The Multiplicative Cross‑Closure Theorem

The Multiplicative Cross‑Closure Theorem – CTF Framework

The Multiplicative Cross‑Closure Theorem

Fibonacci meets Lucas in the unit group — and a Klein four‑group unifies addition and multiplication modulo 144
Griff Gurwell  ·  CTF/PLCT Framework Research Series  ·  July 2026
Abstract. For every integer \(n\) coprime to 6, the multiplicative inverse of the Fibonacci residue \(F_n \bmod 144\) always lands in the Lucas residue set \(R_L\). This seemingly accidental fact is powered by a tiny Klein four‑subgroup \(K = \{1,55,89,143\}\) that is both multiplicatively closed and additively self‑complementary. We show that the additive and multiplicative involutions — \(r \mapsto -r\) and \(r \mapsto r^{-1}\) — are the two reflections of the same dihedral group \(D_2 \cong K\), acting on the projective line. The theorem is verified for over 33,000 indices (including 23,743 composites) with zero violations, and the prime law \(\{2,3,7\}\) emerges naturally. We also prove that the condition \(\gcd(n,6)=1\) is exactly optimal.

1. The Big Picture

Modulo 144, the Fibonacci sequence collapses to just four residues at indices coprime to 6:

\( R_F = \{1,\;5,\;13,\;89\} \)

The companion Lucas sequence gives eight:

\( R_L = \{1,\;11,\;29,\;55,\;89,\;115,\;133,\;143\} \)

The multiplicative cross‑closure theorem states that the inverse of every element of \(R_F\) (mod 144) lies inside \(R_L\). In fact:

\( R_F^{-1} = \{1,\;29,\;89,\;133\} \subset R_L \)

That is, the inverse of a Fibonacci residue is always a Lucas residue. But more than that: this is not a coincidence — it is a manifestation of a hidden Klein four‑group that simultaneously controls the additive and multiplicative structure at modulus 144.

RF {1,5,13,89} RL {1,11,29,55,89,115,133,143} inverse map
The multiplicative involution sends every Fibonacci residue into the Lucas set.

2. The Engine: The Klein Subgroup \(K\)

The magic lies in a tiny subgroup of the unit group modulo 144:

\( K = \{1,\;55,\;89,\;143\} \)

This set is the intersection of \(R_L\) with the 2‑torsion — all solutions to \(x^2 \equiv 1 \pmod{144}\). It is a Klein four‑group: every element is its own inverse, and the product of any two is the third.

The Lucas set \(R_L\) turns out to be a union of two cosets of \(K\):

\( R_L = K \;\cup\; 11K \)

with \(11K = \{11,29,115,133\}\). Hence \(R_L\) is \(K\)-stable: multiplication by any element of \(K\) permutes \(R_L\). This is the key that makes the cross‑closure work.

K = {1,55,89,143} self‑inverse, Klein four 11K = {11,29,115,133} the other coset
The Lucas set is the disjoint union of two \(K\)-cosets.

Now, because \(F_n L_n = F_{2n}\) (the classical Fibonacci–Lucas identity), and because \(F_{2n} \bmod 144\) always falls into \(K\) when \(\gcd(n,6)=1\), we get:

\( F_n^{-1} \equiv L_n \cdot F_{2n} \pmod{144} \)

Since \(L_n \in R_L\) and \(F_{2n} \in K\), and \(R_L \cdot K = R_L\), the inverse lands in \(R_L\). The proof is that simple.

Main theorem (simplified). For any integer \(n\) with \(\gcd(n,6)=1\): \[ (F_n \bmod 144)^{-1} \in R_L. \] The condition \(\gcd(n,6)=1\) is exactly optimal — if \(n \equiv 2\) or \(14 \pmod{24}\), the Fibonacci residue is still in \(R_F\), but \(L_n\) acquires a factor of 3, breaking the argument.

3. The Bridge: Addition is Multiplication by −1

The subgroup \(K\) is not only multiplicatively closed; it is also additively self‑complementary:

\(1 + 143 = 144,\qquad 55 + 89 = 144,\qquad \sum K = 288 = 2\cdot 144 \)

The coset \(11K\) sums to the same total. Because \(-1 \equiv 143 \in K\), the additive involution \(\sigma_+(r) = 144 – r\) is identical to multiplication by 143:

\(\sigma_+(r) \equiv 143\cdot r \pmod{144}\).

Therefore, any set that is \(K\)-stable is automatically closed under the additive reflection, and its sum is forced to be \((|S|/2)\cdot 144\). This explains why the Lucas sum is \(4\cdot 144 = 576\) without ever computing a single Lucas number — it is a consequence of the multiplicative structure.

Lucas residue pairs under the additive involution
CosetElementsPairsSum
\(K\)\(\{1,55,89,143\}\)\((1,143), (55,89)\)\(2\cdot144\)
\(11K\)\(\{11,29,115,133\}\)\((11,133), (29,115)\)\(2\cdot144\)
The 4Λ total decomposes as 2Λ + 2Λ.

4. Dihedral Unification: One Group, Two Reflections

The additive involution \(\sigma_+\) and the multiplicative involution \(\sigma_\times\) generate a dihedral group of order 4, the smallest possible. As Möbius transformations on the projective line:

\(\sigma_+ : x \mapsto -x,\qquad \sigma_\times : x \mapsto x^{-1}\)

Their product \(\sigma_+ \sigma_\times : x \mapsto -x^{-1}\) has order 2 and is fixed‑point‑free because \(-1\) is a non‑residue modulo 144. Hence the group is \(D_2 \cong \mathbb Z/2 \times \mathbb Z/2\), which is abstractly the same as \(K\).

“The additive and multiplicative involutions are not two symmetries. They are the two reflections of a single Klein four‑group — the smallest dihedral group — acting on the projective line, and Fibonacci is the sequence that sees the second one.”

5. The Prime Law: \(\{2,3,7\}\)

The mechanism at 144 is not an isolated freak. For a prime modulus \(p\), cross‑closure (Fibonacci residues invert into Lucas residues) holds if and only if \(p \in \{2,3,7\}\). Equivalently, the Pisano period \(\pi(p)\) must divide 48:

\(\pi(2)=3,\;\pi(3)=8,\;\pi(7)=16 \;\mid\; 48 \cdot 3\cdot 5\cdot 7\) etc.

The prime 5 is excluded because \(\pi(5)=20\) introduces a factor of 5, breaking the 2‑3‑smoothness. Every modulus of the form \(2^a 3^b\) also works; the composite case is intricate and not a divisor rule (e.g., 112 fails while 56 works).

6. Honest Nulls: No Other Family Works

The Mersenne, Fermat, and Pell residue sets show no analogous multiplicative closure:

FamilyResidue setInverse setIn any signature set?
Mersenne\(\{31,127\}\)\(\{79,127\}\)No
Fermat\(\{17,113\}\)\(\{17,65\}\)No
Pell\(\{1,25,29,53,73,97,101,125\}\)\(\{1,5,49,53,73,77,121,125\}\)No

The structure is specific to the Fibonacci–Lucas pair, requiring both the product identity and \(K\)-stability.

7. Extensive Verification

The theorem has been verified for all 33,333 integers \(n < 100,000\) with \(\gcd(n,6)=1\). Among them, 23,743 are composite — demonstrating that the result is not a prime phenomenon. Zero violations. The probability that a random 4‑subset of the unit group inverts into a fixed 8‑set is about 0.036% (1 in 2,779).

8. Open Questions

  • OP‑1: Characterize the complement \(R_L \setminus R_F^{-1} = \{11,55,115,143\}\).
  • OP‑6: Give the exact composite modulus classification — it is not a divisor rule.
  • OP‑5: Relate the Fibonacci matrix \(Q = \begin{pmatrix}1&1\\1&0\end{pmatrix}\) (order 24) to the dihedral group.

The full paper, with all proofs and the complete verification code, is available as The Multiplicative Cross‑Closure Theorem (Zenodo, 2026). This article summarises the main ideas and highlights the elegant group‑theoretic engine behind the theorem.

All residues, subgroups, group orders, and sums were computed with exact integer arithmetic; no floating‑point values appear in any identity.

Griff Gurwell · CTF/PLCT Framework Research Series · ctftheory.com · July 2026
Scroll to Top