Skip to main content
2 events
when toggle format what by license comment
Oct 16, 2020 at 19:48 comment added Qiaochu Yuan The generalization I guess is to observe that when you pass from a bilinear form $B$ to a quadratic form $Q$ to a bilinear form again you end up with $2B$, so for $B$ to induce a quadratic form it suffices for $2B$ to be a bilinear form, not $B$; so you can take pointwise square roots of a bilinear form (my apologies for switching back to thinking multiplicatively here) and the result will still induce a genuine quadratic form. Of course these will only be non-unique if there's $2$-torsion.
Oct 14, 2020 at 23:36 history answered Qiaochu Yuan CC BY-SA 4.0