docs(research): language LEGO grammar ABI to executable multilingual graph queries

[AdaWorldAPI]

Purpose

Capture the language LEGO architecture before it disperses across chat context: grammatical constructions become canonical, addressable semantic functions and an executable natural-language query frontend for the existing SoA/masked-RelAlg/Datalog-AST/R2IL/Quack/NAL substrate. Cypher remains an optional frontend rather than a compulsory translation target.

Added

1. docs/research/linguistic-lego-grammar-to-query-rfc.md

   • End-to-end representation: COCA/DeReKo priors; UD/CoNLL-U and CorefUD annotations; KorAP/Krill/Koral operators; multilingual Bible contrasts; Haspelmath function masks; canonical grammatical SoA; semantic question AST; mask/joins/folds; epistemic witnesses.
   • Existing code vs proposed work cross-checked against deepnsm-v2: position vs frequency for homographs; rule quorum; restored grammar heuristics + WordNet register; German grammar probes (D-LXC-13…23) #1321, docs(research): semantic Morton grammar ABI, Moore pressure and qualia-masked DO #1545, research(deepnsm-v2): Haspelmath grammar harness on frozen 8:8:8 thought ABI #1570, deepnsm-v2 wire: every CE64 field of a parsed triple from its source, Gadamer gate first; plan D-CGW-0..12 #1591, D-CGW-1: CE64 bits 43..45 are the orientation #1594, board: CE64 one layout; Moore tenant as synapse, membrane and local NARS #1595 and docs: epistemic frontier capstone map + O1/O3 decisions (46..49 activation, W = thought layer) #1597.
   • Preserves actual G4D3 4×3-byte TEKAMOLO facet, actual 24×i4 event-locus contract, one CE64 layout/orientation codebook, and no per-verse/tenant register identity.
   • Explains lexical alternatives, scope/quantifiers/negation, evidence binding, no parser confidence promoted into NAL/CE64 truth, and the still-missing text→R2IL execution seam.
   • Gated L-0…L-8 implementation outline.

2. docs/research/linguistic-lego-corpus-and-falsification-matrix.md

   • Links/resources from Universal Dependencies, CorefUD, UniMorph, COCA, DeReKo, KorAP (Kalamar/Kustvakt/Krill/Koral/korapxmltool), Tantivy, ParZu, Grew, nlprule, SMOR, discourse-lab, discoursegraphs, PDTB, Haspelmath, multilingual Bible.
   • Corpus access and tool/model licensing cautions; CoNLL-U≠valid UD when STTS leaks into UPOS.
   • Concrete adversarial grammar/query fixtures and differential oracles for POS/lemma, UD dependencies, coreference, semantic roles, topology, query AST equivalence, reasoning and execution.
   • Explicit stop conditions and reproducible pre-existing probe commands (not claimed run here).

Critical guardrails

• Do not rebuild the existing Haspelmath or TEKAMOLO systems. No new width/layout, no reallocated 24×i4 reserved loci, no per-verse S:P:O dictionaries.
• UD gold labels grammar, not semantic truth; Bible verse correspondences supply approximate alignment, not semantic ground truth or independent NAL witnesses.
• Masks restrict candidate populations; graph joins and Datalog recursion need real typed operators and an oracle. No gather hidden as a mask.
• Preserve current-code / later-decided CE64 interpretation differences until reconciled by their own PR.
• Documentation/research only: no executable behavior change, no benchmarks or tests claimed.

Suggested next action

Review the two docs, then choose L-1 (round-trip UD/CoNLL-U SoA DTO) plus L-2 (Krill span/operator differential harness) as narrowly scoped subsequent PRs; keep an independently annotated cross-language natural-question AST oracle for L-5.

[coderabbitai]

Important

• 🔍 Trigger review

This repository does not receive automatic reviews because it has fewer than 10 stars.

⚙️ Run configuration

• Configuration used: Organization UI
• Review profile: CHILL
• Plan: Advanced
• Run ID: 285017ab-e5ba-4716-ac98-ed9789214be8

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Warning

Billing warning: we have not been able to collect payment for this subscription for more than 72 hours. Please update the payment method or pay any pending invoices in Billing to avoid service interruption.

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-18T14:32:18.865241Z	96747ac	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 96747ac55b

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[chatgpt-codex-connector]

Separate antecedent agreement from relative-pronoun case

For relative clauses where the antecedent and pronoun have different syntactic roles, this fixture requirement incorrectly includes case in their agreement. The relative pronoun agrees with its antecedent in gender and number, but its case is determined within the relative clause—as the RFC's own dem Kind, das ... example demonstrates. Implementing this item literally would reject precisely the valid cross-case construction the experiment is intended to preserve.

Useful? React with 👍 / 👎.