Dévissage¶

Reduce a statement about algebraic objects to finite filtrations of simpler subquotients and a justified transfer rule.

Version: v1 · 2026-10-04 · History
Domain-specific #: 13723
Domain group: Formal Sciences
Origin domain: Mathematics
Subdomains: Algebraic Geometry, Algebraic K Theory → Mathematics
Aliases: Devissage, Dévissage argument

Core Idea¶

Dévissage is a method for proving statements about algebraic objects or categories by filtering a complicated object into finitely many simpler subquotients and then using a valid transfer rule to recover the intended conclusion. The name covers related arguments, not one theorem with one universal hypothesis list. In coherent-sheaf theory, a property of the simple pieces can extend through short exact sequences to all coherent sheaves under specified conditions. In algebraic K-theory, a separate theorem can identify the K-theory of a suitable exact subcategory with that of a larger abelian category whose objects admit finite filtrations by subcategory pieces.^[1]^[2]

The filtration is only half the method. Knowing each quotient has a property does not make the whole object have it unless that property is preserved under the relevant extensions; nor does every filtered subcategory yield a K-theory equivalence without Quillen's categorical hypotheses. Dévissage names the reduction plus justified reconstruction, not a license to treat any list of pieces as a proof.^[1]^[2]

Structural Signature¶

Sig role-phrases:

Exact ambient setting — Coherent sheaves, modules, or an abelian category provide subobjects, quotients, and short exact sequences with typed meanings.
Finite filtration — A chain 0 = F₀ ⊂ F₁ ⊂ ··· ⊂ Fₙ = F divides a target into successive quotients Fᵢ/Fᵢ₋₁.^[1]
Admissible pieces — Each quotient belongs to a class on which the required property or invariant is already understood; membership must be checked, not presumed.
Transfer rule — Extension closure, two-out-of-three, or an exact-subcategory K-theory theorem moves information from pieces to the target in the appropriate version.^[1]^[2]
Specified conclusion — The proof must say whether it establishes a property of objects, an equivalence of categorical invariants, or something else; those conclusions are not interchangeable.

A finite filtration is often built by reducing support or complexity in a Noetherian setting. Its length and piece type depend on the actual theorem, not on the word dévissage.

What It Is Not¶

Not the filtration alone. A nested chain states an organization of an object; dévissage also needs a reason that information on quotients determines the desired statement.
Not arbitrary induction. The induction must be tied to subquotients, exact structure, and an explicit transfer rule.
Not one universal coherent-sheaf lemma. Stacks gives different versions: an extension-stable property on all specified integral-support pieces suffices in one lemma, while a rank-one generic witness version requires stronger two-out-of-three and support hypotheses.^[1]
Not the claim that every invariant is extension-stable. An invariant may behave through an exact sequence in a more intricate way or fail to be determined by the quotients alone.
Not automatic equality of ring K-theory. Quillen's theorem concerns the K-theory of stated abelian categories; it must not be confused with projective-module K-theory of a ring without further comparison.^[2]

Scope of Application¶

The Stacks Project proves that a coherent sheaf on a Noetherian scheme has a finite filtration whose successive quotients are pushforwards of certain coherent ideal sheaves on integral closed subschemes. If a property holds for every piece of that specified form and is closed under extensions, it holds for every coherent sheaf. A more selective one-witness-per-integral-subschema argument is also available, but its two-out-of-three, support, and generic-stalk conditions must be stated separately.^[1]

In algebraic K-theory, Quillen-style dévissage instead uses a full exact abelian subcategory with the required closure under subobjects and quotients, plus finite filtrations of objects of the larger category by pieces in the subcategory. The conclusion is an equivalence of categorical K-theory spaces and corresponding groups. Weibel's K-book gives finite modules over ℤ/pʳ filtered by p-power layers with 𝔽ₚ-module quotients as an example. This illustrates the categorical theorem; it is not a proof that every filtered property of those modules transfers.^[2]

Clarity¶

Dévissage separates which pieces suffice from why they suffice. A proof that shows only that an object has a composition-like series has not yet demonstrated the target statement. The next question is whether the statement is stable across every short exact sequence needed to rebuild the object, or whether a named theorem supplies a different categorical transfer.

It also prevents a subtle hypothesis swap. Stacks Lemma 30.12.4 assumes the property for all listed integral-support ideal-sheaf pieces and only extension closure. Lemma 30.12.6 starts with a specially supported rank-one-generic witness for each integral closed subscheme but demands two-out-of-three and extra stalk conditions. Treating those as the same proposition can make a proof invalid.^[1]

Manages Complexity¶

The method replaces an arbitrary coherent sheaf or category object with a finite sequence of more controlled quotients. Support and exactness are the organizing variables; one need not analyze the entire object at once. The Stacks filtration theorem makes that reduction systematic in a Noetherian scheme.^[1]

The simplification is constrained by what is lost in passing to quotients. The extension data can matter. A valid proof either shows the desired property survives extension or invokes an invariant theorem proved under hypotheses sufficient to account for that loss.

Abstract Reasoning¶

Start with the intended conclusion and ask what transfer rule it requires. Then prove that each target admits a finite filtration by the permitted simpler class. For an extension-stable property P, if P(Fᵢ₋₁) and P(Fᵢ/Fᵢ₋₁) imply P(Fᵢ), finite induction yields P(F) once P(0) and every quotient case are established. This is the logic of the Stacks property-transfer lemma, with its exact piece class and Noetherian premise.^[1]

For K-theory, the reasoning is not merely that each object has a filtration. One must verify the inclusion of categories is exact, that the subcategory has the specified closure, and that finite filtrations by its objects exist. Only then may the dévissage theorem yield the K-theory equivalence. The theorem supplies a conclusion not obtainable from the elementary object-by-object induction alone.^[2]

Knowledge Transfer¶

The filtration-plus-transfer test transfers literally among coherent sheaves, finite modules, and suitable abelian-category settings. The permitted pieces and exact theorem change. A general idea of “solve a complex case by simpler cases” is much broader, but absent algebraic subquotients and a justified exact transfer it is only an analogy to dévissage. Its broader whole-to-parts skeleton resembles Decomposition; the named mathematical method is more restrictive.

Examples¶

Coherent-sheaf property on a Noetherian scheme¶

Take a property P of coherent sheaves on a Noetherian scheme that holds for the integral-support ideal-sheaf pieces specified in Stacks Lemma 30.12.3 and is stable under extensions. Every coherent sheaf admits the finite filtration into those pieces; repeated short exact sequences then carry P to the whole sheaf. The example is schematic: no arbitrary property is presumed to satisfy these conditions.^[1]

Mapped back: ambient = coherent sheaves on a Noetherian scheme; filtration = finite coherent-subsheaf chain; pieces = specified pushforward ideal sheaves on integral closed subschemes; transfer = extension closure; conclusion = P holds for every coherent sheaf.

Finite `p`-power modules in categorical K-theory¶

Finite modules in the relevant ℤ/pʳ setting admit a filtration by powers of p; successive quotients are 𝔽ₚ-vector spaces. Under Quillen's exact-subcategory hypotheses, the smaller category's K-theory determines that of the larger finite-module category. This is a categorical/G-theoretic statement; it is not an unqualified equality of every ring-theoretic K-invariant.^[2]

Mapped back: ambient = the stated finite abelian module category; filtration = p-power submodule chain; pieces = 𝔽ₚ-modules; transfer = Quillen/Weibel dévissage theorem with category hypotheses; conclusion = equivalence of the relevant categorical K-theory.

Structural Tensions¶

Simple pieces versus reconstruction conditions. Simplifying to quotients makes the target tractable but can discard extension information on which the statement depends. Using too weak a transfer rule overclaims; demanding an unnecessarily strong one can obscure an available proof. Diagnostic: which precise exact-sequence or categorical theorem moves the conclusion from these quotients to the original object?

General method versus specialized theorem. The two examples share finite reduction yet have different hypotheses and conclusions. Treating one formal theorem as the definition of every dévissage argument shrinks the method; treating all arguments as interchangeable erases essential conditions. Diagnostic: is the target an object-level property, or an equivalence of invariants attached to categories?^[1]^[2]

Structural–Framed Character¶

Dévissage is predominantly structural within algebraic mathematics. Its evaluative weight is low: a proof either meets its hypotheses or not, independent of its aesthetic appeal. Human practice chooses a useful filtration and desired invariant, but exactness and finite subquotient conditions constrain the argument formally. The term comes from mathematical proof practice. “Peeling layers” vocabulary can travel widely; outside categories with subobjects, quotients, and a valid transfer principle it is an imported metaphor rather than literal algebraic dévissage. The portable skeleton is reduction plus reconstruction from verified pieces, related to Decomposition, but that skeleton alone does not validate a prime classification for this named method. Its character: a domain-specific proof strategy governed by exact filtration and theorem-specific transfer.

Structural Core vs. Domain Accent¶

The general skeleton is finite decomposition into pieces followed by a warranted inference from parts to whole. In algebraic dévissage the pieces are subquotients in an exact setting, and the warrant can be extension closure or a categorical K-theory theorem. Those conditions are constitutive, not ornamental. Remove them and one has ordinary decomposition or induction, not this method. The portable reach belongs to broader prime structures such as Decomposition; the named entry remains domain-specific because its proof obligations depend on algebraic exactness.

This entry presupposes Filtration (algebra).

The staged composition/presupposes parent is Filtration (algebra): dévissage requires finite nested algebraic subobjects and successive quotients, but is not itself a filtration because its residual is the justified transfer argument. Exact sequence names short exact relations that support some transfer steps. Decomposition is a broader pattern, not a substitute for checking theorem hypotheses. Specialist theorem-hypothesis review remains a separate admission gate.

Relationships to Other Abstractions¶

Current abstraction Dévissage Domain-specific

Parents (1) — more general patterns this builds on

Dévissage presupposes Filtration (algebra) Domain-specific

Dévissage requires finite subobject filtrations before justified reconstruction.

Hierarchy paths (3) — routes to 3 parentless roots

Dévissage → Filtration (algebra) → Order → Comparison → Self Checking

Show alternative paths (2)

Neighborhood in Abstraction Space¶

Dévissage sits in a moderately populated region (56^th percentile for distinctiveness): it has near-neighbors but no dense thicket of look-alikes.

Family — Algebraic Substructures & Closures (12 abstractions)

Nearest neighbors

Resolution Theorem (Algebraic K-Theory) — 0.90
Auslander–Reiten theory — 0.86
Closed Immersion — 0.85
Initial and terminal objects — 0.85
Filtration (algebra) — 0.84

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

Not to Be Confused With¶

Filtration (algebra): a nested family of subobjects; dévissage uses a suitable finite filtration in a proof.
Composition series: a particular filtration with simple factors; dévissage may use different classes of pieces.
Noetherian induction alone: an induction principle that can help construct filtrations or prove cases, but does not itself supply the quotient-to-whole transfer.
Quillen's dévissage theorem alone: one powerful categorical result within a broader family of dévissage arguments, with its own exact-subcategory conditions.

References¶

[1] The Stacks Project, §30.12 “Devissage of coherent sheaves”, especially Lemmas 30.12.3–30.12.6. Note the distinct premises of Lemmas 30.12.4 and 30.12.6. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h ↩i ↩j ↩k

[2] Charles A. Weibel, The K-Book: An Introduction to Algebraic K-Theory, AMS, 2013, ISBN 978-0-8218-9132-2. Author-hosted manuscript: https://sites.math.rutgers.edu/~weibel/Kbook/Kbook.pdf, Chapter V §4, Dévissage Theorem 4.1 and its finite p-power module example; manuscript pagination may differ from the published book. The original Quillen proof remains a separate independent-check item. registry ↩a ↩b ↩c ↩d ↩e ↩f ↩g ↩h