Artin conductor¶

The Artin conductor assigns a ramification-sensitive number or ideal to a Galois-group character.

Version: v1 · 2026-09-28 · History
Domain-specific #: 7567
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomain: Algebraic Number Theory → Mathematics

Core Idea¶

The Artin conductor is the ramification invariant assigned to a character of the Galois group of a finite extension of local or global fields.^[1] Locally, for a character \(\chi\) and lower-numbered ramification groups \(G_i\), it is the integer

\[ f(\chi)=\sum_{i\geq 0}\frac{|G_i|}{|G_0|}\bigl(\chi(1)-\chi(G_i)\bigr), \]

where \(\chi(G_i)\) denotes the character's average on \(G_i\). Each summand records the failure of the corresponding ramification subgroup to act trivially.^[2] Consequently, an unramified character has conductor zero, while the terms with \(i>0\) isolate the wild contribution known as the Swan conductor.^[3]

For a global field extension, the local exponents are assembled into the ideal \(\mathfrak f(\chi)=\prod_{\mathfrak p}\mathfrak p^{f(\chi,\mathfrak p)}\).^[4] Unramified primes contribute exponent zero, so only ramified primes affect the product.^[5] This local-to-global construction is what allows the conductor to enter the functional equation of an Artin \(L\)-function and the conductor–discriminant formula.^[6]

The defining operation is therefore not merely attaching a number to a representation. It weights the character's nontrivial action across the ramification filtration and, globally, packages those local defects prime by prime. Removing that filtration-sensitive computation leaves a different representation invariant.

How would you explain it like I'm…

No faithful explanation at this level. All three generators agree any five-year-old version collapses the conductor into a crude count or messiness score, losing that it is a weighted sum over the ramification filtration of a character's failure to act trivially.

No faithful explanation at this level. All three generators agree a ten-year-old version must drop the character and the filtration weighting, teaching a count of bad primes, which is a different invariant.

Filtration-Weighted Ramification Measure

When you extend a number system, for example by adjoining a root to the rational numbers, the extension has a symmetry group called its Galois group, and a character is a function on that group built from a representation of it, so it records how the group acts. For each prime there is a nested chain of subgroups G_0, G_1, G_2, ... called the ramification filtration, which measures, in finer and finer degrees, how much the symmetries fail to leave the prime alone. The Artin conductor of a character is a single integer built from that chain: each subgroup in the chain contributes the difference between the character's value at the identity and its average over that subgroup, scaled by how big the subgroup is relative to G_0. So each layer contributes only when that layer of symmetry acts nontrivially, and a character whose extension is unramified at the prime gets conductor zero. The layers beyond the first isolate what is called the wild part, the Swan conductor. For a global field one computes such an exponent at each prime and multiplies the primes to those powers into a single ideal, where unramified primes contribute nothing, and this local-to-global packaging is what lets the conductor appear in the functional equation of an Artin L-function and in the conductor-discriminant formula.

The Artin conductor is the ramification invariant attached to a character of the Galois group of a finite extension of local or global fields. Locally, for a character chi and the lower-numbered ramification groups G_i, it is the integer f(chi) = sum over i >= 0 of (|G_i|/|G_0|) times (chi(1) - chi(G_i)), where chi(G_i) denotes the character's average value on G_i. Each summand records the failure of the corresponding ramification subgroup to act trivially, so an unramified character has conductor zero, and the terms with i > 0 isolate the wild contribution known as the Swan conductor. Globally, the local exponents are assembled into the ideal f(chi) = product over primes p of p raised to f(chi, p); unramified primes contribute exponent zero, so only ramified primes affect the product. That local-to-global construction is what allows the conductor to enter the functional equation of an Artin L-function and the conductor-discriminant formula. The defining operation is therefore not merely attaching a number to a representation: it weights the character's nontrivial action across the ramification filtration and then packages those local defects prime by prime. Remove the filtration-sensitive computation and you are left with a different representation invariant.

Structural Signature¶

Sig role-phrases:

Local Galois extension — a finite extension of local fields with Galois group G supplies the ramified action being measured.
Galois character — a character χ records the representation whose failure to be trivial along ramification levels is evaluated.
Ramification filtration — the lower-numbered subgroups G_i order inertia from the tame level into successively deeper wild levels.
Character-average defect — χ(1) − χ(G_i) measures how far the averaged action on each subgroup is from triviality.
Filtration weight — the ratio |G_i| / |G_0| scales each level's defect in the local conductor formula.
Local conductor exponent — summing the weighted defects gives the nonnegative integer f(χ) attached to the local character.^[7]
Tame–wild decomposition — the i = 0 term supplies the tame contribution, while the terms for i > 0 form the Swan conductor.
Unramified guarantee — trivial action on the ramification groups makes every defect vanish and forces local conductor zero.
Primewise globalization — restricting a global character at each prime supplies local exponents f(χ, 𝔭) for the global construction.
Conductor ideal — the product ∏𝔭 𝔭^f(χ, 𝔭) packages the finite ramified support and its exponents into an ideal.
Information-loss boundary — the local integer or global ideal records ramification severity and support but does not reconstruct the full filtration or character action.

What It Is Not¶

Not an arbitrary size assigned to a representation. The local exponent is the weighted sum of character-average defects across the ramification filtration; omitting that filtration-sensitive operation yields a different invariant.
Not the Galois character or ramification filtration itself. The conductor is a compressed number or ideal derived from those inputs, and its information-loss boundary prevents reconstruction of the full action from the output alone.
Not identical to the Swan conductor. The Swan conductor contains the terms from the higher groups G_i for i > 0 and records the wild contribution, whereas the full Artin conductor also includes the i = 0 tame term.^[8]
Not one object of the same type locally and globally. The local Artin conductor is a nonnegative integer attached to a local character; the global conductor is an ideal assembled prime by prime from those local exponents.
Not zero only for the trivial character. Conductor zero means that the relevant character is unramified—its ramification-group defects vanish—not that the character must be trivial on the entire Galois group.
Not a complete record of how ramification is distributed. Distinct filtered character actions can share the same exponent, so equal conductors do not imply equal representations or identical higher-group behavior.
Not an Artin L-function or a discriminant. The conductor enters functional equations and conductor–discriminant relations, but those downstream appearances do not turn the measurement into the analytic function or field invariant that consumes it.

Scope of Application¶

The Artin conductor applies within algebraic number theory wherever a Galois character is evaluated against the ramification filtration of a finite local or global field extension under a declared numbering and normalization. Its literal reach is bounded by those typed inputs: the resulting local integer or global ideal records ramified support and severity but does not reconstruct the character or full filtration.

Local ramification measurement — weighted character-average defects across lower-numbered ramification groups yield the local nonnegative conductor exponent.
Unramifiedness testing — vanishing defects force conductor zero, allowing an unramified local character or prime to be recognized within the stated representation setting.
Tame–wild decomposition — separating the i = 0 term from the higher terms distinguishes the tame contribution from the Swan conductor.
Global conductor ideals — local exponents at decomposition groups become prime powers in a finite ideal product, with unramified primes omitted.
Artin L-function analysis — the global conductor supplies the ramification datum that appears in the functional equation of the associated Artin L-function.
Conductor–discriminant relations — conductor exponents enter formulas connecting characters of a global extension with its discriminant.
Modularity and arithmetic-level questions — Artin conductors express ramification-sensitive levels in settings such as Serre-type modularity statements, provided the representation and local restrictions are explicit.^[9]
Related arithmetic conductors — Artin and Swan representations participate in defining conductors for elliptic curves and abelian varieties, while those derived invariants remain distinct from the original character conductor.

Clarity¶

Naming the Artin conductor makes several related invariants separable. The local conductor is an integer attached to a character and a local extension; the global conductor is an ideal assembled from those local exponents. Within the local integer, the \(i=0\) contribution records the tame part, whereas the terms for \(i>0\) form the Swan conductor and isolate wild ramification. Thus “conductor zero,” “Swan conductor zero,” and “unramified character” are not interchangeable assertions without the relevant hypotheses.

The concept licenses the number theorist to ask: At which levels of the ramification filtration does this character fail to act trivially, and what weighted exponent does that failure contribute at each prime? That question prevents the conductor from being treated as an arbitrary size measure of a representation and keeps a local numerical calculation distinct from its global ideal-valued packaging.

Manages Complexity¶

Ramification data comprise an entire descending filtration of subgroups together with a character's action at every level. The local Artin conductor compresses that profile to a weighted integer: at each ramification group G_i, it tracks the defect χ(1) − χ(G_i), weights it by the order ratio |G_i|/|G_0|, and sums. This turns many subgroup-by-subgroup questions into a few legible outcomes. Zero marks the unramified branch; the i=0 term records the tame contribution; the remaining i>0 terms collect into the Swan conductor and expose the wild branch. A comparison of characters can therefore begin with total conductor and tame/wild decomposition instead of their complete action tables.

Globally, the same compression operates prime by prime. Local exponents become the powers in the conductor ideal 𝔣(χ), and the zero contribution of unramified primes reduces an apparently infinite product to the finite ramified support. The resulting ideal is compact enough to enter functional equations and conductor–discriminant relations while retaining where ramification occurs and with what exponent. Its boundary is equally important: the conductor does not reconstruct the ramification filtration, the character, or how nontriviality is distributed among higher groups. Distinct local actions may share an exponent, so any problem needing that finer structure must return to the full filtered representation.

Abstract Reasoning¶

The local formula licenses a diagnostic move from the averaged character values on the ramification groups to the depth and type of nontrivial ramified action. If every defect χ(1) − χ(G_i) vanishes, the character is unramified and the conductor is zero. A nonzero i = 0 contribution with no higher contribution diagnoses tame ramification, whereas positive terms for i > 0 identify the wild part summarized by the Swan conductor.

An interventionist calculation runs from changing the character's action at one filtration level to the corresponding change in the weighted conductor sum, holding the extension and other character averages fixed. Making the action trivial on a higher ramification subgroup removes that subgroup's defect term; moving nontriviality deeper into the filtration changes which weights contribute. This is a formal sensitivity statement about the invariant, not a claim that arbitrary character averages can be varied independently.

A local-to-global move runs from the family of local exponents f(χ, 𝔭) to the support and powers of the global conductor ideal. Primes with exponent zero disappear from the product, so an unramified prime is predicted not to divide the conductor, while a ramified prime contributes its computed power. The reasoning boundary is sharp: the total local integer and global ideal do not reconstruct the complete filtration or character action, and the i > 0 Swan part must not be mistaken for the full conductor.

Knowledge Transfer¶

Within algebraic number theory, the Artin conductor transfers literally from local ramification calculations to global conductor ideals and across characters or representations for which the same filtered action is defined. The mathematical cargo remains fixed: lower-numbered ramification groups, averaged character values, weighted defects of trivial action, the tame i = 0 contribution, the higher Swan contribution, and prime-by-prime assembly. The local diagnostics carry as well—conductor zero for unramified action and positive higher terms for wild ramification—while the global construction transports those exponents into the support and powers of an ideal used in Artin L-function and conductor–discriminant settings.

Its honest wider reach is primarily (C) instrument or measure. The construction transfers literally wherever its algebraic preconditions and normalization are preserved; the boundary is not “mechanism versus metaphor” but what the invariant records and discards. It measures filtration-sensitive nontriviality and local ramified support, yet it does not reconstruct the character, the full filtration, or how action is distributed among groups. Other mathematical quantities called conductors, and nonmathematical scores that compress layered deviation, may share a (B) measurement pattern or an (A) analogy, but they are not Artin conductors without the Galois character and ramification formula. The name must therefore stop when those typed inputs disappear, even if another invariant also assigns a size or ideal to an object.

Examples¶

Canonical¶

Let L/K be a finite Galois extension of local fields and let a character χ be unramified. Then every ramification subgroup acts trivially in the represented action, so its averaged character value equals χ(1) at every level.^[10] Each defect χ(1) − χ(G_i) in the Artin-conductor sum is therefore zero, and the local exponent is f(χ) = 0.^[11] The higher i > 0 terms also vanish, so the Swan conductor is zero. This calculation does not say that χ is the trivial character on the whole Galois group; it says that the character detects no ramification. If nontriviality appeared only at G_0, the conductor could be positive while the higher Swan part remained zero, marking the tame branch.

Mapped back: L/K supplies the Local Galois extension, χ the Galois character, and the subgroups G_i the Ramification filtration. Trivial action makes every Character-average defect vanish despite its Filtration weight, yielding the zero Local conductor exponent and realizing the Unramified guarantee; separating i = 0 from i > 0 preserves the Tame–wild decomposition.

Applied / In Practice¶

In a global Artin L-function calculation, a number theorist restricts the global character at each prime and computes the corresponding local conductor exponent. Every unramified prime contributes exponent zero and therefore disappears from the ideal product; only the finitely many ramified primes remain, each raised to its local exponent.^[12] The resulting conductor ideal records both the ramified support and its compressed severity, and this ideal appears in the functional equation of the Artin L-function.^[13] Two characters can nevertheless have the same conductor ideal while differing in their full local actions, so the ideal is usable analytic input without being a reconstruction of the representation or filtration.

Mapped back: local restrictions reproduce the Galois character, Ramification filtration, and Local conductor exponent prime by prime. Primewise globalization removes zero exponents and assembles the Conductor ideal used in the functional equation, while the possibility of distinct actions sharing that ideal enforces the Information-loss boundary.

Structural Tensions¶

T1: Ramification compression versus action recovery (a useful invariant that forgets its inputs). The local Artin conductor condenses an entire ramification filtration and a character's averaged action into one integer. That compression makes comparison and downstream arithmetic possible, but distinct filtered actions can yield the same weighted sum. Treating the exponent as a complete fingerprint erases where nontriviality occurs; refusing the compression forfeits the compact invariant that functional equations and global assembly require. The conductor is therefore deliberately sufficient for some questions and insufficient for reconstructing the representation. Diagnostic: Does the problem ask only for aggregate ramification severity, or does it require the subgroup-by-subgroup action that equal conductor values can conceal?

T2: Tame total versus wild depth (one conductor and a consequential decomposition). The full local conductor includes the i = 0 contribution, while the Swan conductor isolates the higher i > 0 terms. Reporting only the total preserves overall size but hides whether ramification is tame or wild; focusing only on the Swan part can make a nonzero tame conductor appear unramified. The decomposition adds interpretive precision, yet it still does not recover the detailed higher filtration. The two quantities must remain coordinated rather than substituted for one another. Diagnostic: Is the conclusion justified by the full conductor, by the higher-group Swan contribution, or by their difference, and would the conclusion change if all nontriviality were confined to G_0?

T3: Local specificity versus global packaging (primewise information and ideal-level use). Globalization turns local exponents into an ideal whose support and powers retain where ramification occurs. This makes local calculations usable in global formulas, but the ideal can encourage analysts to treat all primes uniformly even though each exponent arose from a distinct decomposition group and filtration. Keeping every local profile visible impedes global comparison; retaining only the ideal suppresses the reasons behind its factors. The construction therefore joins rather than eliminates the two scales. Diagnostic: Is the global ideal enough for the intended argument, or must one reopen a particular prime's local character and filtration to justify the needed distinction?

T4: Vanishing as unramifiedness versus vanishing as triviality (a sharp guarantee with a narrow scope). A zero conductor certifies trivial action on the relevant ramification groups, not triviality of the character on the whole Galois group. The vanishing criterion is powerful precisely because it ignores unramified action, but that selectivity is easy to overread. Weakening the criterion makes zero less informative; strengthening it to global character triviality asserts information the measurement was designed to discard. Diagnostic: Does the argument need absence of ramified action, which conductor zero supplies, or absence of all nontrivial Galois action, which it does not?

T5: Artin-conductor autonomy versus reduction to Invariance. Every qualifying Artin conductor is a strict arithmetic specialization of the exact parent Prime Invariance (Invariance): under an isomorphic re-realization or basis change preserving the character and ramification action, the character averages, ramification-group orders, weighted local exponent, and global conductor ideal remain unchanged. Reduction preserves that feature–transformation–scope–license structure, but loses the ramification filtration, tame–wild decomposition, unramified-zero guarantee, and local-to-global ideal that make the conductor independently diagnostic.
Diagnostic: Is there merely a feature preserved under a scoped nontrivial transformation, or is that invariant specifically generated by the Artin conductor's Galois-character and ramification formula?

Structural–Framed Character¶

The Artin conductor is structural-leaning. Its evaluative_weight is absent because its exponent or ideal records ramification rather than endorsing an outcome. Its human_practice_bound character is weak: notation, numbering, and normalization are mathematical conventions, but the weighted invariant follows formally once the extension, character, and filtration are fixed. Its institutional_origin is absent because no authority constitutes the preserved value. Its vocab_travels result is restricted: feature, transformation, preservation, and scope generalize, while Galois character, ramification filtration, Swan conductor, and conductor ideal retain arithmetic meanings. Its import_vs_recognize result favors recognition, because equivalent realizations can be tested for exact preservation without importing an evaluative perspective.

The smallest reviewed portable support is Invariance: a named feature remains unchanged under a named nontrivial transformation within a bounded scope and thereby licenses representative-independent reasoning. The Artin conductor is a strict kind of that Prime, adding the character, lower ramification filtration, weighted local sum, tame–wild split, unramified-zero guarantee, and primewise global ideal. Portable and cross-domain reach belongs to that Prime, while the Galois and conductor machinery remains the domain accent.

Its character: a structural-leaning arithmetic invariant whose preservation architecture is portable but whose exact value and inferential role are fixed by ramification-sensitive character data.

Structural Core vs. Domain Accent¶

This decomposition shows why the Artin conductor is a domain-specific abstraction rather than a Prime.

What is skeletal (could lift toward a cross-domain prime). The abstract carrier is a mathematical object together with a named feature extracted from it. A nontrivial change of representation alters the realization while preserving that feature exactly within a declared scope; the invariant licenses computation from any equivalent realization, and recognition fails if the alleged preservation depends on a chosen basis or changes the underlying action. The Artin conductor strictly specializes Invariance: stripping the arithmetic inputs leaves the feature–transformation–preservation–scope structure, whereas removing exact preservation under equivalent realization destroys the proposed parent signature.

What is domain-bound. The carrier is a Galois character evaluated along the lower ramification filtration of a finite local or global field extension. Character-average defects weighted by ramification-group orders yield the local exponent; the tame term and higher Swan terms distinguish depths of ramification, and primewise local exponents assemble the global conductor ideal. Replace the Galois character or filtration with an arbitrary layered score, omit the weighting formula, or treat conductor zero as triviality of the entire character rather than unramified action, and the result is not an Artin conductor.

Why this does not clear the prime bar. The complete Galois-character, ramification-filtration, weighted-defect, and local-to-global ideal signature does not recur literally across three unrelated domains; portable reach belongs to Invariance. Stripping away the algebraic-number-theory accent leaves an invariant feature but not the named conductor. Conversely, retaining conductor or ramification vocabulary while removing scoped preservation and the filtration-sensitive construction leaves a label or basis-dependent quantity rather than the candidate-level structure.

This entry is a kind of Invariance.

Instantiates — Invariance (Invariance). The preserved feature is the local integer f(χ) or the global conductor ideal 𝔣(χ). A nontrivial equivalent re-realization of the Galois representation, including a basis change that preserves its character and ramification action, changes matrices or representatives while leaving the character averages, ramification-group orders, weighted local sum, and primewise global ideal exactly unchanged. The claim is bounded to a declared local or global extension, ramification numbering, normalization, and identified character data. Within that scope the conductor may be computed from any convenient equivalent realization and used unchanged in unramifiedness tests, tame–wild comparison, functional equations, and conductor–discriminant relations. Removing the algebraic vocabulary leaves a named feature preserved under nontrivial equivalence with an explicit scope and inferential license; removing that preservation leaves only a basis-dependent function, not the Artin conductor.

Decline — Measurement (Measurement). The Artin conductor superficially resembles measurement because a declared character and ramification filtration determine a local integer or global ideal. It does not realize Measurement's full signature, however: the filtration is mathematical input rather than an instrument coupled to a target, the formula is an exact construction rather than an observational procedure, and the output has no physical or conventional unit, calibration chain, observer-frame, or uncertainty envelope. Removing algebraic-number-theoretic ramification does not leave Measurement; it leaves no Artin conductor at all. The former frozen subsumption proposal therefore remains declined rather than being rescued by weakening the Prime.

Relationships to Other Abstractions¶

Current abstraction Artin conductor Domain-specific

Parents (1) — more general patterns this builds on

Artin conductor is a kind of Invariance Prime

The preserved feature is the local integer f(χ) or the global conductor ideal 𝔣(χ).

Hierarchy path (1) — routes to 1 parentless root

Artin conductor → Invariance

Neighborhood in Abstraction Space¶

Artin conductor sits in a sparse region of the domain-specific corpus (82^nd percentile for distinctiveness): few abstractions share its structure, so a faithful description tends to retrieve it precisely.

Family — Unclustered & Miscellaneous (2551 abstractions)

Nearest neighbors

Finite extensions of local fields — 0.84
Conductor (ring theory) — 0.83
Local Tate Duality — 0.82
Serre Group — 0.82
Brauer's Induction Theorem — 0.82

Computed from structural-signature embeddings · 2026-10-08

Not to Be Confused With¶

Swan conductor. The Swan conductor is the higher-group, wild contribution formed from the terms with i > 0; it is a component of the full Artin conductor, which also includes the i = 0 tame term. Tell: a quantity omitting the tame contribution is the Swan conductor rather than the complete Artin conductor.
Artin character. The Artin character is the group character assembled as a sum of irreducible characters weighted by their conductor exponents; it uses Artin conductors as coefficients rather than being the number or ideal attached to one character. Tell: if the output is itself a character of the Galois group, not a local exponent or global ideal, it is the Artin character.
Artin representation. The Artin representation is a representation whose character is the Artin character, whereas the Artin conductor is the ramification-sensitive invariant computed from a specified character. Tell: a vector-space action realizing the Artin character is a representation, not the conductor value derived from filtered action.
Field discriminant. A discriminant is an invariant of the field extension that is related to character conductors by the conductor–discriminant formula; it is not the weighted ramification-filtration sum attached to a particular character. Tell: if the object is assigned to the extension without first selecting the character whose subgroup defects are weighted, it is not that character's Artin conductor.
Artin L-function. An Artin L-function is the analytic object whose functional equation contains the global conductor; the conductor is an arithmetic ideal supplying ramification data to that equation. Tell: an Euler-product function with a functional equation consumes the conductor but is not the conductor.
A local Artin conductor. The local form is the nonnegative integer obtained from one local character and its lower ramification groups, while the global form is the ideal assembled from such exponents prime by prime. Tell: inspect whether the output is one filtration-weighted exponent or a prime-power ideal product before comparing conductor claims.
The conductor of an elliptic curve or abelian variety. Such an arithmetic conductor is defined using Artin and Swan representation data but belongs to a different carrier from the Galois-character conductor defined here. Tell: if the invariant is attached to an elliptic curve or abelian variety rather than directly to the specified Galois character, the carrier distinguishes it from the Artin conductor.

References¶

[1] James W. Cogdell, On Artin L-functions, Ohio State University (accessed 2026-09-13). registry ↩ Show verification details

[2] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[3] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[4] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[5] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[6] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[7] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[8] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[9] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[10] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[11] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[12] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩

[13] Unverified encyclopedia synthesis; no authoritative source located for the claim as written. ↩