Linear equations in primes

Abstract: Consider a system Psi of non-constant affine-linear forms psi_1,...,psi_t: Z^d -> Z, no two of which are linearly dependent. Let N be a large integer, and let K be a convex subset of [-N,N]^d. A famous and difficult open conjecture of Hardy and Littlewood predicts an asymptotic, as N -> infty, for the number of integer points n in K for which the integers psi_1(n),...,psi_t(n) are simultaneously prime. This implies many other well-known conjectures, such as the Hardy-Littlewood prime tuples conjecture, the twin prime conjecture, and the (weak) Goldbach conjecture.

In this paper we (conditionally) verify this asymptotic under the assumption that no two of the affine-linear forms psi_1,...,psi_t are affinely related; this excludes the important ``binary cases such as the twin prime or Goldbach conjectures, but does allow one to count ``non-degenerate configurations such as arithmetic progressions. Our result assumes two families of conjectures, which we term the Inverse Gowers-norm conjecture GI(s) and the Mobius and Nilsequences Conjecture MN(s), where s in {1,2,...} is the complexity of the system and measures the extent to which the forms psi_i depend on each other. For s = 1 these are essentially classical, and the authors recently resolved the cases s = 2.

Our results are therefore unconditional in the case s = 2, and in particular we can obtain the expected asymptotics for the number of 4-term progressions p_1 < p_2 < p_3 < p_4 <= N of primes, and more generally for any (non-degenerate) problem involving two linear equations in four prime unknowns.

Review text

Let \(\psi_1,\ldots,\psi_t:\mathbb{Z}^d\rightarrow\mathbb{Z}\) be non-constant affine linear forms, and let \(K\) be a convex sub-region of \([-N,N]^d\). Then there is a natural conjecture, of Hardy--Littlewood type, for the number of \(\mathbf{n}\in K\cap\mathbb{Z}^d\) for which the values \(\psi_i(\mathbf{n})\) are simultaneously prime. This paper gives a conditional proof of the conjecture, under the assumption that no two of the functions \(\psi_i\) have homogeneous parts which are parallel. This condition rules out the pair of functions \(n,n+2\), which would have handled the twin prime conjecture. None the less the paper represents a very substantial step forward in our understanding of such problems. One particular case of interest is that in which the forms are \(n_1,n_1+n_2,n_2+2n_2,\ldots,n_1+(t-1)n_2\), for which one obtains an asymptotic formula for the number of \(t\)-term arithmetic progressions of primes in appropriate regions. Previously [Ann. Math. (2) 167, No. 2, 481--547 (2008; Zbl 1191.11025)] the authors had proved that there are infinitely many \(t\)-term arithmetic progressions of primes, but the method had not produced an asymptotic formula. In general the method of the present paper gives a conditional proof that \[ \sum_{\mathbf{n}\in K\cap\mathbb{Z}^d}\Lambda(\psi_1(\mathbf{n}))\ldots \Lambda(\psi_t(\mathbf{n}))=c\text{Vol}(K)+o(N^d) \] as \(N\rightarrow\infty\), where the constant \(c\) is the usual product of local densities. As mentioned, the proof is a conditional one, subject to two hypotheses, namely that the ``inverse Gowers norm conjecture GI(s) and the ``Möbius and nilsequences conjecture MN(s) each hold for every integer \(s\geq 2\). The conjectures GI(2) and MN(2) had previously been proved by the authors [Proc. Edinb. Math. Soc., II. Ser. 51, No. 1, 73--153 (2008; Zbl 1202.11013)], and [Ann. Inst. Fourier 58, No. 6, 1863--1935 (2008; Zbl 1160.11017)], and these enable systems of linear forms of ``complexity at most 2 to be handled unconditionally. In particular the paper gives an unconditional proof for the case of sets of 4 primes in arithmetic progression. Indeed, in subsequent work by the authors, together with Ziegler, the conjectures GI(s) have been established for all \(s\); and similarly the authors have proved MN(s) in all cases. Thus the results of the present paper are now known unconditionally. The conjectures GI(s) and MN(s) are somewhat technical to describe. Very roughly, the inverse Gowers norm conjecture describes functions for which the Gowers norm \(U^{s+1}\) is large, saying this can only happen if the function correlates with an \(s\)-step nilsequence. The Möbius and nilsequences conjecture states roughly that the Möbius function \(\mu(n)\) does not correlate with \(s\)-step nilsequences. Thus one can think of MN(1) as saying that \(\sum_{n\leq N}\mu(n)e(\theta n)=o(N)\) for every real \(\theta\). The paper builds on the authors' earlier work on arithmetic progressions of primes, using the machinery of the transference principle and the generalized von Neumann theorem from that paper. A second major component involves nilmanifolds, for which the theory is largely developed from scratch so as to suit the particular needs of the current work. Overall this is a long and complicated paper, but the exposition is excellent. Given the importance of the results obtained it is well worth detailed study.

Recommendations

Cites work

Cited in

(only showing first 100 items - show all)

This page was built for publication: Linear equations in primes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q974077)