Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Finite-sum weights and coefficients for the quantum Tet certificate

Definition
QuantumSixJSumCertificate

by igor.kh · Nov 13, 2026 · Mathlib 0df444a (Lean v4.33.1)

Auxiliary rational functions over Q(q)\mathbb Q(q)Q(q) for the finite-sum certificate of the Appendix Tet recurrence. The bundle specifies the raw summand, common factorial weight, cleared three-term coefficients, flux multiplier, flux, raw and finite-sum residuals, finite summation length, Appendix prefactor, and clearing factor. The diagonal and two step-2 entry functions are branch-expanded expressions used for factorial shifts; their identification with the symmetric two-index matrix is a proof obligation, not assumed. All factorials retain their signed-index zero convention and division is total in the rational-function field. No identity or recurrence is assumed.

Independent mathematical read-back

Read-back: finite-sum certificate declarations

Explicit notation used in the paragraphs

All scalar values below belong to the rational-function field K=Q(q)K=\mathbb Q(q)K=Q(q), where qqq is the image of the polynomial indeterminate. For every integer nnn, write

[n]=qn−q−nq−q−1,F(n)={0n<0,∏r=1n[r]n≥0.[n]=\frac{q^n-q^{-n}}{q-q^{-1}},\qquad F(n)=\begin{cases}0&n<0,\\ \prod_{r=1}^{n}[r]&n\geq0.\end{cases}[n]=q−q−1qn−q−n​,F(n)={0∏r=1n​[r]​n<0,n≥0.​

In particular F(0)=1F(0)=1F(0)=1. Powers with integer exponents are field powers. All division and inversion are total field operations: 0−1=00^{-1}=00−1=0 and u/0=0u/0=0u/0=0. All label subtraction is integer subtraction, not truncated subtraction. Every division of integer labels by 222 means Euclidean integer division (for divisor 222, the floor of the half), including when the numerator is negative or odd. No evaluation at a numerical value of qqq is involved.

The predicate T(u,v,w)T(u,v,w)T(u,v,w) means that u,v,wu,v,wu,v,w are nonnegative integers, each is at most the sum of the other two, and u+v+wu+v+wu+v+w is even. Write Hr=T(a,d,r)∧T(b,c,r)H_r=T(a,d,r)\land T(b,c,r)Hr​=T(a,d,r)∧T(b,c,r) and Vs=T(a,b,s)∧T(c,d,s)V_s=T(a,b,s)\land T(c,d,s)Vs​=T(a,b,s)∧T(c,d,s), always with the external labels a,b,c,da,b,c,da,b,c,d of the paragraph in question. The following explicitly specified scalar expressions provide notation for expanded dependency quantities:

δ(r)={(−1)r[r+1]r≥0,0r<0,ϑ(u,v,w)={(−1)zF(z+1)F(z−u)F(z−v)F(z−w)F(u)F(v)F(w),z=(u+v+w)/2T(u,v,w),0otherwise,\delta(r)=\begin{cases}(-1)^r[r+1]&r\geq0,\\0&r<0,\end{cases}\qquad \vartheta(u,v,w)=\begin{cases} \displaystyle\frac{(-1)^z F(z+1)F(z-u)F(z-v)F(z-w)}{F(u)F(v)F(w)},\quad z=(u+v+w)/2&T(u,v,w),\\ 0&\text{otherwise}, \end{cases}δ(r)={(−1)r[r+1]0​r≥0,r<0,​ϑ(u,v,w)=⎩⎨⎧​F(u)F(v)F(w)(−1)zF(z+1)F(z−u)F(z−v)F(z−w)​,z=(u+v+w)/20​T(u,v,w),otherwise,​ Nr={ϑ(a,d,r)ϑ(b,c,r)/δ(r)Hr,0¬Hr,E(u,v,w)={[(u−v+w)/2][(−u+v+w)/2]−[(u+v−w)/2][(u+v+w)/2+2][2]T(u,v,w),0otherwise.N_r=\begin{cases}\vartheta(a,d,r)\vartheta(b,c,r)/\delta(r)&H_r,\\0&\neg H_r,\end{cases}\qquad E(u,v,w)=\begin{cases} \displaystyle\frac{[(u-v+w)/2][(-u+v+w)/2]-[(u+v-w)/2][(u+v+w)/2+2]}{[2]}&T(u,v,w),\\0&\text{otherwise}. \end{cases}Nr​={ϑ(a,d,r)ϑ(b,c,r)/δ(r)0​Hr​,¬Hr​,​E(u,v,w)=⎩⎨⎧​[2][(u−v+w)/2][(−u+v+w)/2]−[(u+v−w)/2][(u+v+w)/2+2]​0​T(u,v,w),otherwise.​

For external integers a,b,c,da,b,c,da,b,c,d and integer channels r,sr,sr,s, let the four-entry and three-entry lists be

αr,s=((a+d+r)/2,(b+c+r)/2,(a+b+s)/2,(c+d+s)/2),βr,s=((b+d+r+s)/2,(a+c+r+s)/2,(a+b+c+d)/2).\alpha^{r,s}=\big((a+d+r)/2,(b+c+r)/2,(a+b+s)/2,(c+d+s)/2\big),\quad \beta^{r,s}=\big((b+d+r+s)/2,(a+c+r+s)/2,(a+b+c+d)/2\big).αr,s=((a+d+r)/2,(b+c+r)/2,(a+b+s)/2,(c+d+s)/2),βr,s=((b+d+r+s)/2,(a+c+r+s)/2,(a+b+c+d)/2).

List indices start at 000. Put

Pr,s=∏i=03∏k=02F(βkr,s−αir,s)F(a)F(b)F(c)F(d)F(r)F(s),P_{r,s}=\frac{\prod_{i=0}^{3}\prod_{k=0}^{2}F(\beta_k^{r,s}-\alpha_i^{r,s})}{F(a)F(b)F(c)F(d)F(r)F(s)},Pr,s​=F(a)F(b)F(c)F(d)F(r)F(s)∏i=03​∏k=02​F(βkr,s​−αir,s​)​, Ur,s=1Nr{Pr,s∑z=0a+b+c+d+r+s(−1)zF(z+1)(∏i=03F(z−αir,s))(∏k=02F(βkr,s−z))Hr∧Vs,0otherwise.U_{r,s}=\frac{1}{N_r}\begin{cases} \displaystyle P_{r,s}\sum_{z=0}^{a+b+c+d+r+s} \frac{(-1)^zF(z+1)}{\big(\prod_{i=0}^{3}F(z-\alpha_i^{r,s})\big)\big(\prod_{k=0}^{2}F(\beta_k^{r,s}-z)\big)}&H_r\land V_s,\\ 0&\text{otherwise}. \end{cases}Ur,s​=Nr​1​⎩⎨⎧​Pr,s​z=0∑a+b+c+d+r+s​(∏i=03​F(z−αir,s​))(∏k=02​F(βkr,s​−z))(−1)zF(z+1)​0​Hr​∧Vs​,otherwise.​

Here an integer-indexed sum from 000 to a negative upper endpoint is empty and equals zero. Thus UUU includes both the admissibility guard on the finite sum and total division by NrN_rNr​; it is not a separate unspecified coefficient.

Declaration paragraphs

QuantumSixJ.diagonalEntry. For all integers a,b,c,d,ja,b,c,d,ja,b,c,d,j, this defines a scalar equal to −Nj[2]E(a,j,d)E(b,j,c)/([j][j+2])-N_j[2]E(a,j,d)E(b,j,c)/([j][j+2])−Nj​[2]E(a,j,d)E(b,j,c)/([j][j+2]) if HjH_jHj​ and j≠0j\ne0j=0, and equal to zero otherwise, with H,N,EH,N,EH,N,E expanded above. In particular the value at j=0j=0j=0 is zero even if both triples in H0H_0H0​ satisfy the triangle and parity conditions; inadmissible and negative horizontal labels also take the zero branch. There is no additional nonvanishing or positivity hypothesis.

QuantumSixJ.Lp2. For all integers a,b,c,d,ja,b,c,d,ja,b,c,d,j, this defines the scalar Nj+2[(a+d−j)/2][(b+c−j)/2]N_{j+2}[(a+d-j)/2][(b+c-j)/2]Nj+2​[(a+d−j)/2][(b+c−j)/2] when both HjH_jHj​ and Hj+2H_{j+2}Hj+2​ hold, and zero otherwise, where NNN is the explicitly guarded product of the two ϑ\varthetaϑ expressions divided by δ\deltaδ given above. Both horizontal channels must satisfy nonnegative triangle and even-sum conditions; no separate hypothesis restricts any input integer, and the formula includes j=0j=0j=0.

QuantumSixJ.Lm2. For all integers a,b,c,d,ja,b,c,d,ja,b,c,d,j, this defines the scalar Nj[(a+d−(j−2))/2][(b+c−(j−2))/2]N_j[(a+d-(j-2))/2][(b+c-(j-2))/2]Nj​[(a+d−(j−2))/2][(b+c−(j−2))/2] when both HjH_jHj​ and Hj−2H_{j-2}Hj−2​ hold, and zero otherwise, where NNN is the explicitly guarded expression above. In particular j=0j=0j=0 and j=1j=1j=1 give zero because the channel j−2j-2j−2 is negative; neither subtraction is natural-number subtraction. There are no hypotheses on the input labels.

QuantumSixJ.Certificate.rawSummand. For any eight integers A,B,C,D,E,F,G,SA,B,C,D,E,F,G,SA,B,C,D,E,F,G,S, this defines the scalar

(−1)SF(S+1)F(S−A)F(S−B)F(S−C)F(S−D)F(E−S)F(F−S)F(G−S).\frac{(-1)^S F(S+1)}{F(S-A)F(S-B)F(S-C)F(S-D)F(E-S)F(F-S)F(G-S)}.F(S−A)F(S−B)F(S−C)F(S−D)F(E−S)F(F−S)F(G−S)(−1)SF(S+1)​.

Here F(⋅)F(\cdot)F(⋅) denotes the explicitly defined signed-index quantum factorial above, while the bare letter FFF inside an argument is the sixth integer parameter. There is no admissibility or support guard and no restriction S≥0S\geq0S≥0; a negative factorial argument in the denominator makes the quotient zero by total field division, and S<−1S<-1S<−1 makes the numerator factorial zero.

QuantumSixJ.Certificate.commonWeight. For any seven integers A,B,C,E,F,j,SA,B,C,E,F,j,SA,B,C,E,F,j,S, set D=E+F−C−jD=E+F-C-jD=E+F−C−j and G=A+B−jG=A+B-jG=A+B−j; this defines the scalar

(−1)SF(S+1) F(S−A+1)−1F(S−B+1)−1F(S−C)−1F(S−D)−1F(E−S+1)−1F(F−S+1)−1F(G−S)−1.(-1)^S F(S+1)\,F(S-A+1)^{-1}F(S-B+1)^{-1}F(S-C)^{-1}F(S-D)^{-1}F(E-S+1)^{-1}F(F-S+1)^{-1}F(G-S)^{-1}.(−1)SF(S+1)F(S−A+1)−1F(S−B+1)−1F(S−C)−1F(S−D)−1F(E−S+1)−1F(F−S+1)−1F(G−S)−1.

The function F(⋅)F(\cdot)F(⋅) is the signed-index quantum factorial, distinct from the integer parameter FFF. The definition has no hypotheses, including no support, admissibility, or positivity restriction, and inverses of zero factorial values are zero.

QuantumSixJ.Certificate.lowerCoefficient. For any integers A,B,jA,B,jA,B,j, this defines the scalar [A−j+1][B−j+1][A+1][B+1][2][j+2][A-j+1][B-j+1][A+1][B+1][2][j+2][A−j+1][B−j+1][A+1][B+1][2][j+2], with every bracket the rational-function quotient specified above. There are no sign, parity, or nonzero assumptions on the parameters; the definition includes j=0j=0j=0 and negative inputs.

QuantumSixJ.Certificate.upperCoefficient. For any integers C,E,F,jC,E,F,jC,E,F,j, set D=E+F−C−jD=E+F-C-jD=E+F−C−j; this defines the scalar [E−C+1][F−D+1][F−C+1][E−D+1][2][j][E-C+1][F-D+1][F-C+1][E-D+1][2][j][E−C+1][F−D+1][F−C+1][E−D+1][2][j]. Every bracket is the specified rational-function quotient; the integer parameter FFF is not a factorial here. There are no hypotheses, and at j=0j=0j=0 the final factor is [0]=0[0]=0[0]=0.

QuantumSixJ.Certificate.centerCoefficient. For any integers A,B,C,E,F,jA,B,C,E,F,jA,B,C,E,F,j, set D=E+F−C−jD=E+F-C-jD=E+F−C−j and G=A+B−jG=A+B-jG=A+B−j; this defines the scalar

−(([A−j][E−C]−[F−D][A+2])([B−j][F−C]−[E−D][B+2])+([F−B][E−A]−[G−D][C+2])[j][j+2])[j+1].-\Big(\big([A-j][E-C]-[F-D][A+2]\big)\big([B-j][F-C]-[E-D][B+2]\big) +\big([F-B][E-A]-[G-D][C+2]\big)[j][j+2]\Big)[j+1].−(([A−j][E−C]−[F−D][A+2])([B−j][F−C]−[E−D][B+2])+([F−B][E−A]−[G−D][C+2])[j][j+2])[j+1].

All brackets are the explicit rational functions above, all parameters range over all integers without further assumptions, and the formula includes zero and negative values of jjj.

QuantumSixJ.Certificate.scaledMultiplier. For any integers A,B,C,E,F,j,tA,B,C,E,F,j,tA,B,C,E,F,j,t, set D=E+F−C−jD=E+F-C-jD=E+F−C−j; this defines the scalar

[2]([j+2][t−E−1]([j][F−A+1][t−B]+[B−j+1][A+1][F−t+1])−[j][F−C+1][F−D+1][t−A][t−B]).[2]\Big([j+2][t-E-1]\big([j][F-A+1][t-B]+[B-j+1][A+1][F-t+1]\big) -[j][F-C+1][F-D+1][t-A][t-B]\Big).[2]([j+2][t−E−1]([j][F−A+1][t−B]+[B−j+1][A+1][F−t+1])−[j][F−C+1][F−D+1][t−A][t−B]).

The brackets are the stated rational-function quotients. No support, sign, parity, or nonzero restrictions are imposed on any of the seven integer parameters.

QuantumSixJ.Certificate.flux. For any integers A,B,C,E,F,j,SA,B,C,E,F,j,SA,B,C,E,F,j,S, set D=E+F−C−jD=E+F-C-jD=E+F−C−j and G=A+B−jG=A+B-jG=A+B−j; this defines the scalar

(−1)SF(S+1)F(S−A)F(S−B)F(S−C−1)F(S−D−1)F(E+1−S)F(F+1−S)F(G−S) [2]([j+2][S−E−1]([j][F−A+1][S−B]+[B−j+1][A+1][F−S+1])−[j][F−C+1][F−D+1][S−A][S−B]).\frac{(-1)^S F(S+1)}{F(S-A)F(S-B)F(S-C-1)F(S-D-1)F(E+1-S)F(F+1-S)F(G-S)} \,[2]\Big([j+2][S-E-1]\big([j][F-A+1][S-B]+[B-j+1][A+1][F-S+1]\big) -[j][F-C+1][F-D+1][S-A][S-B]\Big).F(S−A)F(S−B)F(S−C−1)F(S−D−1)F(E+1−S)F(F+1−S)F(G−S)(−1)SF(S+1)​[2]([j+2][S−E−1]([j][F−A+1][S−B]+[B−j+1][A+1][F−S+1])−[j][F−C+1][F−D+1][S−A][S−B]).

Here F(⋅)F(\cdot)F(⋅) is the signed-index factorial, distinct from the integer parameter FFF, and all bracket and factorial expressions are expanded in the conventions above. There are no hypotheses and no explicit support guard; zero denominators and negative factorial arguments use the total-field and signed-factorial conventions.

QuantumSixJ.Certificate.rawResidual. For any integers A,B,C,E,F,j,SA,B,C,E,F,j,SA,B,C,E,F,j,S, set D=E+F−C−jD=E+F-C-jD=E+F−C−j and G=A+B−jG=A+B-jG=A+B−j, and abbreviate the following explicit products by p−,p0,p+p_-,p_0,p_+p−​,p0​,p+​ solely within this paragraph:

p−=[A−j+1][B−j+1][A+1][B+1][2][j+2],p0=−(([A−j][E−C]−[F−D][A+2])([B−j][F−C]−[E−D][B+2])+([F−B][E−A]−[G−D][C+2])[j][j+2])[j+1],p+=[E−C+1][F−D+1][F−C+1][E−D+1][2][j].\begin{aligned} p_-&=[A-j+1][B-j+1][A+1][B+1][2][j+2],\\ p_0&=-\Big(([A-j][E-C]-[F-D][A+2])([B-j][F-C]-[E-D][B+2])\\ &\hspace{18mm}+([F-B][E-A]-[G-D][C+2])[j][j+2]\Big)[j+1],\\ p_+&=[E-C+1][F-D+1][F-C+1][E-D+1][2][j]. \end{aligned}p−​p0​p+​​=[A−j+1][B−j+1][A+1][B+1][2][j+2],=−(([A−j][E−C]−[F−D][A+2])([B−j][F−C]−[E−D][B+2])+([F−B][E−A]−[G−D][C+2])[j][j+2])[j+1],=[E−C+1][F−D+1][F−C+1][E−D+1][2][j].​

The defined scalar is exactly

p−(−1)SF(S+1)F(S−A+1)F(S−B+1)F(S−C)F(S−D)F(E−1−S)F(F−1−S)F(G−S)+p0(−1)SF(S+1)F(S−A)F(S−B)F(S−C)F(S−D)F(E−S)F(F−S)F(G−S)+p+(−1)SF(S+1)F(S−A−1)F(S−B−1)F(S−C)F(S−D)F(E+1−S)F(F+1−S)F(G−S).\begin{aligned} &p_-\frac{(-1)^S F(S+1)}{F(S-A+1)F(S-B+1)F(S-C)F(S-D)F(E-1-S)F(F-1-S)F(G-S)}\\ {}+{}&p_0\frac{(-1)^S F(S+1)}{F(S-A)F(S-B)F(S-C)F(S-D)F(E-S)F(F-S)F(G-S)}\\ {}+{}&p_+\frac{(-1)^S F(S+1)}{F(S-A-1)F(S-B-1)F(S-C)F(S-D)F(E+1-S)F(F+1-S)F(G-S)}. \end{aligned}++​p−​F(S−A+1)F(S−B+1)F(S−C)F(S−D)F(E−1−S)F(F−1−S)F(G−S)(−1)SF(S+1)​p0​F(S−A)F(S−B)F(S−C)F(S−D)F(E−S)F(F−S)F(G−S)(−1)SF(S+1)​p+​F(S−A−1)F(S−B−1)F(S−C)F(S−D)F(E+1−S)F(F+1−S)F(G−S)(−1)SF(S+1)​.​

The notation F(⋅)F(\cdot)F(⋅) denotes the factorial and bare FFF the integer parameter. There are no hypotheses or support restrictions, and zero denominators give zero quotients, not undefined terms.

QuantumSixJ.Certificate.sumLength. For any integers a,b,c,d,j,la,b,c,d,j,la,b,c,d,j,l, this defines the natural number max⁡(a+b+c+d+j+l,0)+3\max(a+b+c+d+j+l,0)+3max(a+b+c+d+j+l,0)+3, where the maximum is regarded as a natural number before adding 333. The truncation applies to the whole sum of the six integers, not to its individual terms; hence every input gives a length of at least 333, and a nonpositive total gives length exactly 333.

QuantumSixJ.Certificate.totalResidual. For any integers a,b,c,d,j,la,b,c,d,j,la,b,c,d,j,l, put A=(a+d+j)/2A=(a+d+j)/2A=(a+d+j)/2, B=(b+c+j)/2B=(b+c+j)/2B=(b+c+j)/2, C=(a+b+l)/2C=(a+b+l)/2C=(a+b+l)/2, E=(b+d+j+l)/2E=(b+d+j+l)/2E=(b+d+j+l)/2, F=(a+c+j+l)/2F=(a+c+j+l)/2F=(a+c+j+l)/2, D=E+F−C−jD=E+F-C-jD=E+F−C−j, G=A+B−jG=A+B-jG=A+B−j, and n=max⁡(a+b+c+d+j+l,0)+3n=\max(a+b+c+d+j+l,0)+3n=max(a+b+c+d+j+l,0)+3. For these integers form

p−=[A−j+1][B−j+1][A+1][B+1][2][j+2],p0=−(([A−j][E−C]−[F−D][A+2])([B−j][F−C]−[E−D][B+2])+([F−B][E−A]−[G−D][C+2])[j][j+2])[j+1],p+=[E−C+1][F−D+1][F−C+1][E−D+1][2][j].\begin{aligned} p_-&=[A-j+1][B-j+1][A+1][B+1][2][j+2],\\ p_0&=-\Big(([A-j][E-C]-[F-D][A+2])([B-j][F-C]-[E-D][B+2])\\ &\hspace{18mm}+([F-B][E-A]-[G-D][C+2])[j][j+2]\Big)[j+1],\\ p_+&=[E-C+1][F-D+1][F-C+1][E-D+1][2][j]. \end{aligned}p−​p0​p+​​=[A−j+1][B−j+1][A+1][B+1][2][j+2],=−(([A−j][E−C]−[F−D][A+2])([B−j][F−C]−[E−D][B+2])+([F−B][E−A]−[G−D][C+2])[j][j+2])[j+1],=[E−C+1][F−D+1][F−C+1][E−D+1][2][j].​

This defines the scalar

∑S=0n−1(p−(−1)SF(S+1)F(S−A+1)F(S−B+1)F(S−C)F(S−D)F(E−1−S)F(F−1−S)F(G−S)+p0(−1)SF(S+1)F(S−A)F(S−B)F(S−C)F(S−D)F(E−S)F(F−S)F(G−S)+p+(−1)SF(S+1)F(S−A−1)F(S−B−1)F(S−C)F(S−D)F(E+1−S)F(F+1−S)F(G−S)).\sum_{S=0}^{n-1}\left( \frac{p_-(-1)^S F(S+1)}{F(S-A+1)F(S-B+1)F(S-C)F(S-D)F(E-1-S)F(F-1-S)F(G-S)} +\frac{p_0(-1)^S F(S+1)}{F(S-A)F(S-B)F(S-C)F(S-D)F(E-S)F(F-S)F(G-S)} +\frac{p_+(-1)^S F(S+1)}{F(S-A-1)F(S-B-1)F(S-C)F(S-D)F(E+1-S)F(F+1-S)F(G-S)}\right).S=0∑n−1​(F(S−A+1)F(S−B+1)F(S−C)F(S−D)F(E−1−S)F(F−1−S)F(G−S)p−​(−1)SF(S+1)​+F(S−A)F(S−B)F(S−C)F(S−D)F(E−S)F(F−S)F(G−S)p0​(−1)SF(S+1)​+F(S−A−1)F(S−B−1)F(S−C)F(S−D)F(E+1−S)F(F+1−S)F(G−S)p+​(−1)SF(S+1)​).

Factorial F(⋅)F(\cdot)F(⋅) is distinct from the integer FFF. There are no admissibility or positivity hypotheses. The summation uses the natural numbers 0,…,n−10,\ldots,n-10,…,n−1 cast to integers, and contains at least three terms even when the total of the six labels is negative; D,GD,GD,G are the displayed derived integers, without separately replacing them by other half-sums on arbitrary odd inputs.

QuantumSixJ.Certificate.prefactor. For any integers a,b,c,d,j,la,b,c,d,j,la,b,c,d,j,l, using the four and three explicit half-sum lists αj,l,βj,l\alpha^{j,l},\beta^{j,l}αj,l,βj,l above, this defines the scalar

∏i=03∏k=02F(βkj,l−αij,l)F(a)F(b)F(c)F(d)F(j)F(l).\frac{\prod_{i=0}^{3}\prod_{k=0}^{2}F(\beta_k^{j,l}-\alpha_i^{j,l})}{F(a)F(b)F(c)F(d)F(j)F(l)}.F(a)F(b)F(c)F(d)F(j)F(l)∏i=03​∏k=02​F(βkj,l​−αij,l​)​.

There is no admissibility guard: odd or negative labels still have values, using Euclidean division for the half-sums, zero for negative-index factorials, and total field division. In particular a negative one of the six denominator labels makes the value zero.

QuantumSixJ.Certificate.clearingFactor. For every integer jjj, this defines the scalar [2][j][j+1][j+2][2][j][j+1][j+2][2][j][j+1][j+2], with each bracket the specified rational-function quotient. No assumption that this product is nonzero is part of the definition; it equals zero at least for j=0,−1,−2j=0,-1,-2j=0,−1,−2 because one factor is [0][0][0].

QuantumSixJ.Certificate.mainResidual. For any integers a,b,c,d,j,la,b,c,d,j,la,b,c,d,j,l, define, within this paragraph, m−=Nj[(a+d−j+2)/2][(b+c−j+2)/2]m_-=N_j[(a+d-j+2)/2][(b+c-j+2)/2]m−​=Nj​[(a+d−j+2)/2][(b+c−j+2)/2] when Hj∧Hj−2H_j\land H_{j-2}Hj​∧Hj−2​ and zero otherwise; m0=−Nj[2]E(a,j,d)E(b,j,c)/([j][j+2])m_0=-N_j[2]E(a,j,d)E(b,j,c)/([j][j+2])m0​=−Nj​[2]E(a,j,d)E(b,j,c)/([j][j+2]) when Hj∧j≠0H_j\land j\ne0Hj​∧j=0 and zero otherwise; and m+=Nj+2[(a+d−j)/2][(b+c−j)/2]m_+=N_{j+2}[(a+d-j)/2][(b+c-j)/2]m+​=Nj+2​[(a+d−j)/2][(b+c−j)/2] when Hj∧Hj+2H_j\land H_{j+2}Hj​∧Hj+2​ and zero otherwise. The defined scalar is

m−Uj−2,l+(m0−E(a,b,l)Nj)Uj,l+m+Uj+2,l,m_-U_{j-2,l}+\big(m_0-E(a,b,l)N_j\big)U_{j,l}+m_+U_{j+2,l},m−​Uj−2,l​+(m0​−E(a,b,l)Nj​)Uj,l​+m+​Uj+2,l​,

where H,N,E,UH,N,E,UH,N,E,U are the explicitly expanded predicates and rational-function finite-sum expressions above. No admissibility, positivity, or nonzero condition is assumed on the six integers; the neighbor values include negative channels, the zero diagonal at j=0j=0j=0 is retained, and every quotient, including the division by NrN_rNr​ in each Ur,lU_{r,l}Ur,l​, uses total field division.

Definition code
import Definitions.Def_QuantumSixJScalar

/-! Auxiliary finite-sum weights and coefficients for the telescoping proof.
The three entry functions are the diagonal and step-2 entries, written without
two-index minimum/maximum branches for factorial-shift calculations.
-/
noncomputable section
open scoped BigOperators QuantumSixJ Classical
namespace QuantumSixJ
def diagonalEntry (a b c d j : ℤ) : Scalar :=
  if hAdmissible a b c d j ∧ j ≠ 0 then
    -h a b c d j * ⟬2⟭ * Λ a j d * Λ b j c / (⟬j⟭ * ⟬j + 2⟭)
  else 0

def Lp2 (a b c d j : ℤ) : Scalar :=
  if hAdmissible a b c d j ∧ hAdmissible a b c d (j + 2) then
    h a b c d (j + 2) * ⟬(a + d - j) / 2⟭ * ⟬(b + c - j) / 2⟭
  else 0

def Lm2 (a b c d j : ℤ) : Scalar :=
  if hAdmissible a b c d j ∧ hAdmissible a b c d (j - 2) then
    h a b c d j * ⟬(a + d - (j - 2)) / 2⟭ * ⟬(b + c - (j - 2)) / 2⟭
  else 0

end QuantumSixJ
noncomputable section
namespace QuantumSixJ.Certificate
open scoped QuantumSixJ BigOperators

def rawSummand (A B C D E F G S : ℤ) : Scalar :=
  (-1 : Scalar) ^ S * ⟬S + 1⟭! /
    (⟬S - A⟭! * ⟬S - B⟭! * ⟬S - C⟭! * ⟬S - D⟭! *
      ⟬E - S⟭! * ⟬F - S⟭! * ⟬G - S⟭!)

def commonWeight (A B C E F j S : ℤ) : Scalar :=
  let D := E + F - C - j
  let G := A + B - j
  (-1 : Scalar) ^ S * ⟬S + 1⟭! *
    (⟬S - A + 1⟭!)⁻¹ * (⟬S - B + 1⟭!)⁻¹ * (⟬S - C⟭!)⁻¹ *
    (⟬S - D⟭!)⁻¹ * (⟬E - S + 1⟭!)⁻¹ * (⟬F - S + 1⟭!)⁻¹ * (⟬G - S⟭!)⁻¹

def lowerCoefficient (A B j : ℤ) : Scalar :=
  (⟬A - j + 1⟭ * ⟬B - j + 1⟭ * ⟬A + 1⟭ * ⟬B + 1⟭) * ⟬2⟭ * ⟬j + 2⟭

def upperCoefficient (C E F j : ℤ) : Scalar :=
  let D := E + F - C - j
  (⟬E - C + 1⟭ * ⟬F - D + 1⟭ * ⟬F - C + 1⟭ * ⟬E - D + 1⟭) * ⟬2⟭ * ⟬j⟭

def centerCoefficient (A B C E F j : ℤ) : Scalar :=
  let D := E + F - C - j
  let G := A + B - j
  let la := ⟬A - j⟭ * ⟬E - C⟭ - ⟬F - D⟭ * ⟬A + 2⟭
  let lb := ⟬B - j⟭ * ⟬F - C⟭ - ⟬E - D⟭ * ⟬B + 2⟭
  let ll := ⟬F - B⟭ * ⟬E - A⟭ - ⟬G - D⟭ * ⟬C + 2⟭
  let nb := -(la * lb + ll * ⟬j⟭ * ⟬j + 2⟭)
  nb * ⟬j + 1⟭

def scaledMultiplier (A B C E F j t : ℤ) : Scalar :=
  let D := E + F - C - j
  ⟬2⟭ * (⟬j + 2⟭ * ⟬t - E - 1⟭ *
    (⟬j⟭ * ⟬F - A + 1⟭ * ⟬t - B⟭ +
      ⟬B - j + 1⟭ * ⟬A + 1⟭ * ⟬F - t + 1⟭) -
    ⟬j⟭ * ⟬F - C + 1⟭ * ⟬F - D + 1⟭ * ⟬t - A⟭ * ⟬t - B⟭)

def flux (A B C E F j S : ℤ) : Scalar :=
  let D := E + F - C - j
  let G := A + B - j
  rawSummand A B (C + 1) (D + 1) (E + 1) (F + 1) G S *
    scaledMultiplier A B C E F j S

def rawResidual (A B C E F j S : ℤ) : Scalar :=
  let D := E + F - C - j
  let G := A + B - j
  lowerCoefficient A B j * rawSummand (A - 1) (B - 1) C D (E - 1) (F - 1) G S +
    centerCoefficient A B C E F j * rawSummand A B C D E F G S +
    upperCoefficient C E F j * rawSummand (A + 1) (B + 1) C D (E + 1) (F + 1) G S

def sumLength (a b c d j l : ℤ) : ℕ := (a + b + c + d + j + l).toNat + 3

def totalResidual (a b c d j l : ℤ) : Scalar :=
  let A := tetA a b c d j l
  let B := tetB a b c d j l
  ∑ k ∈ Finset.range (sumLength a b c d j l),
    rawResidual (A 0) (A 1) (A 2) (B 0) (B 1) j (k : ℤ)

def prefactor (a b c d j l : ℤ) : Scalar :=
  (∏ i : Fin 4, ∏ k : Fin 3, ⟬tetB a b c d j l k - tetA a b c d j l i⟭!) /
    (⟬a⟭! * ⟬b⟭! * ⟬c⟭! * ⟬d⟭! * ⟬j⟭! * ⟬l⟭!)

def clearingFactor (j : ℤ) : Scalar := ⟬2⟭ * ⟬j⟭ * ⟬j + 1⟭ * ⟬j + 2⟭

def mainResidual (a b c d j l : ℤ) : Scalar :=
  Lm2 a b c d j * normalizedTetSum a b c d (j - 2) l +
    (diagonalEntry a b c d j - Λ a b l * h a b c d j) * normalizedTetSum a b c d j l +
    Lp2 a b c d j * normalizedTetSum a b c d (j + 2) l

end QuantumSixJ.Certificate
end
Source
I. Khavkine, Recurrence relation for the 6j-symbol of su_q(2) as a symmetric eigenvalue problem, https://arxiv.org/abs/1009.2261, Appendix finite Tet formula and Eq. (24); auxiliary factorial-shift and telescoping-certificate calculation for those formulas.

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me