# BCODR-LUC-04 — True Overlap Germ Tangent Decomposition
## Root 127 最早共同重疊點的 Normal–Tangent 分離與二階孤立性
### From a Three-Dimensional Critical Tangent Space to a Zero-Dimensional Birth Germ

**文件代號：** `BCODR-LUC-04-v0.1`  
**版本：** v0.1  
**日期：** 2026-09-21  
**作者：** Neo.K  
**AI 協力：** Aletheia / ChatGPT, GPT-5.6 Sol  
**機構：** EveMissLab／一言諾科技有限公司  
**Target：** `ROOT6217_BARRIER_SHAPE`  
**前置：** BCODR-LUC-03；LESR Paper 09B  
**狀態：** `TRUE-OVERLAP GERM / SECOND-ORDER TANGENT CLOSURE`

---

# 0. Recenter

BCODR-LUC-03 proved that root 127 contains a true balanced common-point overlap germ.

Its earliest synchronized radius is:

$$
\boxed{
T^\star
=
0.126953667790942.
}
$$

The true inward provenance is prune:

$$
\boxed{
155.
}
$$

The next question is no longer:

> Which coarse box should be split?

It is:

> At the first true overlap point, does any nonzero tangent direction remain inside the overlap at the same radius?

This is a local tangent-geometry problem.

---

# 1. Normalized Parameter Coordinates

Use the normalized root-6217 coordinates:

$$
\boxed{
z
=
(\widetilde x_3,
\widetilde y_3,
\widetilde\rho_5,
\widetilde x_5,
\widetilde y_5).
}
$$

The true meeting point is:

$$
\boxed{
z^\star
=
[-0.125, -0.490449386049, -0.429185812844, 0.104804263668, 0.395569590601].
}
$$

---

# 2. Two Distance Fields

Define:

$$
F_+(z)
=
\frac12
d(z,\mathcal S_+)^2
$$

for the analytic critical-ray field.

Define:

$$
F_-(z)
=
\frac12
d(z,Q_{155})^2
$$

for the prune-155 expanded box field.

At the meeting point:

$$
F_+(z^\star)
=
F_-(z^\star)
=
\frac12(T^\star)^2.
$$

---

# 3. Active Root Face

The meeting point lies on the root-127 lower $x_3$ face:

$$
\boxed{
\widetilde x_3=-0.125.
}
$$

Thus the root-domain constraint contributes one active face normal.

---

# 4. Active Distance Gradients

Let:

$$
p^\star
$$

be the nearest analytic-ray point and:

$$
q^\star
$$

the nearest point in prune 155.

Then:

$$
g_+
=
z^\star-p^\star,
$$

$$
g_-
=
z^\star-q^\star.
$$

Numerically:

$$
\boxed{
g_+
=
[0.10559856031, 0, -0.0264085908998, 0, -0.0653357800345],
}
$$

$$
\boxed{
g_-
=
[0.125, 0, 0.00831418715568, 0, 0.0205695906014].
}
$$

Both have support only in:

$$
(\widetilde x_3,\widetilde\rho_5,\widetilde y_5).
$$

---

# 5. Balanced KKT Relation

There exists:

$$
\alpha^\star
=
0.239444757052325
$$

such that:

$$
\boxed{
\alpha^\star g_+
+
(1-\alpha^\star)g_-
-
\eta^\star e_{x_3}
=
0,
}
$$

with:

$$
\boxed{
\eta^\star
=
0.120354426986961
>0.
}
$$

This is the local minimax stationarity relation.

---

# 6. Active Normal Rank

The three rows:

$$
g_+,
\qquad
g_-,
\qquad
e_{x_3}
$$

have rank:

$$
\boxed{
2.
}
$$

The reason is structural:

after removing the active $x_3$ component,

the outward and inward distance gradients are oppositely collinear in the:

$$
(\rho_5,y_5)
$$

subspace.

So the balanced overlap has one fewer independent first-order normal than a generic transverse intersection.

---

# 7. Critical Tangent Space

The first-order critical tangent space therefore has dimension:

$$
\boxed{
5-2=3.
}
$$

It is:

$$
\boxed{
T_{\rm crit}
=
\left\{
\delta z:
\delta x_3=0,
\quad
\delta y_5
=
-0.404197989002394\,
\delta\rho_5
\right\}.
}
$$

---

# 8. Natural Tangent Basis

A convenient orthonormal basis is:

$$
\boxed{
e_{y_3},
}
$$

$$
\boxed{
e_{x_5},
}
$$

and the normalized coupled direction:

$$
\boxed{
e_{\rho y}
\propto
e_{\rho_5}
+
-0.404197989002394
e_{y_5}.
}
$$

So first-order geometry still allows three apparent sliding modes:

1. $B_3$ vertical motion;
2. $B_5$ horizontal motion;
3. coupled $\rho_5/y_5$ motion.

---

# 9. Why First-Order Tangency Is Not Enough

A first-order tangent direction can satisfy all active linear equations and still leave the overlap at second order.

This is the same phenomenon as two tangent circles:

their tangent lines coincide at first contact,

but the exact intersection at the first radius may still be a single point.

Therefore BCODR must inspect the second-order distance geometry.

---

# 10. Outward Reduced Hessian

Let the normalized analytic ray be:

$$
p(r).
$$

Before eliminating $r$:

$$
\widehat F_+(z,r)
=
\frac12
\|z-p(r)\|^2.
$$

At the nearest ray parameter:

$$
r^\star,
$$

the reduced Hessian is:

$$
\boxed{
H_+
=
I
-
\frac{
p'(r^\star)p'(r^\star)^T
}{
D_+
},
}
$$

where:

$$
\boxed{
D_+
=
\|p'\|^2
-
g_+\cdot p''
=
1706.98000390147.
}
$$

This is the Schur complement after minimizing over the ray parameter.

---

# 11. Inward Reduced Hessian

At the root-127/prune-155 meeting point:

- $y_3$ is inside the prune interval;
- $x_5$ is inside the prune interval;
- the active outside coordinates are:
  $$
  x_3,\rho_5,y_5.
  $$

Therefore:

$$
\boxed{
H_-
=
\operatorname{diag}(1,0,1,0,1).
}
$$

---

# 12. Restriction to the Critical Tangent Space

Let:

$$
B_T
$$

be the $5\times3$ tangent-basis matrix.

Define:

$$
\boxed{
Q_+
=
B_T^T H_+ B_T,
}
$$

$$
\boxed{
Q_-
=
B_T^T H_- B_T.
}
$$

Numerically:

$$
\boxed{
Q_+
=
[[0.9917527159, 0, -0.09049364957]; [0, 1, 0]; [-0.09049364957, 0, 0.007054860947]],
}
$$

$$
\boxed{
Q_-
=
[[0, 0, 0]; [0, 0, 0]; [0, 0, 1]].
}
$$

---

# 13. Individual Curvature Is Not the Final Criterion

$Q_+$ alone can contain a very soft or even locally negative curvature direction because the analytic seed is a curved one-dimensional set.

$Q_-$ has zero directions because prune 155 is locally flat in $y_3$ and $x_5$.

Therefore neither quadratic form alone determines overlap isolation.

The correct object is the combined minimax stiffness.

---

# 14. Combined Germ Stiffness

Define:

$$
\boxed{
\kappa_{\rm germ}
=
\min_{
v\in T_{\rm crit},
\ \|v\|=1
}
\max
\{
v^T H_+v,
v^T H_-v
\}.
}
$$

Numerically:

$$
\boxed{
\kappa_{\rm germ}
=
0.45429481027797.
}
$$

Thus:

$$
\boxed{
\kappa_{\rm germ}>0.
}
$$

---

# 15. Softest Tangent Direction

The minimizing unit tangent direction in normalized coordinates is approximately:

$$
\boxed{
v_{\rm soft}
=
[0, -0.738718613358, -0.624897568788, 1.73650131834e-07, 0.252582340636].
}
$$

At this direction:

$$
\boxed{
v_{\rm soft}^T H_+v_{\rm soft}
\approx
0.45429481027797,
}
$$

$$
\boxed{
v_{\rm soft}^T H_-v_{\rm soft}
\approx
0.45429481027797.
}
$$

The two costs are almost balanced.

---

# 16. Second-Order Isolation Theorem

Assume the active-set identities remain valid in a neighborhood of:

$$
z^\star.
$$

For a critical tangent displacement:

$$
z=z^\star+h v,
\qquad
v\in T_{\rm crit},
\qquad
\|v\|=1,
$$

the two squared-distance functions satisfy:

$$
d_+^2
=
(T^\star)^2
+
h^2
v^TH_+v
+
o(h^2),
$$

$$
d_-^2
=
(T^\star)^2
+
h^2
v^TH_-v
+
o(h^2).
$$

Therefore:

$$
\max(d_+^2,d_-^2)
\ge
(T^\star)^2
+
\kappa_{\rm germ}h^2
+
o(h^2).
$$

Since:

$$
\kappa_{\rm germ}>0,
$$

every nonzero critical tangent direction increases the required overlap radius to second order.

Thus:

$$
\boxed{
\text{the earliest true-overlap birth germ is locally isolated to second order.}
}
$$

---

# 17. Linear Tangent Dimension vs Birth Dimension

This gives a useful distinction.

## Linearized critical tangent dimension

$$
\boxed{
3.
}
$$

## Exact birth-germ dimension at $T^\star$

within the local active-set model:

$$
\boxed{
0.
}
$$

So:

$$
\boxed{
\text{first-order freedom}
\not\Rightarrow
\text{finite overlap manifold at birth}.
}
$$

---

# 18. How the Overlap Grows After Birth

Let the common kernel radius increase to:

$$
T^\star+\varepsilon.
$$

Then:

$$
(T^\star+\varepsilon)^2
=
(T^\star)^2
+
2T^\star\varepsilon
+
O(\varepsilon^2).
$$

The softest tangent displacement therefore scales approximately as:

$$
\boxed{
|h|
\lesssim
\sqrt{
\frac{
2T^\star
}{
\kappa_{\rm germ}
}
}
\sqrt{\varepsilon}.
}
$$

Numerically:

$$
\boxed{
|h|
\lesssim
0.747599042164318
\sqrt{\varepsilon}
}
$$

to leading order.

This is the higher-dimensional analogue of a lens growing out of a first circle-circle contact point.

---

# 19. Radius Cost per Tangent Displacement

Equivalently:

$$
\boxed{
\varepsilon
\gtrsim
\frac{
\kappa_{\rm germ}
}{
2T^\star
}
h^2.
}
$$

The coefficient is:

$$
\boxed{
1.78921498757354.
}
$$

Thus any finite tangent excursion has a strictly positive second-order kernel-radius cost.

---

# 20. Consequence for BCODR Cut Strategy

Root 127 does not need indefinite recursive box cutting to discover a hidden thick corridor.

At the true overlap birth:

$$
\boxed{
\text{there is no nonzero second-order-free tangent direction.}
}
$$

So further cutting should focus on:

1. verifying the active-set interval conditions;
2. tracking the post-birth $\sqrt{\varepsilon}$ lens growth;
3. checking topology changes at larger radius.

---

# 21. Consequence for LESR Calculus

Paper 09C originally expected to receive surviving tangent directions from BCODR.

At root 127:

$$
\boxed{
\text{BCODR leaves no zero-cost tangent direction at second order.}
}
$$

Therefore root 127 does not currently require a separate shape-area $\Phi_2$ calculation merely to decide whether the overlap germ has a flat direction.

BCODR has already closed that geometric tangent responsibility.

---

# 22. What Remains for Formal Certification

The local result still needs:

## CERT-04A

Interval enclosure of:

$$
z^\star,
r^\star,
p^\star,q^\star.
$$

## CERT-04B

Interval proof of the active-set pattern:

- root $x_3$ lower face active;
- prune 155 active in $x_3,\rho_5,y_5$;
- $y_3,x_5$ remain interior.

## CERT-04C

Interval enclosure of:

$$
\kappa_{\rm germ}
$$

with a strict lower bound, e.g.:

$$
\boxed{
\kappa_{\rm germ}>0.4.
}
$$

That is a very generous target relative to the observed:

$$
0.454\ldots
$$

---

# 23. True Circle-Overlap Interpretation

The original BCODR idea said:

> let circular expansions meet, overlap, then cut and separate the overlap.

Root 127 now realizes the exact high-dimensional analogue.

At:

$$
T<T^\star
$$

there is no common overlap point.

At:

$$
T=T^\star
$$

the two provenance fields first meet at one isolated point.

At:

$$
T>T^\star
$$

a finite overlap lens grows with tangent scale:

$$
O(\sqrt{T-T^\star}).
$$

So the circle intuition survives the lift to five-dimensional parameter space.

---

# 24. Information Meaning

The first-contact point carries maximum structural information because it simultaneously identifies:

1. the analytic candidate provenance;
2. prune-155 certificate provenance;
3. the active root face;
4. the critical tangent directions;
5. the second-order growth law.

This is a more precise information object than a generic high-overlap cell.

---

# 25. Revised Surviving Spine

For root 127 the BCODR spine now contracts to:

$$
\boxed{
\text{analytic ray}
\rightarrow
z^\star
\leftarrow
\text{prune 155}
}
$$

with:

- one root-face normal condition;
- a three-dimensional linear critical tangent space;
- zero second-order-free tangent dimensions.

This is the first truly local geometric spine extracted by BCODR.

---

# 26. Next Global Responsibility

The selected frontier still contains root 131 as another balanced true-overlap germ.

But:

$$
G_{\rm sep}(131)
$$

is much larger.

So 131 is structurally different from 127.

Before returning to global calculus, the next BCODR task should compare:

$$
\boxed{
\text{isolated near-true germ }127
}
$$

against:

$$
\boxed{
\text{coarse-separated balanced germ }131.
}
$$

This determines whether 131 also collapses to an isolated local contact after its own active-set reduction or instead represents a different topology class.

---

# 27. Conclusion

BCODR-LUC-04 completes the normal–tangent decomposition of root 127.

The first-order active geometry leaves:

$$
\boxed{
3
}
$$

critical tangent dimensions.

But the combined second-order overlap stiffness is:

$$
\boxed{
\kappa_{\rm germ}
\approx
0.454295
>0.
}
$$

Therefore the earliest true common-point overlap is locally isolated to second order.

The overlap only acquires finite tangent extent when the kernel radius grows beyond:

$$
T^\star.
$$

Thus BCODR has reduced:

$$
\text{5D root box}
\rightarrow
\text{1D meeting minimization}
\rightarrow
\text{3D critical tangent space}
\rightarrow
\text{0D birth germ}.
$$

This is exactly the intended cut–separate–recompose–close sequence.

---

# Canonical Source Declaration

本檔為 `BCODR-LUC-04-v0.1` 的 UTF-8 Markdown canonical source。

數學 delimiter 僅使用 `$...$` 與 `$$...$$`。

本篇區分：

- active normal rank；
- linearized critical tangent dimension；
- combined second-order stiffness；
- exact overlap birth dimension；
- post-birth overlap growth。

目前 numerical differential geometry 已獨立 replay；interval certification remains open.

===END===