# Technical companion: operations, constraints and recoverability

This reference consolidates the proved models through C1070. Every formula acts on a declared object: ordered source rows, resolved counts, digit registers, labelled coordinates or section occurrences. An algebraic inverse recovers only the information retained by its map. Rank counts concern the stated variable model; they are not counts of historically independent observations. E=25/23, P=70/69 and J=300/299 denote the Priestly, Prophetic and Enochian Keys. Rounded conventions follow latest File52c; no second decimal inversion is used.

## 1. Ordered regular and cumulative coordinates

For n ordered edges, define the upper-triangular suffix matrix

\[
U_{ij}=\mathbf1_{j\ge i}.
\]

With a fixed terminal coordinate a, an earlier-to-later BC path is a·1+Uw. Regular construction supplies begetting edges b; cumulative construction supplies source-selected lifespans L. Thus the shared topology gives a·1+Ub or a·1+UL, not interchangeable weights. Since U has determinant 1, adjacent differences recover its ordered inputs once the terminal is retained. A single outer total instead loses internal allocation. Mode-specific row values and count conventions remain premises.

For the SP cap model, births satisfy B=N·1+Ub at Noah placement N. Therefore dB=U db only when N is held; generally add 1 dN. Within a fixed cap-support region, the lifespan and cumulative responses are

\[
dL'=A U\,db,\qquad dC=U A U\,db,
\]

where A selects clipped rows, baseline lives and the Noah/Flood interval are fixed, and cumulative placement is held. These derivatives do not cross support thresholds. Sources: File09 §0.4 state controls and cumulative convention; File18 §§3.1,3.1.3; C954,C972–C975,C983.

## 2. The complete SP cap and its conditional inverse

Retain completed begettings b=(130,105,90,70,65,62,65,67,52), Noah 3493, Flood-start distance 600 and inclusive adjustment 1. Lamech’s 52 completed years correspond to his 53rd counted-year station. Put k=601 and capacities c=Ub+k·1. The rule is

\[
L'_i=\min(L_i,c_i).
\]

The complete nine-row field is:

| Row | Baseline L | Capacity c | Output L′ |
|---|---:|---:|---:|
| Adam |930|1307|930|
| Seth |912|1177|912|
| Enosh |905|1072|905|
| Kenan |910|982|910|
| Mahalalel |895|912|895|
| Jared |962|847|847|
| Enoch |365|785|365|
| Methuselah |969|720|720|
| Lamech |777|653|653|

The minimum is the unique coordinatewise greatest vector satisfying both upper bounds. The reductions 115,249,124 sum 488; this is a conditional mechanism, not a historical direction claim. Ordinary and inclusive death tags must remain distinct. The three clipped rows terminate at Flood start 2893.

With capacities fixed, an output y<c recovers L=y, whereas y=c permits every L≥c. Original excess is lost. Holding the begetting chain, the clipped support alone permits 584≤k<716; the full total 7137 gives 5334+3k and hence 601. Holding k, the three late cap equations recover bL=52, bM=67 and bJ+bE=127. Their rank-three matrix has kernel (1,−1,0,0). Retaining Enoch 65 then recovers Jared 62; the first five begettings remain unconstrained by these equations.

Nearest-five count rounding commutes with the minimum because it is monotone. Rounding date labels first need not reproduce rounded counts. Sources: File18 §§3.1,3.1.3; File51a §§1–2 rounding policy; C953–C970,C978–C983.

## 3. One-pass decimal registers and the convergence branch

For a positive original duration x=10ᶻm with m not ending in zero, the declared one-pass operation reverses m and restores its z trailing zeros. It applies separately to source components. It is not generally additive and does not act on a substituted whole-span total.

In the two-component register where each original duration has exactly one trailing zero and a three-digit core, write core digits(h,t,u), with h,u∈{1,…,9} and t∈{0,…,9}. Let H,T,V be the two column sums. Then

\[
S=10(100H+10T+V),\qquad I=10(H+10T+100V).
\]

The bounds are H,V∈[2,18], T∈[0,18]. S=2700 forces (H,T,V)=(2,6,10), hence I=10620. S=12600 permits exactly (11,15,10) and(12,5,10), giving 11610 and10620. The source 917+343 selects the latter: after the units carry, no tens-to-hundreds carry occurs.

The two measurement rows have rank two and primitive integer kernel (10,−101,10). The bounded digit box excludes nonzero kernel moves, so both totals identify aggregate columns. They do not recover component allocation: 1650|1050 and the diagnostic 1550|1150 share both totals. No alternative chronology is thereby admitted.

At the selected junction, R=2700,C=12600 and I=10620 satisfy5I=4C+R. The4/5 weight also preserves the selected shifts ΔC=−2, ΔR=8; it is not another independent observation. Placement gives1406+10620=12026. Continuation specifically reverses the original tail1400→4100, adding2700 and yielding14726. Its cause is the tail rule and retained Nativity 6, not numerical substitution of the regular2700 trunk. Sources: File52c §§3.3–3.4,3.14; C932–C952, especially C952’s provenance correction.

## 4. Complete decimal-path interfaces with Keys

File52c’s four original paths are 1400|1050|1650,1400|1650|1050,1400|1050|600|1050 and1400|3430|9170. After one pass, all total 14720. Their prefix residues modulo23 are respectively (6,2,0),(6,4,0),(6,2,4,0),(6,9,0). Thus E gives the integral outer duration 16000, while every proper source prefix remains fractional.

With an integral held anchor, an integer boundary requires its prefix sum to be divisible by23. Only the final numerical cut qualifies here; source events remain recorded. Within the fixed two-branch register and retained 1400 tail, convergence and integral E completion select the same branch. They are equivalent checks in that domain.

The required E-selection amounts for reproducing P or J totals are 7360/3 and7360/13. Neither can be a sum of whole integer components in these paths. Algebraic allocation identities therefore do not supply source cuts. Uniform Key measurements also add only multiples of existing measurement rows, leaving rank two. Sources: File52c §3.4; C1058–C1068.

## 5. Calibration, allocation and their dependencies

The exact common-volume calibration is

\[
336E=360P=364J=K=8400/23.
\]

Calibration alone leaves K free. Retaining the inherited completion relations E−6P=−5 and E−26J=−25 gives the system with rows(336,−360,0),(336,0,−364),(1,−6,0),(1,0,−26), right side(0,0,−5,−25). Coefficient and augmented ranks are three; the left dependency is(7,−6,−420,84), including constants. Backward solution is an equivalent description of inherited identities, not independent historical derivation.

Selection of fraction f under factor k gives A_f(k)=1+f(k−1), so A_g(A_f(k))=A_gf(k). Here E→P uses1/6, P→J uses computed3/13, and E→J uses1/26. Successive whole conversions instead give PJ=7000/6877. Only source-appointed subdivisions authorize selective chronological execution.

On299|161, every predetermined Key assignment preserves component volumes 109200|58800 when each output carries its matching unit. The assembled 300|175 comparison comes from inherited C533. Its sum 475 is not asserted to be a literal mixed-calendar elapsed duration. Sources: Strategy §3.3; File12 active-Key clarification; File63 §1.3,§§7.3–7.4,§9.7; File60 §5; C984–C992,C1004–C1007,C1070.

## 6. Retained parts and coarsening

For converted u and retained v,

\[
(T_0,T_P,T_E)=(u+v,Pu+v,Eu+v),
\]

with rank two, relation5T₀−6Tₚ+Tₑ=0, and inverse u=69(Tₚ−T₀), v=T₀−u. This recovers all four declared pairs (690,0),(690,30),(690,60),(9660,2940). It does not recover names, order, anchor or route authority. Supplement A §10.3, rather than File52c, controls the 9660|2940 family.

For a partition-summing incidence map C, universal commutation

\[
C\operatorname{diag}(k_i)=\operatorname{diag}(k_B)C
\]

holds exactly when factors are constant within each block. Necessity follows by testing each coordinate vector; sufficiency follows by factoring the common multiplier. Equality on one weighted input is weaker. The inherited 483 partial/uniform comparison has component defect (−35/6,+35/6), invisible in total 490. Sources: Supplement A §§1.2,10.3; File63 §1.3; C993–C997.

## 7. Affine pivots and complete integer domains

Let D_{k,a}(x)=a+k(x−a). For reduced ratios k₁=p₁/q₁,k₂=p₂/q₂ and integer x,a,b, both stages are integral exactly when

\[
x=a+q_1t,\qquad p_1t\equiv b-a\pmod{q_2}.
\]

The domain is empty unless g=gcd(p₁,q₂) divides b−a; otherwise it is one coset of period q₁q₂/g. A whole field belongs precisely when one point has the required residue and all differences are period multiples. Translation of points and both pivots shifts the residue equally.

At a=14006,b=4836 the eight nonempty diagnostic domains are E→E:68+529ℤ; E→P:1126+1587ℤ; E→J:68+6877ℤ; P→E:206+1587ℤ; P→P:206+4761ℤ; P→J:4967+20631ℤ; J→E:2345+6877ℤ; J→J:43607+89401ℤ. J→P is empty because3∤−9170. These are not newly authorized routes.

The complete File46 head field has indices (92,89,46,43,0), anchor 14006 and mesh 10. E preserves its indices on rational mesh 250/23. Integer-grid failure does not invalidate that exact rational comparison. Likewise 529n→575n→625n describes inherited duration widths without supplying compatible pivots. Sources: File46 §§2–3,6A.5; C998–C1003; inherited C505–C507.

A difference-only obstruction can be checked before considering placement. For a reduced Key denominator q, two integral images about one integer pivot require q to divide the original difference. The source paired gaps 990 and 1980 fail this necessary condition for all three denominators 23,69,299. No common integer pivot can repair that failure. Conversely, a divisible difference supplies no particular pivot residue or source authorization. The complete paired source field has difference mesh 5, compared with 115 for the inherited nine-member field; E sends these to 125/23 and125. This compares an exact rational field with an existing integer subfield without enlarging either chronology. Sources: File52c paired endpoint tables and §3.7.3; C1064–C1066.

## 8. Covenant constraints and forced clutch joins

The labelled twelve-node regular incidence map has determinant 1. In the declared thirteen-variable claim model, nine selected relations have rank eight. Retaining two magnitudes and one placement raises rank to eleven; source rows qI=1 and Isaac→Jacob60 restore rank thirteen. Rank depends on which biographies are held.

For Covenant Q, u=Q−Levi-death and w=Q−Joseph-death, the landings Q−Ju and Q−Ew agree iff13w=12u. Source-generated 299 and276 give 1566. Introducing a common-point variable adds a definition plus one agreement condition, not two independent source constraints.

For the clutch, let B,D be full-state regular Levi birth/death, M Moses birth, X Exodus, and ℓ,k,a the Levi, Kohath and Amram lifespans. Then

\[
d=[\ell+k+a+(M-X)]-(B-X)=M+k+a-D=C_K-D.
\]

The source base (2435,2260,2080,1933,1796,1663,1526) gives d=14 and h=7. Hence C_K−d=D is definitional. Shared Levi life gives C_L−d=B automatically. The adjacent Jacob/Isaac join reuses the source 147 condition, rather than following from the clutch alone.

The 42-coordinate array X(i,j,r)=b_i+jp−rh, with seven rows, j∈{0,1},r∈{0,1,2}, has unrestricted rank nine. Constraints h=2p and the repeated 137 life each remove one freedom. Retain exact p=3.5 and one Nisan-derived clutch across phases; recomputing a separate full clutch would cancel that phase at the full-shift landing. Whole-year phase 3 is separate. The +215 all-coordinate transformation is formal covariance, not the admitted state change retaining terminal 1406. Sources: File60 §§2–6; File61 §§6–8; File62 §§1–2; C1032–C1057,C1070.

The source register also retains event status: the author-designated Covenant 2081 and proposed offering 2051/Isaac age 15 are different premises with different claim status. They are not interchangeable numerical normalizations. C1057 records the coupled model as 55 outputs of rank 19 in its declared local inputs. Neither this compression nor the clutch moves Moses’ historical birth: once the clutch is recomputed from its defining relation, the original cumulative anchor cancels from translated coordinates, whose placement is controlled by the regular Levi death and shared lifespans. Formal null directions illustrate information left unresolved by selected pattern equations; they do not constitute admitted alternative biographies.

These conditions distinguish a forced join from a source compatibility check. The first join follows from the definition of d; the second follows from the shared lifespan; the next requires the separately retained 147 relation. Counting all three as independent coincidences would discard their dependency structure.

## 9. Count registers, metric slots and quotient symmetry

**Esau.** Preserve the nine labelled counts and declared register orders. Six clean prefixes plus five female prefixes have rank eight; only male donkeys remain unmeasured. Adding total 550 gives rank nine. Six clean prefixes, female F3/F5 and the total form determinant +1 basis; clean plus unclean prefixes form determinant −1 basis. These are integral coordinate changes, not reduction of nine numerical premises. The fourteen-prefix field has five structural dependencies; source-specific 20/50 agreements are additional value equalities. Source: File58 §§13.1–13.4; C1009–C1013.

**Tishri.** Seven retained ritual templates include Sukkot’s 70:14:98:7. Species margins (75,18,176,11) have rank four on seven amplitudes. Two festival-equality conditions and daily 44 raise rank to seven, recovering (44,6,1,1,1,1,1). Removing daily 44 returns one freedom; removing both equalities returns two. Total 189, seven days and decrement 1 force totals 30,…,24; fixed nonbull 17 gives bulls 13,…,7. Templates and decrement remain source inputs. Source: File58 §§2.2–3.2,7.1; C1014–C1016.

**NT metric.** Equations (P−1)w=u and(E−1)w=6u have rank one, kernel (69,1). Adding 35u=2450 fixes u=70,w=4830. Alternatively A+55u=3856 and A+20u=1406 have determinant −35 and recover A=6,u=70 within the already fitted display. File54 §6 controls the fork; §§8.6,13.2 control 2450/35. Also File43 §§2.1–2.3,3.4; C1019–C1022,C1070.

**Toledot and reflection.** A quotient q supports induced reflection R exactly when q(x)=q(y) implies q(Rx)=q(Ry). Identifying only occurrences 9,10 fails: R(i)=12−i sends them to distinct 3,2. Retained multiplicity restores eleven occurrences and Terah’s 5|1|5 measure, but not descended reflection. By contrast, reflection permutes the eleven seven-edge blocks of the primary NT 77-edge path. The AD 65–135 extension is excluded. Sources: File70 Appendix B.1; File43/File54 primary path; C1023–C1029.