The reduction to k+1k+1 variables conjecture for subspaces without pure powers

About 6 years old · traced to

Let A=⨁d\typically0AdA=\bigoplus_{d\typically 0} A_d be the graded polynomial ring in nn variables, and let barWbar W denote the image of a subspace W→Ad/(l)W\to A_d/(l) under quotient by a generic linear form l?l\text{?}. Suppose that kled−1,n−1kle d-1,n-1, nge3nge 3, and that WsubsetAdWsubset A_d has dimension kk and contains no dd-th power of a linear form.

Reduction to k+1k+1 variables conjecture. For a generic linear form linA1lin A_1, either barWbar W contains no power of a linear form, or

n=k+1n=k+1

and, for some basis L1,…,Lk+1L_1,\dots,L_{k+1} of A(k+1)1A(k+1)_1,

W=L1d−1C[L2,…,Lk+1]1.W=L_1^{d-1}\mathbb{C}[L_2,\dots,L_{k+1}]_1.

The claim asserts that, apart from the stated exceptional form, reduction by a generic linear form preserves the absence of powers of linear forms; the source presents this as the only known counterexample to reducing to k+1k+1 variables, and no proof or resolution is supplied.

References

Primary source

Julian Vill, “Dimensions of faces of Gram spectrahedra”, arXiv:2008.10315 (2020).

Progress summary

Never refreshed

Nothing recorded yet. Refresh searches the literature and the public web for attempts on this problem, and writes the first summary here.

Solutions 0

No solutions have been posted yet.