Moduli space of \(G\)-connections on an elliptic curve (Q494733)

Let \(X\) be a smooth complex elliptic curve and \(G\) a connected reductive affine algebraic group defined over \(\mathbb{C}\). Let \({\mathcal M}_X(G)\) denote the moduli space of topologically trivial algebraic \(G\)-connections on \(X\), that is, pairs of the form \((E_G, D)\), where \(E_G\) is a topologically trivial algebraic principal \(G\)-bundle on \(X\) and \(D\) is an algebraic connection on \(E_G\). Since the fundamental group \(\pi_1(X)\) can be identified with \(\mathbb Z\oplus \mathbb Z\), one identifies \(\mathrm{Hom}(\pi_1(X), G)\) with \(G\times G\). Let \(\mathcal R_X(G)\subset \mathrm{Hom}(\pi_1(X), G)=G\times G\) be the connected component containing the identity \((e, e)\). The group \(G\) acts by conjugations on \(G\times G\) (on both components), this action restricts to \(\mathcal R_X(G)\), the geometric quotient \(M_X^B(G)=\mathcal R_X(G)/\!\!/G\), which is an affine variety, is called the Betti moduli space. The Riemann-Hilbert correspondence sends a flat connection on \(X\) to its monodromy representation. This gives a biholomorphic map \({\mathcal M}_X(G)\to M_X^B(G)\). Since \(M_X^B(G)\) is an affine variety, this implies that \(\mathcal{M}_X(G)\) possesses many nonconstant holomorphic functions. The main result of the paper states however that \({\mathcal M}_X(G)\) does not have any nonconstant algebraic functions (Theorem 1.1). The paper under review consists of four sections. The first section is an introduction. The main theorem is formulated here. The definition of an algebraic connection is given in Section 2, a characterization of principal \(G\)-bundles that admit an algebraic connection is given here in Proposition 2.3. In Section~3 the author demonstrates that the moduli space \(\mathcal M_X\) of rank one algebraic connections on \(X\) does not have any nonconstant algebraic functions. The main result of the paper is proven in Section 4. The author considers a maximal torus \(\mathcal T\subset G\). The inclusion gives a surjective morphism \({\mathcal M}_X(\mathcal T)\subset {\mathcal M}_X(G)\). Since \({\mathcal M}_X(\mathcal T)\) is isomorphic to a finite direct sum of \(\mathcal M_X\), this leads to the required statement.