Let $P,Q,R$ be the Fourier series of the Eisenstein series $E_2,E_4,E_6$, that is,

$$ P(q)=1-24\sum_{n=1}^{\infty}\sigma_1(n)q^n, $$

$$ Q(q)=1+240\sum_{n=1}^{\infty}\sigma_3(n)q^n, $$

$$ R(q)=1-504\sum_{n=1}^{\infty}\sigma_5(n)q^n. $$

I heard that $P,Q,R$ are algebraically independent over $\mathbb C(q)$, due to Mahler, but I cannot find a reference. Could you help me if you know the reference?

Meanwhile, I found 'On Algebraic Differential Equations Satisfied by Automorphic Functions', by Mahler, 1969. I do not think that this paper addresses this topic, but if it does, why does it imply the algebraic independence of $P,Q,R$?

  • 1
    $\begingroup$ The problem for just $Q$ and $R$ is settled at mathoverflow.net/questions/282554/… $\endgroup$ – Gerry Myerson Oct 30 '18 at 11:55
  • 3
    $\begingroup$ What about $\mathbb C (z) $? $\endgroup$ – LWW Oct 30 '18 at 12:05
  • 5
    $\begingroup$ @GerryMyerson OP's question is about independence over $\mathbb{C}(z)$, not $\mathbb{C}$, so a harder question. It may be that using the functional equation for the Eisenstein series will reduce to the easier problem but it's not obvious. A harder question, that probably gives an answer to OP's question, is the Mahler-Manin conjecture, proved by K. Barré-Sirieix et al., (Invent. Math. 124 (1996), 1-9). That paper might contain a proof or a reference to OP's question. $\endgroup$ – Felipe Voloch Oct 30 '18 at 16:08
  • $\begingroup$ Voloch // why Mahler-Manin conjecture implies the question?? I didn't understand. $\endgroup$ – LWW Oct 31 '18 at 2:25
  • $\begingroup$ If you find $z$ such that $P(z),Q(z),R(z)$ are all algebraic, then by Mahler-Manin, $z$ is transcendental. This wouldn't happen if the functions were algebraically dependent. Have a look at the comment after Theorem 4 of Waldschmidt's Bourbaki talk n° 824. I think it completely answers your question. $\endgroup$ – Felipe Voloch Oct 31 '18 at 4:53

I think the follwing is the answer of this question. It is not depend on the above comments because I didn't understand them. This is just my approach. (But not fully my idea because it depends on a strong proposition which is already known, and the remaining part is just a corollary.)

Proposition. Let $f$ be a modular form of weight $k$, defined over $\mathbb Q^{\mathrm{alg}}$. If $f$ is non constant, then the functions $f, Df$ and $D^2 f$ are algebraically independent over the function field $\mathbb C(q)$.

The proof is in the book, 'Introduction to Algebraic Independence Theory', Y.V.Nesterenko and P.Philippon, (Eds), Ch.1 prop 1.1. Here $q=e^{2\pi i \tau}$ and $D=\frac{1}{2\pi i}{\tau}=q\frac{d}{dq}$.

Let $\Delta = E_4^3-E_6^2$, cusp form of weight 12. Due to Ramanujan, we have the following formulas :

$$ DP=\frac{1}{12}(P^2-Q), \quad DQ=\frac{1}{3}(PQ-R), \quad DR=\frac{1}{2}(PR-Q^2). $$

Thus we have $D\Delta = P\Delta$. Take $D$ in both side and do some elementary algebra, we have

$$ Q=13\frac{(D\Delta)^2}{\Delta^2}-12\frac{D^2 \Delta}{\Delta}. $$

By the proposition, $\{\Delta, D\Delta, D^2\Delta\}$ is algebraically independent over $\mathbb C(q)$, so $\{Q^3-R^2,P,Q\}$ is algebraically independent over $\mathbb C(q)$, so $\{P,Q,R^2\}$ is algebraically independent over $\mathbb C(q)$. This is equivalent to say that $R^2$ is transcendental over $\mathbb C(q)[P,Q]$, and this implies that $R$ is transcendental over $\mathbb C(q)[P,Q]$, i.e. $\{P,Q,R\}$ is algebraically independent over $\mathbb C(q)$.

| cite | improve this answer | |
  • $\begingroup$ It's worth mentioning Zagier's generalization of Jacobi and Ramanujan's D-algebraic dependence, see people.mpim-bonn.mpg.de/zagier/files/doi/10.1007/… $\endgroup$ – Igor Pak Nov 4 '18 at 20:06
  • $\begingroup$ Pak // Could you let me know where can I find those contents? I tried to find it in the textbook you refered but fail... $\endgroup$ – LWW Nov 4 '18 at 22:46
  • $\begingroup$ Don't remember the exact place. Start with p. 49 for D-alg eq. but also see short section on p. 85 for alg independence. $\endgroup$ – Igor Pak Nov 4 '18 at 22:52

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.