Let $f\in k[x_0,...,x_n]_d$ be a degree $d$ homogeneous polynomial in $n+1$ variables.

Is there a way to associate to $f$ a form $g(y_1,...,y_m)$ which is symmetric in the sets of binary variables $y_i = [y_0^i:y_1^i]$ of degree $d_i$ in each set for a suitable choice of $m$ and $d_i$?

For instance we could associate to a polynomial $f\in k[x_0,...,x_3]_2$ the form $g(y_1,y_2) = \sum_{0\leq i_1,i_2\leq 3}a_{i_1,i_2}(y_0^1)^{3-i_1}(y_1^1)^{i_1}(y_0^2)^{3-i_2}(y_1^2)^{i_2}$.

  • $\begingroup$ I see the beginning of an interesting question but it would be nice to know what exactly do you mean by "associated". $\endgroup$ May 25 '17 at 21:06

Not yet an answer but you can see $f$ as an element of the symmetric power $S^d(V^{\vee})$ for $V$ a vector space of dimension $n+1$. You can take $V=S^n(W)$ with $W$ of dimension 2. That should give something like your representation for $g$. The indices labeling the $x$ basis of $V^{\vee}$ become labels for monomials of degree $n$ in your $y$ variables.

More explicitly, if $$ f(x_0,\ldots,x_n)=\sum_{i_1,\ldots i_d=0}^n f_{i_1,\ldots,i_d}x_{1_1}\ldots x_{i_d} $$ with the (physicsy) tensor $f_{i_1,\ldots,i_d}$ being symmetric, then the associated form is $$ g(y^1,\ldots,y^d)=\sum_{i_1,\ldots i_d=0}^n f_{i_1,\ldots,i_d} (y_0^1)^{n-i_1}(y_1^1)^{i_1}\ldots (y_0^d)^{n-i_d}(y_1^d)^{i_d}\ . $$ It is of course essential to now see $f$ as a symmetric multilinear form rather than a homogeneous polynomial, i.e., to work with $n$ distinct pairs of binary variables $y$. This is the idea behind the symbolic method of classical invariant theory (see my answer to this MO question for more details).

  • $\begingroup$ So it was an answer after all. Glad to be of help. BTW may I ask about the motivation for the question? I am interested in everything that has to do with the classical symbolic method. $\endgroup$ May 26 '17 at 13:16
  • 1
    $\begingroup$ In this book "Algebraic geometry and theta functions" Coble refers sevaral times to such a correspondence without making it explicit. I was just trying to understand what he was talking about. Thanks again. $\endgroup$
    – F_L
    May 26 '17 at 17:45

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.