First appearance value conjecture for standard Vandermonde matrices

About 1 year old · traced to

Let VnV_n be the standard Vandermonde matrix associated with (1,2,…,n)(1,2,\ldots,n), and let fVf_V be its associated permutation-sum function.

Standard-Vandermonde value conjecture. For n≥2n \geq 2,

APD⁡n−1(fV)=(n−1)!∏k=1n−1k!.\operatorname{APD}_{n-1}(f_V)=(n-1)!\prod_{k=1}^{n-1}k!.

The formula completely matches exact calculation results for n≤7n \leq 7, but no general proof is supplied.

References

Primary source

Kenichi Takemura, “Alternating Power Difference and Matrix Symmetry: Closed-Form Formulas for the First Appearance Degree m_1”, arXiv:2512.18169 (2025).

Progress summary

Refreshed
Claimed solved

A reader-written argument claims a complete proof of the conjecture, but it has not been independently checked.

Kenichi Takemura’s 2025 paper conjectures that the first nonzero alternating power difference of the standard Vandermonde matrix has the stated factorial value for every dimension. The paper explicitly leaves a rigorous general proof open.

Known results

  • Exact calculations verify the value formula for n≤7n\leq 7 (Takemura, 2025).
  • The same paper rewrites the conjectured value as (n−1)!det⁡(Vn)(n-1)!\det(V_n), using det⁡(Vn)=∏k=1n−1k!\det(V_n)=\prod_{k=1}^{n-1}k!.

Posted attempt

A reader-written argument claims a complete proof: it derives the universal identity APD⁡n−1(A)=(n−1)!1Tadj⁡(A)1\operatorname{APD}_{n-1}(A)=(n-1)!\mathbf{1}^{\mathsf T}\operatorname{adj}(A)\mathbf{1} and applies Vn−11=e1V_n^{-1}\mathbf{1}=e_1. The argument has not been independently verified.

Current status (as of August 2026): The conjecture has a complete but unverified posted proof claim; the published source still treats it as open, so independent verification remains outstanding.

Sources

Solutions 1

ProofThis solution needs a summarySee full solutionHide full solution

For every n×nn\times n matrix AA, write 1=(1,…,1)T\mathbf1=(1,\ldots,1)^{\mathsf T}. The determinant expansion gives

∑m≥0APD⁡m(A)zmm!=det⁡(ezAij).\sum_{m\ge0}\operatorname{APD}_m(A)\frac{z^m}{m!} =\det(e^{zA_{ij}}).

As z→0z\to0,

(ezAij)=Jn+zA+O(z2),Jn=11T.(e^{zA_{ij}})=J_n+zA+O(z^2), \qquad J_n=\mathbf1\mathbf1^{\mathsf T}.

Since JnJ_n has rank one, determinant multilinearity implies

det⁡(ezAij)=zn−11Tadj⁡(A)1+O(zn).\det(e^{zA_{ij}}) =z^{n-1}\mathbf1^{\mathsf T}\operatorname{adj}(A)\mathbf1 +O(z^n).

The coefficient also follows directly from

det⁡(zA+11T)=zndet⁡A+zn−11Tadj⁡(A)1.\det(zA+\mathbf1\mathbf1^{\mathsf T}) =z^n\det A +z^{n-1}\mathbf1^{\mathsf T}\operatorname{adj}(A)\mathbf1.

Hence, universally,

APD⁡n−1(A)=(n−1)!1Tadj⁡(A)1.(1)\operatorname{APD}_{n-1}(A) =(n-1)!\mathbf1^{\mathsf T}\operatorname{adj}(A)\mathbf1. \tag{1}

For the standard Vandermonde matrix

(Vn)ij=ij−1,(V_n)_{ij}=i^{j-1},

its first column equals 1\mathbf1, so

Vn−11=e1.V_n^{-1}\mathbf1=e_1.

Consequently

1Tadj⁡(Vn)1=det⁡(Vn)1TVn−11=det⁡(Vn).\mathbf1^{\mathsf T}\operatorname{adj}(V_n)\mathbf1 =\det(V_n)\mathbf1^{\mathsf T}V_n^{-1}\mathbf1 =\det(V_n).

The Vandermonde product gives

det⁡(Vn)=∏1≤i<j≤n(j−i)=∏k=1n−1k!.\det(V_n) =\prod_{1\le i<j\le n}(j-i) =\prod_{k=1}^{n-1}k!.

Substituting into (1) proves both equivalent formulas:

APD⁡n−1(Vn)=(n−1)!det⁡(Vn)=(n−1)!∏k=1n−1k!.\operatorname{APD}_{n-1}(V_n) =(n-1)!\det(V_n) =(n-1)!\prod_{k=1}^{n-1}k!.

The positive leading coefficient also proves m1(Vn)=n−1m_1(V_n)=n-1.