Counting points of fixed degree and bounded height on linear varieties (Q982529)

Given a number field \(k\), for any bound \(B\) and any dimension \(N\) the number of points in \(\mathbb{P}^{N-1}_k(k)\) whose height is bounded above by \(B\) is easily seen to be finite. In [\textit{S. H. Schanuel}, Bull. Soc. Math. Fr. 107, 433--449 (1979; Zbl 0428.12009)] Schanuel gave an estimate for the number of such points, showing that as \(B\rightarrow\infty\) this number grows like \(C_1B^N+O(B^{N-1/m})\) where \(m=[k:\mathbb{Q}]\) and \(C_1\) is a constant depending on \(k\) and \(N\) only. This was generalized by \textit{J. L. Thunder} in [J. Number Theory 43, No. 2, 228--250 (1993; Zbl 0773.11022)] to more general linear subvarieties \(\mathbb{V}\subseteq \mathbb{P}^{N-1}_k\) to an estimate of the form \(C_1H(\mathbb{V})^{-1}B^N+O(B^{N-1/m})\), where \(H(\mathbb{V})\) is the height of the subvariety; this estimate was eventually improved by \textit{C. Christensen} and \textit{W. Gubler} in 2008 [Manuscr. Math. 126, No. 4, 505--525 (2008; Zbl 1155.11034)]. More generally, Northcott's theorem says that the number of points in \(\mathbb{P}^{N-1}_k(\bar{k})\) of bounded degree \(e\) is finite (the previous discussion being the case of degree \(e=1\)) and it is natural to ask for a generalization of the above results. This was firstly done by \textit{W. M. Schmidt} in [Monatsh. Math. 115, No.1-2, 169--181 (1993; Zbl 0784.11054)] in the case \(k=\mathbb{Q}\) and eventually generalized by several authors -- see [\textit{M. Widmer}, Acta Arith. 140, No. 2, 145--168 (2009; Zbl 1252.11054)] for a nice historical survey and references. In the same spirit, the main result of the paper under review is the theorem below, for which we introduce the following notation: for a number field \(K\) we set \(r_K\) and \(s_K\) to be the number of real and imaginary places, respectively, and for \(n\in\mathbb{N}\) we put \[ S_K(n)=\frac{h_KR_K}{w_K\zeta_K(n+1)}\Big(\frac{2^{r_K}(2\pi)^{s_K}}{\sqrt{|\Delta_k|}}\Big)^{n+1}(n+1)^{r_K+s_K-1} \] where \(h_K,R_K,w_K,\zeta_K,\Delta_K\) are the class number, the regulator, the number of roots of unity, the Dedekind zeta function and the discriminant of \(K\), respectively. Moreover, for \(n\in\mathbb{N}\) we define \(V(n)\) to be the volume of the Euclidean ball in \(\mathbb{R}^n\) of radius \(1\). Then we have \textbf{Theorem 1.1.} Let \(k\) be a number field of degree \(m\), let \(n,e\) and \(N\geq n+1\) be positive integers, and let \(\mathbb{V}\) be a linear subvariety of \(\mathbb{P}_k^{N-1}\) of dimension \(n\) defined over \(k\). Suppose that either \(e=1\) or \[ n>5e/2+4+2/(me)\;. \] Then the sum \[ \alpha(k,e,n)=\sum_{[K:k]=e}(2^{-r_K}\pi^{-s_K})^{n+1}V(n+1)^{r_K}V(2n+2)^{s_k}S_K(n) \] converges and as \(X>0\) tends to infinity we have \[ Z_2(\mathbb{V};k,e,X)=\alpha(k,e,n)H_2(\mathbb{V})^{-me}X^{me(n+1)}+O(X^{me(n+1)-1}\mathfrak{L}_0)\;. \] Here \(\mathfrak{L}_0=\log\max\{2,2X\}\) if \((me,n)=(1,1)\) and \(\mathfrak{L}_0=1\) otherwise; \(Z_2\) is the counting function of points on \(\mathbb{V}(\bar{k})\) of bounded degree \(e\) and height \(H_2\leq X\); and the implicit constant in \(O\) depends only on \(k,e,n\). Observe that the height \(H_2\) occurring in the theorem is a \(\ell^2\) modification of the usual \(\ell^\infty\) one. The paper starts with a very well-written introduction (Section 1.) providing for nice concrete examples and comparisons with previous works. Then Sections 2. to 5. are devoted to defining \(H_2\) and other suitable ``adelic-Lipshitz'' heights. Section 7. reduces the bound for \(H_2\) to a bound for a cleverly chosen adelic-Lipschitz height and Section 8. deduces this lower bound from the author's previous work (loc. cit.) (recalled in Section 6.).