> Full-text translation from Chinese. The original remains authoritative for its research claims; this translation is not independent mathematical verification.

# After the First Solution: Arithmetic Freedom of the Perfect Cuboid

## After the First Solution: Arithmetic Freedom of the Perfect Cuboid

**Series: Global Closure Theory of the Perfect Cuboid**  
**Paper: 07 / 08**  
**Status: Research introductory draft / Phase 0 — Conceptual Closure**  
**Date: 2026-09-28**

---

## Abstract

The preceding six papers have formalized the reformulation of the Perfect Cuboid, the continuous shape space, the common initial state and constraint-induced differentiation, Radix–Carry Global Closure, the transition from floating-point values to exact rational closure, and the first primitive manifestation. This paper enters the second stage: assuming that at least one canonical primitive Perfect Cuboid already exists, the actual question to study next is not its integer multiples, but whether nontrivial arithmetic freedom exists among different primitive shapes.

If:

$$
P=(a,b,c,d,e,f,g)
$$

is a primitive solution, then:

$$
kP
$$

for every:

$$
k\in\mathbb Z_{>0}
$$

is still an integer solution. This infinitude comes only from the action of scaling, and this paper calls it:

$$
\boxed{
\text{scaling infinity}
}
$$

What actually matters is:

$$
\boxed{
\text{shape infinity}
}
$$

that is, whether the canonical primitive solution space:

$$
\mathcal P_{\mathrm{can}}
$$

contains:

- A single point;
- Finitely many points;
- Infinitely many discrete points;
- A one-parameter family of rational solutions;
- A two-parameter family of rational solutions;
- Or other, more complex arithmetic structures.

This paper defines multiple layers of arithmetic freedom, including set-theoretic freedom, parametric freedom, transformational freedom, growth freedom and local deformation freedom. Geometrically, after removing scale, the positive-real shape world of the Perfect Cuboid has two continuous degrees of freedom, so ‘2’ can be viewed as a natural geometric upper bound; arithmetic freedom, however, does not automatically equal geometric dimension. Rational points may be extremely sparse, and even the existence of a first primitive shape does not guarantee a second.

This paper therefore establishes three core questions:

$$
\boxed{
\text{Multiplicity}
}
$$

$$
\boxed{
\text{Generation}
}
$$

$$
\boxed{
\text{Growth}
}
$$

Multiplicity asks how many primitive shapes there are; Generation asks whether there is a non-scaling solution-preserving transformation:

$$
\mathcal T(P)\not\sim P
$$

for which, nevertheless:

$$
F(P)=0
\Rightarrow
F(\mathcal T(P))=0;
$$

Growth studies:

$$
N(H)
=
\#\{P\in\mathcal P_{\mathrm{can}}:g(P)\le H\}
$$

and how it evolves as:

$$
H\rightarrow\infty
$$

tends to infinity.

This paper also proposes quantities such as the seed-expansion graph, orbit decomposition, family rank, arithmetic branching factor and candidate growth dimension, as a second-stage research framework that can be initiated immediately if the first solution is discovered.

---

## 1. The First Solution Is Not the End: The Research Object Changes

Suppose we have obtained an exact canonical primitive solution:

$$
P_1
=
(a_1,b_1,c_1,d_1,e_1,f_1,g_1).
$$

It satisfies:

$$
F(P_1)=0,
$$

$$
\gcd(a_1,b_1,c_1)=1,
$$

and:

$$
a_1<b_1<c_1.
$$

The Perfect Cuboid existence question is then settled:

$$
\boxed{
\exists P\in\mathcal P_{\mathrm{can}}.
}
$$

The next question should no longer be merely:

> Can we find another integer multiple?

This is because:

$$
2P_1,3P_1,4P_1,\ldots
$$

provide no new shape.

The actual question becomes:

> Does there exist:

$$
P_2\in\mathcal P_{\mathrm{can}}
$$

such that:

$$
P_2\neq P_1?
$$

---

## 2. Scaling Infinity and Shape Infinity

If:

$$
P_1,
$$

exists, then:

$$
\{kP_1:k\in\mathbb Z_{>0}\}
$$

is already an infinite set.

Thus the raw integer solution set:

$$
\mathcal I
=
\{X\in\mathbb Z_{>0}^{7}:F(X)=0\}
$$

once nonempty,

must be infinite.

Therefore:

$$
|\mathcal I|=\infty
$$

says almost nothing about primitive diversity.

This paper strictly distinguishes the two kinds of infinitude.

### 2.1 Scaling infinity

Infinitely many integer multiples of the same primitive shape:

$$
P,
2P,
3P,
\ldots
$$

### 2.2 Shape infinity

There exist infinitely many pairwise inequivalent primitive shapes:

$$
P_1,P_2,P_3,\ldots
$$

where:

$$
P_i\not\sim P_j
$$

for:

$$
i\neq j.
$$

Therefore:

$$
\boxed{
\text{scaling infinity}
\neq
\text{shape infinity}.
}
$$

---

## 3. Canonical primitive solution space

Following Paper 6:

$$
\mathcal P_{\mathrm{can}}
=
\left\{
P\in\mathbb Z_{>0}^{7}
:
F(P)=0,
\;
\gcd(a,b,c)=1,
\;
a<b<c
\right\}.
$$

The actual research in the second stage is:

$$
\boxed{
\text{understand }\mathcal P_{\mathrm{can}}.
}
$$

The first question is the most basic cardinality:

$$
|\mathcal P_{\mathrm{can}}|
\in
\{0,1,2,\ldots,\infty\}.
$$

If the existence hypothesis of Paper 6 holds,

then:

$$
|\mathcal P_{\mathrm{can}}|\ge1.
$$

but:

$$
|\mathcal P_{\mathrm{can}}|=1
$$

remains entirely possible.

---

## 4. The First Layer of Arithmetic Freedom: Multiplicity

Define:

$$
M
=
|\mathcal P_{\mathrm{can}}|.
$$

We can then distinguish:

### M0

$$
M=0.
$$

No Perfect Cuboid exists.

### M1

$$
M=1.
$$

There exists a unique primitive shape.

### MF

$$
1<M<\infty.
$$

There exist finitely many primitive shapes.

### M∞

$$
M=\infty.
$$

There exist infinitely many primitive shapes.

This is only the set-theoretic layer.

It does not yet answer:

> Do infinitely many shapes exhibit a pattern?

---

## 5. The Second Layer of Arithmetic Freedom: Parametric Freedom

Suppose:

$$
\mathcal P_{\mathrm{can}}
$$

is infinite.

We must still ask whether a rational parameterization exists.

For example, does there exist:

$$
P(t),
\qquad
t\in\mathbb Q
$$

such that:

$$
P(t)\in\mathcal P_{\mathrm{can}}
$$

for infinitely many:

$$
t.
$$

If so,

we can say it has at least one dimension of parametric freedom.

More strongly,

does there exist:

$$
P(s,t),
\qquad
s,t\in\mathbb Q
$$

forming a two-parameter family?

This would approach the geometric upper bound on degrees of freedom in the continuous shape world.

---

## 6. Geometric Freedom and Arithmetic Freedom

From Paper 2, we know that:

$$
\mathcal S_{\mathbb R}
$$

has, after removing scale:

$$
2
$$

continuous degrees of freedom.

That is:

$$
\dim_{\mathbb R}\mathcal S_{\mathbb R}=2.
$$

But the Perfect Cuboid requires special rational points in:

$$
\mathcal S_{\mathbb Q}
$$

this space.

Thus:

$$
\boxed{
\text{geometric freedom}=2
}
$$

does not imply:

$$
\boxed{
\text{arithmetic freedom}=2.
}
$$

What may actually occur is:

$$
0,
1,
2
$$

or even some discrete situation that cannot be described simply by a single dimension.

---

## 7. A Preliminary Classification of Arithmetic Freedom

This paper provisionally defines:

### AF-0: Isolated arithmetic points

Primitive shapes form, at most, a discrete set of points,

with no known nontrivial rational family.

### AF-1: One-parameter arithmetic family

There exists:

$$
P(t)
$$

generating infinitely many inequivalent primitive shapes.

### AF-2: Two-parameter arithmetic family

There exists:

$$
P(s,t)
$$

generating a two-dimensional rational family.

This classification is research terminology,

not a substitute for formal dimension in algebraic geometry.

It describes:

$$
\boxed{
\text{Generatable arithmetic freedom}
}
$$

rather than merely the dimension of the underlying variety.

---

## 8. The Third Layer of Arithmetic Freedom: Transformational Freedom

Even without an explicit parameterization,

a solution-preserving transformation may exist:

$$
\mathcal T:
\mathcal P_{\mathrm{can}}
\rightarrow
\mathcal P_{\mathrm{can}}.
$$

Ideally:

$$
F(P)=0
\Rightarrow
F(\mathcal T(P))=0.
$$

If:

$$
\mathcal T(P)\not\sim P
$$

for some:

$$
P,
$$

then:

$$
\mathcal T
$$

produces a new primitive shape.

If it can also be iterated:

$$
P,
\mathcal T(P),
\mathcal T^2(P),
\ldots
$$

to obtain infinitely many distinct primitive shapes,

shape infinity follows immediately.

---

## 9. Three Levels of Solution-Preserving Transformations

### T0: Trivial scaling map

$$
\mathcal T_k(P)=kP.
$$

This does not change the shape,

so it does not count as genuine arithmetic freedom.

### T1: Permutation map

Permute:

$$
a,b,c
$$

as edge labels.

Canonicalization returns the same shape,

so this does not count as a new primitive solution either.

### T2: Nontrivial shape transformation

Require:

$$
\operatorname{Can}(\mathcal T(P))
\neq
P.
$$

Only this is a genuinely meaningful solution-preserving transformation.

---

## 10. The Seed-Expansion Problem

The first primitive solution:

$$
P_1
$$

can be viewed as an arithmetic seed.

The actual question is:

$$
\boxed{
\text{Can }P_1\text{ generate }P_2?
}
$$

If there exist:

$$
\mathcal T_1,\ldots,\mathcal T_k,
$$

we can form:

$$
P_1
\rightarrow
\mathcal T_i(P_1)
\rightarrow
\mathcal T_j\mathcal T_i(P_1)
\rightarrow\cdots.
$$

This forms a:

$$
\boxed{
\text{seed-expansion graph}.
}
$$

---

## 11. Solution graph

Define the graph:

$$
G_{\mathrm{sol}}
=
(V,E).
$$

where:

$$
V=\mathcal P_{\mathrm{can}}.
$$

If there exists an allowed transformation:

$$
\mathcal T
$$

such that:

$$
Q=\operatorname{Can}(\mathcal T(P)),
$$

add the edge:

$$
P\rightarrow Q.
$$

If the transformation is invertible,

it can be viewed as an undirected edge.

Using this graph, we can study:

- Connected components;
- Orbits;
- Branching;
- Cycles;
- Isolated vertices;
- Growth rate.

---

## 12. Orbit decomposition

Given a transformation semigroup:

$$
\Gamma=
\langle
\mathcal T_1,\ldots,\mathcal T_k
\rangle,
$$

for:

$$
P
$$

define the orbit:

$$
\operatorname{Orb}_\Gamma(P)
=
\{\gamma(P):\gamma\in\Gamma\}.
$$

We may have:

$$
|\operatorname{Orb}_\Gamma(P)|=1,
$$

or it may be finite,

or infinite.

If:

$$
|\operatorname{Orb}_\Gamma(P)|=\infty,
$$

we immediately obtain infinitely many primitive shapes.

---

## 13. Arithmetic component

If:

$$
G_{\mathrm{sol}}
$$

is not connected,

there may be multiple arithmetic families:

$$
\mathcal C_1,
\mathcal C_2,\ldots.
$$

The first solution:

$$
P_1
$$

may lie in only one of these components.

Thus:

$$
\boxed{
\text{one seed}
}
$$

does not guarantee that all solutions can be generated.

This means that even finding an elegant recurrence

does not allow us to declare immediately that ‘all Perfect Cuboids are generated by it’.

---

## 14. Family completeness

If we propose a parameterization:

$$
P(t),
$$

we need to distinguish:

### Family-valid

Every admissible:

$$
t
$$

gives a solution.

### Family-infinite

It gives infinitely many inequivalent primitive solutions.

### Family-complete

Every canonical primitive solution can be generated by some:

$$
t
$$

parameter.

The third condition is the strongest,

and also the hardest.

Therefore:

$$
\boxed{
\text{parametric family}
\neq
\text{complete classification}.
}
$$

---

## 15. Local deformation freedom

Suppose:

$$
P
$$

corresponds to a normalized rational shape:

$$
Q\in\mathcal S_{\mathbb Q}.
$$

In the real shape space,

there is always a local two-dimensional deformation:

$$
Q+\delta.
$$

But the vast majority of:

$$
\delta
$$

will not preserve the rational-square conditions.

We can therefore define the question:

> Near a rational solution, does there exist a nontrivial rational path that stays entirely within the solution set?

If there exists:

$$
Q(t)
$$

with:

$$
Q(0)=Q,
$$

and for infinitely many:

$$
t\in\mathbb Q
$$

we have:

$$
Q(t)\in\mathcal S_{\mathbb Q},
$$

then this means:

$$
Q
$$

is not an arithmetically isolated point.

---

## 16. A Research Definition of the Arithmetic Tangent

If a rationally parameterized path exists:

$$
Q(t),
$$

we can consider:

$$
V=
\left.
\frac{dQ}{dt}
\right|_{t=0}.
$$

Then:

$$
V
$$

is an underlying real tangent direction.

But the existence of:

$$
V
$$

is insufficient to prove that rational solutions exist nearby.

Thus this paper uses:

$$
\boxed{
\text{arithmetic tangent}
}
$$

as a research concept:

> A direction that is admissible not only under the linearized real constraints, but can also be realized by an actual rational solution sequence.

This requires verification using:

$$
Q_n\in\mathcal S_{\mathbb Q}
$$

with:

$$
Q_n\rightarrow Q
$$

as the limiting condition.

---

## 17. The Possibility of an Isolated Rational Point

Even if the real solution surface is two-dimensional,

a rational Perfect Cuboid shape may still be isolated in the arithmetic sense.

That is, there exists a neighborhood:

$$
U(Q)
$$

such that:

$$
U(Q)\cap\mathcal S_{\mathbb Q}
=
\{Q\}.
$$

In the ordinary Euclidean topology,

if:

$$
\mathcal S_{\mathbb Q}
$$

is itself dense, this will not happen;

but the rational-square subconditions of the Perfect Cuboid may make the relevant arithmetic set extremely sparse.

Whether the first solution is isolated is therefore a question to test immediately after finding it.

---

## 18. Neighborhood mining

If we find:

$$
Q_1,
$$

we can construct a rational neighborhood around it:

$$
\mathcal N_H(Q_1)
=
\left\{
Q\in\mathbb Q^2:
\operatorname{height}(Q)\le H,
\;
\|Q-Q_1\|<\varepsilon
\right\}.
$$

Then check exactly which:

$$
Q
$$

still satisfy the Perfect Cuboid constraints.

If we discover:

$$
Q_2,Q_3,\ldots
$$

forming a clear trend,

we can attempt to reconstruct:

- A rational curve;
- A recurrence;
- A Möbius map;
- An elliptic-like group law;
- Another algebraic transformation.

---

## 19. Transformation discovery by pair matching

Suppose multiple primitive solutions have been found:

$$
P_1,\ldots,P_n.
$$

We can examine their normalized ratios:

$$
r_i=\frac{b_i}{a_i},
$$

$$
s_i=\frac{c_i}{a_i}.
$$

Investigate whether:

$$
(r_i,s_i)
\mapsto
(r_j,s_j)
$$

has a low-complexity rational relation.

For example:

$$
r_j=
\frac{\alpha r_i+\beta}{\gamma r_i+\delta},
$$

or:

$$
s_j=
R(r_i,s_i)
$$

where:

$$
R
$$

is a low-degree rational function.

Only if multiple solution pairs support the same map

is it worth promoting to a conjectured transformation.

---

## 20. Formal Requirements for a Solution-Preserving Map

If we propose:

$$
\mathcal T,
$$

and want to upgrade it to a theorem,

we must prove that:

$$
\forall P\in D,
$$

if:

$$
F(P)=0,
$$

then:

$$
F(\mathcal T(P))=0.
$$

where:

$$
D
$$

is an explicit domain.

Moreover, if we claim to generate a new primitive shape,

we must also address:

$$
\gcd,
$$

$$
S_3
$$

canonicalization,

and:

$$
\mathcal T(P)\not\sim P.
$$

---

## 21. Invariants of Transformations

Any genuinely solution-preserving transformation may preserve or systematically change certain quantities.

For example, if:

$$
I(P)
$$

satisfies:

$$
I(\mathcal T(P))=I(P),
$$

it is an orbit invariant.

Possible objects of study include:

- Parity profile;
- Prime-support class;
- Valuation pattern;
- Normalized ratios;
- Residue signatures;
- Radix carry signatures;
- Denominator type.

These invariants can help explain why:

$$
G_{\mathrm{sol}}
$$

may split into multiple components.

---

## 22. Primitive-preserving transformation

If:

$$
\mathcal T(P)
$$

first produces a nonprimitive solution,

then its primitive reduction:

$$
\operatorname{Prim}(\mathcal T(P))
$$

may still be a new shape.

Thus what should actually be studied is:

$$
\boxed{
\widetilde{\mathcal T}
=
\operatorname{Can}
\circ
\operatorname{Prim}
\circ
\mathcal T.
}
$$

As long as:

$$
\widetilde{\mathcal T}(P)\neq P,
$$

a new canonical primitive shape is produced.

---

## 23. The Fourth Layer of Arithmetic Freedom: Branching Freedom

If a seed:

$$
P
$$

has multiple different transformations:

$$
\mathcal T_1(P),
\ldots,
\mathcal T_k(P),
$$

capable of producing multiple distinct child shapes,

there is branching.

Define the local branching factor:

$$
b(P)
=
\#\left\{
\widetilde{\mathcal T}_i(P)
:
\widetilde{\mathcal T}_i(P)\neq P
\right\}.
$$

If, on average:

$$
b(P)>1,
$$

the solution graph may expand rapidly.

If:

$$
b(P)=1,
$$

it is closer to a chain-like family.

---

## 24. Arithmetic tree

If the transformations are primarily forward-generating,

they can form:

$$
T_{\mathrm{arith}}.
$$

root：

$$
P_1.
$$

children：

$$
P_{1,1},
P_{1,2},
\ldots
$$

The next level:

$$
P_{2,1},
P_{2,2},
\ldots
$$

We then study:

$$
N_{\mathrm{depth}}(n)
$$

which denotes how many inequivalent primitive shapes can be generated within depth:

$$
n
$$

of this tree.

This is another measure of freedom.

---

## 25. Height growth along an orbit

Suppose:

$$
P_{n+1}
=
\widetilde{\mathcal T}(P_n).
$$

Study:

$$
g_n=g(P_n).
$$

We may have:

$$
g_{n+1}\sim Cg_n,
$$

or:

$$
g_{n+1}\sim g_n^\alpha,
$$

or:

$$
\log g_n\sim cn.
$$

This determines how sparse a family is in height space.

Even if an orbit is infinite,

if:

$$
g_n
$$

grows extremely rapidly,

then within a finite:

$$
H
$$

only very few solutions may still be visible.

---

## 26. Counting function

Define:

$$
N(H)
=
\#\left\{
P\in\mathcal P_{\mathrm{can}}
:
g(P)\le H
\right\}.
$$

This is one of the most important global quantities of the entire second stage.

If:

$$
|\mathcal P_{\mathrm{can}}|<\infty,
$$

then:

$$
N(H)
$$

eventually stabilizes.

If:

$$
|\mathcal P_{\mathrm{can}}|=\infty,
$$

then:

$$
N(H)\rightarrow\infty.
$$

But the growth rates may differ greatly.

---

## 27. Several Possible Growth Regimes

### Finite

$$
N(H)=O(1).
$$

### Logarithmic

$$
N(H)\sim C\log H.
$$

### Polylogarithmic

$$
N(H)\sim C(\log H)^k.
$$

### Power-law

$$
N(H)\sim CH^\alpha.
$$

### Sub-power irregular

$$
N(H)\rightarrow\infty
$$

but grows more slowly than any:

$$
H^\varepsilon
$$

power of this form.

There is currently no reason to assume in advance that any particular regime must hold.

---

## 28. Growth freedom

Define the research quantity:

$$
D_g
=
\limsup_{H\rightarrow\infty}
\frac{\log N(H)}{\log H}.
$$

If:

$$
N(H)\sim CH^\alpha,
$$

then:

$$
D_g=\alpha.
$$

If:

$$
N(H)\sim C\log H,
$$

then:

$$
D_g=0.
$$

Thus:

$$
D_g=0
$$

does not mean that there are only finitely many solutions.

It means only that growth is slower than any positive power.

We therefore need to retain both:

$$
M
$$

and:

$$
D_g.
$$

---

## 29. Parametric Freedom and Growth Freedom Differ

Even if a one-parameter family exists:

$$
P(t),
$$

its primitive height:

$$
g(P(t))
$$

may grow extremely rapidly.

Thus:

$$
\boxed{
\text{parametric dimension}
\neq
\text{height growth exponent}.
}
$$

Conversely,

superimposing several different families may also make:

$$
N(H)
$$

grow faster.

Arithmetic freedom therefore needs at least a multidimensional description.

---

## 30. Arithmetic freedom vector

This paper proposes a research summary:

$$
\mathfrak F_{\mathrm{arith}}
=
(M,D_{\mathrm{par}},D_g,B_{\mathrm{avg}},C_{\mathrm{orb}}).
$$

where:

- $$M$$: primitive multiplicity;
- $$D_{\mathrm{par}}$$: the greatest known parametric freedom;
- $$D_g$$: height growth freedom;
- $$B_{\mathrm{avg}}$$: the average branching indicator;
- $$C_{\mathrm{orb}}$$: orbit/component structure.

This is not a single scalar,

because ‘freedom’ itself may have different aspects.

---

## 31. What Should Be Done First after One Solution?

If:

$$
P_1
$$

actually appears,

this paper recommends immediately carrying out four groups of experiments.

### Experiment A: Local rational neighborhood

Near:

$$
[P_1]
$$

search for other exact rational shapes.

### Experiment B: Low-complexity transformation search

Search for a low-degree rational map:

$$
P_1\mapsto P_2
$$

between these solutions.

### Experiment C: Height-shell mining

Expand into:

$$
g>g_1
$$

this height range,

searching for the second and third primitive solutions.

### Experiment D: Invariant profiling

Construct the prime, valuation, radix, carry and ratio signature of:

$$
P_1
$$

this solution.

These four efforts can proceed in parallel.

---

## 32. The Second Solution Has Very High Value

The first solution tells us only:

$$
M\ge1.
$$

As soon as a second inequivalent primitive solution:

$$
P_2
$$

appears,

we immediately obtain:

$$
M\ge2.
$$

More importantly, for the first time we can study the differences between:

$$
P_1\leftrightarrow P_2
$$

these two solutions.

Without a second point,

it is hard to conjecture a family.

With two points,

we can start looking for:

- A common parameter;
- A shared invariant;
- A transformation;
- A denominator relation.

But two points are still insufficient to prove a pattern.

---

## 33. The Third and Fourth Solutions Begin to Enable Structural Identification

If we obtain:

$$
P_1,P_2,P_3,P_4,
$$

we can start checking whether:

$$
P_{n+1}=\mathcal T(P_n)
$$

has a consistent recurrence.

We can also examine whether:

$$
(r_n,s_n)
$$

lies on a low-degree algebraic curve.

If:

$$
R(r_n,s_n)=0
$$

holds for multiple solutions,

we can regard:

$$
R=0
$$

as a candidate family relation.

---

## 34. Family Discovery and Overfitting

Any finite collection of points can be interpolated by some high-degree polynomial.

Thus:

$$
R(P_i)=0
$$

holding for a small amount of data

cannot directly be regarded as a family theorem.

A complexity penalty should be included:

- Low degree;
- Low coefficient height;
- Cross-validation;
- Prediction of new solutions;
- Exact proof.

Only when the relation can:

$$
\boxed{
\text{predict unseen exact solutions}
}
$$

does it have genuine structural value.

---

## 35. The Verification Procedure for Transformation Discovery

If, from:

$$
P_1,P_2,\ldots
$$

we conjecture:

$$
\mathcal T,
$$

the formal procedure is:

1. Derive symbolically:
   $$
   F(\mathcal T(P));
   $$
2. Use:
   $$
   F(P)=0
   $$
   to simplify;
3. Prove:
   $$
   F(\mathcal T(P))=0;
   $$
4. Analyze the denominator;
5. Perform primitive reduction;
6. Perform canonicalization;
7. Prove nontriviality;
8. Study whether iteration is infinite.

---

## 36. Nontriviality gate

Define:

$$
\operatorname{NTG}(\mathcal T,P)=\mathrm{PASS}
$$

if and only if:

$$
\operatorname{Can}
\circ
\operatorname{Prim}
(\mathcal T(P))
\neq P.
$$

If:

$$
\operatorname{NTG}=\mathrm{FAIL},
$$

the transformation is merely:

- Scaling;
- A permutation;
- Or another canonically equivalent operation.

---

## 37. Infinite-orbit gate

Even if:

$$
\mathcal T(P)\neq P,
$$

we may still have:

$$
\mathcal T^2(P)=P.
$$

forming a two-cycle.

To prove an infinite shape family,

we therefore also need:

$$
\mathcal T^m(P)\neq\mathcal T^n(P)
$$

for all:

$$
m\neq n.
$$

This can be proved through a monotone invariant:

$$
H(\mathcal T(P))>H(P)
$$

of this form.

For example, if:

$$
g(\mathcal T(P))>g(P)
$$

holds for every:

$$
P
$$

solution,

then the orbit cannot be periodic.

---

## 38. Monotone height transformation

If we find:

$$
\mathcal T
$$

such that:

$$
g(\mathcal T(P))>g(P),
$$

and it always preserves an exact primitive shape,

then from any seed:

$$
P_0
$$

we can obtain:

$$
g_0<g_1<g_2<\cdots.
$$

Thus all:

$$
P_n
$$

are distinct.

This immediately implies:

$$
\boxed{
|\mathcal P_{\mathrm{can}}|=\infty.
}
$$

This would be a very strong second-stage result.

---

## 39. The Reverse Transformation

If:

$$
\mathcal T
$$

is invertible,

and its inverse map:

$$
\mathcal T^{-1}
$$

lowers height,

then each orbit may have a minimal seed:

$$
P_{\min}^{\mathrm{orbit}}.
$$

This decomposes the solution space into:

$$
\boxed{
\text{minimal seeds}
+
\text{forward expansions}.
}
$$

This would closely resemble the seed-world perspective in world-generation models.

---

## 40. Orbit seed

Define:

$$
P
$$

to be an orbit seed

if there exists no:

$$
Q
$$

such that:

$$
\widetilde{\mathcal T}(Q)=P
$$

and:

$$
g(Q)<g(P).
$$

Each component may have:

- One seed;
- Multiple seeds;
- No minimal seed;
- A cycle.

If the transformation has height monotonicity,

a minimal seed can usually be obtained.

---

## 41. A Semigroup of Multiple Transformations

If we find:

$$
\mathcal T_1,\mathcal T_2,
$$

they may be noncommutative:

$$
\mathcal T_1\mathcal T_2(P)
\neq
\mathcal T_2\mathcal T_1(P).
$$

Solution generation is then no longer a single path,

but a semigroup action.

Its complexity may far exceed that of a one-parameter recurrence.

This may produce:

$$
\boxed{
\text{branching arithmetic world}.
}
$$

---

## 42. The First Source of an Upper Bound on Freedom

The normalized real shape:

$$
(x,y,z)
$$

satisfies:

$$
x^2+y^2+z^2=1,
$$

and has only two continuous degrees of freedom.

Thus, if a smooth/rational family exists,

we should generally not expect more than:

$$
2
$$

independent continuous parameters.

Thus:

$$
\boxed{
D_{\mathrm{par}}\le2
}
$$

can be viewed as a natural geometric upper bound.

But discrete branching complexity can exceed the ‘two-parameter’ intuition,

because discrete graph structure is not continuous dimension.

---

## 43. $D_{\mathrm{par}}=0$ May Still Allow Infinitely Many Solutions

For example:

$$
\mathcal P_{\mathrm{can}}
$$

may be an infinite discrete set,

without a single rational parameter family covering it.

We may then have:

$$
M=\infty,
$$

but:

$$
D_{\mathrm{par}}=0
$$

under our operational definition.

Thus:

$$
\boxed{
\text{infinite}
\neq
\text{positive parametric freedom}.
}
$$

---

## 44. $D_{\mathrm{par}}=1$ Does Not Guarantee That a Single Family Covers All Solutions

There may exist:

$$
\mathcal F_1,
\mathcal F_2,\ldots
$$

many one-parameter families.

Each family is one-dimensional,

but their union may have a complex component structure.

We should therefore separately record:

$$
C_{\mathrm{fam}}
=
\text{number of detected families}.
$$

---

## 45. Primitive family

Define a primitive family:

$$
\mathcal F
=
\{P(t):t\in D\subseteq\mathbb Q\}
$$

with the requirements:

1. $$P(t)$$ exactly satisfies the Perfect Cuboid conditions;
2. It is nonempty after primitive reduction;
3. Infinitely many $$t$$ give distinct canonical shapes.

If we have only:

$$
P(t)=k(t)P_0,
$$

it is not a primitive family,

but merely a scaling family.

---

## 46. Family normalization

To prevent scaling from being mixed in,

it is best to work in the:

$$
g=1
$$

shape world and define:

$$
Q(t).
$$

Require:

$$
g(Q(t))=1.
$$

and then:

$$
Q(t)\in\mathcal S_{\mathbb Q}.
$$

In this way, any variation of:

$$
t
$$

must be a shape change,

not a scale change.

---

## 47. Family rank

If the family:

$$
Q(t)
$$

can be controlled by one free rational parameter,

define the rank:

$$
r_{\mathrm{fam}}=1.
$$

If we have:

$$
Q(s,t),
$$

then:

$$
r_{\mathrm{fam}}=2.
$$

If there is algebraic dependence among the parameters,

the actual rank should subtract those dependencies.

Thus family rank is:

$$
\boxed{
\text{independent rational parameter count}.
}
$$

---

## 48. Constraint Rank and Family Rank

The original normalized real world:

$$
x^2+y^2+z^2=1
$$

is two-dimensional.

Adding three rational-square conditions

does not mean adding ordinary continuous equalities,

but arithmetic conditions.

We therefore cannot directly use dimension counting such as:

$$
2-3=-1
$$

this calculation.

This is why arithmetic freedom is difficult to predict directly from geometric freedom.

---

## 49. Reintroducing the Rational-Square Conditions

For the:

$$
g=1
$$

normalized shape:

$$
x,y,z,
$$

require the existence of:

$$
u,v,w\in\mathbb Q
$$

such that:

$$
x^2+y^2=u^2,
$$

$$
x^2+z^2=v^2,
$$

$$
y^2+z^2=w^2.
$$

Thus, to find a family,

we can directly seek a rational map:

$$
t
\mapsto
(x(t),y(t),z(t),u(t),v(t),w(t))
$$

satisfying four polynomial identities.

This may reveal parametric freedom more readily than starting from primitive integers.

---

## 50. Family Search Can Start Near the First Solution

If:

$$
Q_1
$$

exists,

let:

$$
x=x_1+\Delta x,
$$

$$
y=y_1+\Delta y,
$$

$$
z=z_1+\Delta z.
$$

and then require:

$$
F(Q_1+\Delta)=0.
$$

Tangent analysis can be performed at the real level.

Next ask whether:

$$
\Delta
$$

can be expressed using a rational parameter:

$$
t
$$

of this form.

This is a form of:

$$
\boxed{
\text{local-to-family lifting}.
}
$$

---

## 51. The Jacobian's Role in the Second Stage

For the full normalized system:

$$
G(x,y,z,u,v,w)
=
\begin{pmatrix}
x^2+y^2-u^2\\
x^2+z^2-v^2\\
y^2+z^2-w^2\\
x^2+y^2+z^2-1
\end{pmatrix},
$$

at a rational solution:

$$
Q
$$

compute:

$$
J_G(Q).
$$

The real tangent space:

$$
\ker J_G(Q)
$$

describes local continuous deformation.

In general, it has:

$$
6-4=2
$$

dimensions,

consistent with shape freedom.

But whether an arithmetic family can expand along these tangent directions

needs separate verification.

---

## 52. A Tangent Basis and Arithmetic Sampling

Take:

$$
V_1,V_2
$$

as a basis of:

$$
\ker J_G(Q)
$$

this kernel.

Consider:

$$
Q+\alpha V_1+\beta V_2.
$$

First perform a real correction to return to:

$$
G=0.
$$

Then perform rational reconstruction near:

$$
\alpha,\beta\in\mathbb Q
$$

these parameters.

This may reveal:

- A rational branch;
- Multiple discrete branches;
- No nearby rational closure at all.

---

## 53. If No Second Solution Is Found Near the First

We still cannot infer:

$$
M=1.
$$

Other primitive families may lie far away.

Thus local isolation says only that:

$$
Q_1
$$

has no other low-height rational closure nearby;

it is not global uniqueness.

To prove:

$$
M=1
$$

requires a global classification.

---

## 54. Global Uniqueness Is a Very Strong Theorem

To prove:

$$
|\mathcal P_{\mathrm{can}}|=1,
$$

we need to prove:

$$
\forall P\in\mathcal P_{\mathrm{can}},
\quad
P=P_1.
$$

This is far harder than finding the first solution.

Second-stage research should therefore avoid implying uniqueness because a second solution has not been found for a long time.

The correct status is:

$$
\boxed{
\text{only one known}
}
$$

not:

$$
\boxed{
\text{only one exists}.
}
$$

---

## 55. Height-shell exploration

Define the shell:

$$
\mathcal P(H_1,H_2)
=
\left\{
P\in\mathcal P_{\mathrm{can}}
:
H_1<g(P)\le H_2
\right\}.
$$

If the height of the first solution:

$$
g_1
$$

is known,

we can search logarithmic shells such as:

$$
(g_1,2g_1],
$$

$$
(2g_1,4g_1],
$$

$$
(4g_1,8g_1],
$$

these intervals.

This facilitates estimating the growth of:

$$
N(H)
$$

this counting function.

---

## 56. Empirical Growth Can Only Be a Conjecture

Even observing:

$$
N(H)\approx C\log H
$$

up to a very large:

$$
H,
$$

does not allow us to declare an asymptotic law directly.

At a minimum, we should distinguish:

$$
N_{\mathrm{obs}}(H)
$$

from the actual:

$$
N(H).
$$

Only exhaustive exact enumeration yields the true counting function.

---

## 57. Certified counting

If we perform complete exact enumeration for all:

$$
g\le H
$$

within this height range,

then:

$$
N(H)
$$

can be certified.

This makes:

$$
N(H_1),N(H_2),\ldots
$$

rigorously known data points.

Conjecture an asymptotic law from these data,

then seek a theorem.

This is a more reliable growth program than heuristic sampling.

---

## 58. Arithmetic Density

We can define density relative to a candidate shell.

For example:

$$
S(H)
=
\#\left\{
(a,b,c):
a<b<c,
\;
a^2+b^2+c^2\le H^2
\right\}.
$$

Then:

$$
\rho(H)
=
\frac{N(H)}{S(H)}.
$$

This describes the sparsity of Perfect Cuboid primitive shapes among integer shape candidates.

Even if:

$$
N(H)\rightarrow\infty,
$$

we may still have:

$$
\rho(H)\rightarrow0.
$$

---

## 59. Arithmetic Sparsity and Arithmetic Freedom Can Coexist

For example, we may have:

$$
N(H)\sim\log H,
$$

so:

$$
N(H)\rightarrow\infty,
$$

but the density:

$$
\rho(H)\rightarrow0.
$$

There are then infinitely many solutions,

yet they are extremely sparse.

Thus:

$$
\boxed{
\text{infinite freedom}
}
$$

and:

$$
\boxed{
\text{dense occurrence}
}
$$

are entirely different.

---

## 60. The Generation Rule May Matter More than Density

If a simple:

$$
\mathcal T
$$

can generate infinitely many solutions from a seed,

then even if:

$$
N(H)
$$

grows extremely slowly,

the structure is still very strong.

Second-stage research should therefore not pursue only larger solution counts,

but prioritize finding:

$$
\boxed{
\text{compressible generation law}.
}
$$

---

## 61. Kolmogorov-like intuition

If all primitive solutions can be generated by:

$$
P_0
$$

together with a short transformation:

$$
\mathcal T
$$

acting on that seed,

then the entire infinite solution family has low description complexity.

If, conversely, each solution seems to require an independent description,

the arithmetic world resembles an irregular discrete set more closely.

This paper does not treat this as a formal Kolmogorov complexity theorem,

but only as research intuition about:

$$
\boxed{
\text{solution-family compressibility}
}
$$

this concept.

---

## 62. Solution-family compression ratio

If the raw data length of the first:

$$
n
$$

primitive solutions is:

$$
L_{\mathrm{raw}}(n),
$$

and the total description length of a generation rule and its seeds is:

$$
L_{\mathrm{gen}}(n),
$$

we can observe:

$$
C(n)
=
\frac{L_{\mathrm{gen}}(n)}
{L_{\mathrm{raw}}(n)}.
$$

If:

$$
C(n)\rightarrow0,
$$

this indicates a highly compressible pattern.

This is a computational research quantity,

not a purely number-theoretic invariant.

---

## 63. Uses of the Radix Signature in the Second Stage

The:

$$
\mathfrak R_B
$$

of Paper 4 is not used only to find the first solution.

For:

$$
P_1,P_2,\ldots
$$

we can compare the carry signatures of:

$$
\mathcal O_B(P_i)
$$

these observations.

If solutions in the same family exhibit a stable pattern,

it may be used to identify:

$$
\text{family membership}.
$$

For example, certain:

$$
p
$$

-adic valuation profiles may be preserved within an orbit.

---

## 64. Cross-Base Family Fingerprint

Define:

$$
\Sigma(P)
=
\left(
\Sigma_2(P),
\Sigma_3(P),
\Sigma_5(P),
\Sigma_7(P)
\right)
$$

where:

$$
\Sigma_p(P)
$$

may contain:

- A valuation vector;
- Low-digit residues;
- A carry-state summary;
- A denominator-free normalized signature.

If:

$$
P_i
$$

and:

$$
P_j
$$

belong to the same family,

then:

$$
\Sigma(P_i)
$$

may exhibit a recognizable transformation pattern.

This provides another route for transformation discovery.

---

## 65. Prime-support branching

Different solution families may require different prime support.

For example, an orbit may allow only certain:

$$
v_p
$$

parity classes.

Another orbit may differ.

If such differences can be proved not to be interconvertible,

we can prove that:

$$
G_{\mathrm{sol}}
$$

has at least two arithmetic components.

---

## 66. Arithmetic Invariants Can Prove Inequivalence

Suppose there exists:

$$
I(P)
$$

satisfying, under all allowed transformations:

$$
I(\mathcal T(P))=I(P).
$$

If:

$$
I(P_1)\neq I(P_2),
$$

then:

$$
P_1
$$

and:

$$
P_2
$$

are not in the same orbit.

Thus invariants are not used only to exclude candidates,

but can also classify the solution space.

---

## 67. The Complete Second-Stage Research Tree

Once the first solution appears,

we can immediately develop:

### Branch A: Multiplicity

Find the second, third and further primitive shapes.

### Branch B: Local Geometry

Study real / rational deformations near the first solution.

### Branch C: Transformation

Find solution-preserving maps.

### Branch D: Counting

Estimate and certify:

$$
N(H).
$$

### Branch E: Invariant

Find orbit / family signatures.

### Branch F: Classification

Attempt to describe all arithmetic components.

---

## 68. Switching from Existence Mode to Expansion Mode

Papers 1 through 6 are mainly in:

$$
\boxed{
\text{Existence Mode}.
}
$$

Once:

$$
P_1
$$

receives PASS,

the research system can switch to:

$$
\boxed{
\text{Expansion Mode}.
}
$$

The objective function then no longer seeks only:

$$
F(P)=0.
$$

because:

$$
P_1
$$

already satisfies it.

The new objective is:

$$
\boxed{
\text{novel exact shape}
}
$$

that is:

$$
P\neq P_i
$$

for every known:

$$
P_i.
$$

---

## 69. Novelty gate

Define:

$$
\operatorname{NG}(P)=\mathrm{PASS}
$$

if:

1. $$P$$ is exact;
2. Primitive;
3. Canonical;
4. Not in the known solution database.

Thus:

$$
\operatorname{NG}
$$

can prevent the generator from repeatedly returning:

$$
kP_1
$$

or permutation duplicates.

---

## 70. Distance-Based Diversity Is Suitable Only as a Heuristic

In the normalized shape space, we can define:

$$
d(Q_i,Q_j).
$$

If a candidate is far from the known solutions,

its exploration priority can be increased.

But:

$$
d>0
$$

cannot prove arithmetic inequivalence.

Actual novelty is still determined by the canonical primitive tuple.

---

## 71. Exploration pressure

If the generator always returns near:

$$
P_1
$$

this solution,

we can add repulsion:

$$
R(Q)
=
\sum_i
\frac{1}{d(Q,Q_i)^2+\epsilon}.
$$

so that continuous search avoids the known basins.

This belongs only to the discovery layer.

The exact verifier remains unchanged.

---

## 72. Arithmetic repulsion

A stronger approach than Euclidean repulsion is to add denominator / residue diversity pressure to the known primitive ratios:

$$
(r_i,s_i)
$$

of these solutions.

For example, prioritize exploration of different:

$$
p
$$

-adic signatures.

This may make it easier to jump to a different arithmetic component.

---

## 73. The Second Expansion of the Solution Space

The first-stage ‘expansion’ is:

$$
\text{real world}
\rightarrow
\text{rational point}
\rightarrow
\text{integer manifestation}.
$$

The second-stage ‘expansion’ is:

$$
\boxed{
P_1
\rightarrow
\mathcal P_{\mathrm{can}}.
}
$$

That is:

> From a known arithmetic world, infer the freedom, generation laws and boundaries of the entire solution space.

These are two entirely different expansions.

---

## 74. Comparing the First and Second Expansions

### First Expansion

Study:

$$
\mathcal S_{\mathbb R}
\rightarrow
\mathcal S_{\mathbb Q}.
$$

The core question:

$$
\exists?
$$

### Second Expansion

Study:

$$
P_1
\rightarrow
\mathcal P_{\mathrm{can}}.
$$

The core question:

$$
\text{how many / how generated / how growing?}
$$

Thus:

$$
\boxed{
\text{existence}
\rightarrow
\text{freedom}.
}
$$

---

## 75. Upper and Lower Bounds on Arithmetic Freedom

After finding one solution,

the minimal lower bound is:

$$
M\ge1.
$$

Parametric freedom is at least:

$$
D_{\mathrm{par}}\ge0.
$$

The geometric upper bound:

$$
D_{\mathrm{par}}\le2.
$$

Thus:

$$
\boxed{
0\le D_{\mathrm{par}}\le2.
}
$$

But:

$$
M
$$

may range over:

$$
1,\ldots,\infty.
$$

Thus ‘freedom’ must at least be described by the two quantities:

$$
(M,D_{\mathrm{par}})
$$

taken together.

---

## 76. If a Solution-Preserving One-Parameter Family Is Found

If we can prove that:

$$
Q(t)\in\mathcal S_{\mathbb Q}
$$

holds for infinitely many:

$$
t\in\mathbb Q
$$

parameters,

with distinct shapes,

we immediately obtain:

$$
M=\infty,
$$

and:

$$
D_{\mathrm{par}}\ge1.
$$

This is stronger than individually listing many solutions.

---

## 77. If a Two-Parameter Family Is Found

If:

$$
Q(s,t)
$$

is valid over a two-dimensional rational parameter domain,

then:

$$
D_{\mathrm{par}}=2
$$

attains the geometric upper bound.

We can then say:

> Arithmetic realization uses the full geometric shape freedom.

But we must still note whether:

$$
Q(s,t)
$$

covers all solutions;

this is a separate completeness question.

---

## 78. If Only Discrete Points Are Ever Found

We may have:

$$
D_{\mathrm{par}}=0
$$

but:

$$
M=\infty.
$$

Counting:

$$
N(H)
$$

and the transformation graph then matter more than family dimension.

The final paper will therefore specifically address:

$$
\boxed{
\text{upper/lower bounds and growth}.
}
$$

---

## 79. The Core Research Hypotheses Proposed in This Paper

### Hypothesis AF1: Nontrivial Multiplicity

If a Perfect Cuboid exists,

then:

$$
|\mathcal P_{\mathrm{can}}|>1.
$$

### Hypothesis AF2: Infinite Primitive Shape Hypothesis

If a Perfect Cuboid exists,

then:

$$
|\mathcal P_{\mathrm{can}}|=\infty.
$$

### Hypothesis AF3: Transformational Closure

There exists a nontrivial:

$$
\widetilde{\mathcal T}
$$

capable of mapping some primitive solutions to new primitive solutions.

### Hypothesis AF4: Positive Parametric Freedom

There exists:

$$
D_{\mathrm{par}}\ge1.
$$

These are all only subsequent hypotheses,

and cannot follow automatically from the first solution.

---

## 80. The Most Conservative Research Position

If the first solution appears,

the only immediately certain fact is:

$$
\boxed{
M\ge1.
}
$$

Beyond that:

$$
M=1?
$$

$$
M<\infty?
$$

$$
M=\infty?
$$

$$
D_{\mathrm{par}}=0?
$$

$$
D_{\mathrm{par}}=1?
$$

$$
D_{\mathrm{par}}=2?
$$

must all be regarded as open questions.

This avoids over-extrapolating from one witness to the entire arithmetic world.

---

## 81. The Minimal Data Structure for the Second Stage

For each known primitive solution, store:

```text
solution_id
a
b
c
d
e
f
g
normalized_ratios
prime_signature
valuation_signature
radix_signatures
discovery_method
parent_solution_id
transformation_id
exact_verification_hash
```

Here:

$$
parent\_solution\_id
$$

and:

$$
transformation\_id
$$

are used to construct:

$$
G_{\mathrm{sol}}.
$$

If the solution was found by an independent search,

the parent may be empty.

---

## 82. Family record

If a candidate family is discovered:

```text
family_id
parameter_domain
rational_map
seed_solutions
predicted_solutions
verified_solutions
invariants
completeness_status
proof_status
```

explicitly distinguish:

$$
\text{candidate}
$$

from:

$$
\text{proved family}.
$$

---

## 83. Transformation record

For:

$$
\mathcal T
$$

store:

```text
transformation_id
domain
formula
exact_preservation_status
primitive_preservation_status
nontriviality_status
invertibility
height_behavior
orbit_tests
proof_status
```

This lets transformation discovery accumulate,

rather than requiring a fresh conjecture each time.

---

## 84. Formalized Success Levels

### AF-L0

There is only:

$$
P_1.
$$

### AF-L1

There exists:

$$
P_2\neq P_1.
$$

### AF-L2

An infinite sequence is known:

$$
P_1,P_2,\ldots
$$

but there is not yet a complete generation theorem.

### AF-L3

There exists a proved nontrivial transformation generating an infinite orbit.

### AF-L4

There exists a proved one-parameter family.

### AF-L5

There exists a proved two-parameter family.

### AF-L6

The primitive solution-space classification is complete.

This ladder can serve as a roadmap for the second stage.

---

## 85. After the First Solution, Structure Is Worth Pursuing More than Minimality

Paper 6 has already separated out minimality.

The core claim of Paper 7 is:

> Once existence is settled, the next step with the highest mathematical value is usually not to reduce the height by a few more digits, but to seek generatable structure in the solution space.

That is, prioritize asking:

$$
\boxed{
\text{Can one solution explain another?}
}
$$

If the answer is yes,

the problem develops from a single number-theoretic witness into arithmetic dynamics.

---

## 86. Correspondence with World-Generation Models

The first primitive solution:

$$
P_1
$$

can be viewed as:

$$
\boxed{
\text{seed world}.
}
$$

A nontrivial transformation:

$$
\mathcal T
$$

is then:

$$
\boxed{
\text{world evolution operator}.
}
$$

The solution graph:

$$
G_{\mathrm{sol}}
$$

is:

$$
\boxed{
\text{reachable arithmetic worlds}.
}
$$

This is not merely a metaphor,

because:

$$
P_1
$$

can indeed serve as a state,

$$
\mathcal T
$$

can indeed serve as an exact rule,

and:

$$
\mathcal T(P_1)
$$

if it has exact closure,

is an actual new world state.

---

## 87. The Actual Question of the Second Expansion

Thus the second expansion is not:

$$
P_1
\rightarrow
2P_1
\rightarrow
3P_1.
$$

but:

$$
P_1
\rightarrow
P_2
\rightarrow
P_3
\rightarrow\cdots
$$

where:

$$
P_i\not\sim P_j.
$$

That is:

$$
\boxed{
\text{shape evolution}
}
$$

rather than:

$$
\boxed{
\text{scale evolution}.
}
$$

---

## 88. The Core Conclusions of This Paper

This paper obtains the following framework.

### Conclusion 1

As soon as one primitive solution exists,

there are automatically infinitely many raw integer solutions,

but this is only scaling infinity.

### Conclusion 2

The actual object to study is:

$$
\mathcal P_{\mathrm{can}}.
$$

### Conclusion 3

Arithmetic freedom includes at least:

$$
\text{multiplicity},
$$

$$
\text{parametric freedom},
$$

$$
\text{transformational freedom},
$$

$$
\text{growth freedom}.
$$

### Conclusion 4

The geometric shape freedom:

$$
2
$$

provides only a natural upper bound for parametric arithmetic freedom;

it does not determine the actual structure of rational points.

### Conclusion 5

If a nontrivial height-increasing solution-preserving transformation exists,

infinitely many primitive shapes can be proved directly.

### Conclusion 6

The first solution provides only:

$$
M\ge1.
$$

Everything else requires a new theorem or certified data.

---

## 89. Next Paper: Upper and Lower Bounds of the Solution Space and a Second Expansion

Paper 8 will turn the qualitative freedom of this paper into quantitative bounds.

Core quantities include:

$$
N(H)
=
\#\{P\in\mathcal P_{\mathrm{can}}:g(P)\le H\},
$$

and:

$$
D_g
=
\limsup_{H\rightarrow\infty}
\frac{\log N(H)}{\log H}.
$$

with further study of:

- Lower bounds on primitive solution counts;
- Counting upper bounds;
- Family-induced lower bounds;
- Transformation-induced growth;
- Exhaustive shell enumeration;
- Certified asymptotics;
- The differences between geometric freedom and arithmetic growth;
- A formal division of the second expansion into stages.

---

## 90. Conclusion

If the first exact primitive solution of the Perfect Cuboid is discovered in the future,

it will not merely be the endpoint of the classical existence problem.

It will also become:

$$
\boxed{
\text{The seed of the second-stage arithmetic world}.
}
$$

The first stage asks:

$$
\boxed{
\exists P?
}
$$

The second stage asks:

$$
\boxed{
\mathcal P_{\mathrm{can}}
\text{ How large is it?}
}
$$

This ‘how large’ cannot be answered by a single dimension alone.

At a minimum, we must separate:

$$
M
=
|\mathcal P_{\mathrm{can}}|,
$$

$$
D_{\mathrm{par}},
$$

$$
D_g,
$$

and the orbit / branching structure of the solution graph.

This paper therefore defines arithmetic freedom as a multilayer structure:

$$
\boxed{
\text{Multiplicity}
+
\text{Parametric Freedom}
+
\text{Transformational Freedom}
+
\text{Growth Freedom}.
}
$$

The genuinely most important distinction is:

$$
\boxed{
\text{scaling infinity}
\neq
\text{shape infinity}.
}
$$

Once the first Perfect Cuboid exists,

its infinitely many integer multiples provide no new shape information.

What is actually worth expanding is:

$$
\boxed{
P_1
\rightarrow
P_2
\rightarrow
P_3
\rightarrow\cdots
}
$$

and whether there exists:

$$
\boxed{
\mathcal T
}
$$

that generates this sequence through a simple exact rule.

Thus Paper 7 completes the transition from ‘first manifestation’ to ‘arithmetic freedom’.

The next paper will conclude this introductory series:

# Upper and Lower Bounds of the Solution Space and a Second Expansion

It will integrate primitive multiplicity, height counting, growth dimension, family lower bounds and global upper bounds into a formal second-stage research program.

---

## Contents of the Series

1. **Reframing the Perfect Cuboid: It Is Not Primarily an Integer Search Problem**
2. **The Continuous World and the Scale Quotient Space: The Topological Starting Point of the Perfect Cuboid**
3. **Common Initial State and Constraint-Induced Differentiation**
4. **The Radix as an Observation Operator: Radix–Carry Global Closure**
5. **From Floating-Point Approximation to Exact Rational Closure**
6. **The First Manifestation Problem: From a Rational Solution to the Minimal Primitive Cuboid**
7. **After the First Solution: Arithmetic Freedom of the Perfect Cuboid**
8. **Upper and Lower Bounds of the Solution Space and a Second Expansion**

---

**End of Paper 07**