MathLabs

Worked solution: Gelfond–Schneider transcendence proof via auxiliary functions (1934)

Step 5 of 7: The analytic upper bound: Jensen's formula makes Φ(s)(0)\Phi^{(s)}(0) tiny
In plain words

Complex analysis has a classical tool, Jensen's formula, that relates the value of a holomorphic function at the origin to how large it grows on a big circle and where its zeros sit inside that circle. Because our Φ\Phi has zeros of high order at every one of the points z0,…,zmz_0,\dots,z_m, this formula essentially says: since Φ\Phi vanishes so much inside the circle, its value (or the value of its ss-th derivative) at the origin has to be very small — smaller, in fact, than any fixed power of LL can compensate for, once the radius of the circle is chosen just right.

Carrying out the estimate carefully shows that ∣Φ(s)(0)∣|\Phi^{(s)}(0)| decays roughly like e−cLlog⁡Le^{-cL\log L} for some constant c>0c>0 depending only on KK — a genuinely tiny quantity once LL is large.

∣log⁡(s!)−log⁡∣Φ(s)(0)∣∣≲[K:Q] Llog⁡L\Big|\log\big(s!\big) - \log|\Phi^{(s)}(0)|\Big| \lesssim [K:\mathbb{Q}]\, L\log L
Detailed analysis

Jensen's formula states that for ff holomorphic on a disk of radius rr with zeros a1,…,ana_1,\dots,a_n inside (and f(0)≠0f(0)\ne0), log⁡∣f(0)∣=−∑jlog⁡(r/∣aj∣)+12π∫02πlog⁡∣f(reiθ)∣ dθ\log|f(0)|=-\sum_j\log(r/|a_j|)+\frac{1}{2\pi}\int_0^{2\pi}\log|f(re^{i\theta})|\,d\theta; when ff vanishes to order ss at 00 the formula generalizes to relate log⁡∣f(s)(0)/s!∣\log|f^{(s)}(0)/s!| to the same quantities (Siu, 'Jensen's Formula'). Applied to F(z)=Φ(z)F(z)=\Phi(z), but restricted only to the mm points z1,…,zm≠z0z_1,\dots,z_m\ne z_0 (each contributing a term log⁡(r/∣zj∣)\log(r/|z_j|), using that FF has order ≥L\ge L zeros there) gives an inequality, since these are not literally all the zeros of Φ\Phi in the disk (Siu, inequality labeled (‡)(\ddagger)).

Bounding the growth of Φ\Phi on the circle ∣z∣=r|z|=r: since eze^z and eβze^{\beta z} have order 11 (i.e. ∣ez∣≤Cϵe∣z∣1+ϵ|e^z|\le C_\epsilon e^{|z|^{1+\epsilon}}), and GG has coefficients and degree of size ≲L\lesssim L, one gets log⁡∣Φ(reiθ)∣≲L+Jr1+ϵ\log|\Phi(re^{i\theta})|\lesssim L+Jr^{1+\epsilon} (Siu, growth estimate via the Appendix's derivative bounds ∥Dλ(fjgk)∥≤(2J+L)LC2J+L\|D^\lambda(f^jg^k)\|\le(2J+L)^LC^{2J+L}). Choosing r=L1/2r=L^{1/2} balances the two terms and gives log⁡∣Φ(reiθ)∣≲L\log|\Phi(re^{i\theta})|\lesssim L.

Combining both sides of Jensen's formula (the left side contributing ≳mLlog⁡r=m2Llog⁡L\gtrsim mL\log r=\frac{m}{2}L\log L from the mm known zeros, the right side bounded by [K:Q]Llog⁡L[K:\mathbb{Q}]L\log L plus O(L)O(L), since the 'size' of the algebraic number Φ(s)(0)\Phi^{(s)}(0) costs an extra factor of [K:Q][K:\mathbb{Q}] to account for its conjugates) yields the final analytic bound: log⁡∣Φ(s)(0)∣≲−c Llog⁡L\log|\Phi^{(s)}(0)|\lesssim -c\,L\log L for an explicit constant c>0c>0 depending on [K:Q][K:\mathbb{Q}] once mm is chosen large enough relative to [K:Q][K:\mathbb{Q}] (Siu, proof of Main Theorem, final comparison of growth orders).

Terms in this step
Jensen's formula
A formula in complex analysis expressing log⁡∣f(0)∣\log|f(0)| for a holomorphic function ff in terms of the locations of its zeros inside a disk and the average of log⁡∣f∣\log|f| on the boundary circle — the key tool for turning 'many zeros' into a numerical upper bound on a function's value.