7. Generalized Riemann Hypothesis Path (Dirichlet L-functions & Broader $L$-functions)

The Generalized Riemann Hypothesis (GRH) asserts that all L-functions of number fields (starting with Dirichlet $L$-series) have zeros only on the critical line. This path broadens RH’s scope. Studying GRH often sheds light on RH itself, since many techniques (and consequences) are analogous, and sometimes working in the generalized setting gives additional insights (e.g. comparing different $L$-functions).

Base Camp 1: Dirichlet Characters and $L$-functions – Start with Dirichlet $L$-series, the simplest extension of $\zeta(s)$. GRH for Dirichlet $L(s,\chi)$ (nontrivial zeros on $\Re(s)=1/2$ for every non-principal character $\chi$) is a natural next step after RH[71]. Key references:

“Introduction to Analytic Number Theory” by Tom Apostol – Chapters 7–9 introduce Dirichlet characters, $L$-functions, and prove Dirichlet’s theorem on primes in arithmetic progressions. Apostol’s development, while assuming only classical analysis, lays the groundwork for GRH: the $L(s,\chi)$ have Euler products and functional equations much like $\zeta(s)$. He discusses their zeros in passing, and one sees how GRH generalizes RH.

“Multiplicative Number Theory” by Davenport – Davenport’s treatment of Dirichlet $L$-functions is more advanced. He proves zero-free regions (and the Deuring–Heilbronn phenomenon about a possible real exceptional zero) for $L(s,\chi)$, paralleling the zeta case. Understanding these results gives insight into why GRH is hard: even for $L(s,\chi)$, we can’t eliminate the tiny possibility of a Siegel zero just above $\Re(s)=1$ without assuming GRH[72]. Davenport’s discussion of the Generalized RH and its implications (e.g. the least quadratic non-residue bound) is crisp and classic.

“A Course in Arithmetic” by J-P. Serre – Chapter 3 covers Dirichlet characters and $L$-functions in a very elegant way, connecting them to characters of Galois groups (using class field theory lightly) and to complex multiplication. Serre doesn’t prove new results on zeros, but his conceptual viewpoint helps: one sees Dirichlet $L$-functions as Artin L-functions for abelian extensions of $\mathbb{Q}$. This hints at the broader realm of GRH (for all Artin L-functions, etc.) and provides a segue into algebraic number theory perspectives.

Base Camp 2: Equivalences and Consequences of GRH – Learn what GRH would imply in number theory. There are many statements known to be equivalent or following from GRH (some even more general than RH’s consequences). For instance, GRH implies the classical conjecture about prime gaps in progressions, bounds on class numbers, etc. References:

“Problems in Analytic Number Theory” by M. Ram Murty – This problem-driven book has an entire section on Conditional Results assuming GRH. Through exercises, it leads the reader to prove that GRH implies tight bounds for the Chebotarev density theorem, class number estimates, and so on. For example, one exercise shows GRH for Dirichlet $L$ implies the Pólya–Vinogradov bound for character sums can be sharpened to $O(\sqrt{q}\log q)$. Another shows GRH implies an explicit bound in the prime number theorem for arithmetic progressions that is “best possible”[60]. Working through these solidifies why GRH is so powerful and desirable.

“The Grand Riemann Hypothesis” by G. Cornell (1997) – A survey in AMS Contemporary Math 216 that describes GRH for various L-functions and collects consequences. It’s helpful for a broad view: it states the Generalized RH (for Dirichlet L and beyond) and mentions things like: “GRH for Dirichlet L implies no exception to quadratic reciprocity beyond known bounds,” or “GRH for Dedekind zeta of number fields implies effective Chebotarev”. Having these in one place, one appreciates that GRH is a linchpin for many theorems.

“Equivalents of the Riemann Hypothesis, Vol. 2” by Broughan (again) – The second volume includes analytic equivalents not just of RH but also discusses analogues for Dirichlet L-functions. For instance, Broughan includes the Extended Riemann Hypothesis (ERH) for Dedekind zeta of number fields and notes statements like: “GRH is true iff for every primitive character $\chi$, the formula $\sum_{n\le x}\chi(n) = O(x^{1/2}\log^2 x)$ holds”. These help one see patterns – often replacing $\mu(n)$ or primes with characters in known RH criteria yields an equivalent GRH criterion. Mastering these analogues can guide one to attempt proofs that might unify all cases (via, say, an approach in the Langlands generality).

Base Camp 3: Advanced $L$-functions and Deep Analogies – Dive into the general theory of $L$-functions (GL(2) and beyond) and analogies from other realms. Since GRH extends to all nice L-functions, one eventually confronts the general conjectures (which link to Langlands program). This is highly advanced, but key sources:

“Automorphic Forms and $L$-functions for the General Linear Group” by D. Bump – Introduces $L$-functions of higher degree (e.g. $L$-functions of modular forms, which are GL(2) L-functions). It explains the analytic properties (functional equation, Euler product, etc.) and states the generalized RH for them. Bump’s text shows the breadth of GRH: it’s not just Dirichlet or Dedekind zeta, but all automorphic L-functions. It also covers tools like Rankin–Selberg integrals which are used to prove partial results towards GRH (like subconvexity bounds).

“L-functions and Random Matrix Theory” by J. Keating & N. Snaith – This is a more specialized monograph linking GRH (and beyond, like Montgomery–Odlyzko laws) to random matrix theory for general L-functions. It postulates that all high-degree L-functions have their zeros statistics modeled by random matrices (of appropriate symmetry types). While this is speculative, it’s widely believed and provides heuristic support for GRH in general. It also unifies the approach: rather than tackling RH one L-function at a time, the random matrix philosophy suggests a universality underlying all GRH cases. This is an active research area blending with quantum chaos path.

“Zeta Functions of Picard Modular Surfaces” by Charles D. Frohman – For a very algebraic-geometric angle, this book (and similar ones on Shimura varieties) study zeta or L-functions of higher-dimensional varieties where analogues of RH can sometimes be proved (via generalizations of Weil’s work). While not directly about GRH for number fields, it reminds us that in the realm of function fields or shimura varieties many analogues of GRH are true by Deligne’s proof[46]. Drawing parallels from these successes is part of the strategy: understand how RH is solved in these cases to inform the number field case. For example, one learns about Hasse–Weil L-functions of elliptic curves, which satisfy an analogue of GRH over function fields (proven by Weil), but whose number field analog (the L-function of an elliptic curve over $\mathbb{Q}$) satisfies GRH only conjecturally.

Base Camp 4: Cutting-Edge Partial Results towards GRH – Finally, see what modern number theory has achieved toward GRH. This includes subconvexity bounds (which can be viewed as an approach to bounding zeros’ real parts) and averaged results:

“Subconvexity for L-functions” (Lectures by Iwaniec and Sarnak) – Subconvexity results give bounds on $L(1/2+it)$ that are stronger than the convexity bound. They indirectly tell us about zero distributions (e.g. no too-large gaps or some control on spacing). Michel and Venkatesh’s “Heights in Families of L-functions” (2006) and Iwaniec–Kowalski’s text[51] both describe known subconvexity results and their relation to GRH. These show we can often get “GRH-like results on average”. For example, the Bombieri–Vinogradov theorem (unconditional PNT in AP on average) is often called a “weak form of GRH” since GRH would imply the individual PNT in AP up to $x^{1/2}$ log factors[60]. Modern improvements like the Polymath project on primes in AP push these boundaries further.

Recent Breakthroughs (e.g. Zhang, Maynard, Polymath) – The 2013 breakthrough on bounded prime gaps (Zhang) and subsequent improvements can be framed in terms of GRH for various related L-functions. For instance, proofs of bounded gaps did not assume GRH, but if GRH were true they’d immediately imply infinitely many prime pairs with gap 2 (the twin prime conjecture). Studying these developments via a GRH lens is instructive: it highlights exactly where not having GRH forces compromises (e.g. working with “Level of Distribution $<1$” vs. “=1” which GRH would allow). A great account is “Gaps between Primes: The Impacts of GRH” by Andrew Granville (2015), which isn’t a book but a set of notes explaining how much easier the prime gaps problem would be if GRH held, and conversely what new ideas were needed since GRH isn’t known. This indirectly measures progress toward GRH – it tells us what unconditional tools are approaching the power of GRH-based tools.

In following the Generalized RH path, one essentially generalizes one’s knowledge at each step: from $\zeta(s)$ to $L(s,\chi)$ to global $L$-functions of higher degree. The recommended books are commonly used in graduate courses and seminars on analytic number theory and automorphic forms, ensuring they are pedagogically sound. By the end, one would see RH as one piece of a much larger puzzle – the Grand Riemann Hypothesis – and be equipped with the broad perspective needed to contribute to these general conjectures.