06 — Axis / Object-Grain Boundary (Objective F)

Verdict: both NO-GO. Verified from live state.

Axis coverage — substrate ABSENT

Live to_regclass checks:

table	result
axis_registry	NULL (absent)
axis_assignment	NULL (absent)
coverage_rule	NULL (absent)

With no axis tables, any "axis coverage" is fake / test-only — there is no schema to register axes, assign objects to axes, or define per-axis coverage rules. The only real coverage surface remains object × scope (the 210 model). Axis-grain stays a design-stage item.

Blockers to axis-grain GO:

1. axis_registry + axis_assignment + coverage_rule must be designed, ratified, and built (none exist).
2. A topic/axis governance model (which axes exist, who owns axis definitions) must be ratified.
3. Until then there is nothing to scan at axis grain.

Object-grain coverage — UNBOUNDED, NO-GO

Current coverage is collection-grain: 35 governed collections. Object-grain would mean per-entity ownership across birth_registry:

• birth_registry rows (live) = 1,084,816.
• Object-grain cells = 1,084,816 × 6 scopes ≈ 6.5 million vs 210 today — a ~31,000× explosion.

Blockers to object-grain GO:

1. Cardinality explosion — 6.5 M coverage cells; the recursive effective-owner climb (depth < 64, CROSS JOIN scopes) over millions of nodes is not viable as an unindexed view; would need materialization + bounding rules.
2. No bounding substrate — coverage_rule (absent) is exactly what would scope object-grain to a governed subset; without it an object-grain scan is unbounded (explicitly forbidden by the mission).
3. Containment at object grain undefined — current containment is collection→group (depth-1); there is no entity→entity containment to inherit ownership through at object grain.
4. No demonstrated need — Phase-1 governs collections; per-entity accountability is a later phase if ever.

Conclusion

Both axis-grain and object-grain are NO-GO, blocked by absent substrate and (for object-grain) unbounded cardinality. The collection×scope 210 model is the correct and only real Phase-1 coverage surface. No fixture or test-only axis/object-grain artifact was created (that would be a hidden governance island).