The ground migration, executed — session record, citation maps, projection edits

apparatus/encode-verify.md

@@ -18,7 +18,8 @@ there, pinned at a version and a status in `graph/upstream.yaml`:
 > **You can encode ground you control.**
 > **You must mechanically verify ground you don't.**
 
-Ground is the read-only surface an actor inspects in order to act (`core/00`). Some of it you own —
+Ground (`core/00`) is what a determination is resolved against; this apparatus reads it as the
+read-only surface an actor inspects in order to act. Some of it you own —
 your source, your config, your schemas. Some of it you do not — a third-party API, a live
 environment, another team's service. The rule says these get different treatment, and the difference
 is not stylistic.

apparatus/prefix-stability.md

@@ -101,7 +101,7 @@ adversarial ones:
 | random mix | 1312 (fails) | 557 ✓ |
 | realistic prefix | optimal ✓ | optimal ✓ |
 
-**Why the simple rule usually works anyway.** In a typical prefix, stable content *is* long (ground,
+**Why the simple rule usually works anyway.** In a typical prefix, stable content *is* long (delivered ground,
 a large corpus of settled decisions) and volatile content *is* short (the task). Rate and rate/length
 then agree, which is why ascending-rate looks right in practice and why the shipped INV-6 ordering
 (ground → decisions → task) is correct. **But the rule that holds in general is Smith's**, and any

applications/sdlc/production-as-ground.md

@@ -150,7 +150,7 @@ exception matters:
 The middle tier is why the naive two-way cut fails. *"Does this query hold at production traffic"* is
 a **technical** question whose predicate does not close statically — you can benchmark against
 substitutes, but the real answer requires real load. Likewise correctness against production's actual
-`P(input)`, which is `core/09`'s ground-distribution caveat arriving in practice: you do not have
+`P(input)`, which is `core/09`'s deployment-distribution caveat arriving in practice: you do not have
 production's input distribution in staging.
 
 **And this is what progressive delivery is *for*.** Canaries and shadow traffic are not a

core/assets/measure-mdl-volume.py

@@ -28,7 +28,7 @@
 correspondence (coefficients stipulated).
 
 Task (identical to the upstream measure-toy.py figures): (M, D), M in 1..4,
-D in 1..31, uniform ground, verdict VALID iff D <= days(M),
+D in 1..31, uniform deployment distribution, verdict VALID iff D <= days(M),
 days = {31, 28, 31, 30}.
 """
 

meta/migration-plan-ground.md

@@ -140,3 +140,21 @@ is a convenience for a reader rather than an instruction to an instrument. `w1-e
 **Corollary, and it is the wave-planning fact:** a generated table's rows are not the wave's to
 move at all. They carry canon's words and they move when canon moves. W1's own scope shrinks once
 that is applied — see the counts correction below.
+
+### A fourth, found by the execution session and added on Emil's Gate 1 ruling
+
+> **A figure whose instrument was not committed is not reproducible, and the ledger should have
+> been carried rather than the number.**
+
+This plan's own headline — 380 qualified, 187 bare — was measured at the predecessor's GATE 4 and
+the instrument behind it never entered the record. The figure survived; the row list did not. The
+execution session had to rebuild the instrument (anchored, clause-clipped, its lexicon frozen
+after two warrant-disciplined revisions, per-rule precision reported), and the rebuilt instrument
+reads the same baseline as **389/178** — a nine-row divergence that can never be located row-wise,
+because there is nothing to locate it against. Ruled at the execution session's Gate 1: the
+committed instrument's ledger governs
+(`meta/sessions/2026-08-31-ground-migration-exec/g1-head-ledger.json`), 380/187 stands recorded as
+a prediction that did not verify, and hunting nine rows to make the old number true would be
+reconciling a divergence rather than recording it. The mechanism is the same class as the other
+three and was learned the same way: by an instrument's consumer, after the fact. **A count that
+matters ships with the instrument that counted it and the rows it counted, or it is prose.**

meta/sessions/2026-08-31-ground-migration-exec/bootstrap.md

@@ -0,0 +1,43 @@
+Unzip migration-package.zip and read migration-package/README.md, then read
+prompt-migration-exec.md in its entirety — this session follows that prompt exactly, including every
+gate and all seven standing rulings.
+
+This is the ground migration, executing. Its predecessor planned it in full and deferred it whole for
+one stated reason: each of the 187 prose re-expressions is a ruling, and the first twenty would set
+precedents the remaining 167 inherit unreviewed. This session exists to spend those rulings
+deliberately — G2 is the whole answer to why the deferral happened, and if G2 fails to produce a
+citable precedent set, the migration defers again.
+
+First act, before anything else: commit this prompt and bootstrap to
+meta/sessions/2026-08-31-ground-migration-exec/ in decision-driven-design, per DDD-dec-20.
+
+Read meta/migration-plan-ground.md in decision-driven-design before anything else in the repos. It is
+the seed: wave structure, every pin and firing named by id and hash, the four things not to
+re-decide, and the method rule with its three mechanisms. The repo copy is authoritative over
+anything in this package.
+
+Fetch all three repos — actor-indexed-determination at v5.12.0 (verify and report the tag found),
+decision-driven-design at head, product-cli read-only for W4 assessment.
+
+Two things moved since the seed was written, both in the migration's favour: W1 is partly done
+(Paper A's revision moved 15 of 31 in the two manuscripts, deferred 16 with reasons; 73 of 88
+remain), and the measure note's pin advanced so the two papers agree on the moved sense.
+
+Rules that override anything you might infer:
+- Interactive execution. Stop at every gate for Emil's ruling. Merge nothing.
+- The seven standing rulings are inherited and not re-litigated. SR-2 in particular: 380 of 567 S2/S3
+  prose occurrences need no edit, because canon already names what is done to the ground.
+- G2's twenty are selected for coverage, not convenience — every distinct sentence shape, hardest
+  first. Every later re-expression cites which precedent it follows; one citing none is a new ruling
+  and stops at a gate.
+- The three method rules are what the predecessor learned by getting them wrong: clause-crossing
+  windows are not anchored rules; report per-rule precision never one figure; classification data
+  cites content, never position.
+- State predicted W6/W7/W5 results before every pin operation, verify after, record divergence rather
+  than reconciling it.
+- W4 (product-cli's serialised ground: field, 660 occurrences) is assessed and reported at Gate 6,
+  never executed here.
+- Commit drafts before reporting at each gate, bodies marked draft-pending-ruling.
+
+Begin with G1 and end your first report at GATE 1: the classification reconciled against head, which
+rows the two merged sessions changed, any sense boundary that moves, and the corrected wave sizes.

meta/sessions/2026-08-31-ground-migration-exec/g1-added-rulings.py

@@ -0,0 +1,113 @@
+#!/usr/bin/env python3
+"""G1 — proposed senses for the live rows the corpus gained since the
+classification was fixed. PROPOSED, draft-pending-ruling: these enter the
+working ledger only on Emil's Gate 1 ruling.
+
+Keyed by content, never by position (the seed's third method mechanism): each
+key is a distinctive substring of the occurrence's context, unique within its
+file at head. The decision rule is the audit's — read the predicate applied
+to the word, never the compound. Each entry: (repo, path, anchor) -> (sense,
+reason). Rows in audit output, session records and release descriptors are
+not ruled here — they are counted, never migrated.
+"""
+
+R = {
+# ---- upstream README.md (19 rows; the predecessor ruled the file into the
+# migration at its GATE 1; its per-row senses were reported only in aggregate
+# — S1x8 S2x4 S3x4 S5x1 U-x1 — so these are re-derived and the two
+# aggregate divergences are flagged in the gate report) ----
+('upstream','README.md','determines choices against ground: programs'): ('S1','world predicate: determined against; the ident flag on this row is a false positive of the ground:-colon rule'),
+('upstream','README.md','What ground does the determination depend on?'): ('S1','world predicate: depended on by the determination'),
+('upstream','README.md','**Ground** — what they are determined against'): ('S1','the definition itself'),
+('upstream','README.md','resolving of a decision against ground'): ('S1','determined against'),
+('upstream','README.md','prior commitments, ground channels, checks'): ('S3','channel: the delivery path; verb-shaped under SR-2'),
+('upstream','README.md','⟨task, ground, acceptance relation'): ('S1','the index tuple names the case coordinate'),
+('upstream','README.md','an actor reading ground, **with an accountable party named**'): ('S3','act predicate: reading'),
+('upstream','README.md','relevant ground is observable and adequacy'): ('S2','store predicate: observable within declared bounds (term:closure)'),
+('upstream','README.md','left open at the act, holding ground fixed'): ('S2','held fixed for the quantification'),
+('upstream','README.md','verdict induced by the task over the ground distribution'): ('S5','population predicate: distribution over cases'),
+('upstream','README.md','verdict function with the required ground truth'): ('U-ordinary','"ground truth" is ordinary English of measurement, none of the five'),
+('upstream','README.md','how much governing ground it can have available at once'): ('S2','store predicate: available at once (hold capacity)'),
+('upstream','README.md','output decoupled from correct ground because'): ('S3','act predicate: absent/false/present at the act'),
+('upstream','README.md','because that ground is absent, false, or present'): ('S3','same predication, second token'),
+('upstream','README.md','Actorhood means resolving choices against ground'): ('S1','determined against'),
+('upstream','README.md','**Ground and judgment dependence** Situated'): ('S2','H2 heading; the predication is availability, made explicit in the next sentence'),
+('upstream','README.md','advantage as ground becomes unavailable'): ('S2','store predicate: unavailable'),
+('upstream','README.md','|decisions, ground, determination, actor'): ('U-mention','doc-map row citing the primitive by name'),
+('upstream','README.md','ask what must be determined, what ground it depends on'): ('S1','depended on'),
+
+# ---- papers/paper-a/paper-a.md (25 rows added by the revision) ----
+('downstream','papers/paper-a/paper-a.md','leave open at the act, held at fixed ground.** [DDD-frame-02'): ('S2','held fixed'),
+('downstream','papers/paper-a/paper-a.md','leave open at the act, **held at fixed ground**: the alternatives'): ('S2','held fixed'),
+('downstream','papers/paper-a/paper-a.md','configuration and the ground at the act are both given'): ('S3','at the act'),
+('downstream','papers/paper-a/paper-a.md','It is not variation > across ground — a cryptographic hash'): ('S1','world predicate: variation across the case'),
+('downstream','papers/paper-a/paper-a.md','The held-at-fixed-ground clause is a repair'): ('U-mention','names the clause, not the object'),
+('downstream','papers/paper-a/paper-a.md','conflating four phenomena: outcome variation across ground'): ('S1','variation across'),
+('downstream','papers/paper-a/paper-a.md','Quantifying at *fixed ground* is what excludes it'): ('S2','fixed'),
+('downstream','papers/paper-a/paper-a.md','**Variation across ground is not discretion**'): ('S1','variation across'),
+('downstream','papers/paper-a/paper-a.md','standing configuration and this act\'s ground are given'): ('S3','this act\'s, given at the act'),
+('downstream','papers/paper-a/paper-a.md','no residual discretion at fixed ground is *predetermined*'): ('S2','fixed'),
+('downstream','papers/paper-a/paper-a.md','standing configuration and the ground jointly fix the output'): ('S1','the case jointly fixing the outcome'),
+('downstream','papers/paper-a/paper-a.md','together with > the ground at the act determines the resolution'): ('S3','at the act'),
+('downstream','papers/paper-a/paper-a.md','configuration and the ground at the act, together, fix the resolution'): ('S3','at the act'),
+('downstream','papers/paper-a/paper-a.md','asks whether standing configuration and ground jointly determine'): ('S1','jointly determine'),
+('downstream','papers/paper-a/paper-a.md','whether standing configuration and ground determine the resolution [DDD-frame-17, falsifier]'): ('S1','determine'),
+('downstream','papers/paper-a/paper-a.md','decides acceptability over accessible ground within declared bounds'): ('S3','accessible'),
+('downstream','papers/paper-a/paper-a.md','demand time, memory, permissions or ground no arrangement has'): ('S2','store predicate: has/holds; bare of a delivery verb'),
+('downstream','papers/paper-a/paper-a.md','| Ground | Assignment | What would overturn it |'): ('U-mention','column header naming the coordinate'),
+('downstream','papers/paper-a/paper-a.md','ground missing to one arrangement is available to another'): ('S3','missing/available to an arrangement'),
+('downstream','papers/paper-a/paper-a.md','does any of those has ground the walk assumed missing'): ('S3','missing/has at the act'),
+('downstream','papers/paper-a/paper-a.md','feedback density, ground access, checker cost'): ('S3','access'),
+('downstream','papers/paper-a/paper-a.md','its ground by provenance, its predicates by closure rung'): ('S2','coded by provenance — SR-4 attribute on the held object; boundary case, flagged'),
+('downstream','papers/paper-a/paper-a.md','ground provenance (§2.4, this paper\'s analysis and not canon)'): ('S2','provenance attribute; boundary case, flagged'),
+('downstream','papers/paper-a/paper-a.md','The gap is grounded in a **control condition'): ('U-ordinary','"grounded in" — a different word'),
+('downstream','papers/paper-a/paper-a.md','the gap, grounded in a control condition'): ('U-ordinary','"grounded in"'),
+
+# ---- papers/paper-a/paper-a-supplement.md (15 rows; the appendix rows are
+# generated from the pinned upstream — they carry canon's words and move when
+# canon moves, never as wave work) ----
+('downstream','papers/paper-a/paper-a-supplement.md','keep citing a commit whose prose says `ground distribution`'): ('U-mention','quotes the phrase as vocabulary'),
+('downstream','papers/paper-a/paper-a-supplement.md','five-way ground provenance (controlled, observed'): ('U-mention','cites the passage by its name'),
+('downstream','papers/paper-a/paper-a-supplement.md','## S5. Vocabulary: `ground` in its population sense'): ('U-mention','word-as-word'),
+('downstream','papers/paper-a/paper-a-supplement.md','The word `ground` had been carrying five senses'): ('U-mention','word-as-word'),
+('downstream','papers/paper-a/paper-a-supplement.md','canon still reads `ground distribution` at `v5.12.0`'): ('U-mention','quotes canon\'s current wording'),
+('downstream','papers/paper-a/paper-a-supplement.md','commitments leave open at the act, held at fixed ground. | | `DDD-frame-03`'): ('S2','generated appendix row quoting DDD-frame-02; moves when canon moves'),
+('downstream','papers/paper-a/paper-a-supplement.md','together with the ground at the act determines the resolution; exercised'): ('S3','generated appendix row quoting DDD-frame-17; moves when canon moves'),
+('downstream','papers/paper-a/paper-a-supplement.md','| `DDD-ground-01` | normative | projected'): ('U-idname','node identifier; SR-6'),
+('downstream','papers/paper-a/paper-a-supplement.md','| `DDD-ground-02` | conceptual | projected'): ('U-idname','node identifier; SR-6'),
+('downstream','papers/paper-a/paper-a-supplement.md','| `DDD-ground-03` | conceptual | projected'): ('U-idname','node identifier; SR-6'),
+('downstream','papers/paper-a/paper-a-supplement.md','| `DDD-ground-05` | conceptual | projected'): ('U-idname','node identifier; SR-6'),
+('downstream','papers/paper-a/paper-a-supplement.md','nor the estimability of the ground distribution. | | `DDD-measure-17`'): ('S5','generated appendix row quoting DDD-measure-16; W1-bound with the claim, moves when canon moves'),
+('downstream','papers/paper-a/paper-a-supplement.md','commitments leave open at the act, **held at fixed ground**: the alternatives'): ('S2','quotes the amended term:residual-discretion draft'),
+('downstream','papers/paper-a/paper-a-supplement.md','admissible once the standing configuration and the ground at the act are both given. It is not variation'): ('S3','same quotation, at-the-act clause'),
+('downstream','papers/paper-a/paper-a-supplement.md','are both given. It is not variation across ground — a cryptographic hash'): ('S1','same quotation, variation-across clause'),
+
+# ---- papers/paper-a/response-to-review.md (15 rows; the response is the
+# paper's record of what it conceded — quotes the review and the repairs) ----
+('downstream','papers/paper-a/response-to-review.md','### 2.2 Raw ground collapses the allocation measure'): ('S1','the review\'s object: world facts fed raw to the measure'),
+('downstream','papers/paper-a/response-to-review.md','mutual information reports the ground as carrying the entire verdict'): ('S1','world facts carrying the verdict'),
+('downstream','papers/paper-a/response-to-review.md','no rule for turning that ground into an answer'): ('S1','same object, next clause'),
+('downstream','papers/paper-a/response-to-review.md','together with the ground at the act fix the resolution'): ('S3','at the act'),
+('downstream','papers/paper-a/response-to-review.md','amended to quantify **at fixed ground**: the alternatives'): ('S2','fixed'),
+('downstream','papers/paper-a/response-to-review.md','configuration *and this act\'s ground* are both given'): ('S3','this act\'s'),
+('downstream','papers/paper-a/response-to-review.md','phenomena are separated: variation across ground, epistemic uncertainty'): ('S1','variation across'),
+('downstream','papers/paper-a/response-to-review.md','no residual discretion at fixed ground is exactly what'): ('S2','fixed'),
+('downstream','papers/paper-a/response-to-review.md','### 3.3 Split relevant ground from accessible ground'): ('S1','relevant: the world-facts half of the review\'s split'),
+('downstream','papers/paper-a/response-to-review.md','Split relevant ground from accessible ground (review §6)'): ('S3','accessible: the delivered half'),
+('downstream','papers/paper-a/response-to-review.md','versus `G_A` (ground accessible and delivered to arrangement'): ('S3','accessible and delivered'),
+('downstream','papers/paper-a/response-to-review.md','the delivery distinction applied to ground. It is **canon work'): ('S1','the undivided object the distinction applies to'),
+('downstream','papers/paper-a/response-to-review.md','the tuple\'s `ground` coordinate is a settled term'): ('U-mention','names the coordinate as a term'),
+('downstream','papers/paper-a/response-to-review.md','§8.1 now says that ground missing to one arrangement is available'): ('S3','missing/available'),
+('downstream','papers/paper-a/response-to-review.md','on the explicit ground that those sentences are about warrant'): ('U-ordinary','"on the ground that" — ordinary English of reasons'),
+
+# ---- the rest of the live additions ----
+('downstream','papers/measure-note/measure-note.md','Worked by `measure-nonuniform-ground.py`, named before the vocabulary moved'): ('U-idname','asset filename; the note itself says why it keeps the name'),
+('downstream','core/decisions/DDD-dec-34.yaml','meta/migration-plan-ground.md predicted three W6'): ('U-idname','file path'),
+('downstream','core/decisions/DDD-dec-34.yaml','repairs and are not ground-migration work'): ('U-mention','names the programme'),
+('upstream','core/decisions/DDD-dec-31.yaml','DDD-ground-0'): ('U-idname','node identifiers; SR-6 (three rows, one anchor: all three tokens sit in DDD-ground-0N ids)'),
+}
+
+if __name__ == '__main__':
+    import collections
+    print(f'{len(R)} proposed rulings')
+    print(dict(collections.Counter(v[0].split("-")[0] for v in R.values())))