Abstract. Let R be a commutative ring, and let l ≥ 2; for l = 2 it is assumed additionally that R has no residue fields of two elements. The subgroups of the general linear group GL(n, R) that contain the elementary symplectic group Ep(2l, R) are described. In the case where R = K is a field, similar results were obtained earlier by Dye, King, and Shang Zhi Li.In the present paper we consider a description of the subgroups in the general linear group G = GL(2l, R) over a commutative ring R that contain the elementary symplectic group Ep(2l, R). It turns out that for every such group H there exists a unique ideal A in R such that H lies between the group EEp(2l, R, A) = Ep(2l, R) E(2l, R, A) and its normalizer in GL(2l, R). More specifically, our main objective in the present paper is a proof of the following result.
Theorem 1. Let R be a commutative ring. Suppose that either l ≥ 3, or l = 2 and the ring R has no residue fields of two elements. Then for every subgroup H in G = GL(2l, R) that contains the elementary symplectic group Ep(2l, R), there exists a unique ideal A R such thatEEp(2l, R, A) ≤ H ≤ N G (EEp(2l, R, A)).An important supplement to Theorem 1 is the following result, in which we explicitly calculate the normalizer of EEp(2l, R, A). Namely, consider the reduction homomorphism ρ A : GL(2l, R) −→ GL(2l, R/A) and denote by CGSp(2l, R, A) the complete preimage of the group GSp(2l, R) with respect to ρ A . Then the condition for a matrix to belong to CGSp(2l, R, A) is described by obvious quadratic congruences on its entries. Now we are in a position to state the second major result of the present paper.
Theorem 2. Under the assumptions of Theorem 1, for any ideal A R we have N G (EEp(2l, R, A)) = CGSp(2l, R, A).Thus, combining Theorems 1 and 2, we see that for any subgroup H with Ep(2l, R) ≤ H ≤ GL(2l, R) there exists a unique ideal such that
EEp(2l, R, A) ≤ H ≤ CGSp(2l, R, A).2000 Mathematics Subject Classification. Primary 20G35. The present paper has been written in the framework of the RFBR projects nos. 01-01-00924 and 00-01-00441, and INTAS 00-566. The theorem on decomposition of unipotents mentioned in §13 is a part of first author's joint work with A. Bak and was carried out at the University of Bielefeld with the support of AvH-Stiftung, . At the final stage, the work of the authors was supported by express grants of the Russian Ministry of Higher Education 'Geometry of root subgroups' PD02-1.1-371 and 'Overgroups of semisimple groups' E02-1.0-61. In the case where R = K is a field, there are only two possibilities for A, namely, A = 0 or A = K. We have EEp(2, K, 0) = Ep(2l, K) = Sp(2l, K), CGSp(2l, K, 0) = GSp(2l, K).On the other hand, EEp(2, K, K) = E(2l, K) = SL(2l, K), CGSp(2l, K, K) = GL(2l, K).In this special case, Theorem 1 asserts that any overgroup of Sp(2l, K) in GL(2l, K) either is contained in GSp(2l, K), or contains SL(2l, K), so that this theorem boils down to a theorem of R. Dye [48]. Subsequently, R. Dye, O. King, and Shang Zhi Li generalized this result to overgroups of ot...