Quiz: Part II
Ten questions on Masullo's argument as a whole.
These questions ask you to combine ideas from several chapters. Choose an answer to see whether it is right and why. Your best score is kept in this browser only.
The circle example and the real case have the same structure. What plays the role of the circle's factor ?
- The test function
- The L-function (for the lab, ): its zeros are the slots
- The Mellin transform
On the circle one divides out multiples of and gets one slot per mode. In the paper one divides out -functions, whose transforms are multiples of the L-function, and gets one slot per zero.
For the circle with and a test function that is zero outside , the trace of is:
- It depends on the shape of
- 0
No power of 3 lies in , so the geometric side is 0. The eigenvalues are not individually zero; they cancel. The same mechanism later removes all but finitely many primes.
Why is the thin strip introduced before the trace is computed?
- Without it, the eigenvalues of off-line zeros are amplified for test functions supported at , and nothing shows the sum over all zeros is the trace of an operator on a Hilbert space
- To make the labs faster
- Because the zeros are only defined inside a strip
The completion in the norm with weight gives a Hilbert space on which the operator is trace class, at the price of keeping only the slots inside the strip.
With strip half-width , which hypothetical zero would still contribute to the spectral side?
Only keeps a slot. 0.6 is exactly on the edge, where the bound of chapter 15 no longer exists.
A test function is zero outside . Which describes the geometric side?
- Every prime contributes a little
- Only primes with a power inside contribute; 23 and all larger primes contribute 0
- Only the prime 2 contributes
Each prime is a circle of length , contributing . This is the circle example's "trace is zero when is supported in " applied prime by prime.
Masullo's formula and the classical explicit formula have the same prime side. What does that imply?
- That the paper is correct
- Nothing
- That the zeros outside the strip must contribute 0 for every test function, which is possible only if there are none
Subtracting the two leaves only the far zeros. So the restricted formula, for every , is equivalent to the hypothesis. This comparison is this site's reading, not the paper's.
Suppose the Riemann Hypothesis were false. Which part of the argument would have to be wrong?
- The classical explicit formula
- A claim about the completed spaces: the exact sequence after completion, or the independent computation of the geometric side
- The step with the test functions
The classical formula is a theorem, and the last step is a short deduction from the trace formula. The weight of the proof is on the two completed-space claims, neither of which this site has checked.
In the last step, why does a zero with lead to a contradiction?
- It would be a pole of in a half-plane where every other term of the formula is holomorphic
- It would make the prime sum negative
- The functional equation forbids it
An identity cannot have a pole on one side only. The functional equation merely mirrors zeros; it forbids nothing.
What do the labs in Part II establish?
- That the proof is correct
- That the proof is wrong
- That the classical identities the construction is built to reproduce hold numerically; they do not test the new steps
Poisson summation on the circle, the factor in front of a -function's transform, and the explicit formula are all established mathematics. The labs make them visible.
Where does the paper say the of the hypothesis comes from?
- From the first zero
- From converting between the multiplicative measure of the norm and the additive measure of the Fourier transform, which needs the weight with
- From the factor in
In the paper's words, this "forces the 1/2 in the very statement of GRH". On this site the same shows up as the in a -function.