A note on the relative class number in function fields (Q4336583)

Let \(\mathbb{F}_q\) be the finite field with \(q= p^m\) elements, let \({\mathbf A}=\mathbb{F}_q[T]\), and \({\mathbf k}=\mathbb{F}_q(T)\). Let \(f\in{\mathbf A}\) be a monic prime of degree \(d\) and let \({\mathbf k}(f)\) be the abelian extension of \({\mathbf k}\) obtained by adjoining the \(f\)-division points of the Carlitz module; it is well known that the class number of this extension splits up into a product \(h(f)= h(f)^+h(f)^-\). In the paper being reviewed, the author develops an elegant analog for \(h(f)^-\) of the Maillet determinant formula. In particular, the author establishes the following beautiful result. Let \(\mathcal M\) be the set of monic polynomials of degree less than \(d\) and set \(t=(1-q^d)/(1- q)\). For \(\{a,b\}\) in \(\mathcal M\) write \(ab= sf+r\) in line with the division algorithm; as \(f\) is irreducible one must have \(r\neq 0\). Set \(c(a,b)\) to be the leading coefficient of \(r\). Theorem: We have \(h(f)^-\) is divisible by \(p\) if and only if there is an integer \(i\), \(1\leq i\leq q-2\), such that the \(t\times t\) matrix \([c(a,b)^i]\) has \(0\) determinant.