Let $\Delta(z)$ be the modular form associated with Ramanujan $\tau$-function.

For any $k=2,3,...$, $Sym^k\Delta$ is conjectured to be an automorphic form on $\mathrm{GL}(k+1)$ and $L(s, Sym^k\Delta)$ is conjectured to be holomorphic and satisfies a functional equation.

Question 1: For which $k$, has $Sym^k\Delta$ been proved to be automorphic form on $\mathrm{GL}(k+1)$?

Question 2: For which $k$, has $L(s,Sym^k\Delta)$ been proved to be homorphic and satisfies a functional equation?

I know certainly that both questions is YES for $k=2,3,4$ (Gelbart-Jacquet, Kim-Shahidi, Shahidi). But I heard there is progress beyond that because $\Delta$ is a modular form rather than arbitrary automorphic representation on $\mathrm{GL}(2)$.

  • $\begingroup$ Is my answer below sufficient? If so, could you accept it? $\endgroup$ – 2734364041 Nov 12 '20 at 23:00

It is indeed true that substantially more is known for holomorphic cusp forms than for general automorphic representations, as a consequence of modularity lifting theorems.

The strongest result so far seems to be due to Clozel and Thorne, who have proved automorphy of $Sym^k(f)$, for $f$ a holomorphic modular form of weight $\ge 2$, for all $k \le 8$. This answers your Questions 1 and 2 for these $k$.

I'm not aware of any cases where Question 2 is known without also answering Question 1 (assuming you want holomorphy on all of $\mathbf{C}$, rather than some smaller domain). But if you're willing to settle for $L(s, Sym^k \Delta)$ being meromorphic on $\mathbf{C}$ with the expected functional equation, then this is known for all $k$, as a consequence of the potential automorphy theorems of Barnet-Lamb, Geraghty, Harris and Taylor. This is discussed in Clozel and Thorne's article.

  • $\begingroup$ Perhaps it is useful to say briefly why 8 is the current limit. E.g., in the Shahidi method, they ran out of the exceptional Lie groups ($E_8$ provided the 4th symmetric power). Here, I am no expert, but I think the limitation is the first sentence of the last paragraph on page 2 of CT: namely that you have a decomposition into essentially orthogonal reps of degrees 4 and 5, which has been considered/classified by Arthur/Mok/(Moeglin-Waldspurger?), and the analogous does not exist for higher powers? $\endgroup$ – just-a-guest Feb 25 '18 at 23:04
  • $\begingroup$ (cont.) There's also the 4th paragraph of pg 3 CT, which seems to suggest that the level-raising argument only works for $l=5,7$, whereas they actually skirt the above Arthur issue (though similar others remain), by saying that "Schur reps" of unitary type can be handled by deformation theory in $GL_9(\bar Z_7)$. Alternatively, in Math Reviews, Wiese suggests that the failure to handle 10 is more "random", in the sense one needs $k-1$ to be prime, noting that the main result is: $C(l-1)$ implies $C(l+1)$ for $l$ prime, where $C$ is the principal Conjecture in the realm, provable for $l=5,7$. $\endgroup$ – just-a-guest Feb 26 '18 at 4:42

According to this exciting new preprint of Newton and Thorne (https://arxiv.org/abs/1912.11261), the answer to both Questions 1 and 2 is "all $k\geq 1$".


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.