Arithmetic lattices in unipotent algebraic groups (Q2303705)

In 1988, a remarkable paper by \textit{F. J. Grunewald} et al. [Invent. Math. 93, No. 1, 185--223 (1988; Zbl 0651.20040)] flagged off the juggernaut of subgroup growth and the subject has grown to vast proportions now. The current paper which pursues and extends these ideas is also inspired by another beautiful paper that appeared in 2012; the authors say ``We were moved to pursue this subject after reading the work of \textit{N. Avni} et al. [Isr. J. Math. 188, 259--279 (2012; Zbl 1260.22003)] on the related concept of commensurator growth.'' The latter work is the first refined study of commensurator growth for lattices in certain locally compact groups. If \(A\) is a subgroup of a group \(G\), the commensurator of \(A\) in \(G\) is the group \[\mathrm{Comm}(A,G) := \{g \in G: [A : A \cap gAg^{-1}], [gAg^{-1} : A \cap gAg^{-1}] < \infty \}\] It is well-known that when \(A\) is a lattice in a semisimple Lie group \(G\), the commensurator is an important object which helps in determining whether \(A\) is arithmetic or not. In their paper, Avni, Lim and Nevo [loc. cit.] consider a natural filtration of \(\mathrm{Comm}(A,G)\) as follows: For any positive integer \(n\), consider the set \[\mathrm{Comm}_n(A,G) = \{g \in Comm(A,G) : [A : A \cap gAg^{-1}] = n \}.\] Under the action of \(N_G(A)\), this set breaks into classes and, one looks at the number \(c_n(A,G)\) of elements in the quotient set \(N_G(A) \backslash \mathrm{Comm}_n(A,G)\). Note that \(c_1(A,G)=1\) as \(\mathrm{Comm}_1(A,G) = N_G(A)\). They studied the growth of \(c_{\leq n}(A,G)\) as \(n \rightarrow \infty\) for a number of examples of \(A\) and \(G\). In the present paper, the authors consider unipotent algebraic \(\mathbb{Z}\)-groups \(G\) and the groups \(\mathrm{Comm}(G(\mathbb{Z}), \Delta)\) for arithmetic lattices \(\Delta\) in \(G(\mathbb{R})\) (for the same \(\mathbb{Z}\)-structure). If \(c_n\) denotes the number of \(\Delta\) such that the products of the indices of \(G(\mathbb{Z}) \cap \Delta\) in the two groups equals \(n\), the `commensurability zeta function' \(\zeta_G^c\) of \(G\) is defined by the Dirichlet series \(\sum_n c_n n^{-s}\). Following the outline of the work of Grunewald, Segal and Smith [loc. cit], the authors of the present paper show that this commensurability zeta function admits an Euler product where each local factor \(\zeta_{G,p}^c\) is a rational function in \(p^{-s}\) over \(\mathbb{Q}\) and the numerators and denominators are bounded independent of \(p\). While the GSS-paper used some model-theoretic results of \textit{J. Denef} [Invent. Math. 77, 1--23 (1984; Zbl 0537.12011)] to prove rationality of the local factors, the present paper uses results of \textit{A. Macintyre} [Ann. Pure Appl. Logic 49, No. 1, 31--74 (1990; Zbl 0731.12015)] that appeared after the GSS paper. This is necessitated by the fact that the authors require to consider integrals over noncompact sets. The additive group \(G=G_a\) (with \(G(\mathbb{Z})= \mathbb{Z}\)) has the commensurability zeta function \(\frac{\zeta(s)^2}{\zeta(2s)}\). The authors mention that even the two-dimensional case is too difficult for them to obtain the zeta function explicitly.