Remarks on symplectic twistor spaces (Q731401): Difference between revisions
From MaRDI portal
Normalize DOI. |
Normalize DOI. |
||
Property / DOI | |||
Property / DOI: 10.1007/S10231-008-0082-5 / rank | |||
Property / DOI | |||
Property / DOI: 10.1007/S10231-008-0082-5 / rank | |||
Normal rank |
Latest revision as of 02:23, 10 December 2024
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Remarks on symplectic twistor spaces |
scientific article |
Statements
Remarks on symplectic twistor spaces (English)
0 references
2 October 2009
0 references
The author studies a symplectic analogue of the Penrose twistor space, where the usual metric connection is replaced by a symplectic connection. Following \textit{N. O'Brian} and \textit{J. Rawnsley} [Ann. Global Anal. Geom. 3, 29--58 (1985; Zbl 0526.53057)], he begins by defining a twistor space to be an almost-complex fibre bundle \(Z\) over a real manifold \(M\) of dimension \(2n\). Now define \({\mathcal Z}^0\) to be \(FM/GL(n,{\mathbb C})\), where \(FM\) is the frame bundle of \(M\) and \(GL(n,{\mathbb C})\) acts through its natural embedding in \(GL(2n,{\mathbb R})\). Given a linear connection \(\nabla\) on \(M\), the bundle \({\mathcal Z}^0\) acquires a Penrose almost complex structure \({\mathcal J}^\nabla\) and we have a fibre-preserving map \(f:Z\to{\mathcal Z}^0\). The author deduces the integrability of the almost complex structure on \(Z\) in certain cases (Theorem 1) and studies the self-holomorphic sections of \({\mathcal Z}^0\), using these to define a new moduli space of complex structures. In the second part of the paper, the author begins by considering a real vector space \(V\) of dimension \(2n\) and the space \({\mathcal J}(V)\) of complex structures on \(V\). When \(V\) admits a symplectic form \(\omega\), one can then define, for \(0\leq l\leq n\), a subspace \(J(V,\omega,l)\) as the space of all those complex structures \(J\in {\mathcal J}(V)\) which are compatible with \(\omega\) and for which the symmetric form \(g_J:=\omega(\;,J)\) has signature \((2n-2l,2l)\) (the signature is always even). One can check that \(J(V,\omega,l)= \text{Sp}(V,\omega)/U(n-l,l)\). In fact, there exists an anti-biholomorphic \(\text{Sp}(2n,{\mathbb R})\)-equivariant map \({\mathcal D}_n\to J(V,\omega,0)\), where \({\mathcal D}_n\) is the Siegel upper half-space (Proposition 4). One can embed the disjoint union of the \(J(V,\omega,l)\) holomorphically in the Grassmannian \(\text{Gr}(n,V^{\mathbb C})\) and the image is the locally closed manifold of real-Lagrangian subspaces (Proposition 5). Now let \(M\) be a real manifold of dimension \(2n\) endowed with a non-degenerate \(2\)-form \(\omega\). Associated with \(M\), we have a bundle \({\mathcal Z}^{\omega,l}\) of complex linear structures which one may call the symplectic twistor space. Given a linear connection \(\nabla\) such that \(\nabla\omega=0\), one can then define the Penrose almost complex structure \({\mathcal J}^\nabla\) on \({\mathcal Z}^{\omega,l}\). The author shows that, if \({\mathcal J}^\nabla\) is integrable on one \({\mathcal Z}^{\omega,l}\), it is integrable on all of them (Proposition 6). Finally, section 3 contains more explicit results for the case of a Riemann surface with its canonical symplectic-metric connection.
0 references
linear connections
0 references
twistor spaces
0 references
almost complex structures
0 references
moduli spaces
0 references
0 references