This question is a follow-up from About Goldbach's conjecture.

As $N_{2}(n)=\sum_{r\leq n}1_{\mathbb{P}}(n-r)1_{\mathbb{P}}(n+r)$, Chebotarev's theorem allows to write:

$$\dfrac{N_{2}(n)}{\pi(n)}\sim \dfrac{N_{1}(n)}{\varphi(P_{ord_C(n)})}$$

So that

$$N_2(n)\sim \pi(n) \dfrac{N_{1}(n)}{\prod_{2\lt p\le \sqrt{2n-3}}(p-1)}$$.

One can easily show that $$N_{1}(n)={\prod_{2\lt p\le \sqrt{2n-3}}(p-2)}\prod_{p\vert n \\ p>2}\dfrac{p-1}{p-2}$$.

Thus one gets

$$N_{2}(n)\sim \pi(n)\dfrac{C}{\log\sqrt{2n-3}}\prod_{p\vert n \\ p>2}\dfrac{p-1}{p-2}$$ and, through the prime number theorem:

$$N_{2}(n)\sim \dfrac{Kn}{\log^{2} n}\prod_{p\vert n \\ p>2}\dfrac{p-1}{p-2}$$ with $K>0$.

As $N_{2}(n)$ is the number $G(2n)$ of couples $(p,q)$ such that $p+q=2n$, $p\leq q$, $p$ and $q$ primes, this shows that

$$G(2n)\sim \dfrac{Kn}{\log^{2} n}\prod_{p\vert n \\ p>2}\dfrac{p-1}{p-2}$$.

I have been told on a French maths forum that Hardy and Littlewood rigorously proved that if such a $K$ exists, then its value is such that their conjecture sometimes knowns as "extended Goldbach's conjecture" is true. Would you have a reference where this proof is given?
Thanks in advance.


1 Answer 1


This is from Section 4 of Hardy and Littlewood's "On some problems of partitio numerorum III: On the expression of a number as a sum of primes" 1923 Acta Math. 44: 1–70.

  • $\begingroup$ Thank you but I can't access. $\endgroup$ Aug 4, 2012 at 21:56
  • $\begingroup$ Is there no library that would do an interlibrary loan for you? $\endgroup$ Aug 4, 2012 at 22:57
  • $\begingroup$ Finally, I could get the whole article. Thank you again. $\endgroup$ Aug 5, 2012 at 11:58
  • $\begingroup$ I removed the link given in the answer, as it (inadvertently I suppose) was the link through some proxy of some specific institution, which as such is useful for hardly anybody and might rather cause confusion. $\endgroup$
    – user9072
    Aug 5, 2012 at 12:38
  • $\begingroup$ That was indeed inadvertent. I replaced the link with a generic link to the Springerlink access point (paywalled).This, however, should work on many institutions networks. $\endgroup$
    – Mark Lewko
    Aug 5, 2012 at 17:20

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.