betaFeedback

WorksMathematics

Excluding 54 of the 76 Nontrivial Factorization Patterns of the Cuboid Polynomial Pa,b,u(t)P_{a,b,u}(t)

Abstract

Let Pa,b,u(t)P_{a,b,u}(t) be Sharipov's even monic degree-1212 third cuboid polynomial, with positive pairwise distinct rational parameters. We prove a uniform normal form for its complete factorization over Q\mathbb{Q}. Every factorization consists of at most two even quadratic factors, even irreducible factors of higher degree, and pairs of irreducible factors exchanged by t↦−tt\mapsto-t. After dividing all degrees by the contribution of this symmetry, the normal form satisfies one degree equation, one sign equation, and an obligatory-factor alternative: an even irreducible factor of degree at least six, or an even irreducible quartic with positive rational-square constant term. The arithmetic input is a uniform construction over Q2\mathbb{Q}_2, using Newton polygons, Hensel lifting, and a norm obstruction from Q2(i)\mathbb{Q}_2(i). Valuations at all finite primes constrain rational factor constants to signed squares, while the real root equation gives exactly two simple negative roots of the auxiliary sextic. The normal form leaves exactly 2222 of the 7676 nontrivial partitions of 1212, thereby excluding 5454 complete irreducible degree patterns by one argument. Of these exclusions, 3737 follow from evenness alone and 1717 from the additional arithmetic restrictions. All local cases are proved explicitly, and an exact verification script accompanies the algebraic identities and exhaustive partition enumeration. The remaining patterns are necessary possibilities; their realization is not asserted, and full irreducibility remains open.

Introduction

The perfect cuboid problem asks for a rectangular box with positive integer edges for which all three face diagonals and the space diagonal are integers. Sharipov’s polynomial approach leads to three irreducibility conjectures [1]. The third concerns a degree-12 even polynomial depending on three positive integer parameters; its exceptional parameter relations and its connection with a Diophantine equation are described in [2].

Goal. We establish a common structural obstruction to complete factorizations of the third polynomial. Instead of excluding degree patterns individually, we obtain a normal form governed by the involution t↦−tt \mapsto-t, square classes of constant terms, and one uniform local factor. The resulting conditions exclude 54 of the 76 reducible degree partitions of twelve. Evenness alone excludes 37 of these partitions, since every odd degree must occur an even number of times. The arithmetic restrictions proved here reduce the remaining 39 patterns to 22. Sharipov’s structural theorem [2 Theorem 6.1] characterizes integer roots through 48 positive integer variables. Our result instead constrains the complete irreducible factorization; it does not exclude rational roots.

Method. The polynomial is a composition P(t)=F(t2)P(t)=F(t^2), where FF is a sextic given by a signed sum of three squared cubic expressions. Its root equation implies that every monic rational factor of FF has constant term in Q×2∪−Q×2\mathbb{Q}^{\times2}\cup-\mathbb{Q}^{\times2}. Over the reals, the same equation has precisely two negative roots. Over Q2\mathbb{Q}_2, an exhaustive valuation argument supplies an even irreducible factor of degree six or eight, or an even irreducible quartic with square constant term. An even local factor forces its containing rational irreducible factor to be even. If that rational factor is quartic, its signed-square constant must be positive. These inputs give a signed partition of six, from which the complete degree inventory follows.

The arguments apply to every positive pairwise distinct rational triple. In particular they apply in the ordered domain 0<u<a<b0<u<a<b. They do not require coprimality or the additional exclusions a2≠bua^2\ne bu and b2≠aub^2\ne au. Thus the conclusions also constrain the factorization on those two exceptional relations, without claiming irreducibility there. No result on the first or second cuboid conjecture is used.

Section 2 fixes the polynomial and states the normal-form theorem. Sections 3 and 4 prove its global and real inputs. Sections 5–8 give the full local construction. Section 9 proves the normal form and its exclusion consequences. Section 10 then collects all 54 excluded patterns in Table 1, grouped by the condition that rules them out. Only after this proved status has been recorded does Section 11 present the 22 patterns not excluded by the argument, in Table 2, and verify their formal compatibility with the normal form. The appendix records a reproducible exact verification script and its computed transcript.

Largest degreeComplete patterns not excludedCount
1010+210 + 2, 10+1[2]10 + 1^{[2]}2
88+48 + 4, 8+2[2]8 + 2^{[2]}; 8+2+1[2]8 + 2 + 1^{[2]}, 8+1[4]8 + 1^{[4]}4
66[2]6^{[2]}, 6+4+26 + 4 + 2, 6+4+1[2]6 + 4 + 1^{[2]}; 6+3[2]6 + 3^{[2]}, 6+2[3]6 + 2^{[3]}; 6+2[2]+1[2]6 + 2^{[2]} + 1^{[2]}, 6+2+1[4]6 + 2 + 1^{[4]}, 6+1[6]6 + 1^{[6]}8
44[3]4^{[3]}, 4[2]+2[2]4^{[2]} + 2^{[2]}, 4[2]+2+1[2]4^{[2]} + 2 + 1^{[2]}; 4[2]+1[4]4^{[2]} + 1^{[4]}, 4+2[4]4 + 2^{[4]}; 4+2[2]+1[4]4 + 2^{[2]} + 1^{[4]}, 4+3[2]+1[2]4 + 3^{[2]} + 1^{[2]}, 4+1[8]4 + 1^{[8]}8

Table 2. The 22 nontrivial complete degree patterns not excluded by Theorem 1. Their occurrence is not asserted.

The polynomial and the normal-form theorem

Throughout the global argument, let a,b,u∈Q>0a,b,u\in\mathbb{Q}_{>0} be pairwise distinct. Put

A=a2,B=b2,U=u2.(1)A=a^2,\qquad B=b^2,\qquad U=u^2. \tag*{(1)}

For arbitrary nonzero A,B,UA,B,U, define

FA,B,U(x)=((x−A)(x+B)(x+U))2+((x−B)(x+A)(x+U))2−((x−U)(x+A)(x+B))2.(2)F_{A,B,U}(x)=\bigl((x-A)(x+B)(x+U)\bigr)^2 +\bigl((x-B)(x+A)(x+U)\bigr)^2 -\bigl((x-U)(x+A)(x+B)\bigr)^2. \tag*{(2)}

We abbreviate this polynomial to FF when the parameters are fixed. It is monic of degree six. With S=A+BS=A+B and T=ABT=AB, its expansion is

F(x)=x6+(6U−2S)x5+(S2+4US+U2−14T)x4(3)F(x)=x^6+(6U-2S)x^5+(S^2+4US+U^2-14T)x^4 \tag*{(3)}
+(6US2−20UT−2S(U2+T))x3+(T2+U2S2−14TU2+4UST)x2+(6UT2−2U2ST)x+U2T2.(4)\begin{aligned} &+ (6US^{2}-20UT-2S(U^{2}+T))x^{3}\\ &+ (T^{2}+U^{2}S^{2}-14TU^{2}+4UST)x^{2}\\ &+ (6UT^{2}-2U^{2}ST)x+U^{2}T^{2}. \tag*{(4)} \end{aligned}

Sharipov’s third polynomial is

Pa,b,u(t)=Fa2,b2,u2(t2).(5)P_{a,b,u}(t)=F_{a^{2},b^{2},u^{2}}(t^{2}). \tag*{(5)}

Expansion of (3) gives the polynomial in [2 Eq. (1.1)]. In particular,

P(−t)=P(t),P(0)=F(0)=(ABU)2=(abu)4>0.(6)P(-t)=P(t),\qquad P(0)=F(0)=(ABU)^{2}=(abu)^{4}>0. \tag*{(6)}

Homogeneity gives

Pλa,λb,λu(λt)=λ12Pa,b,u(t),λ≠0.(7)P_{\lambda a,\lambda b,\lambda u}(\lambda t)=\lambda^{12}P_{a,b,u}(t),\qquad\lambda\ne0. \tag*{(7)}

Definition 1 (Reflection and complete degree pattern). For a monic polynomial hh of degree dd, its monic reflection is h†(t)=(−1)dh(−t)h^{\dagger}(t)=(-1)^{d}h(-t). A complete degree pattern is the partition of 12 formed by the degrees of all monic irreducible factors of PP over Q\mathbb{Q}, counted with multiplicities. The one-part pattern 12 means irreducibility and is excluded from the list of nontrivial patterns. The notation d[q]d^{[q]} means qq occurrences of the degree dd.

We use the standard terminology of integer partitions; see [3 Chapter 1].

Theorem 1 (Uniform normal form). For positive pairwise distinct rational aa, bb, uu, the complete factorization of P=Pa,b,uP=P_{a,b,u} over Q\mathbb{Q} can be written as

P(t)=∏i=1k(t2+si2)∏j=1rgj(t2)H(t)H†(t),(8)P(t)=\prod_{i=1}^{k}(t^{2}+s_{i}^{2})\prod_{j=1}^{r}g_{j}(t^{2})H(t)H^{\dagger}(t), \tag*{(8)}

where multiplicities are retained and the following statements hold:

  1. 0≤k≤20\leq k\leq2 and si∈Q×s_{i}\in\mathbb{Q}^{\times}.

  1. Each gjg_{j} is monic over Q\mathbb{Q}, has degree dj≥2d_{j}\geq2, and gj(t2)g_{j}(t^{2}) is irreducible. Moreover,

gj(0)=εjρj2,εj∈{1,−1},ρj∈Q×.g_{j}(0)=\varepsilon_{j}\rho_{j}^{2},\qquad\varepsilon_{j}\in\{1,-1\},\qquad\rho_{j}\in\mathbb{Q}^{\times}.
  1. HH is a product containing one member of each non-even reflected pair, with its multiplicity. Writing m=deg⁡Hm=\deg H, one has

k+∑j=1rdj+m=6,∏j=1rεj=(−1)m.(9)k+\sum_{j=1}^{r}d_{j}+m=6,\qquad\prod_{j=1}^{r}\varepsilon_{j}=(-1)^{m}. \tag*{(9)}
  1. There is an obligatory even factor:

some dj≥3orsome dj=2 with εj=1.(10)\text{some }d_{j}\geq3\quad\text{or}\quad\text{some }d_{j}=2\text{ with }\varepsilon_{j}=1. \tag*{(10)}

:::`

In particular r≥1r \ge1 and m≤4m \le4.

Corollary 1 (Complete numerical inventory). The necessary conditions in Theorem 1 leave exactly 22 nontrivial degree patterns. Consequently 54 of the 76 nontrivial partitions of 12 cannot be complete irreducible degree patterns of Pa,b,uP_{a,b,u}. Table 1 records all 54 exclusions, and Table 2 lists the 22 patterns not excluded by these conditions.

The count describes compatibility with proved necessary conditions. It does not assert that any remaining reducible pattern occurs for a permitted parameter triple. Arbitrary groupings of irreducible factors into larger reducible factors are not complete degree patterns.

Root valuations, square classes, and reflection

For V≠0V \ne0 write RV(x)=(x−V)/(x+V)R_V(x)=(x-V)/(x+V). Direct evaluation gives

F(−A)=4A2(A−B)2(A−U)2,F(−B)=4B2(A−B)2(B−U)2,F(−U)=−4U2(A−U)2(B−U)2.(11)\begin{aligned} F(-A)&=4A^2(A-B)^2(A-U)^2,\\ F(-B)&=4B^2(A-B)^2(B-U)^2,\\ F(-U)&=-4U^2(A-U)^2(B-U)^2. \tag*{(11)} \end{aligned}

Under pairwise distinctness all these values are nonzero. Hence every root of FF satisfies

RA(x)2+RB(x)2=RU(x)2.(12)R_A(x)^2+R_B(x)^2=R_U(x)^2. \tag*{(12)}

Lemma 1 (Valuations of roots). Let a,b,u∈Q×a,b,u \in\mathbb{Q}^{\times} have pairwise distinct squares. For every prime pp and every root θ\theta of FF in Q‾p\overline{\mathbb{Q}}_p, with vp(p)=1v_p(p)=1, one has

vp(θ)∈{2vp(a),2vp(b),2vp(u)}.(13)v_p(\theta)\in\{2v_p(a),2v_p(b),2v_p(u)\}. \tag*{(13)}

Every monic rational divisor gg of FF therefore satisfies

g(0)∈Q×2∪−Q×2.(14)g(0)\in\mathbb{Q}^{\times2}\cup-\mathbb{Q}^{\times2}. \tag*{(14)}

Proof. If vp(θ)v_p(\theta) differs from all three parameter-square valuations, each RV(θ)R_V(\theta) is an integral unit reducing to 11 or −1-1. Reduction of (12) then gives 1+1=11+1=1 in the residue field, a contradiction in every characteristic, including two. This proves (13). The valuation of g(0)g(0) is the sum of the valuations of its roots, with multiplicities, so it is even at every finite prime. A nonzero rational number having this property is a square up to sign. □

Lemma 2 (Even quadratics). Every even monic irreducible quadratic factor of PP over Q\mathbb{Q} has the form t2+s2t^2+s^2, with s∈Q×s\in\mathbb{Q}^{\times}.

Proof. If t2+ct^2+c divides P=F(t2)P=F(t^2), then −c-c is a rational root of FF. Lemma 1 gives −c=±s2-c=\pm s^2. The positive sign would split t2+c=t2−s2t^2+c=t^2-s^2 over Q\mathbb{Q}. Thus c=s2c=s^2. □

Lemma 3 (Reflection and descent of evenness). Every non-even monic irreducible factor hh of PP occurs with its distinct reflection h′h', with the same multiplicity. If a rational irreducible factor of PP contains a nonconstant even factor over Q2\mathbb{Q}_2, that rational irreducible factor is even. Moreover, no non-even rational irreducible factor has a purely imaginary root.

Proof. Reflection preserves irreducibility and multiplicity since PP is even. An irreducible factor fixed by reflection cannot have odd degree: it would be an odd polynomial divisible by tt, contrary to P(0)≠0P(0) \ne0. Thus it is even. For the second assertion, let κ\kappa be an even local factor of hh. It also divides h†h^\dagger. If h≠h†h \ne h^\dagger, their rational Bezout identity persists over Q2\mathbb{Q}_2 and rules out a common nonconstant factor. Hence h=h†h=h^\dagger is even. Finally, if hh had a purely imaginary root β\beta, its rational coefficients would make β‾=−β\overline{\beta}=-\beta a root as well. Then hh and h†h^\dagger would share a root, forcing h=h†h=h^\dagger by irreducibility.

Exactly two negative roots

Lemma 4 (Real root count). For positive pairwise distinct real AA, BB, UU, the sextic FA,B,UF_{A,B,U} has exactly two negative real roots, and both are simple. Consequently, for the rational-square parameters in (1), the polynomial PP has at most two even irreducible quadratic factors, counted with multiplicities.

Proof. By symmetry take A<BA < B. For z>0z>0 outside {A,B,U}\{A,B,U\} put

qV(z)=RV(z)RU(z),L(z)=qA(z)−2+qB(z)−2.q_V(z)=\frac{R_V(z)}{R_U(z)}, \qquad L(z)=q_A(z)^{-2}+q_B(z)^{-2}.

Then

qV′(z)=2(V−U)(z2+UV)(z+V)2(z−U)2,(15)q'_V(z)=\frac{2(V-U)(z^2+UV)}{(z+V)^2(z-U)^2}, \tag*{(15)}
F(−z)=(z−A)2(z−B)2(z+U)2(L(z)−1).(16)F(-z)=(z-A)^2(z-B)^2(z+U)^2(L(z)-1). \tag*{(16)}

The function LL extends across UU with value zero. The points AA, BB, UU themselves do not give roots of F(−z)F(-z), by (11).

If U<A<BU<A<B, both summands of LL strictly decrease on (0,U)(0,U), with LL going from 22 to 00, and strictly increase on (U,A)(U,A), with LL going from 00 to +∞+\infty. For z>Az>A one has 0<qA(z)<10<q_A(z)<1, so L(z)>1L(z)>1. If A<U<BA<U<B, both summands decrease on (A,U)(A,U), where LL goes from +∞+\infty to 00, and increase on (U,B)(U,B), where it goes from 00 to +∞+\infty. For 0<z<A0<z<A one has 0<qA(z)<10<q_A(z)<1, and for z>Bz>B one has 0<qB(z)<10<q_B(z)<1, excluding further roots. If A<B<UA<B<U, one has 0<qB(z)<10<q_B(z)<1 for 0<z<B0<z<B. Both summands decrease on (B,U)(B,U) from a total limit +∞+\infty to 00, and increase on (U,∞)(U,\infty) from 00 to 22.

These signs follow directly from (15) and (qV−2)′=−2qV′/qV3(q_V^{-2})'=-2q'_V/q_V^3. There are precisely two crossings of L=1L=1 in each ordering. At a crossing the two summand derivatives have the same nonzero sign. Thus the crossings, and hence the negative roots of FF, are simple.

By Lemma 2, an even irreducible quadratic of PP corresponds to a negative rational root −s2-s^2 of FF. There are at most two such roots, and they are simple, proving the final assertion.

The local alternative and its tools

In the local argument a,b,u∈Q2a,b,u\in\mathbb{Q}_2, and their squares are pairwise distinct. Write v=v2v=v_2, with v(2)=1v(2)=1, extended to the algebraic closure, and put K=Q2(i)K=\mathbb{Q}_2(i). We use the elementary Newton polygon and Hensel principles for local fields [4, 5]. Weierstrass preparation over Z2\mathbb{Z}_2 is used to isolate roots with positive valuation; the formal version needed here is given in [6 Corollary 1.2].

A Newton edge of length nn and slope −h-h supplies nn roots of valuation hh, counted with multiplicities. The reduced denominator of hh divides the ramification index of the field generated by any such root. Root clusters separated by valuation give factors over the ground field. For an integral power series with first unit coefficient in degree nn, Weierstrass preparation gives an integral power-series unit times a monic distinguished polynomial of degree nn. Its roots are the positive-valuation roots under consideration. In each use below the series is rational with denominator an analytic unit, so no root is lost on its domain.

Lemma 5 (Square class and degree doubling). Let M/Q2M/\mathbb{Q}_2 be finite. If y∈My \in M and v(y)>2v(y) > 2, then 1−y1-y is a square in MM. In particular, if

θ=−u2+q,M=Q2(q),v(q/u2)>2,\theta=-u^2+q,\qquad M=\mathbb{Q}_2(q),\qquad v(q/u^2)>2,

then θ=−u2s2\theta=-u^2s^2 for some s∈M×s \in M^\times. If i∉Mi \notin M and d=[M:Q2]d=[M:\mathbb{Q}_2], the minimal polynomial of a square root of θ\theta is even of degree 2d2d and has square constant term in Q2\mathbb{Q}_2.

Proof. Hensel’s lemma applied at 11 to Z2−(1−y)Z^2-(1-y) gives the first assertion because the error has valuation greater than twice the derivative valuation; see [5 Chapter II, §1.5]. Let ff be the minimal polynomial of θ\theta. The fields Q2(θ)\mathbb{Q}_2(\theta) and Q2(q)\mathbb{Q}_2(q) agree. If β2=θ=−u2s2\beta^2=\theta=-u^2s^2, then Q2(β)=M(i)\mathbb{Q}_2(\beta)=M(i) has degree 2d2d. Hence f(t2)f(t^2) is its minimal polynomial. Its constant term is

f(0)=(−1)dNM/Q2(θ)=u2dNM/Q2(s)2.(17)f(0)=(-1)^dN_{M/\mathbb{Q}_2}(\theta)=u^{2d}N_{M/\mathbb{Q}_2}(s)^2. \tag*{(17)}

Lemma 6 (Two facts about Q2(i)\mathbb{Q}_2(i)). The field K=Q2(i)K=\mathbb{Q}_2(i) has value group 12Z\frac{1}{2}\mathbb{Z}, residue field F2\mathbb{F}_2, and ring of integers Z2[i]\mathbb{Z}_2[i]. Every unit zz in this ring satisfies

z4≡1(mod8OK).(18)z^4\equiv1\pmod{8\mathcal{O}_K}. \tag*{(18)}

Furthermore −1-1 is not a norm from KK to Q2\mathbb{Q}_2.

Proof. The polynomial of 1+i1+i is X2−2X+2X^2-2X+2, which is Eisenstein; see also [5 Chapter II, §4.5] for the Gaussian 2-adic extension. For a unit z=r+isz=r+is, r,s∈Z2r,s\in\mathbb{Z}_2, exactly one of r,sr,s is odd. The real part of z2z^2 is odd and its imaginary part is divisible by 44, giving (18) after squaring. If r2+s2=−1r^2+s^2=-1 with a negative valuation among r,sr,s, unequal valuations give a uniquely minimal square, and equal valuations m<0m<0 give valuation 2m+1<02m+1<0 for their sum. Both are impossible. For integral r,sr,s reduction modulo four is impossible. Thus −1-1 is not a norm. Multiplication by a nonzero rational square does not change membership in the norm group.

Theorem 2 (Uniform local alternative). For nonzero a,b,u∈Q2a,b,u\in\mathbb{Q}_2 with pairwise distinct squares, Pa,b,uP_{a,b,u} has either an even irreducible factor of degree six or eight, or an even monic irreducible quartic with square constant term in Q2\mathbb{Q}_2. In the quartic case supplied by the construction, the root field is biquadratic over Q2\mathbb{Q}_2.

The proof occupies § 6–§ 8. By (6), dividing all parameters and tt by a common power of two preserves factor degrees and evenness. A quartic constant changes by a fourth power, preserving its square class.

A unique distinguished valuation

Lemma 7. If v(u)v(u) differs from both v(a)v(a) and v(b)v(b), then PP has an even irreducible quartic over Q2\mathbb{Q}_2 with root field Q2(i,2)\mathbb{Q}_2(i,\sqrt{2}) and square constant term.

Proof. Put t=u(1+z)t=u(1+z). For w=a,bw=a,b, choose

cw={w/u,v(w)>v(u),u/w,v(w)<v(u),ew=cw2/4∈Z2.c_w = \begin{cases} w/u, & v(w)>v(u),\\ u/w, & v(w)<v(u), \end{cases} \qquad e_w=c_w^2/4\in\mathbb{Z}_2.

Choose the integral numerator and denominator

(Nw,Dw)={((1+z)2−cw2,(1+z)2+cw2),v(w)>v(u),(cw2(1+z)2−1,cw2(1+z)2+1),v(w)<v(u).(N_w,D_w)= \begin{cases} ((1+z)^2-c_w^2,(1+z)^2+c_w^2), & v(w)>v(u),\\ (c_w^2(1+z)^2-1,c_w^2(1+z)^2+1), & v(w)<v(u). \end{cases}

Then Rw(t2)=Nw/DwR_w(t^2)=N_w/D_w, Dw(0)D_w(0) is a unit, and Nw2=Dw2−16ew(1+z)2N_w^2=D_w^2-16e_w(1+z)^2. With

D=Da2Db2,T=(1+z)2(eaDb2+ebDa2),D=D_a^2D_b^2,\qquad T=(1+z)^2(e_aD_b^2+e_bD_a^2),

the root equation becomes

G(z)=D(z)Q(z)+16T(z)(z2+2z+2)2=0,(19)G(z)=D(z)Q(z)+16T(z)(z^2+2z+2)^2=0, \tag*{(19)}

where

Q(z)=z2(z+2)2−2(z2+2z+2)2=−(z4+4z3+12z2+16z+8).Q(z)=z^2(z+2)^2-2(z^2+2z+2)^2=-(z^4+4z^3+12z^2+16z+8).

Writing G=∑gjzjG=\sum g_jz^j, one obtains

v(g0)=3,v(g1)≥3,v(g2)≥2,v(g3)≥2,v(g4)=0,v(g_0)=3,\qquad v(g_1)\geq3,\qquad v(g_2)\geq2,\qquad v(g_3)\geq2,\qquad v(g_4)=0,

and all higher coefficients are integral. For example, the constant correction is divisible by 64, and only −D(0)-D(0) contributes an odd degree-four coefficient. Thus the Newton polygon has an edge from (0,3)(0,3) to (4,0)(4,0), supplying exactly four roots with v(z)=3/4v(z)=3/4. All cleared denominators are nonzero there: each DwD_w is a unit, and v(z2+2z+2)=1v(z^2+2z+2)=1.

Each root has degree divisible by four. Since there are exactly four roots of this valuation, they constitute one irreducible quartic block. Reflection sends zz to −2−z-2-z and preserves this cluster, so its polynomial in tt is even.

For a root β=t\beta=t put θ=β2\theta=\beta^2 and M=Q2(θ)M=\mathbb{Q}_2(\theta). The irreducible even quartic gives [M:Q2]=2[M:\mathbb{Q}_2]=2. Set R=(θ−U)/(θ+U)R=(\theta-U)/(\theta+U) and w=T(z)/D(z)w=T(z)/D(z). Although written in zz, this last expression belongs to MM, since 16w=2−RA(θ)2−RB(θ)216w=2-R_A(\theta)^2-R_B(\theta)^2; it is integral. At a root of (18),

R2=2(1−8w),R2−1=1−16w.R^2=2(1-8w),\qquad R^2-1=1-16w.

Both 1−8w1-8w and 1−16w1-16w are squares in MM by Hensel’s lemma. Hence M=Q2(2)M=\mathbb{Q}_2(\sqrt{2}), and

(1+R)2R2−1=−θU\frac{(1+R)^2}{R^2-1}=-\frac{\theta}{U}

gives θ=−u2s2\theta=-u^2s^2 in MM. The field MM does not contain ii, so the root field of the quartic is M(i)=Q2(i,2)M(i)=\mathbb{Q}_2(i,\sqrt{2}). Its constant term is a square by the norm calculation (16). □

Exactly two parameters share the distinguished valuation

Lemma 8. Suppose v(a)=v(u)≠v(b)v(a)=v(u)\ne v(b) and a2≠u2a^2\ne u^2. After scaling a,ua,u to units put r=v(a2−u2)≥3r=v(a^2-u^2)\ge3. The polynomial PP has an even irreducible local factor of degree eight if r=3r=3, degree six if r>3r>3 and 3∤r3\nmid r, and degree four with square constant term if r>3r>3 and 3∣r3\mid r. In the last case its root field is biquadratic.

Proof. Take a,ua,u to be units, write C=A−UC=A-U, and set q=x+Uq=x+U. Since v(B)v(B) is a nonzero even integer,

RB(−U+q)2=1+16f(q),f(q)∈Z2[[q]].R_B(-U+q)^2=1+16f(q),\qquad f(q)\in\mathbb{Z}_2[[q]].

For v(B)>0v(B)>0 this follows from RB(x)2=1−4Bx/(x+B)2R_B(x)^2=1-4Bx/(x+B)^2; for v(B)<0v(B)<0 use the same expression after dividing numerator and denominator by BB. The denominators are analytic units on v(q)>0v(q)>0. The root equation is equivalent there to

g(q)+16f(q)q2(q+C)2=0,(20)g(q)+16f(q)\frac{q^2}{(q+C)^2}=0, \tag*{(20)}

where

g(q)=q4−2Cq3+C(C+12U)q2+4UC(C−2U)q−4U2C2.(21)g(q)=q^4-2Cq^3+C(C+12U)q^2+4UC(C-2U)q-4U^2C^2. \tag*{(21)}

The valuations of its coefficients in increasing degree are

2r+2,r+3,r+2,r+1,0.2r+2,\qquad r+3,\qquad r+2,\qquad r+1,\qquad0.

The perturbation in (20) is above the relevant lower edges. The full integral power series reduces to q4q^4 modulo two; Weierstrass preparation therefore supplies a degree-four factor capturing exactly its positive-valuation roots. The selected roots below satisfy q≠0q\ne0, −C-C, and x+B≠0x+B\ne0, so no cleared pole produces a root.

If r=3r=3, the Newton polygon is the segment (0,8)(0,8)–(4,0)(4,0). Set q=4zq=4z and divide by 282^8. The normalized equation is integral and its reduction is

z4+z+1∈F2[z].z^4+z+1\in\mathbb{F}_2[z].

It has no linear root and is not divisible by z2+z+1z^2+z+1, so it is irreducible. Hensel lifting gives a root in the unramified quartic extension LL, and the residue of zz generates its residue field; therefore Q2(q)=L\mathbb{Q}_2(q)=L. For θ=q−U\theta=q-U, one has θ/U≡−1(mod4)\theta/U\equiv-1\pmod{4} in LL. If θ/U\theta/U were a square, its square root would have residue 11, and hence be 1+2h1+2h because LL is unramified. Its square would be 11 modulo four, a contradiction. Thus adjoining θ\sqrt{\theta} doubles the degree to eight, and its minimal polynomial is the even polynomial obtained by composing that of θ\theta with t2t^2.

If r>3r>3, the lower vertices are

(0,2r+2),(1,r+3),(4,0).(0,2r+2),\qquad(1,r+3),\qquad(4,0).

When 3∤r3\nmid r, the last segment has length three and root valuation (r+3)/3(r+3)/3 of denominator three. It gives an irreducible cubic factor and a field LL of degree three. Its selected roots have valuation greater than two, and i∉Li\notin L. Lemma 5 gives an even degree-six factor of PP. Finally let r=3m>3r = 3m > 3, so m≥2m \ge2. Substitute q=2m+1zq = 2^{m+1}z and divide (19) by 24m+42^{4m+4}. All coefficients are integral. The coefficients of degrees zero through four have valuations 2m−2,0,m,2m,02m - 2, 0, m, 2m, 0, respectively; all coefficients of the normalized perturbation are divisible by two. The reduction is

z4+z=z(z+1)(z2+z+1).z^{4} + z = z(z + 1)(z^{2} + z + 1).

The simple irreducible quadratic factor lifts in the unramified quadratic extension LL; its nontrivial residue guarantees that Q2(q)=L\mathbb{Q}_{2}(q) = L. The substitutions in the power series converge on integral zz, so Hensel lifting is applicable. Now v(q)=m+1≥3v(q) = m + 1 \ge3 and i∉Li \notin L. Lemma 5 gives an even quartic with square constant term and root field L(i)L(i), which is biquadratic.

Three equal parameter valuations

Scale the parameters to units. Then AA, BB, U≡1(mod8)U \equiv1 \pmod{8}. Put

C=A−U,D=B−U,c=v(C),d=v(D),C = A - U,\qquad D = B - U,\qquad c = v(C),\qquad d = v(D),

where c,d≥3c,d \ge3. By symmetry assume c≤dc \le d. For q=x+Uq = x + U, write

F(−U+q)4U2=∑j=06ajqj.\frac{F(-U + q)}{4U^{2}} = \sum_{j=0}^{6} a_jq^j.

Exact expansion gives

a6=1/(4U2),a5=−(C+D+2U)/(2U2),a4=1+(C+D)/U+(C2−12CD+D2)/(4U2),a3=−CD(C+D−16U)/(2U2),a2=CD(CD+12U(C+D)−16U2)/(4U2),a1=CD(CD−2U(C+D))/U,a0=−C2D2.(22)\begin{aligned} a_6 &= 1/(4U^2),\\ a_5 &= -(C + D + 2U)/(2U^2),\\ a_4 &= 1 + (C + D)/U + (C^2 - 12CD + D^2)/(4U^2),\\ a_3 &= -CD(C + D - 16U)/(2U^2),\\ a_2 &= CD(CD + 12U(C + D) - 16U^2)/(4U^2),\\ a_1 &= CD(CD - 2U(C + D))/U,\\ a_0 &= -C^2D^2. \tag*{(22)} \end{aligned}

In particular,

v(a6)=−2,v(a5)=0,a4≡1(mod8),v(a3)≥c+d+2,v(a2)=c+d+2,v(a1)=2c+d+1(c<d),v(a1)≥3c+2(c=d).(23)\begin{aligned} v(a_6) &= -2,\qquad v(a_5) = 0,\qquad a_4 \equiv1 \pmod{8},\\ v(a_3) &\ge c + d + 2,\qquad v(a_2) = c + d + 2,\\ v(a_1) &= 2c + d + 1\quad(c < d),\qquad v(a_1) \ge3c + 2\quad(c = d). \tag*{(23)} \end{aligned}

The rightmost Newton edge is (4,0)(4,0)–(6,−2)(6,-2); all earlier points lie on or above the line of slope −3-3 through (4,0)(4,0). Thus there are two roots with v(q)=1v(q) = 1 and four with v(q)≥3v(q) \ge3. Let their monic cluster polynomials be VV, WW, so that

F(−U+q)=V(q)W(q),deg⁡V=2,deg⁡W=4.(24)F(-U + q) = V(q)W(q),\qquad\deg V = 2,\quad\deg W = 4. \tag*{(24)}

Lemma 9 (The four-root cluster does not split in KK). The four roots of WW do not all belong to K=Q2(i)K = \mathbb{Q}_{2}(i).

Proof. Let g=d−cg=d-c. From (22)–(23), the valuations of the four roots are

We consider these four possibilities.

If g=1g=1, write C=2cC0C=2^{c}C_{0}, D=2c+1D0D=2^{c+1}D_{0}, e=C0D0e=C_{0}D_{0}. A hypothetical root in KK has the form q=2c(1+i)zq=2^{c}(1+i)z with zz a unit. By (18),

v(q4−C2D2)=v(−24c+2(z4+e2))=4c+3.v(q^{4}-C^{2}D^{2})=v(-2^{4c+2}(z^{4}+e^{2}))=4c+3.

But v(a1q)=4c+52v(a_{1}q)=4c+\frac{5}{2}. All remaining terms, after grouping q4+a0q^{4}+a_{0}, have larger valuations, including (a4−1)q4(a_{4}-1)q^{4}. The minimum is unique, so the equation cannot hold.

If g=2g=2, substitute q=2c+1zq=2^{c+1}z and divide by 24c+42^{4c+4}. The reduction is z4+z+1z^{4}+z+1. A unit of KK reduces to 11, which is not a root of that polynomial. None of these four roots is in KK.

If g>2g>2, put h=c+g+13h=c+\frac{g+1}{3}. If hh is not integral, its denominator is three, so h∉Zh\notin\mathbb{Z}; the three roots of valuation hh cannot be in KK. If hh is integral, substitute q=2hzq=2^{h}z and divide by 24h2^{4h}. The resulting integral polynomial reduces to z4+zz^{4}+z. The three roots of valuation hh give units zz, but the only nonzero residue in F2\mathbb{F}_{2} is 11, a simple root. Hensel uniqueness permits only one such root in KK, so these three roots cannot all belong to KK.

It remains to treat g=0g=0. Put

C=2cC0,D=2cD0,e=C0D0,j=C0+D02,h=2c,C=2^{c}C_{0},\qquad D=2^{c}D_{0},\qquad e=C_{0}D_{0},\qquad j=\frac{C_{0}+D_{0}}{2},\qquad h=2^{c},

and

T(z)=F(−U+hz)4U2h4.T(z)=\frac{F(-U+hz)}{4U^{2}h^{4}}.

There is a useful direct identity

T(z)=[z(z+D0)(1+h(C0−z)2U)]2+[z(z+C0)(1+h(D0−z)2U)]2−[(z+C0)(z+D0)(1−hz2U)]2.(25)\begin{aligned} T(z)={}&\left[z(z+D_{0})\left(1+\frac{h(C_{0}-z)}{2U}\right)\right]^{2}\\ &+\left[z(z+C_{0})\left(1+\frac{h(D_{0}-z)}{2U}\right)\right]^{2}\\ &-\left[(z+C_{0})(z+D_{0})\left(1-\frac{hz}{2U}\right)\right]^{2}. \tag*{(25)} \end{aligned}

Since v(h)≥3v(h)\ge3, each additional factor in parentheses has square congruent to 11 modulo eight. Thus

T(z)≡z4−e2−4ez2−4ejz(mod8OK[z]).(26)T(z)\equiv z^{4}-e^{2}-4ez^{2}-4ejz\pmod{8\mathcal{O}_{K}[z]}. \tag*{(26)}

For a unit root z∈Kz\in K, equation (18) and e2≡1(mod8)e^{2}\equiv1\pmod{8} imply z+j∈2OKz+j\in2\mathcal{O}_{K}. If jj is even, this contradicts that zz is a unit. If jj is odd, it implies v(z−1)≥1v(z-1)\ge1. On the other hand (26) gives

T(1+w)≡w4+2w2(mod4OK[w]).T(1+w)\equiv w^{4}+2w^{2}\pmod{4\mathcal{O}_{K}[w]}.

The coefficient valuations, in increasing degree, satisfy

(≥2, ≥2, 1, ≥2, 0, ≥3, ≥4).(\ge2,\ \ge2,\ 1,\ \ge2,\ 0,\ \ge3,\ \ge4).

Thus every coefficient point (ℓ,v2([wℓ]T(1+w)))(\ell,v_2([w^\ell]T(1+w))) is on or above the supporting line y=2−ℓ/2y=2-\ell/2, which passes through (2,1)(2,1) and (4,0)(4,0). The Newton polygon therefore has an edge of slope −1/2-1/2 through (2,1)(2,1) and (4,0)(4,0), possibly extended left to (0,2)(0,2). There are at least two roots with v(w)=1/2v(w)=1/2. They give unit roots z=1+wz=1+w, hence roots of the four-root cluster, but cannot be in KK because roots there required v(z−1)≥1v(z-1)\ge1. This completes all cases. □\square

Lemma 10 (The norm of the four-root cluster). For the factorization (24),

V(0)∈Q2×2,W(0)∈−Q2×2.(27)V(0)\in\mathbb{Q}_2^{\times2},\qquad W(0)\in-\mathbb{Q}_2^{\times2}. \tag*{(27)}

There is a root qq of WW whose field M=Q2(q)M=\mathbb{Q}_2(q) has degree d0∈{2,3,4}d_0\in\{2,3,4\} and does not contain ii.

Proof. Define

H(Y)=F(U+4Y)256U4.\mathcal{H}(Y)=\frac{F(U+4Y)}{256U^4}.

The original three-square identity gives

H(Y)=[(Y−C/4)(1+D+4Y2U)(1+2YU)]2+[(Y−D/4)(1+C+4Y2U)(1+2YU)]2−[Y(1+C+4Y2U)(1+D+4Y2U)]2.(28)\begin{aligned} \mathcal{H}(Y) ={}&\left[(Y-C/4)\left(1+\frac{D+4Y}{2U}\right)\left(1+\frac{2Y}{U}\right)\right]^2\\ &+\left[(Y-D/4)\left(1+\frac{C+4Y}{2U}\right)\left(1+\frac{2Y}{U}\right)\right]^2\\ &-\left[Y\left(1+\frac{C+4Y}{2U}\right)\left(1+\frac{D+4Y}{2U}\right)\right]^2. \tag*{(28)} \end{aligned}

Every bracket is an integral polynomial congruent to YY modulo two. Its square is therefore congruent to Y2Y^2 modulo four, and

H(Y)≡Y2(mod4Z2[Y]).(29)\mathcal{H}(Y)\equiv Y^2\pmod{4\mathbb{Z}_2[Y]}. \tag*{(29)}

Weierstrass preparation gives

H(Y)=J(Y)(Y2+βY+e),\mathcal{H}(Y)=J(Y)(Y^2+\beta Y+e),

where JJ is an integral power-series unit and initially β,e∈2Z2\beta,e\in2\mathbb{Z}_2. The constant coefficient in (29) gives e∈4Z2e\in4\mathbb{Z}_2; comparison of linear coefficients then gives β∈4Z2\beta\in4\mathbb{Z}_2. The two distinguished roots Y1,Y2Y_1,Y_2 have positive valuation. They give the two roots qj=2U+4Yjq_j=2U+4Y_j of valuation one, and hence

V(0)=q1q2=4U2−8Uβ+16e=4U2(1−2β/U+4e/U2).V(0)=q_1q_2=4U^2-8U\beta+16e=4U^2\left(1-2\beta/U+4e/U^2\right).

The parenthesis is 11 modulo eight, so it is a square. Using (10) and (24),

W(0)=F(−U)V(0)=−4U2C2D2V(0)∈−Q2×2.W(0)=\frac{F(-U)}{V(0)}=\frac{-4U^2C^2D^2}{V(0)}\in-\mathbb{Q}_2^{\times2}.

Choose a root q∉Kq \notin K using Lemma 9, and put M=Q2(q)M = \mathbb{Q}_2(q), d0=[M:Q2]≤4d_0 = [M : \mathbb{Q}_2] \le4. Then d0≥2d_0 \ge2. For d0=2d_0 = 2, the fact that q∉Kq \notin K excludes M=KM = K; for d0=3d_0 = 3, degree parity excludes K⊂MK \subset M. If d0=4d_0 = 4 and K⊂MK \subset M, then WW is the minimal polynomial of qq and

W(0)=NM/Q2(q)=NK/Q2(NM/K(q)).W(0) = N_{M/\mathbb{Q}_2}(q) = N_{K/\mathbb{Q}_2}\left(N_{M/K}(q)\right).

This is impossible by (27) and Lemma 6. Thus i∉Mi \notin M in every case.

Proof of Theorem 2. If v(u)v(u) differs from both other parameter valuations, apply Lemma 7. If it agrees with exactly one, interchange a,ba,b if necessary and use Lemma 8. If all three agree, scale to units and choose q,M,d0q,M,d_0 as in Lemma 10. Since v(q)≥3v(q) \ge3 and UU is a unit, Lemma 5 applied to θ=−U+q\theta=-U+q gives an even irreducible factor of degree 2d02d_0 with square constant term. For d0=3,4d_0=3,4 this gives degree six or eight. For d0=2d_0=2, its root field M(i)M(i) is the compositum of two distinct quadratic extensions, hence biquadratic. The three valuation cases exhaust all parameter triples.

Globalization and proof of the normal form

Proposition 1 (The obligatory global factor). For positive pairwise distinct rational a,b,ua,b,u, the polynomial PP has an even irreducible factor of degree at least six or an even irreducible quartic t4+pt2+ρ2t^4+pt^2+\rho^2, with p∈Qp\in\mathbb{Q} and ρ∈Q×\rho\in\mathbb{Q}^{\times}.

Proof. Choose a local factor κ\kappa from Theorem 2, and a monic rational irreducible factor hh of PP divisible by κ\kappa over Q2\mathbb{Q}_2. Lemma 3 forces hh to be even. If deg⁡κ≥6\deg\kappa\ge6 or deg⁡h>4\deg h>4, then deg⁡h≥6\deg h\ge6. Otherwise deg⁡h=deg⁡κ=4\deg h=\deg\kappa=4, so h=κh=\kappa is an even quartic with square constant term in Q2\mathbb{Q}_2. Write h(t)=g(t2)h(t)=g(t^2), g=x2+px+cg=x^2+px+c. Polynomial division shows that g∣Fg\mid F: a remainder of degree less than two would, after composition, be divisible by hh and have degree less than four. Lemma 1 gives c=±ρ2c=\pm\rho^2 over Q\mathbb{Q}. The negative sign is incompatible with c∈Q2c\in\mathbb{Q}^{2} because −1-1 is not a square in Q2\mathbb{Q}_2. Thus c=ρ2>0c=\rho^2>0.

Proof of Theorem 1. Separate the monic rational irreducible factors fixed by reflection from the distinct reflected pairs. Choose one member from each pair, with multiplicities, and let HH be their product. The other members have product H†H^\dagger. The fixed irreducibles are even by Lemma 3. Their quadratic members have the form t2+si2t^2+s_i^2 by Lemma 2, and their number kk is at most two by Lemma 4.

Write every remaining even irreducible as gj(t2)g_j(t^2), with gjg_j monic of degree dj≥2d_j\ge2. The same polynomial-division argument as above shows gj∣Fg_j\mid F; it is irreducible since a factorization of gjg_j would factor gj(t2)g_j(t^2). Lemma 1 gives gj(0)=εjρj2g_j(0)=\varepsilon_j\rho_j^2. This proves (8). Degree comparison gives the first equation in (9).

Since

H(0)H†(0)=(−1)mH(0)2H(0)H^\dagger(0)=(-1)^m H(0)^2

and P(0)>0P(0)>0, comparison of constant terms gives the second equation. Finally Proposition 1 gives (10). It forces ∑dj≥2\sum d_j\ge2, hence r≥1r\ge1 and m≤4m\le4.

Corollary 2 (Degree and parity consequences). Every non-even irreducible rational factor of PP has degree at most four. In particular no irreducible odd degree above three occurs. If all factors have degree at most four and exactly one quartic occurs, then the number of quadratic factors is even.

Proof. The first statements follow from m≤4m \le4, since each non-even factor appears in HH or H†H^{\dagger}, and an even irreducible cannot have odd degree. Under the last hypothesis, the unique quartic is even by reflection. It is the only factor gj(t2)g_j(t^2) in (7); thus r=1r = 1, d1=2d_1 = 2, and ε1=1\varepsilon_1 = 1 by (9). The sign equation makes mm even, and k+2+m=6k + 2 + m = 6 makes kk even. Every other quadratic occurs in a reflected pair. Hence the total number of quadratics is even. □

Corollary 3 (A common sign obstruction). The three complete patterns

4+3+3+2,4+2+2+2+1+1,4+2+1[6]4 + 3 + 3 + 2,\qquad4 + 2 + 2 + 2 + 1 + 1,\qquad4 + 2 + 1^{[6]}

are impossible for every positive pairwise distinct rational triple.

Proof. Each has a unique quartic, no larger factor, and an odd number of quadratics, contrary to Corollary 2. □

Remark 1 (The role of all finite primes). A rational number that is a square in Q2\mathbb{Q}_2 need not be positive; for example −7-7 is a square there. The positivity in Proposition 1 uses both the global signed-square statement (13), proved at every finite prime, and the local square class at two. Similarly, a local degree partition is not automatically a rational degree partition. Reflection and the persistent Bezout identity are what allow the local even factor to control a global irreducible.

Excluded complete degree patterns

We now translate the proved factor restrictions into a complete inventory of exclusions. Throughout this inventory, a pattern records all irreducible factors over Q\mathbb{Q}, with multiplicities, as in the definition in Section 2. Each exclusion holds for every positive pairwise distinct rational triple. The three steps below are applied successively, so that a pattern already excluded at one step is not counted again at a later step.

(I) Reflection. By Lemma 3, every irreducible factor of odd degree occurs together with a distinct reflected factor of the same degree and multiplicity. Consequently each odd degree must occur an even number of times in a complete pattern. Among the 76 nontrivial partitions of twelve, exactly 37 fail this condition. They are already excluded by evenness, before any additional arithmetic input is used.

(II) The obligatory even factor. Of the 39 patterns that pass the reflection test, a further 14 contain no even degree of size at least four. Proposition 1 excludes all of them: every complete factorization must contain an even irreducible factor of degree at least six or an even irreducible quartic with positive square constant term. In particular, the degree pattern must contain an even degree at least four. Passing this degree test does not yet establish the required evenness or constant term of a particular factor; it only identifies the patterns that this step has not excluded.

(III) The sign condition for a unique quartic. There are now 25 patterns left. Corollary 2 requires an even number of quadratic factors whenever there is exactly one quartic and no larger factor. Corollary 3 explicitly excludes the three patterns violating this requirement. This last step uses the sign equation and the positive-square constant of the obligatory quartic, rather than reflection alone.

Thus the successive exclusions number 37+14+3=5437+14+3=54, leaving 76−54=2276-54=22 nontrivial degree patterns to examine for compatibility with the normal form. These are successive counts of abstract degree patterns, not independent Diophantine obstructions. Table 1 records the full proved exclusion status: every displayed pattern is impossible, and the heading of each group identifies the result that excludes it. Within each group, patterns are arranged by their largest degree.

ggroot valuations
00c,c,c,cc,c,c,c
11c+12,c+12,c+12,c+12c+\frac{1}{2},c+\frac{1}{2},c+\frac{1}{2},c+\frac{1}{2}
22c+1,c+1,c+1,c+1c+1,c+1,c+1,c+1
g>2g>2d−1,c+g+13,c+g+13,c+g+13d-1,c+\frac{g+1}{3},c+\frac{g+1}{3},c+\frac{g+1}{3}

Table 1.

Strictly excluded complete patternsCount
(I) Reflection: Lemma 337
11+111+1, 9+39+3, 9+2+19+2+1
9+1[3]9+1^{[3]}, 8+3+18+3+1, 7+57+5
7+4+17+4+1, 7+3+27+3+2, 7+3+1[2]7+3+1^{[2]}
7+2[2]+17+2^{[2]}+1, 7+2+1[3]7+2+1^{[3]}, 7+1[5]7+1^{[5]}
6+5+16+5+1, 6+3+2+16+3+2+1, 6+3+1[3]6+3+1^{[3]}
5+4+35+4+3, 5+4+2+15+4+2+1, 5+4+1[3]5+4+1^{[3]}
5+3[2]+15+3^{[2]}+1, 5+3+2[2]5+3+2^{[2]}, 5+3+2+1[2]5+3+2+1^{[2]}
5+3+1[4]5+3+1^{[4]}, 5+2[3]+15+2^{[3]}+1, 5+2[2]+1[3]5+2^{[2]}+1^{[3]}
5+2+1[5]5+2+1^{[5]}, 5+1[7]5+1^{[7]}, 4[2]+3+14^{[2]}+3+1
4+3+2[2]+14+3+2^{[2]}+1, 4+3+2+1[3]4+3+2+1^{[3]}, 4+3+1[5]4+3+1^{[5]}
3[3]+2+13^{[3]}+2+1, 3[3]+1[3]3^{[3]}+1^{[3]}, 3+2[4]+13+2^{[4]}+1
3+2[3]+1[3]3+2^{[3]}+1^{[3]}, 3+2[2]+1[5]3+2^{[2]}+1^{[5]}, 3+2+1[7]3+2+1^{[7]}
3+1[9]3+1^{[9]}
(II) Obligatory even factor: Proposition 114
5[2]+25^{[2]}+2, 5[2]+1[2]5^{[2]}+1^{[2]}, 3[4]3^{[4]}
3[2]+2[3]3^{[2]}+2^{[3]}, 3[2]+2[2]+1[2]3^{[2]}+2^{[2]}+1^{[2]}, 3[2]+2+1[4]3^{[2]}+2+1^{[4]}
3[2]+1[6]3^{[2]}+1^{[6]}, 2[5]+1[2]2^{[5]}+1^{[2]}
2[4]+1[4]2^{[4]}+1^{[4]}, 2[3]+1[6]2^{[3]}+1^{[6]}, 2[2]+1[8]2^{[2]}+1^{[8]}
2+1[10]2+1^{[10]}, 1[12]1^{[12]}
(III) Unique-quartic sign: Corollary 33
4+3[2]+24+3^{[2]}+2, 4+2[3]+1[2]4+2^{[3]}+1^{[2]}, 4+2+1[6]4+2+1^{[6]}
Total strictly excluded54

Table 1 (PDF p. 14). All 54 strictly excluded complete irreducible degree patterns. The three groups are successive and disjoint.

The twenty-two remaining degree patterns

After all exclusions in Table 1, exactly 22 nontrivial partitions remain. Table 2 lists them separately to distinguish the proved impossibility of the first table from the fact that the patterns in the second are not excluded. Each entry is compatible with the necessary conditions of Theorem 1; this is a statement about abstract factor data, not the existence of a permitted parameter triple realizing the pattern. The enumeration and this formal compatibility are verified immediately after the table.

Proof of Corollary 1. The normal form provides a direct finite enumeration. Choose k∈{0,1,2}k \in\{0,1,2\}, a nonempty list of degrees dj≥2d_j \ge2, and a partition (h1,…,hs)(h_1,\ldots,h_s) of

m=6−k−∑dj≥0.m = 6-k-\sum d_j \ge0.

The corresponding degree pattern is the multiset

2,…,2⏟k times,2d1,…,2dr,h1,h1,…,hs,hs.(30)\underbrace{2,\ldots,2}_{k\ \mathrm{times}},\quad2d_1,\ldots,2d_r,\quad h_1,h_1,\ldots,h_s,h_s. \tag*{(30)}

Keep only lists admitting signs satisfying (8) and (9); remove the one-part pattern 12. If some dj≥3d_j \ge3, the obligatory-factor condition imposes no sign restriction. If all dj=2d_j = 2 and r≥2r \ge2, their signs can have either prescribed product while at least one sign is positive. If r=1r=1 and d1=2d_1=2, its sign must be positive and mm must be even. These observations completely describe the finite enumeration.

For an explicit count, let MM be the largest degree in a remaining nontrivial pattern. Corollary 2 and the obligatory factor give M∈{4,6,8,10}M \in\{4,6,8,10\}. For M=10M=10, the remaining degree two is 2 or 1+11+1. For M=8M=8, reflection leaves the four possibilities 4, 2+22+2, 2+1+12+1+1, and 1[4]1^{[4]}. For M=6M=6, there may be a second sextic; otherwise the remaining degree six has paired odd degrees and gives seven possibilities.

For M=4M=4, three quartics give one pattern, and two quartics leave 2+22+2, 2+1+12+1+1, or 1[4]1^{[4]}. For a unique quartic, let njn_j be the number of degree-jj factors. Then n1,n3n_1,n_3 are even by reflection and n2n_2 is even by Corollary 2. Solving 3n3+2n2+n1=83n_3 + 2n_2 + n_1 = 8 gives four possibilities: 3+3+1+13 + 3 + 1 + 1, 2[4]2^{[4]}, 2[2]+1[4]2^{[2]} + 1^{[4]}, or 1[8]1^{[8]}. Together, these cases give the eight entries with largest degree four listed in Table 2.

For completeness, all listed patterns are compatible with the abstract normal-form constraints. For largest degree at least six, take the displayed factors of degree at least four to be even and pair all cubics and linears. If the number of quadratics is odd, take one to be even and pair the others; if it is even, pair them all. The signs can be chosen as above because a dj≥3d_j \ge3 is present. For largest degree four, the three patterns with multiple quartics and small remaining degrees, together with 4[3]4^{[3]}, admit the same choice and appropriate signs. For a unique quartic take its sign positive; in each listed case choose even kk and pair the remaining quadratics. This satisfies both equations and the obligatory-factor condition. These are formal compatibility choices, not constructions of parameter triples.

There are 77 partitions of twelve, including the one-part partition 12, and hence 76 nontrivial partitions. Table 2 has 2+4+8+8=222 + 4 + 8 + 8 = 22 distinct entries. Together with the 54 disjoint exclusions in Table 1, these entries account for all 76 nontrivial partitions. Since each of the 22 entries admits the formal choices just described, none is excluded by the stated normal-form conditions. This proves the exact numerical inventory without asserting that any remaining pattern occurs for the polynomial family. Appendix § 13 independently reproduces the enumeration from (30) and verifies the completeness of the underlying partition list.

Scope of the result

The normal form is a uniform necessary condition for complete rational factorization. Its proof uses the same arithmetic mechanism for all parameter triples: root valuations at finite primes, a real root count, and a forced even local factor at two. In particular, the degree bound on non-even factors follows from the obligatory block and does not require a separate quintic-factor argument. The unique-quartic parity restriction follows from the same normal form and its sign equation.

No global squarefreeness theorem is needed: reflection preserves multiplicities, the normal form retains them, and the two negative roots used to bound the even quadratics are simple. Nor is the complete classification of all local factors required. The local proof selects one factor with sufficient degree and, when needed, sufficient control of its constant square class.

Table 1 therefore records 54 proved impossibilities, whereas Table 2 records 22 possibilities not excluded by the present argument. The latter include patterns with rational roots, quadratic factors, and cubic factors. Their compatibility with the stated conditions does not imply their existence for this family. Excluding them requires additional arithmetic information. The argument proves neither pointwise irreducibility of the third polynomial nor nonexistence of perfect cuboids.

Exact verification script and transcript

The proof in the main text is independent of a bounded parameter search. The following script uses Python and SymPy [7] with exact rational arithmetic. The script has been verified with Python 3.12.14 and SymPy 1.14.0. It checks the sextic identities, the shifted coefficients, the congruence divisibilities used in the local proof, and the signed normal-form enumeration. The partition list is generated exhaustively and checked against two independent counting procedures; all reported counts and lists are computed. Newton polygon deductions, Hensel lifting, and the field and norm arguments are proved in the text. The partition enumeration concerns abstract degree data and is finite; it does not sample the parameters aa, bb, uu.

Script: Complete exact verification

Code.

"""
Exact identities and exhaustive degree-pattern verification.
Requires Python 3.8+ and SymPy. All counts and lists are computed.
"""

from collections import Counter
from dataclasses import asdict, dataclass
from itertools import product
from math import prod
import sympy as s

def zero(expression):
    assert s.cancel(expression) == 0

def integral_coefficients(expression, *variables):
    polynomial = s.Poly(s.expand(expression), *variables, domain=s.QQ)
    assert all(s.denom(coefficient) == 1
               for coefficient in polynomial.coeffs())

x, z, A, B, U, V = s.symbols("x z A B U V")
F = (((x-A)*(x+B)*(x+U))**2 + ((x-B)*(x+A)*(x+U))**2 - ((x-U)*(x+A)*(x+B))**2)
S, Tpar = A+B, A*B
expanded_F = (x**6+(6*U-2*S)*x**5
             +(S*S+4*U*S+U*U-14*Tpar)*x**4
             +(6*U*S*S-20*U*Tpar-2*S*(U*U+Tpar))*x**3
             +(Tpar*Tpar+U*U*S*S-14*Tpar*U*U+4*U*S*Tpar)*x*x
             +(6*U*Tpar*Tpar-2*U*U*S*Tpar)*x+U*U*Tpar*Tpar)
zero(F-expanded_F)
lam = s.symbols("lam")
zero(F.subs({x:lam**2*x, A:lam**2*A,
             B:lam**2*B, U:lam**2*U}, simultaneous=True)-lam**12*F)
zero(s.Poly(F, x).LC()-1)
zero(F.subs(x, 0)-(A*B*U)**2)
zero(F.subs(x, -A)-4*A*A*(A-B)**2*(A-U)**2)
zero(F.subs(x, -B)-4*B*B*(A-B)**2*(B-U)**2)
zero(F.subs(x, -U)+4*U*U*(A-U)**2*(B-U)**2)
RA, RB, RU = [(x-W)/(x+W) for W in (A, B, U)]
zero(F-(x+A)**2*(x+B)**2*(x+U)**2*(RA**2+RB**2-RU**2))

# The real-root identity and its derivative, valid in all three orders.
RVz, RUz = (z-V)/(z+V), (z-U)/(z+U)
qV = RVz/RUz
zero(s.diff(qV, z) - 2*(V-U)*(z*z+U*V)/((z+V)**2*(z-U)**2))
RAz, RBz = (z-A)/(z+A), (z-B)/(z+B)
L = (RUz/RAz)**2+(RUz/RBz)**2
zero(F.subs(x, -z)
     - (z-A)**2*(z-B)**2*(z+U)**2*(L-1))
# The sign of q_V-1 supplies the outside-interval exclusions.
zero(qV-1-2*z*(U-V)/((z+V)*(z-U)))

# Exact expansion near the distinguished point x=-U.
q, C, D = s.symbols("q C D")
FCD = s.expand(F.subs({A:U+C, B:U+D}))
shifted = s.expand(FCD.subs(x, q-U))
expected = [
    -4*C*C*D*D*U*U,
    4*C*D*U*(C*D-2*C*U-2*D*U),
    C*D*(C*D+12*C*U+12*D*U-16*U*U),
    -2*C*D*(C+D-16*U),
    C*C-12*C*D+4*C*U+D*D+4*D*U+4*U*U,
    -2*(C+D+2*U),
    1,
]
zero(shifted-sum(value*q**j for j, value in enumerate(expected)))

# The shared-valuation quartic before the integral perturbation.
E = ((x-U)**2*(x+U+C)**2
    - (x-U-C)**2*(x+U)**2
    - (x+U+C)**2*(x+U)**2)

g = (q**4-2*C*q**3+C*(C+12*U)*q*q
     +4*U*C*(C-2*U)*q-4*U*U*C*C)
zero(E.subs(x, q-U)+g)

# The unique-valuation quartic and its root-field identity.
Q = z*z*(z+2)**2-2*(z*z+2*z+2)**2
zero(Q+z**4+4*z**3+12*z*z+16*z+8)
R = (x-U)/(x+U)
zero((1+R)**2/(R**2-1)+x/U)

# A short exact formula for the outer two-root preparation.
Y = s.symbols("Y")
H = s.cancel(FCD.subs(x, U+4*Y)/(256*U**4))
Hshort = (
    ((Y-C/4)*(1+(D+4*Y)/(2*U))*(1+2*Y/U))**2
    + ((Y-D/4)*(1+(C+4*Y)/(2*U))*(1+2*Y/U))**2
    - (Y*(1+(C+4*Y)/(2*U))*(1+(D+4*Y)/(2*U)))**2
)
zero(H-Hshort)
c, d = s.symbols("c d")
Hunit = s.Poly(s.expand(H.subs({C:8*c, D:8*d})), Y)
# Multiplication by U^4 preserves integrality for an odd unit U.
for j, divisor in enumerate([4, 4, 1, 8, 8, 32, 16]):
    integral_coefficients(Hunit.nth(j)*U**4/divisor, U, c, d)
integral_coefficients((Hunit.nth(2)-1)*U**4/4, U, c, d)
b, e = s.symbols("b e")
outer_factor = (q-2*U)**2+4*b*(q-2*U)+16*e
zero(s.resultant(Y*Y+b*Y+e, q-2*U-4*Y, Y)-outer_factor)
zero(outer_factor.subs(q, 0)-(4*U*U-8*U*b+16*e))

# Equal valuation of C and D: the product identity makes the mod-8
# congruence transparent. Put C=h*c, D=h*d, with v_2(h)>=3.
h, eta = s.symbols("h eta")
T = s.cancel(FCD.subs({x:-U+h*z, C:h*c, D:h*d})
             /(4*U*U*h**4))
Tshort = (
    (z*(z+d)*(1+h*(c-z)/(2*U)))**2
    + (z*(z+c)*(1+h*(d-z)/(2*U)))**2
    - ((z+c)*(z+d)*(1-h*z/(2*U)))**2
)
zero(T-Tshort)
en, jn = c*d, (c+d)/2
T0 = z**4-en**2-4*en*z*z-4*en*jn*z
zero(Tshort.subs(h, 0)-T0)
# h=8*eta proves every coefficient of T-T0 is divisible by eight
# over Z_2[U,U^{-1},c,d]; this is not a sample parameter check.
integral_coefficients((T-T0).subs(h, 8*eta)*U*U/8,
                      U, eta, c, d, z)
t, p, constant = s.symbols("t p constant")
zero(s.discriminant(t**4+p*t*t+constant, t)
    -16*constant*(p*p-4*constant)**2)
print("PASS: exact polynomial, real-root, and dyadic coefficient identities.")

# Exhaustive inventory and independent completeness checks.
DEGREE = 12
HALF_DEGREE = DEGREE // 2
MAX_EVEN_QUADRATICS = 2
NEGATIVE_ROOTS_OF_SEXTIC = 2

class VerificationError(Exception):
    pass

def require(condition, message):
    # Deliberately not an assert: these checks also run with python -O.
    if not condition:
        raise VerificationError(message)

def partitions(total, minimum=1):
    if total == 0:
        yield ()
        return
    for first in range(minimum, total + 1):
        for tail in partitions(total - first, first):
            yield (first,) + tail

def count_partitions_by_length(total):
    counts = [[0] * (total + 1) for _ in range(total + 1)]
    counts[0][0] = 1
    for n in range(1, total + 1):
        for k in range(1, n + 1):
            counts[n][k] = counts[n - 1][k - 1] + counts[n - k][k]
    return counts[total]

def count_partitions_by_product(total):
    coefficients = [1] + [0] * total
    for part in range(1, total + 1):
        for n in range(part, total + 1):
            coefficients[n] += coefficients[n - part]
    return coefficients[total]

def checked_partitions(total):
    for n in range(total + 1):
        generated = list(partitions(n))
        require(len(generated) == len(set(generated)),
                'Repeated partition at degree {}'.format(n))
        for pattern in generated:
            require(sum(pattern) == n and all(d > 0 for d in pattern),
                    'Invalid partition: {!r}'.format(pattern))
            require(tuple(sorted(pattern)) == pattern,
                    'Unordered partition: {!r}'.format(pattern))
        observed = Counter(map(len, generated))
        counted = count_partitions_by_length(n)
        require(all(observed[k] == counted[k] for k in range(n + 1)),
                'Counts by length disagree at degree {}'.format(n))
        require(len(generated) == sum(counted) == count_partitions_by_product(n),
                'Independent partition counts disagree at degree {}'.format(n))
    # Convert to descending notation only AFTER the completeness checks.
    return tuple(sorted((tuple(reversed(p)) for p in generated), reverse=True))

@dataclass(frozen=True)
class Profile:
    k: int
    ds: tuple
signs: tuple
hs: tuple
negative_roots: tuple

@property
def m(self):
    return sum(self.hs)

def reflection_assignments(pattern):
    multiplicities = Counter(pattern)
    degrees = sorted(multiplicities, reverse=True)
    choices = []
    for degree in degrees:
        count = multiplicities[degree]
        if degree % 2:
            if count % 2:
                return
            choices.append((0,))
        else:
            limit = min(count, MAX_EVEN_QUADRATICS) if degree == 2 else count
            choices.append(tuple(f for f in range(limit + 1) if (count - f) % 2 == 0))
    for fixed_counts in product(*choices):
        k, ds, hs = 0, [], []
        for degree, fixed in zip(degrees, fixed_counts):
            if degree == 2:
                k = fixed
            elif fixed:
                ds.extend([degree // 2] * fixed)
                hs.extend([degree] * ((multiplicities[degree] - fixed) // 2))
        yield k, tuple(ds), tuple(hs)

def real_signature(ds, signs, k):
    for negatives in product(*(range(d + 1) for d in ds)):
        if (sum(negatives) == NEGATIVE_ROOTS_OF_SEXTIC - k
                and all(sign == (-1) ** (d + nu)
                        for d, sign, nu in zip(ds, signs, negatives))):
            return tuple(negatives)
    return None

def normal_form_profiles(pattern):
    profiles = []
    for k, ds, hs in reflection_assignments(pattern):
        m = sum(hs)
        require(k + sum(ds) + m == HALF_DEGREE,
                'Degree bookkeeping failed for {!r}'.format(pattern))
        if not ds:
            continue
        for signs in product((-1, 1), repeat=len(ds)):
            if prod(signs) != (-1) ** m:
                continue
            obligatory_factor = any(d >= 3 or (d == 2 and sign == 1)
                                    for d, sign in zip(ds, signs))
            if not obligatory_factor:
                continue
            negatives = real_signature(ds, signs, k)
            # This independently verifies that the exact real root count
            # does not further reduce the signed profiles in this inventory.
            require(negatives is not None,
                    'A signed profile fails the real-root test: {!r}'.format(
                        (pattern, k, ds, signs, hs)))
            profiles.append(Profile(k, ds, tuple(signs), hs, negatives))
    return tuple(profiles)

def exclusion_rule(pattern):
    counts = Counter(pattern)
    unpaired = sorted((d for d, count in counts.items() if d % 2 and count % 2),
reverse=True)
if unpaired:
    return 'I', 'odd multiplicity of odd degree(s): ' + ', '.join(map(str, unpaired))
if not any(d >= 4 and d % 2 == 0 for d in pattern):
    return 'II', 'no even degree >= 4'
if max(pattern) <= 4 and counts[4] == 1 and counts[2] % 2:
    return 'III', 'unique quartic; odd number of quadratics'
return None, None

def verify_profile(pattern, profile):
    require(len(profile.ds) == len(profile.signs) == len(profile.negative_roots),
            'Witness vector lengths disagree')
    require(all(sign in (-1, 1) for sign in profile.signs),
            'Witness has an invalid constant-term sign')
    require(all(h > 0 for h in profile.hs), 'Witness has a nonpositive H degree')
    reconstructed = ([2] * profile.k + [2 * d for d in profile.ds]
                     + [h for h in profile.hs for _ in range(2)])
    require(tuple(sorted(reconstructed, reverse=True)) == pattern,
            'Witness does not reconstruct {}'.format(pattern))
    require(0 <= profile.k <= MAX_EVEN_QUADRATICS and bool(profile.ds),
            'Invalid quadratic bound or missing obligatory factor')
    require(all(d >= 2 for d in profile.ds), 'Invalid higher even degree')
    require(profile.k + sum(profile.ds) + profile.m == HALF_DEGREE,
            'Witness degree equation failed')
    require(prod(profile.signs) == (-1) ** profile.m, 'Witness sign equation failed')
    require(any(d >= 3 or (d == 2 and sign == 1)
                for d, sign in zip(profile.ds, profile.signs)),
            'Witness obligatory-factor condition failed')
    require(sum(profile.negative_roots) + profile.k == NEGATIVE_ROOTS_OF_SEXTIC,
            'Witness real-root count failed')
    require(all(0 <= nu <= d and sign == (-1) ** (d + nu)
                for d, sign, nu in zip(profile.ds, profile.signs,
                                        profile.negative_roots)),
            'Witness real-root signs failed')

def audit():
    universe = checked_partitions(DEGREE)
    records = []
    for pattern in universe:
        profiles = normal_form_profiles(pattern)
        stage, reason = exclusion_rule(pattern)
        require(bool(profiles) == (stage is None),
                'The two classification methods disagree for {!r}'.format(pattern))
        for profile in profiles:
            verify_profile(pattern, profile)
            records.append({
                'pattern': pattern,
                'status': ('irreducible' if len(pattern) == 1
                           else 'retained' if profiles else 'excluded'),
                'exclusion_stage': stage,
                'reason': reason,
                'witness': asdict(profiles[0]) if profiles else None,
            })
    trivial = [r for r in records if r['status'] == 'irreducible']
    retained = [r for r in records if r['status'] == 'retained']
    excluded = [r for r in records if r['status'] == 'excluded']
    require(len(trivial) == 1 and trivial[0]['pattern'] == (DEGREE,),
            'The one-part pattern has not been handled correctly')
    require(len(records) == len(trivial) + len(retained) + len(excluded),
            'The final classes do not cover the universe')
    stages = Counter(r['exclusion_stage'] for r in excluded)
    summary = {
        'degree': DEGREE,
        'all_partitions': len(universe),
        'partition_count_by_length': dict(sorted(Counter(map(len, universe)).items())),
'irreducible_patterns': len(trivial),
'nontrivial_patterns': len(universe) - len(trivial),
'excluded_patterns': len(excluded),
'retained_patterns': len(retained),
'excluded_by_stage': {stage: stages[stage] for stage in ('I', 'II', 'III')},
'retained_by_largest_degree': dict(sorted(
    Counter(r['pattern'][0] for r in retained).items(), reverse=True)),
}
return {'summary': summary, 'records': records}

def format_pattern(pattern):
    return '+'.join(map(str, pattern))

result = audit()
summary, records = result["summary"], result["records"]
print("Partition counts by number of factors:")
print(", ".join("({}: {})".format(length, count)
    for length, count in summary["partition_count_by_length"].items()))
print("All partitions: {}".format(summary["all_partitions"]))
print("Nontrivial partitions: {} - {} = {}".format(
    summary["all_partitions"], summary["irreducible_patterns"],
    summary["nontrivial_patterns"]))
stages = summary["excluded_by_stage"]
print("Excluded by I, II, III: {} + {} + {} = {}".format(
    stages["I"], stages["II"], stages["III"],
    summary["excluded_patterns"]))
print("Nontrivial = excluded + retained: {} = {} + {}".format(
    summary["nontrivial_patterns"], summary["excluded_patterns"],
    summary["retained_patterns"]))
print("Retained counts by largest degree:")
print(", ".join("({}: {})".format(degree, count)
    for degree, count in summary["retained_by_largest_degree"].items()))
print("Complete degree inventory:")
for index, record in enumerate(records, 1):
    status = ("ONE-PART" if record["status"] == "irreducible"
        else record["status"].upper())
    if record["exclusion_stage"]:
        status += " ({}): {}".format(
            record["exclusion_stage"], record["reason"])
    print("{:2d}. {:25s} {}".format(
        index, format_pattern(record["pattern"]), status))
print("Formal compatibility witnesses (k, ds, eps, hs, nu):")
for record in records:
    if record["status"] != "retained":
        continue
    w = record["witness"]
    print("{}: {}".format(format_pattern(record["pattern"]),
        (w["k"], w["ds"], w["signs"],
        w["hs"], w["negative_roots"])))

Transcript

PASS: exact polynomial, real-root, and dyadic coefficient identities.
Partition counts by number of factors:
(1: 1), (2: 6), (3: 12), (4: 15), (5: 13), (6: 11), (7: 7), (8: 5), (9: 3), (10: 2),
    (11: 1), (12: 1)
All partitions: 77
Nontrivial partitions: 77 - 1 = 76
Excluded by I, II, III: 37 + 14 + 3 = 54
Nontrivial = excluded + retained: 76 = 54 + 22
Retained counts by largest degree:
(10: 2), (8: 4), (6: 8), (4: 8)
Complete degree inventory:
 1. 12                      ONE-PART
 2. 11+1                    EXCLUDED (I): odd multiplicity of odd degree(s): 11, 1
 3. 10+2                    RETAINED
4. 10+1+1                  RETAINED
5. 9+3                      EXCLUDED (I): odd multiplicity of odd degree(s): 9, 3
6. 9+2+1                    EXCLUDED (I): odd multiplicity of odd degree(s): 9, 1
7. 9+1+1+1                  EXCLUDED (I): odd multiplicity of odd degree(s): 9, 1
8. 8+4                      RETAINED
9. 8+3+1                    EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
10. 8+2+2                   RETAINED
11. 8+2+1+1                 RETAINED
12. 8+1+1+1+1               RETAINED
13. 7+5                     EXCLUDED (I): odd multiplicity of odd degree(s): 7, 5
14. 7+4+1                   EXCLUDED (I): odd multiplicity of odd degree(s): 7, 1
15. 7+3+2                   EXCLUDED (I): odd multiplicity of odd degree(s): 7, 3
16. 7+3+1+1                 EXCLUDED (I): odd multiplicity of odd degree(s): 7, 3
17. 7+2+2+1                 EXCLUDED (I): odd multiplicity of odd degree(s): 7, 1
18. 7+2+1+1+1               EXCLUDED (I): odd multiplicity of odd degree(s): 7, 1
19. 7+1+1+1+1+1             EXCLUDED (I): odd multiplicity of odd degree(s): 7, 1
20. 6+6                     RETAINED
21. 6+5+1                   EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
22. 6+4+2                   RETAINED
23. 6+4+1+1                 RETAINED
24. 6+3+3                   RETAINED
25. 6+3+2+1                 EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
26. 6+3+1+1+1               EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
27. 6+2+2+2                 RETAINED
28. 6+2+2+1+1               RETAINED
29. 6+2+1+1+1+1             RETAINED
30. 6+1+1+1+1+1+1           RETAINED
31. 5+5+2                   EXCLUDED (II): no even degree >= 4
32. 5+5+1+1                 EXCLUDED (II): no even degree >= 4
33. 5+4+3                   EXCLUDED (I): odd multiplicity of odd degree(s): 5, 3
34. 5+4+2+1                 EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
35. 5+4+1+1+1               EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
36. 5+3+3+1                 EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
37. 5+3+2+2                 EXCLUDED (I): odd multiplicity of odd degree(s): 5, 3
38. 5+3+2+1+1               EXCLUDED (I): odd multiplicity of odd degree(s): 5, 3
39. 5+3+1+1+1+1             EXCLUDED (I): odd multiplicity of odd degree(s): 5, 3
40. 5+2+2+2+1               EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
41. 5+2+2+1+1+1             EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
42. 5+2+1+1+1+1+1           EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
43. 5+1+1+1+1+1+1+1         EXCLUDED (I): odd multiplicity of odd degree(s): 5, 1
44. 4+4+4                   RETAINED
45. 4+4+3+1                 EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
46. 4+4+2+2                 RETAINED
47. 4+4+2+1+1               RETAINED
48. 4+4+1+1+1+1             RETAINED
49. 4+3+3+2                 EXCLUDED (III): unique quartic; odd number of quadratics
50. 4+3+3+1+1               RETAINED
51. 4+3+2+2+1               EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
52. 4+3+2+1+1+1             EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
53. 4+3+1+1+1+1+1           EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
54. 4+2+2+2+2               RETAINED
55. 4+2+2+2+1+1             EXCLUDED (III): unique quartic; odd number of quadratics
56. 4+2+2+1+1+1             RETAINED
57. 4+2+1+1+1+1+1           EXCLUDED (III): unique quartic; odd number of quadratics
58. 4+1+1+1+1+1+1+1         RETAINED
59. 3+3+3+3                 EXCLUDED (II): no even degree >= 4
60. 3+3+3+2+1               EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
61. 3+3+3+1+1+1             EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
62. 3+3+2+2+2               EXCLUDED (II): no even degree >= 4
63. 3+3+2+2+1+1             EXCLUDED (II): no even degree >= 4
64. 3+3+2+1+1+1+1           EXCLUDED (II): no even degree >= 4
65. 3+3+1+1+1+1+1+1         EXCLUDED (II): no even degree >= 4
66. 3+2+2+2+2+1             EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
67. 3+2+2+2+1+1+1           EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
68. 3+2+2+1+1+1+1+1         EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
69. 3+2+1+1+1+1+1+1+1       EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
70.  3+1+1+1+1+1+1+1+1+1        EXCLUDED (I): odd multiplicity of odd degree(s): 3, 1
71.  2+2+2+2+2+2                  EXCLUDED (II): no even degree >= 4
72.  2+2+2+2+2+1+1                EXCLUDED (II): no even degree >= 4
73.  2+2+2+2+1+1+1+1              EXCLUDED (II): no even degree >= 4
74.  2+2+2+1+1+1+1+1+1            EXCLUDED (II): no even degree >= 4
75.  2+2+1+1+1+1+1+1+1+1          EXCLUDED (II): no even degree >= 4
76.  2+1+1+1+1+1+1+1+1+1+1        EXCLUDED (II): no even degree >= 4
77.  1+1+1+1+1+1+1+1+1+1+1+1      EXCLUDED (II): no even degree >= 4
Formal compatibility witnesses (k, ds, eps, hs, nu):
10+2: (1, (5,), (1,), (), (1,))
10+1+1: (0, (5,), (-1,), (1,), (2,))
8+4: (0, (4, 2), (-1, -1), (), (1, 1))
8+2+2: (0, (4,), (1,), (2,), (2,))
8+2+1+1: (1, (4,), (-1,), (1,), (1,))
8+1+1+1+1: (0, (4,), (1,), (1, 1), (2,))
6+6: (0, (3, 3), (-1, -1), (), (0, 2))
6+4+2: (1, (3, 2), (-1, -1), (), (0, 1))
6+4+1+1: (0, (3, 2), (-1, 1), (1,), (0, 2))
6+3+3: (0, (3,), (-1,), (3,), (2,))
6+2+2+2: (1, (3,), (1,), (2,), (1,))
6+2+2+1+1: (0, (3,), (-1,), (2, 1), (2,))
6+2+1+1+1+1: (1, (3,), (1,), (1, 1), (1,))
6+1+1+1+1+1+1: (0, (3,), (-1,), (1, 1, 1), (2,))
4+4+4: (0, (2,), (1,), (4,), (2,))
4+4+2+2: (0, (2, 2), (1, 1), (2,), (0, 2))
4+4+2+1+1: (1, (2, 2), (-1, 1), (1,), (1, 0))
4+4+1+1+1+1: (0, (2, 2), (1, 1), (1, 1), (0, 2))
4+3+3+1+1: (0, (2,), (1,), (3, 1), (2,))
4+2+2+2+2: (0, (2,), (1,), (2, 2), (2,))
4+2+2+1+1+1+1: (0, (2,), (1,), (2, 1, 1), (2,))
4+1+1+1+1+1+1+1+1: (0, (2,), (1,), (1, 1, 1, 1), (2,))

AI Assistance

OpenAI GPT-6.1 Sol assisted with the development of the argument, symbolic calculations, verification code, and the preparation of the exposition. The author is responsible for the mathematical content and the final presentation.

References

  1. [1]R. A. Sharipov, Perfect cuboids and irreducible polynomials, Ufa Math. J. 4 (2012), no. 1, 153–160.arxiv.org/abs/1108.5348
  2. [2]R. A. Sharipov, A note on the third cuboid conjecture. Part I, arXiv:1203.2567 [math.NT] (2012).arxiv.org/abs/1203.2567
  3. [3]G. E. Andrews, The Theory of Partitions, Cambridge Mathematical Library, Cambridge University Press, Cambridge, 1998.
  4. [4]J.-P. Serre, Local Fields, Graduate Texts in Mathematics, vol. 67, Springer-Verlag, New York–Berlin, 1979.
  5. [5]A. M. Robert, A Course in p-adic Analysis, Graduate Texts in Mathematics, vol. 198, Springer-Verlag, New York, 2000.DOI
  6. [6]L. Berger, The Weierstrass preparation theorem and resultants of p-adic power series, Münster J. Math. 14 (2021), 155–163.
  7. [7]A. Meurer, C. P. Smith, M. Paprocki, et al., SymPy: symbolic computing in Python, PeerJ Comput. Sci. 3 (2017), e103.

Paper details

vlnc:W_fmggzdwe
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents