In Titchmarsh's book "The theory of the Riemann Zeta-function" his Lemma $3.12$ is one of the main tools. Lemma $3.12$ is a version of Perron formula. Lemma $3.12$ starts by observing that, with $c>0$ and with $n<x$,

$$ \frac{1}{2\pi i}\int\limits_{c-iT}^{c+iT}\left(\frac{x}{n}\right)^w \frac{\mathrm dw}{w}=1+ O\left( \frac{(x/n)^c}{T\log (x/n)} \right) $$

This stems from the fact that $c>0$, and so if one completes the integration contour to the left by encircling the point $w=0$ completely in the complex $w$ plane by a rectangle whose right-most side is the original integration line $(c-iT,c+iT)$, the residue is simply $1$. This is the reason for the requirement $c>0$. The requirement $n<x$ comes from letting the left-most side of the integration rectangle be at $\Re(w)=-\infty$ and so $(x/n)^w\to 0$ if and only if $n<x$ as $\Re(w)\to -\infty$.

And so, in Theorem $3.13$ the starting point is the application of Lemma $3.12$,

$$ \sum\limits_{n<x} \frac{\mu(n)}{n^s}=\frac{1}{2\pi i}\int\limits_{c-iT}^{c+iT} \frac{x^w \mathrm dw}{w\zeta(s+w)}+ O\left( \frac{x^c}{cT} \right)+ O\left( \frac{\log x}{T} \right) $$

On the LHS the condition $n<x$ is the condition from Lemma $3.12$, and $c$ is still required to obey $c>0$.

Now Titchmarsh enlarges the contour of integration and encloses the $w=0$ completely, as earlier. Do notice that $c>0$ should still hold, since $c$ is nothing else but $\Re(w)$ on the right-most side the integration contour. No matter how the original line $(c-iT,c+iT)$ is deformed, the deformed closed integration contour must enclose the point $w=0$, and hence $c>0$ on the RHS part of the contour that intersects the real axis in $w$ plane.

With $c>0$ the pole at $w=0$ produces the residue $1/\zeta(s)$ and the result is now

$$\sum\limits_{n<x} \frac{\mu(n)}{n^s}=\frac{1}{\zeta(s)}+O$$

The $O$ term tends to $0$ in the end, after Titchmarsh adjusts $T$ and $x$ appropriately.

And so then Titchmarsh requires

$$c=\frac{1}{\log x}$$

Additionally, he then lets $x\to\infty$.

But in this limit, $c\to 0$. And this makes the original starting equation useless, since the pole of the integrand at $w=0$ is no longer enclosed in the integration countour, but lies on the integration contour instead, if one interprets the limit as $\lim\limits_{x\to\infty}x=\infty$.

On the other hand, if one interprets the limit $\lim\limits_{x\to\infty}x$ in the way that $x$ grows large but never hits the point at infinity, then $c$ never reaches $0$, and this would do. But aren't there problems with this interpretation of a limit in standard mathematics?

Maybe Titchmarsh uses the Sokhotski–Plemelj theorem, but then the result misses the summand $\frac{1}{2\zeta(s)}$...

Or maybe I'm missing some detail completely here...

So my question is:

What exactly enables Titchmarsh to take the limit $x\to\infty$?


Let $\epsilon >0$; by the above (standard stuff you seem to agree with), there is a large enough half-integer $x(\epsilon)$ (depending on $s=1+it, \epsilon$), s.t. $|\sum\limits_{n<x} \frac{\mu(n)}{n^s}-\frac{1}{\zeta(s)}| < \epsilon$ for any other half-integer $x \geq x(\epsilon)$ - as usual there is no restriction in using half-integers (i.e. $odd/2$) beacause the Dirichlet sum is constant between integers and the jump goes to zero since $s=1+it$

(choosing in Perron, $c=\frac{1}{\log x}, \log T= (\log x)^{\frac{1}{10}}, \delta = A(\log T)^{-9}=A(\log x)^{-\frac{9}{10}},x$, hence $T$ large enough, $A$ positive absolute constant coming from zero-free regions of $\zeta$)

But then the relation $|\sum\limits_{n<x} \frac{\mu(n)}{n^s}-\frac{1}{\zeta(s)}| < \epsilon$ for $x>x(\epsilon)$ is precisely what is needed to conclude that the limit as $x$ goes to $\infty$ of $\sum\limits_{n<x} \frac{\mu(n)}{n^s}$ is precisely $\frac{1}{\zeta(s)}$, by the usual definition of limit.

So the point is that you apply Perron with large but finite $x$ and then in "shorthand" you let $x$ go to $\infty$ meaning the above $\epsilon - x(\epsilon)$ limit relation, which has nothing to do with making $x$ infinite as a number

  • $\begingroup$ Awesome, thank You, Conrad! $\endgroup$ – anonymous Mar 27 '19 at 18:04
  • $\begingroup$ You are welcome $\endgroup$ – Conrad Mar 27 '19 at 19:03
  • $\begingroup$ $A$ is a constant that comes from the zero-free region - it probably has a value depending on an inferior bound on $x$ but not sure what is though if you allow large enough $x$ you can probably make it as big as you want (eg if you use $x >1000$ then the choice of $T$ gives $T > T_0=(\log 1000)^{.01}$ and that gives a zero free region constant $A$ valid for those $T$). In general $A$ is irrelevant being just a constant that depends only on your intiial starting point but has no bearing on large values of $x,T$ $\endgroup$ – Conrad Jan 15 '20 at 20:25
  • $\begingroup$ @Conrad, what if we set $c = \eta + 1/\log(x)$ where $\eta$ is another small positive number. We will need to change $T = x$ So that all terms in RHS go to zero. And we don’t have to worry about $c$ going to zero. $\frac{x^c}{Tc} = \frac{x^{(\eta+1/\log(x))}}{ (x\eta+ x/\log(x))}$ goes to 0 as $x \rightarrow \infty$ etc.... $\endgroup$ – Shree Feb 5 '20 at 23:29
  • $\begingroup$ @shree the question makes sense only in the context of titchmarsh book as there is more going on and one needs to move the integral through the pole at $0$ coming from $1/w$ and still stay in a non zero area of zeta which is a function of $T$ so this determines $\delta(T)$ which in turn forces what $x$ can be in terms of $T$ etc; we then reverse engineer and get $T$ in terms of $x$ so they are not free parameters $\endgroup$ – Conrad Feb 6 '20 at 0:41

To see the idea of the proof, let $M(x) = \sum_{n \le x} \mu(n)$ and we'll look instead at $f(x) = \int_1^x M(y)dy$ to obtain absolutely convergent integrals.

Then Titchmarsh shows a lower bound $|\zeta(1+it)|\ge B/\log(2+| t|)$ plus an upper bound $|\zeta'(s)|\le B\log(2+| t|)$ to obtain some bound $|\zeta(s)|>1/(A\log(2+|t|)) ,|\frac1{\zeta(s)}|<\log(2+|t|)$ for $s=\sigma+it, \sigma\ge 1-\frac{1}{A\log^2(2+ |t|)}$.

And this is what we need to conclude $$f(x) = \frac{1}{2i\pi} \int_{2-i\infty}^{2+i\infty} \frac{1}{\zeta(s)} \frac{x^{s+1}}{s(s+1)}ds= \frac{1}{2i\pi} \int_{\Re(s) = 1-\frac{1}{A\log^2 (2+|\Im(s)|)}} \frac{1}{\zeta(s)} \frac{x^{s+1}}{s(s+1)}ds \\= O(\int_{-\infty}^\infty \frac{\log (2+|t|)}{1+t^2 }x^{2-\frac{1}{A \log^2(1+ |t|)}}dt)\\ = O(\int_0^T \frac{x^{2-\frac{1}{A \log^2(1+ |T|)}}\log (2+|t|)}{1+t^2 }dt)+O(\int_T^\infty \frac{x^{2}\log T}{1+t^2 }dt)\\ =O(x^{2-\frac{1}{A \log^2(1+ |T|)}}) + O(\frac{x^2 \log T}{T})\\=O(x^{2-\frac{1}{A \log^2(1+ e^{\log^{1/4} x})}}) +O(\frac{x^2\log^{1/4} x}{e^{\log^{1/4} x}})=O(\frac{x^2}{e^{\log^{1/8} x}})=o(x)$$

If $M(x)> cx$ infinitely often, as $M(x+y)\ge M(x)-y$ then $|f(x+cx/2)-f(x) | \ge \sum_{n=0}^{cx/2} (cx-n) \ge x^2 c^2/8$ infinitely often, contradicting that $f(x)=o(x^2)$. Thus we proved $M(x)=o(x)$, the PNT.

  • $\begingroup$ An interesting proof of an equivalent of PNT, thanks! There's a typo at the end of the long equation, the end result is $o(x^2)$. It's all visible though, no problems here. Thanks again! $\endgroup$ – anonymous Mar 27 '19 at 18:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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