On a Hyperdeterminant Computation of Cayley

76 Views Asked by At

I recently read a paper by Cayley entitled "On the theory of linear transformations" (it can be found in volume I of his collected works as number 13), wherein Cayley computes a "hyperdeterminant" for a $p=3$, $m=2$, $n=4$, which is his was of saying he's computing a function in the entries of a $2\times 2 \times 2\times 2$ tensor that can be expressed as a homogeneous polynomial of degree 3 in the maximal minors of certain matrices listing all the entries of the tensor in question. More precisely, if we let $$a = 1111, b=2111, c =1211,d = 2211, e = 1121, f = 2121, g = 2211, h = 2221$$ $$ i = 1112, j = 2112, k=1212, l= 2212, m = 1122, n = 2122, o = 2212, p = 2222$$ the matrices in question (whose maximal minors we're interested in) would look something like (they're only well defined up to a permutation of their columns): $$M^{(1)} = \begin{pmatrix} a & c & e & g & i & k & m & o \\ b & d & f & h & j & l & n & p \\ \end{pmatrix}\:,\: M^{(2)} = \begin{pmatrix} a & b & e & f & i & j & m & n \\ c & d & g & h & k & l & o & p \\ \end{pmatrix}$$ $$M^{(3)} = \begin{pmatrix} a & b & c & d & i & j & k & l \\ e & f & g & h & m & n & o & p \\ \end{pmatrix}\:,\: M^{(4)} = \begin{pmatrix} a & b & c & d & e & f & g & h \\ i & j & k & l & m & n & o & p \\ \end{pmatrix}$$ where the first matrix lists all possible columns of the form $\begin{pmatrix}1rst\\2rst\end{pmatrix}$, the second lists all columns of the form $\begin{pmatrix}r1st\\r2st\end{pmatrix}$, and similarly for the third and fourth matrices with the third and fourth indices, respectively, where we always have $$1\leq r,s,t\leq 2$$ To recap, Cayley's definition of a hyperdeterminant of order $p$ of a tensor is a polynomial in the coefficients which can be written as a (usually distinct) polynomial homogeneous of degree $p$ in the maximal minors of each of the matrices listed above (these matrices obviously being only for this particular case). For the case in question, Cayley writes: enter image description here

and the stated values continue for another half page or so. It's not clear to me what exactly Cayley is computing, and how. If someone could enlighten me, I would greatly appreciate it.