Primary Directions External-Authority Ledger

1. Record status

• Record date: 2026-07-19
• Scope: narrow external numerical evidence for Campanus, Topocentric, zodiacal fixed-star, reflected-point, Ptolemaic-parallel, Placidian rapt-parallel, and Morinus aspect-context branches
• Authority fixtures: tests/fixtures/primary_directions_campanus_topocentric_authority.json, tests/fixtures/primary_directions_special_target_authority.json, and the corrected source labels in tests/fixtures/primary_directions_ptolemy_examples.json
• Executable comparisons: tests/unit/test_primary_direction_campanus_topocentric_authority.py, tests/unit/test_primary_direction_special_target_authority.py, and tests/unit/test_primary_direction_ptolemy.py
• Purpose: state exactly which published products the executable corpus compares, which checks are only internal invariants or regression evidence, and which branches remain unevaluable.

The evidence below is deliberately product-specific. A published worked example can attest its named low-level law without validating every method, target family, motion doctrine, epoch, or public search combination that uses that law.

2. Evidence classes

This ledger keeps the following evidence classes separate:

• External authority validation: a current computation is compared with a numerical fact from an original or method-origin source under named input and product semantics.
• External published-example comparison: a later published worked example supplies the external numerical fact. This is useful external evidence, but is not silently promoted to a primary historical authority.
• Cross-engine corroboration: a secondary reconstruction or independent software calculation corroborates an original historical row.
• Secondary-source corroboration: a modern technical derivation or calculator adaptation supplies a bounded worked example.
• Invariant or regression evidence: Moira is compared with its own formula, branch rule, prior fixture value, engine result, or transport serialization. This detects drift and proves internal consistency; it is not external astronomical or historical validation.

The corpus-schema tests, the 1e-12 Campanus/Regiomontanus conjunction agreement, the internal contra-parallel sign test, and direct-engine-versus-REST parity all belong to the last class. They do not widen the external evidence recorded below.

3. Campanus and Topocentric source-law evidence

The Campanus/Topocentric fixture preserves printed equatorial inputs rather than silently substituting a modern reconstruction of the natal charts.

3.1 Campanus-Regiomontanus mundane conjunction

Bob Makransky's worked Prince Charles example directs the Sun as promissor to Mercury as significator in mundo. The tested object is the narrow conjunction law shared by the Campanus and Regiomontanus paths: form the significator pole, place both objects under that pole, and subtract the significator's W from the promissor's W.

Quantity (degrees)	Published	Moira	Absolute gate
Significator pole	21.27	21.266884721880288	0.03
Significator W	210.41	210.405005826081	0.03
Promissor W under significator pole	222.58	222.57942742883125	0.03
Signed direct arc	12.17	12.174421602750243	0.03

The source prints inputs and results to 0.01 degree. A complete +/-0.005-degree corner sweep of the printed inputs places the reconstructed arc in [12.15846, 12.19038] degrees; the 0.03-degree gate encloses that rounding domain. The exact equality of Moira's Campanus and Regiomontanus results at 1e-12 is a separate internal invariant. This row does not validate the wider Campanus mundane-aspect family.

3.2 Polich oblique ascension under a named pole

Vendel Polich's origin-text example supplies an eastern oblique ascension for an ecliptic point at longitude 165.02666666666667 degrees, zero latitude, under pole 51.53333333333333 degrees and obliquity 23.439166666666665 degrees.

Quantity (degrees)	Published	Moira	Absolute gate
Eastern oblique ascension	158.75	158.74082783607466	0.02

The published result is tabular and printed to one arcminute. The 0.02-degree gate allows 1.2 arcminutes for printed-table interpolation and rounding. It attests this named under-the-pole operation, not a complete chart or every Topocentric direction.

3.3 Topocentric zodiacal aspect under the significator pole

Makransky's worked example uses Saturn as significator and a zero-latitude Moon trine point as promissor. The tested signed object is again promissor W under the significator pole minus significator W.

Quantity (degrees)	Published	Moira	Absolute gate
Significator pole	29.89	29.896316011196287	0.02
Significator W	151.26	151.26140957829284	0.02
Promissor W under significator pole	145.89	145.89100748389953	0.02
Signed primary-motion arc	-5.37	-5.370402094393285	0.02

The source prints equatorial inputs and results to 0.01 degree. A complete +/-0.005-degree input corner sweep places the signed arc in [-5.38644, -5.35436] degrees; the 0.02-degree gate encloses that domain.

Motion-convention boundary

The source labels the -5.37-degree result converse. Moira preserves two different, explicitly named products rather than collapsing their meanings:

• the established topocentric_zodiacal_aspect preset retains traditional converse by exchanging promissor and significator roles; its converse arc for the printed entries is 5.4084796290214285 degrees
• the additive topocentric_zodiacal_aspect_signed_primary_motion preset constructs one ordered arc, wraps it to the signed shortest circular displacement, and classifies its sign; positive is direct and negative is converse

For the low-level printed equatorial entries, the ordered raw result is 354.6295979056067 degrees and its signed form is the source-matching -5.370402094393285 degrees. The public-preset witness reconstructs Saturn's ecliptic position from the printed right ascension and declination, reconstructs the Moon source point 120 degrees before the published trine, deliberately supplies a non-zero native lunar latitude, and lets the preset assign zero latitude to the aspect point. It emits one converse arc of 5.359051191104641 degrees, within the source's 0.02-degree printed-input gate. This small difference from the low-level reconstruction reflects the public ecliptic-to-equatorial reconstruction path over rounded source inputs; it is not a second converse doctrine.

Coincidence within the declared 1e-12-degree numerical tolerance produces no event. An exact or tolerance-coalesced antipodal 180-degree displacement has no unique sign and fails closed. This narrow sign classifier is assigned-zero, projected-perfection, Topocentric, zodiacal, and aspect-specific. It does not change traditional role exchange and is not the deferred global neo-converse doctrine. Public searches require explicit, non-empty significator and promissor filters; the unrestricted target set is not admitted because it necessarily includes a directionally ambiguous MC/IC antipode.

4. Special-target external comparisons

4.1 Lilly/Kolev Vega zodiacal projection to the Ascendant

The fixture converts Lilly's 1616-09-19 Julian chart date to 1616-09-29 Gregorian, reconstructs the stated UT1-proxy instant, resolves Vega through Moira's sovereign star engine, explicitly suppresses the resolved stellar latitude, and directs the resulting zodiacal point to the Ascendant by the admitted Ptolemaic semi-arc law. This is a historical zero-latitude zodiacal product, not a modern astrometric oracle for Lilly's catalog.

Comparison (degrees)	Published/reconstructed	Moira	Absolute gate
Vega longitude, Lilly	280.0	279.94671064908306	0.06
Direction arc, Lilly	3.2333333333333334	3.15867227080156	0.10
Vega longitude, Kolev reconstruction	279.96666666666664	279.94671064908306	0.03
Direction arc, Kolev reconstruction	3.1666666666666665	3.15867227080156	0.02

The Lilly residuals are 0.05328935091694 degrees (3.1974 arcminutes) in longitude and 0.07466106253177 degrees (4.4797 arcminutes) in arc. The round-number 0.06/0.10 gates enclose those historical-catalog and chart- reconstruction residuals; they are cross-model regression envelopes, not uncertainty estimates. The Kolev residuals are 0.01995601758358 degrees (1.1974 arcminutes) and 0.00799439586511 degrees (0.4797 arcminutes); the 0.03/0.02 gates enclose the minute-printed reconstruction. Because the complete computational inputs depend on Kolev's reconstruction, the effective evidence class is cross-engine corroboration even though Lilly remains the original authority for his published row.

4.2 Lilly Jupiter antiscion to the Ascendant

This row reflects Lilly's Jupiter longitude across the Cancer-Capricorn axis, assigns zero latitude, and directs the antiscion to the Ascendant.

Quantity (degrees)	Published	Moira	Absolute gate
Reflected longitude	278.0833333333333	278.0833333333333	0.02
Direction arc	1.4	1.377903256028162	0.03

The reflection is exact at the stored precision. The arc residual is 0.02209674397184 degrees (1.3258 arcminutes); the 0.03-degree gate encloses the minute-printed historical/reconstruction residual. This is external authority validation of the named planetary, zero-latitude antiscion product only.

4.3 Sepharial Ptolemaic zodiacal parallel

Sepharial's illustration selects the Cancer ecliptic point whose declination equals Uranus's published north declination (23 degrees 24 arcminutes). The branch-selecting longitude is explicit; it is not presented as Uranus's radical longitude.

Quantity (degrees)	Published	Moira/source arithmetic	Absolute gate
Declination-equivalent longitude	94.0	93.22323381173652	1.0
Arc from the two published rising sidereal times	80.5	80.5	1e-9

The source rounds the equivalent point to a whole degree, which governs the 1.0-degree longitude gate. The arc check is exact arithmetic over the two published sidereal times, not a separate end-to-end direction search. This row validates a parallel, not a contra-parallel.

The corrected Saturn row in primary_directions_ptolemy_examples.json likewise preserves the source's 6 degrees 54 arcminutes south declination and labels the projection a parallel. Moira computes 342.4210029835662 degrees against the published 342.6166666666667 degrees under the existing 0.25-degree printed-table gate; published rising-time arithmetic gives 11.5 degrees. The source calls this Saturn to the Sun's place and never identifies it as a contra-parallel.

4.4 Leo/Griscti Placidian rapt parallels

The direct example uses Saturn's diurnal and the Moon's nocturnal semi-arcs with the Moon taken by opposition in right ascension. The converse example uses the source's forward converse construction.

Product (degrees)	Published	Moira	Absolute gate
Direct rapt-parallel arc	37.7	37.697679078635076	0.02
Converse rapt-parallel arc	36.86666666666667	36.86546721498743	0.02

The residuals are 0.00232092136492 degrees (0.1393 arcminute) and 0.00119945167924 degrees (0.0720 arcminute). The 0.02-degree gates allow for the minute-rounded source inputs. These rows are secondary-source corroboration through a modern calculator adaptation, not primary validation from a page- and glyph-correlated Placidus transcription.

4.5 Borealis Morinus aspect context

The executable modern example starts from Jupiter at longitude 203 degrees 34 arcminutes and latitude +1 degree 10 arcminutes, departing a northern maximum latitude of 1 degree 34 arcminutes. It applies a positive 60-degree sinister sextile along the Morinus aspect circle.

Quantity (degrees)	Published	Moira	Absolute gate
Projected longitude	263.55	263.5517101359027	0.2
Projected latitude	-0.3333333333333333	-0.32207177424366357	0.2

The residuals are 0.0017101359027 degrees (0.1026 arcminute) and 0.01126155908967 degrees (0.6757 arcminute). The conservative 0.2-degree gates are cross-source envelopes for a minute-rounded secondary example, not claimed uncertainties. The worked equations use 60 degrees; a closing prose reference to 30 degrees is treated as a source typo. This single sextile does not externally validate quadrant continuation through squares, trines, or opposition.

5. Explicitly unevaluable or deferred branches

No synthetic expected values are assigned to these branches:

Branch	Why current evidence is insufficient	Evidence required
Fixed star, true latitude in mundo	Lilly's Vega row suppresses latitude; Vega proper is circumpolar in the reconstructed chart.	A named non-circumpolar star, epoch, complete chart geometry, true-latitude doctrine, and published arc.
Fixed star to planet	Accessible Lilly/Kolev rows do not provide a complete independently reproducible historical star coordinate set.	Page-correlated star identity/coordinates, chart geometry, latitude doctrine, significator coordinates, and published arc.
Node- or angle-sourced antiscia	The recovered Lilly corpus supplies planetary antiscia only.	An original worked direction explicitly reflecting a node or angle.
Antiscia cum latitudine	Lilly's latitude-aware doctrine differs from Moira's admitted zero-latitude primary-direction reflection.	A separate latitude-aware public doctrine and source-matched tests.
Ptolemaic contra-parallel	The recovered Sepharial Saturn row is explicitly south declination but is not labeled contra-parallel.	A published worked example that names the contra-parallel and supplies hemisphere and branch semantics.
Placidian rapt parallel from a primary text	Executable rows come from the Leo/Griscti adaptation.	A complete page-correlated Placidus transcription with identities, semi-arcs, RAs, meridian distances, and arc.
Morinus primary-source aspect context	The executable Churchill row is a modern derivation; Morin Book XXII has not been numerically transcribed.	A page-correlated primary-text example or licensed authoritative translation.
Morinus quadrant continuation	The published sextile remains on one arctangent branch.	Published worked values beyond the principal branch at square, trine, or opposition.
Wider Campanus mundane aspects	The Makransky row is a conjunction using the law shared with Regiomontanus.	Published Campanus-specific aspect examples with complete inputs.

Internal unit tests do exercise node/angle reflections, contra-parallel negation, Morinus quadrant continuity, fixed-star-to-planet paths, and wider policy composition. Those remain invariant or regression evidence until the external products named above are available.

6. Sources and rights/provenance

• Bob Makransky, Primary Directions: A Primer of Calculation, parts 1 and 2, Dear Brutus Press, 1988, ISBN 0-9677315-0-X. The repository stores only minimal bibliographic metadata and numerical facts, not scans or substantive prose.
• Vendel Polich, The Topocentric System, translated by Jose M. Lebron, Editorial Regulus, 1975 (hosted scan). The repository stores only minimal bibliographic metadata and numerical facts.
• William Lilly, Christian Astrology, 1659, pp. 765-768 (public-domain scan). The fixture cites original numerical facts and does not reproduce page images.
• Rumen Kolev, William Lilly and His Method of Primary Directions, 1998 (hosted PDF). The hosted PDF does not state a license; the fixture retains only limited numerical facts and a citation.
• Sepharial, Directional Astrology, chapter X, 1921 (hosted excerpt). The underlying United States pre-1929 work is public domain; the host URL is retained for provenance.
• Alan Leo, calculator adaptation by Jane Griscti, Primary Directions, 2010 (hosted PDF). The adaptation bears a copyright notice; only limited numerical facts and a citation are retained.
• Alexey Borealis, A Circle of Mundane Aspects, 2024 (technical article). The site does not state a license; only limited numerical facts and a citation are retained.
• Jean-Baptiste Morin, Astrologia Gallica, 1661, Book XXII, Section II, Chapters 2-3 (ETH Bibliothek scan, DOI). The scan carries the Public Domain Mark. It is currently provenance for a deferred primary-source comparison, not an executable numerical oracle.

No third-party scan or long source passage is bundled in these fixtures.

7. Executed evidence receipt

With downloads disabled and strict known-issue expiry enabled, the focused authority slice was executed through the project Python 3.14 environment:

$env:MOIRA_TEST_MODE='1'
$env:MOIRA_STRICT_KNOWN_ISSUES='1'
$env:MOIRA_NO_DOWNLOAD='1'
.\.venv\Scripts\python.exe -m pytest `
  tests\unit\test_primary_direction_campanus_topocentric_authority.py `
  tests\unit\test_primary_direction_special_target_authority.py -q

Result: 13 passed. This receipt exercises the exact fixture products and internal visibility covenants described above. It does not run a live network oracle, reconstruct every historical chart, validate the additive REST vessels, or turn the listed deferrals into supported branches.

8. Supportable statement

Moira now has externally anchored, product-specific numerical evidence for one shared Campanus-Regiomontanus mundane conjunction law, two Topocentric under-the-pole laws, and the source-labeled Topocentric converse product through its narrow signed-primary-motion preset. The traditional role-exchange preset remains a distinct compatible product. Moira also has evidence for one historical zero-latitude Vega direction, one planetary antiscion, one Ptolemaic parallel, direct and converse Placidian rapt-parallel examples, and one Morinus sextile projection. The evidence classes and model differences remain explicit. It does not establish full method-family parity, primary-source coverage for every special target, or a general neo-converse law.