On a rationality problem for fields of cross-ratios

From MaRDI portal



Abstract: Let k be a field, ngeqslant5 be an integer, x1,dots,xn be independent variables and Ln=k(x1,dots,xn). The symmetric group Sn acts on Ln by permuting the variables, and the projective linear group mPGL2 acts by applying (the same) fractional linear transformation to each varaible. The fixed field LnmPGL2 is called "the field of cross-ratios". Let SsubsetSn be a subgroup. The Noether Problem asks whether the field extension LnS/k is rational, and the Noether Problem for cross-ratios asks whether KnS/k is rational. In an effort to relate these two problems, H. Tsunogai posed the following question: Is LnS rational over KnS? He answered this question in several situations, in particular, in the case where S=Sn. In this paper we extend his results by recasting the problem in terms of Galois cohomology. Our main theorem asserts that the following conditions on a subgroup SsubsetSn are equivalent: (a) LnS is rational over KnS, (b) LnS is unirational over KnS, (c) S has an orbit of odd order in 1,dots,n.


Let \(k\) be a field, \(n\geq 5\) an integer, \(x_1, \ldots, x_n\) independent variables, and \(L_n = k\left( x_1, \ldots, x_n \right)\). There is a natural action by the projective general linear group \(\text{PGL}_2\) on \(L_n\) via the fractional linear transformation \[\begin{pmatrix} a&b \\ c&d \end{pmatrix} \cdot x_i = \frac{ax_i + b }{cx_i + d}.\] Denote the fixed field of this action by \(K_n = L_n^{\text{PGL}_2}\). This is called the field of cross-ratios and is generated by the \(n-3\) elements \[\left[ x_1, x_2, x_3, x_i \right] = \frac{(x_i - x_1)(x_3-x_2)}{x_i - x_2)(x_3 - x_1)}, \qquad i= 4,\ldots, n.\] Fix a subgroup \(S\subseteq \Sigma_n\). \(S\) acts on both \(K_n\) and \(L_n\) by permuating the variables. The Noether Problem was first posed in [\textit{E. Noether}, Math. Ann. 78, 221--229 (1917; JFM 46.0135.01)] in connection with the inverse problem of Galois theory. It asks whether the fixed field \(L_n^S\) is rational, stably rational, or retract rational over \(k\). Furthermore, the Noether Problem for cross-ratios inquires if the fixed field \(K_n^S\) is rational, stably rational, or retract rational. In [Tokyo J. Math. 39, No. 3, 901--922 (2017; Zbl 1432.14014)], \textit{H. Tsunogai} asks if \(L_n^S\) is rational over \(K_n^S\), connection the Noether Problem and the Noether Problem for cross-ratios. In the present paper, the author reformulates this question in the language of algebraic groups, torsors, and Galois cohomology (see Proposition 3), which allows the author to provide a complete answer to Tsunogai's question. In Theorem 2, the author proves that the following are equivalent \begin{itemize} \item[1.] \(L_n^S\) is rational over \(K_n^S\), \item[2.] \(L_n^S\) is unirational over \(K_n^S\), \item[3.] \(S\) has an orbit of odd order in \(\{1, \ldots, n\}\). \end{itemize} The second section of the paper is devoted to the formulation of Tsugai's question in the language of Galois cohomology. In the third section, the author reviews necessary background on Galois cohomology. Section 4 contains the proof of Proposition 3. Sections 5 and 6 are devoted to the proof of Theorem 2, which is proved by first reducing to the case where \(S\) is a \(2\)-group.











This page was built for publication: On a rationality problem for fields of cross-ratios

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q784837)