Zeta Functions Expose the Hidden Architecture of Mathematical Singularities
30 May 2026, Yanjiang
The Archimedean zeta function’s poles expose the hidden architecture of singularities, each spectral line corresponding to a deep invariant of the minimal exponent.
Think of a mathematician staring at a surface that folds into a sharp point — a singularity. The surface is smooth everywhere else, but at that one point, calculus breaks down. Derivatives fail. Ordinary geometry retreats. For decades, algebraic geometers have asked the same practical question: how do you measure just how bad that singularity really is? Not with adjectives — “mild,” “severe,” “catastrophic” — but with numbers that carry analytic weight. A preprint (arXiv:2412.07849) from Dougal Davis, András C. Lőrincz, and Ruijie Yang now provides an unexpectedly crisp answer: you listen to it with a zeta function.
The Archimedean zeta function is not the zeta function most people know — not the one that encodes prime numbers and connects to the Riemann Hypothesis. This zeta function lives entirely in the domain of singularity theory. Feed it a polynomial with a singularity, and out comes a complex function whose poles — those values where the function blows up to infinity — turn out to be like spectral lines in a physical measurement. Each pole corresponds to some deep invariant of the singularity. The sharper the singularity, the more dramatic the poles. The spectrograph analogy is not perfect — a zeta function is not a physical instrument but a specific integral transform defined by $f$ — yet it captures the essential logic. Davis and colleagues have now shown that this analytic spectrograph reveals far more structure than anyone had previously proved.
The landscape of a silent war
Singularity theory has long been a battlefield between algebra and analysis. On the algebraic side, you have invariants like the log canonical threshold — a number that tells you, roughly, at what power the singularity becomes “tolerable” enough that certain integrals converge. On the analytic side, there is the Bernstein–Sato polynomial $bf(s)$, a polynomial that encodes how the function $f$ behaves when you raise it to a complex power and apply differential operators. The two sides speak to each other, but the conversation has been fragmentary. Results that connect them are hard-won and often restricted to isolated singularities — the simplest kind, where the bad point stands alone in a sea of smoothness.
The minimal exponent occupies a special place in this landscape. Introduced through the Bernstein–Sato polynomial, it measures the severity of the singularity in a way that unifies the algebraic and analytic perspectives. Kollár and Litchin connected it to the log canonical threshold. Loeser, in the 1980s, connected it to poles of the Archimedean zeta function. But gaps remained. Loeser proved that for isolated singularities, the negative of the minimal exponent is the largest pole — but he could not determine its order, the multiplicity that tells you how many independent ways the zeta function can blow up at that point. For non-isolated singularities — the messy, real-world kind where entire curves or surfaces go bad — the question was open.
Davis, Lőrincz, and Yang have now closed it. Their work proves that the largest nontrivial pole of the reduced zeta function is precisely the negative of the minimal exponent, and the order of that pole is exactly the multiplicity of the corresponding root of the Bernstein–Sato polynomial. The result holds in full generality — no isolation assumption, no special pleading. “This simultaneously generalizes a result of Loeser for isolated singularities and of Kollár–Litchin for the log canonical threshold,” the authors write, “and improves them by accounting for the multiplicity.”
The spectrograph yields a surprise
But the paper does more than complete a known program. It also delivers a surprise — the kind that makes the mathematical community sit up straighter. Loeser had asked in 1985 whether every root of the Bernstein–Sato polynomial must appear as a pole of the Archimedean zeta function. The intuition was reasonable: if the zeta function truly encodes all the analytic information about the singularity, then every algebraic fingerprint ought to leave an analytic mark. Davis and colleagues construct an explicit counterexample. They exhibit a polynomial $f$ where a root of $bf(s)$ is simply not a pole of the zeta function. The spectrograph, it turns out, is selective. It does not record every frequency — only those that survive a subtle filtration process.
This was not a loose end but a genuine refutation of a long-standing conjecture. It means the relationship between the algebraic and analytic sides is richer and more nuanced than the field had assumed. The zeta function is not a passive mirror; it makes choices. Figuring out why it makes those choices — what hidden structure governs the selection — becomes a new question for the next generation.
An important line of questioning arising from recent work on Hodge and V-filtrations concerns what happens when you push the analytic description further. The authors themselves acknowledge a limitation: their characterization of Hodge ideals via zeta functions requires a condition $\alpha + k < \widetilde{\alpha}_f + 1$, where $\widetilde{\alpha}_f$ is the minimal exponent. Beyond that threshold, the zeta function’s analytic machinery does not yet capture the full Hodge-theoretic picture. This is not a flaw in the proof — it is a frontier, a place where the current bridge between analysis and Hodge theory runs out of paving stones. The paper’s fourth cited reference, a recent work by Davis himself on Hodge and V-filtrations of mixed Hodge modules, raises the question of Hodge-V compatibility — a delicate condition linking the Hodge filtration to the Kashiwara–Malgrange V-filtration. The current paper does not address this condition, leaving it as an open challenge for future work.
Yet the achievements are substantial. The team proves that the Hodge filtration on vanishing cycles — a central construction in Saito’s theory of mixed Hodge modules — determines the poles of the zeta function. This sharpens an earlier result of Barlet. They also obtain analytic descriptions of the V-filtration itself, of Hodge ideals, and of higher multiplier ideals. The latter result refines the picture advanced by Schnell and collaborators in their 2023 work on higher multiplier ideals. Where earlier work had provided partial characterizations within restricted parameter ranges, Davis, Lőrincz, and Yang give a uniform analytic expression — provided the key condition on $\alpha + k$ is satisfied. The methodology leans heavily on a positivity property of the polarization on the lowest piece of the Hodge filtration of a complex Hodge module, a tool developed by Sabbah and Schnell.
The deeper meaning
At its core, this paper is about something more than technical classification. It is about the dream of a unified description. Singularity theory has accumulated a menagerie of invariants over the decades: log canonical thresholds, minimal exponents, Bernstein–Sato roots, Hodge ideals, multiplier ideals, higher multiplier ideals, the V-filtration. Each came with its own definition, its own computational challenges, its own partial overlap with the others. Davis, Lőrincz, and Yang are building a dictionary. They are showing that the Archimedean zeta function — a single analytic object — can serve as a Rosetta Stone that translates between these different languages.
Not everything translates yet. The condition $\alpha + k < \widetilde{\alpha}_f + 1$ is a reminder that the dictionary is still incomplete. The counterexample to Loeser’s question is a reminder that the translation is not always straightforward — sometimes the zeta function simply declines to render a term that algebra says should be there. But the trajectory is unmistakable. The field is moving toward a synthesis in which the zeta function is the central organizing principle.
There is a philosophical satisfaction in this. Mathematicians have long sensed that the most profound connections are the ones that link objects defined in completely different ways — algebra, analysis, geometry — and show them to be shadows of the same underlying reality. Davis, Lőrincz, and Yang have added a substantial piece to that vision for singularity theory. The zeta function, once a curious analytic accessory, now appears as a deep structural probe.
Perhaps one day, when mathematicians want to understand the geometry of a singular space, they will reach first not for algebraic invariants but for this analytic spectrograph. They will compute the poles, read off the minimal exponent and the multiplicity, and know immediately how severe the singularity is — in a language that analysis provides and geometry respects. What Davis and colleagues have done is not the final word but a decisive step toward making that vision precise. The bridge between Hodge theory and Archimedean zeta functions is now stronger than it has ever been. The remaining gaps are not discouragements but invitations. The spectrograph has more to reveal.
— Yanjiang
Yanjiang is an online editor of LoomSci.com.
References
- Dougal Davis et al., Archimedean zeta functions, singularities, and Hodge theory, arXiv:2412.07849
- M. Saito, Hodge ideals and microlocal V-filtration, arXiv:1612.08667
- C. Schnell et al., Higher multiplier ideals, arXiv:2309.16763
- D. Davis et al., On the Hodge and V-filtrations of mixed Hodge modules, arXiv:2503.16619