Galois groups of Mori trinomials and hyperelliptic curves with big monodromy (Q253178): Difference between revisions

In the paper under review, the author gave refinement and extension of results obtained by \textit{S. Mori} [Jpn. J. Math., New Ser. 3, 105--109 (1977; Zbl 0379.14011)] concerning a class of polynomials (in fact, trinomials) of degree \(2g+1\) and the Jacobian of the genus \(g\) hyperelliptic curve \(C_f:y^2=f(x)\) by a detailed study of the corresponding Galois group \(\mathrm{Gal}(f/{\mathbb Q})\) of \(f\) over \({\mathbb Q}\). More precisely, let \(u\) be a polynomial of an odd degree \(n=2g+1\) over a field \(K\) such that \(\mathrm{char}(K)\neq 2\). Then there is the genus \(g=(n-1)/2\) hyperelliptic curve \[ C_u:y^2=u(x) \] whose Jacobian \(J(C_u)\) is a \(g\)-dimensional abelian variety over \(K\). In [loc. cit.], \textit{S. Mori} observed that if furthermore the Galois group of \(u\) over \(K\) is a \textit{doubly transitive} permutation group, then \(\mathrm{End}_K(J(C_u))={\mathbb Z}\). Mori constructed explicit trinomials \(f(x)=f_{g,p,b,c}(x)=x^{2g+1}-bx-\frac{pc}4\) over \({\mathbb Q}\) and showed that the Galois group of these trinomials is doubly transitive, hence \(\mathrm{End}(J(C_f))={\mathbb Z}\) (the endomorphism ring over \(\overline{\mathbb Q};\) see below for some more details). Here the parameters \(g,p,b\) and \(c\) are integers satisfying some divisibility and congruent properties that were recalled in the paper under review, and the results of Mori were restated as Theorem 1.4 in the paper under review as well. The first main result of the paper under review is Theorem 1.7. For simplicity, the reviewer mentions that the main point is to show that in Mori's examples \(\mathrm{Gal}(f/{\mathbb Q})={\mathbf S}_{2g+1}.\) When combined with an earlier result of the author [Math. Res. Lett. 7, No. 1, 123--132 (2000; Zbl 0959.14013)] (i.e. Theorem 1.2 below), this gives another proof of the theorem of Mori [loc. cit., Theorem 1, p.105]. Theorem 1.2 Suppose that \(\mathrm{char}(K)=0\) and \(\mathrm{Gal}(u/K)={\mathbf S}_n.\) Then \(\mathrm{End}(J(C_u))={\mathbb Z}.\) In fact, Theorem 1.7 of the paper under review extends Mori's results by dropping one of the congruence conditions, namely the condition that \(c\) is congruent to \(-p\) modulo \(4\). This condition was used by Mori to show first that \(J(C_f)\) has everywhere semi-stable reduction over \({\mathbb Z}\), and then to derive that \(\mathrm{End}(J(C_f))=\mathrm{End}_{\mathbb Q}(J(C_f))\) by using a result of \textit{K. A. Ribet} [Ann. Math. (2) 101, 555--562 (1975; Zbl 0305.14016), Corollary 1.4]. Using results of \textit{C. Hall} [Bull. Lond. Math. Soc. 43, No. 4, 703--711 (2011; Zbl 1225.11083)] regarding semistable reduction with toric dimension \(1\), the author of the paper under review applied Theorem 1.7 and his recent results [Trans. Am. Math. Soc. 368, No. 5, 3651--3672 (2016; Zbl 1343.14025)] to show that the images of the absolute Galois group \(\mathrm{Gal}({\mathbb Q})\) in the automorphism groups of Tate modules of \(J(C_f)\) are almost as large as possible (see statements of Theorem 1.11 of the paper under review). The reviewer wants to note that in most situations, the author of the paper under review follows Mori by assuming that \(g\) is a natural number, but in the statements involving endomorphism rings, the author assumes further that \(g>1\). The reason may be because the author studies only hyperelliptic curves as the title suggests. However Mori's original paper appears to contain results for all \(g\geq 1\). In this sense, we leave the task of determining whether the corresponding statements with \(g=1\) in the paper under review remain true to an interested reader. To end the review, the reviewer briefly mentions the organization of the paper. Section 2 gives the proof of Theorem 1.11. Section 3 studies reduction of certain trinomials. Section 4 studies discriminants of Mori trinomials. Section 5 gives the proof of Theorem 1.7. Section 6 gives generalization in the case of number field \(K\). The generalized Mori trinomials are now parameterized by the generalized Mori quadruples \((g,\mathfrak{p},{\mathbf b},{\mathbf c}),\) where \(g\) is a positive integer, \(\mathfrak{p}\) a maximal ideal in the ring of integers \(\mathcal{O}\) of \(K\) while \({\mathbf b}\) and \({\mathbf c}\) are elements of \(\mathcal{O}\) satisfying suitable conditions. Section 7 contains corrigendum to the author's earlier paper [Cent. Eur. J. Math. 12, No. 5, 659--674 (2014; Zbl 1296.11068)].