to every reading kind

Part of #747. Phase 1 kept the ambiguity markers as explicit flags (may_be_given, may_be_name, may_be_initial). Phase 2 derives ambiguity from clashes between readings instead, and makes the gate a property of a reading rather than a rule written per class of word. Today's gates are six separate rules, each reading its own evidence:

ambiguous acronyms by periods, capitals and words to spare (rules/S3, S4, S6)
one-letter roman numerals by a word to spare (rules/S10)
a title that may be the name, by the title run's floor (rules/H3)
particles by position (rules/P4)
bound given names by a word to spare (rules/P5)
single-letter conjunctions by case (rules/P8)

A listed word none of them reaches is read wherever its position allows, which is why fils is excluded rather than gated. The vocabulary model RFC (docs/design/vocabulary-model-rfc.md) has the overview.

This is the first phase that changes behavior, so it needs its own spec, measurements and snapshot moves.

Questions it has to answer:

What cue the gate reads for each kind, starting from the rules above. An acronym's cues are periods and capitals, while a lowercase generation word like fils needs another, and an initial-shaped one (H., regnal I.) has S10's count against a period, which marks a one-letter word as an initial (rules/T5).
Whether a reading with no reliable cue is off unless a pack turns it on (default = false, e.g. Czech st.).
What a gate at the FRONT of a name looks like (a title that is also a given name).
How readings surface on a ParsedName (the reading not taken, on the ambiguity).

Decided already (decisions/vocabulary-model): vocabulary readings settle a clash; name readings only keep the gate on; bearer counts are provenance and never compute a reading.

Standing blocks for the gate files S3, S4, S6, S10, H3 and P5 (#762) (#…

Read a bare Wm as the given name it abbreviates, not a title (#789) (#…

Add a Policy switch for names with two family names (#754) (#875)

Record vocabulary phase 2's agreed design in the RFC and decisions (#752

Derive the two name-clash markers from name readings (#752) (#883)

Read King as a given name outside a title phrase (#752 phase 2j, #348) (

Collapse the deferred field's unreachable guards (#752, phase 2m)

Fix the PR review's findings on Ambiguity.alternative (#752, phase 2m)

Read the given names Sultana and Sultan as names outside a title phra…

Compare alternatives only where both sides record them (#752, phase 2m)

Count master's credential-comma row among the sweep's skips (#752, ph…

Keep the titles 2.6.0 listed in v1's Constants().titles (#752) (#904)

Carry the reading a report did not take, as Ambiguity.alternative (#752…

Name an alternative's field through one slot and one name-piece helper (

Read the name Shaikh as a name outside a title phrase (#752 phase 2j, #…

Read the given names Prince and Princess as names outside a title phr…

Read the name Sheikh and its spellings as names outside a title phrase (

Phase 2 is done: every sub-phase of the agreed design (#881, decisions/vocabulary-model's 2026-10-23 #752 entry) has landed, each as its own PR.

Sub-phase	What	PR
design	the RFC's Phase 2 section and the decisions entry	#881
2a	the two name-clash markers derived from name readings	#883
2b	one predicate per question: split credentials, numerals, paired initials	#884
2c, 2d	credentials joined by a hyphen or a slash	#887
2e	a credential split across a space (rules/S12)	#891
2f	one predicate for every roman-numeral rule	#886
2g	capitals settle a listed credential	#895
2h	generation readings gated by kind, `default = false` (rules/S11)	#890
2i	the title phrase (rules/H9), `attaches`	#894
2j	titles that are names: `wing` #896, `mahdi`/`bard` #898, `buddha`/`siddha`/`gyani` #900, `king` #901, `sultana`/`sultan` #903, `shaikh` #906, `prince`/`princess` #908, `sheikh` and its spellings #912
2k	a given name written short over rules/H2's shape	#897
2l	the bracket escape reads the derived clash (rules/S13)	#899
2m	the reading not taken, as `Ambiguity.alternative`	#902

Two follow-ups landed beside them: #904 keeps v1's Constants().titles as 2.6.0 listed it, and #905 is a helper refactor.

The issue's four questions:

The cue each kind reads: numerals by one predicate (2b, 2f), credentials by their spelling (2c–2e) and their capitals (2g), fils by case and the count, the agnomens by the count alone (2h).
A reading with no reliable cue: default = false records it and compiles to nothing; a pack turns it on (2h).
A gate at the front of a name: a title with a given reading is a title only inside a title phrase, and a name word elsewhere, reported title-or-name where it opens the name, no comma marks the given name and a listed phrase holds the word (2i, then 2j's batches; decisions/H9).
Surfacing: the reading not taken is Ambiguity.alternative (2m). The reading taken, Token.reading, waits for 3.0.

The title words 2j did not reach (baron, bishop, judge, marquis and the rest) continue in #913. #918 updates the decisions Open: lines that pointed here, and records the last open question parked on this issue: two_family_names reports nothing and surfaces no alternative, since the caller declared the convention (decisions/O6).

Close #752's open questions in decisions/ and point the title words l…

Navigation Menu

Repository navigation

Phase 2: derive ambiguity from clashing readings and extend the gate to every reading kind #752

30 remaining items

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Phase 2: derive ambiguity from clashing readings and extend the gate to every reading kind #752

Description

Activity

30 remaining items

derek73 commented on Oct 27, 2026

Metadata

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions