# ABEYANCE pilot v0.1 — source audit and publication decision

Author/project owner: Jinkai Liu. AI-assisted implementation under explicit approval, 2026-09-19. Scope: minimal additive schema branch and ABEYANCE only. First ABEYANCE pilot; legacy migration only after Jinkai Liu review.

## Decision

**Featured Mapping: Pending.** This publishes a research pilot, not a winning Chinese correspondence. 巴 has the strongest currently supported DESIRE comparison among the four requested tracks; its historical source pronunciation and dated Chinese sense still have gaps. 擘 adds a valuable dialect candidate but locality and original-character analysis remain qualified. 盼 is a strong semantic comparator. 闭 remains a weak modern cognitive candidate; temporary suspension is not literal closure.

Publication gate for a Featured: a clearly frozen sense; adequate documentation for the selected Chinese sense and actual pronunciation layer; no unresolved critical mismatch in the advertised sound comparison; explicit contextual limits. A candidate can be described in a stage card without passing this Featured gate. There is no weighted total and no automatic promotion from the former Featured 闭.

## Historical freeze

- A: reconstructed Vulgar Latin *batare, gape/yawn. Retained as source context; Chinese Mapping Not selected. It has no independent historical-sound record and does not need an additional character. Do not equate GAPE and YAWN in the B search.
- B: Old French baer/bayer, opening wide/mouth. TLFi reports 1121–35, with later overlapping usage. Mapping Pending despite promising dialect research.
- C: same historical forms, figurative aspire/desire, reported around 1190. EXPECT means hopeful desire here, not neutral prediction. 巴 and 盼 recorded; no Old French IPA invented.
- D: Anglo-Norman abeiance and variants; AND legal attestations 1410, 1425–26. Already concerns suspension without a claimant/owner. Mapping Pending; 待 is a meaning control, not an equivalent legal term. The exact intervening derivation is unresolved rather than a fabricated direct C→D edge.
- E: modern English abeyance, temporary inactivity/suspension. 闭 retained as cognitive candidate; standard translation 暂缓／搁置. Exact earliest English sense chronology remains pending.

Sources: [TLFi bayer](https://www.cnrtl.fr/definition/bayerai), [Académie](https://www.dictionnaire-academie.fr/article/DNP1179), [AND abaiance](https://www.anglo-norman.net/entry/abaiance), [Cambridge](https://dictionary.cambridge.org/us/dictionary/english/abeyance), [Cornell LII](https://www.law.cornell.edu/wex/abeyance). No strict antonymization. Stop at onomatopoeic/uncertain deeper *batare source; the PIE root of the English gloss gape is irrelevant to this genealogy.

## Four requested Chinese tracks

### 巴 / 巴望 / 巴不得

[识典古籍：巴](https://www.shidianguji.com/character/巴) labels its electronic entry 《汉语大词典》; sense 5 concerns hope/waiting, sense 11 opening/cracking, and sense 13 cheek/jaw. These are different senses, not an attested historical progression. This is mediated dictionary evidence; original edition and pages were not directly obtained.

[曹秀玲、顾建身《汉语愿望表达与愿力调节》](https://hyxx.ybu.edu.cn/__local/7/C2/B4/B9FFCC77DBF06FCD7ED14B8B6FF_46734805_83316.pdf), 《汉语学习》2025(6):3–13, directly treats 巴不得、巴望 and 盼/盼望 as wish expressions (p.3). Discussion and literary examples of 巴望 occur at p.8. Definition footnotes at p.6 partly cite 《倒序现代汉语词典》; do not label all as directly checked 《现代汉语词典》第七版 entries. Modern literary examples are not medieval first attestations.

The [汉典 巴望](https://zdic.net/hans/巴望) and [巴不得](https://zdic.net/hans/巴不得) basic word explanations support desire; embedded 国语辞典 sections were not used as formal project evidence. Claimed early 《金线池》 examples surfaced in secondary search, but no inspected Mainland critical edition was obtained, so first attestation stays Pending. A search hit containing 刘巴 followed by 不得 is a personal-name sequence, not evidence of the lexical word 巴不得.

[《沂水县志》语法章](https://shandong-chorography.org/database/cg5/section/4/article/749/) gives 巴巴 open/hope uses. Its web phonetic symbols are not reliable enough to transcribe. [宋开玉's suffix study](https://www.rxgdyjy.sdu.edu.cn/info/1053/4206.htm) motivates the morphological caution: 嘴巴 does not establish standalone 巴 = mouth/gape. Conclusion: C stronger than A; B general opening only. Do not attach a snake/place-name Old Chinese reconstruction to the desire sense.

### 擘 / m- dialect candidate

[林亦《广西粤方言塞音、塞擦音鼻音化小规律所辖字的考察》](https://dialects.jnu.edu.cn/_upload/article/files/4c/5b/04d24f9c48caaa0f9d6d0a5190a9/aaa61f39-5630-413f-aa62-50c565bfba4c.pdf), 《粤语研究》21, pp.54–66, inspected PDF text. Table p.55 and discussion p.59 report [mak7b], large opening movements including 擘大嘴. The author proposes 擘 as the original character; the wording is analytical, not a claim of certain identity. Smaller movements have [mɐk7a]/[mek7a]; they must not replace the large-mouth form.

Geographical scope: Guangxi Yue, discussion of 沿江白话 literary/colloquial strata, 廉州话 and 桂南平话 variants. The text says those points retain [mak7b] for the large-movement sense, but does not bind the mouth example to a single named survey speaker/location. Thus no invented precise village. The paper uses language-protection-project material begun in 2015; this does not establish each token's recording year. Publication year and exact collection date remain pending.

Numerals are explicitly tone categories. 7b is 下阴入, not a contour like 55 or 21. No Mandarin bò substitution. Modern reference comparison: m versus b only shares labiality; -k is unmatched in modern bayer. Old French sound fit not evaluated. A modern regional Chinese p→m analysis is not evidence of cross-language common origin.

### 盼

Modern pàn hope/desire is supported by the wish-expression paper, [识典古籍](https://www.shidianguji.com/character/盼) sense 3 and [汉典](https://zdic.net/hans/盼) basic/detailed explanations. The old bright-eyes/looking sense must not be silently relabelled DESIRE and dated to the 詩經. Mainland edition/first hope attestation Pending. Modern p is [pʰ]; aspiration, -n and syllable structure differ from modern French bayer. Strong semantic expansion beyond b-only; historical phonetic fit not evaluated.

### 闭

[汉典：闭](https://zdic.net/hans/闭) supports closing/stopping; dictionary-native notation preserved in the archived legacy observation. The pilot uses the standard pinyin bì, with [p]/[i] broad comparison in prose; does not claim acoustic measurement. English /əˈbeɪ.əns/ has extra syllables, /beɪ/ differs from bì in voicing and vowel. It is a modern stressed syllable, not a historical root. “Temporary / awaiting resumption” is not inherent in 闭. Keep as research candidate; no literal CLOSE translation.

## Controls, search record and cost

Run ID: ABEY-PILOT-AUDIT. Sources/queries consulted during Discovery and this pilot: frozen open-wide/open-mouth, wish/desire, waiting/title suspension and modern temporary suspension; searches for 巴望/巴不得 historical quotations, 擘大嘴/mak, and the cited institutional texts. Modern Mandarin, historical dictionary senses and Guangxi/Shandong dialect material inspected. Bounded targeted search, not an exhaustive corpus run or blind experiment; no pre-registered budget or precision denominator.

Keep 张/开 for opening, 待 for waiting; 八 bā tests similar sound with unrelated meaning. Keep 慕 as a reserve desire candidate. Reject 喷/吠 for the frozen OPEN target: production of air/voice is not opening a mouth as a lexical sense. The full previous 21-item discovery ledger is preserved in this repository's accompanying discovery record; its scores are leads, not newly promoted results.

b-p-m-f adds the useful m- dialect 擘 lead; 盼 is a useful p- comparator but was already suggested before the first discovery run. No equally strong f- candidate emerged. Additional mouth-action false positives require semantic rejection. Report qualitative gain, not a fabricated false-positive rate.

All candidate Historical Relation fields remain Not claimed. Main Mainland original dictionaries/critical editions, exact historical-sense dates, Old French IPA and point-specific dialect audio remain Pending where not obtained. Missing access does not imply None found; no current stage is falsely marked exhaustive None found. The schema tests that state using a bounded synthetic search fixture.

## Implementation boundary

Existing schema container reused with optional model_version=0.1 branch, source stage refs, temporal/phonological metadata, separate fits/evidence, nullable Featured selection and max-three display refs. Legacy branch remains accepted. Only ABEYANCE record migrated. Old author prose/raw note unchanged; previous Featured and combined numeric assessment retained under legacy_calibration. Entry Mapping Level is Unrated so the former 闭 C score cannot grade all stages. Shared renderer detects model_version; legacy records retain previous output. Manifest release advances; dataset's existing format/version pin remains for compatibility, while entry version/model_version identify this pilot.

Featured Pending is an editorial implementation decision under the user's explicit authorization, not a statement that Jinkai independently verified every linguistic claim. Subsequent candidate promotion or legacy migration awaits a separate review.