Scully's refined splitting-pattern conjecture for quadratic forms

Let FF be a field of characteristic different from 22, and let pp and qq be anisotropic quadratic forms over FF of dimension at least 22. Let h(q)h(q) be the height of the splitting pattern of qq, and let q0,q1,,qh(q)q_0,q_1,\ldots,q_{h(q)} be its successive higher anisotropic kernels. Set k=dim(q)2iW(qF(p))k=\operatorname{dim}(q)-2\mathfrak{i}_W(q_{F(p)}), and choose ll with 0lh(q)0\leq l\leq h(q) and dim(ql)=k\operatorname{dim}(q_l)=k. Let ss be the unique non-negative integer satisfying

2s<dim(p)2s+1,2^s<\operatorname{dim}(p)\leq 2^{s+1},

and suppose that k<2sk<2^s. For each rr with 0r<l0\leq r<l, Scully's refined conjecture. There exist integers ar,br1a_r,b_r\geq1 and integers ϵr\epsilon_r with kϵrk-k\leq\epsilon_r\leq k such that

dim(qr)=2s+1ar+ϵr,\operatorname{dim}(q_r)=2^{s+1}a_r+\epsilon_r,

and, with at most one exception, either

ir+1(q)12(k+ϵr)\mathfrak{i}_{r+1}(q)\leq\frac{1}{2}(k+\epsilon_r)

or

ir+1(q)=br2s+1+ϵr.\mathfrak{i}_{r+1}(q)=b_r2^{s+1}+\epsilon_r.

The possible exception is when dim(qr)=2s+1k\operatorname{dim}(q_r)=2^{s+1}-k; then r=l1r=l-1 and ir+1(q)=2sk\mathfrak{i}_{r+1}(q)=2^s-k. This conjecture refines the dimension-and-Witt-index conjecture to the full splitting pattern of qq. The source proves it when k7k\leq7 or when dim(q)2s+2+k\operatorname{dim}(q)\leq2^{s+2}+k, while the general statement was presented as conjectural.

Sources & referencesView supporting material

Primary source

Stephen Scully, “Hyperbolicity and near hyperbolicity of quadratic forms over function fields of quadrics”, arXiv:1609.07100 (2017).

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.