Part I · Quiz needs review

Quiz: Part I

Eight questions on the chapters of Part I.

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.

  1. Which chain of ideas correctly links the primes to the zeros of ζ\zeta?

    • Primes → Euler product → −ζ′/ζ=∑Λ(n)/ns-\zeta'/\zeta = \sum \Lambda(n)/n^{s} → poles at the zeros → one wave per zero in ψ(x)−x\psi(x) - x
    • Primes are the zeros of ζ\zeta → counting zeros counts primes
    • ζ(n)=0\zeta(n) = 0 when nn is prime → the explicit formula

    Unique factorization gives the product; its logarithmic derivative has the prime weights as coefficients; and the singularities of that function, at the zeros, become the oscillating terms.

  2. Suppose it were known only that every zero has real part at most 0.6. What would follow for the error ψ(x)−x\psi(x) - x?

    • Nothing
    • It would be at most about x0.6x^{0.6}, better than anything known today but weaker than the hypothesis
    • It would be at most about x\sqrt{x}

    The error is governed by the largest wave, xmax⁡βx^{\max \beta}. Each improvement in the zero-free region is an improvement in the error term; the hypothesis is the endpoint 12\tfrac12.

  3. Why is the sum ∑1/ns\sum 1/n^{s} equal to a product with one factor per prime?

    • Because there are infinitely many primes
    • Because the series converges
    • Because every whole number factors into primes in exactly one way

    Multiplying out the factors 1+p−s+p−2s+⋯1 + p^{-s} + p^{-2s} + \cdots produces each 1/ns1/n^{s} exactly once. Convergence is needed for the manipulation but is not the reason.

  4. Billions of zeros have been computed and all are on the line. Why is that not a proof?

    • The computations might contain errors
    • There are infinitely many zeros, and a zero off the line at great height would have no visible effect in any range that can be computed
    • Because the zeros computed were the wrong ones

    A finite check cannot cover an infinite set. Chapter 3 shows the amplitude of a zero's wave falling with its height, so a distant exception leaves no trace at accessible xx.

  5. If ζ\zeta had a zero at 0.3+14i0.3 + 14i, how many zeros would that force in total, and why?

    • Two: the zero and its complex conjugate
    • One
    • Four: conjugation gives 0.3−14i0.3 - 14i, and the mirror symmetry s↔1−ss \leftrightarrow 1 - s gives 0.7±14i0.7 \pm 14i

    Two of the four lie to the right of the line. That is why the hypothesis can be stated as "no zeros with σ>12\sigma \gt \tfrac12".

  6. Why do the chapters work with ψ\psi and not directly with the prime count π\pi?

    • The weights of ψ\psi are exactly the coefficients of −ζ′/ζ-\zeta'/\zeta, so its formula in terms of zeros is exact and simple
    • ψ\psi is easier to compute
    • π\pi is not related to the zeros

    Both carry the same information, and results transfer between them. But the explicit formula for ψ\psi has one clean term xρ/ρx^{\rho}/\rho per zero.

  7. In the explicit formula, what do the real part β\beta and the height γ\gamma of a zero each control?

    • β\beta the frequency, γ\gamma the size
    • β\beta the size of the wave (xβx^{\beta}), γ\gamma its frequency in ln⁡x\ln x
    • Both control the size

    xρ=xβ⋅eiγln⁡xx^{\rho} = x^{\beta} \cdot e^{i\gamma \ln x}. The hypothesis is a statement about the sizes only: every wave has size x\sqrt{x}.

  8. The Riemann Hypothesis is about zeros with real part between 0 and 1, where the sum ∑1/ns\sum 1/n^{s} diverges. How can ζ\zeta have zeros there at all?

    • The sum converges there after all, if enough terms are taken
    • ζ\zeta there means the analytic continuation: the one analytic function that agrees with the sum where the sum converges
    • The zeros are zeros of the partial sums

    As with 1/(1−x)1/(1 - x) and its geometric series, the function exists on a larger domain than the formula first used to define it, and rigidity makes the extension unique.