Centers of higher degree forms. (Q1406299)

A \(d\)-linear space of dimension \(n\) over a field \(k\) is a pair \((V,\Theta)\), where \(V\) is a vector space of dimension \(n\) over \(k\) and \(\Theta\) is a symmetric \(d\)-linear form on \(V\). \((V,\Theta)\) is regular if \(\Theta(w,V,\dots,V)=0\) implies \(w=0\), and its centre \(\text{Cent}(V,\Theta)\) is the commutative subalgebra of \(\text{End}_k(V)\) consisting of all \(f\) such that \(\Theta(f(v_1),v_2,\dots,v_d)= \Theta(v_1,f(v_2),\dots,v_d)\) for all \(v_i\in V\). The motivating question for this paper is: Which subalgebras of \(\text{End}_k(V)\) can be realized as the centre of some regular form on \(V\)? As a step towards addressing this question, the authors conjecture that if \(d>2\) and \((V,\Theta)\) is regular, then \(\dim _k\text{Cent}(V,\Theta)\leq n\), and prove this conjecture for \(n\leq 5\). The cases \(n\leq 3\) and \(n=4\) are easily proved, but the case \(n=5\) requires a lengthy argument using the theory of elementary divisors and Jordan forms. The authors conclude by remarking that the same techniques do not appear strong enough to attack the case \(n=6\), and add that they ``do not really have much evidence or intuition about Conjecture 2.5 for large values of \(n\)''. We point out that \textit{A. Keet} [Quaest. Math. 16, 413--442 (1993; Zbl 0797.11043)] has defined, for any \(d\geq 2\) and any vector space \(V\) of dimension \(n\), the hyperbolic \(d\)-linear space \(\mathbf{hyp}(V)\), which has dimension \(n+\binom {n+d-2}{n-1}\) while its centre has dimension \(1+\binom {n+d-2}{n-1}\).