Finite representation type #
A quiver has finite representation type when it has only finitely many isomorphism classes of
finite-dimensional indecomposable representations. This file defines TauCeti.IsFiniteRepType, the
finiteness of the skeleton of the full subcategory of finite-dimensional indecomposables, whose
objects are the representations that are pointwise finite-dimensional
(TauCeti.IsFinDim, from TauCeti.RepresentationTheory.Quiver.Representation.FiniteDimensional)
and indecomposable; and it proves the criterion by which the property is refuted: an infinite
family of pairwise non-isomorphic finite-dimensional indecomposables.
Both directions of that criterion are proved, because both are used. The refuting direction is what exhibits a quiver of infinite representation type; the affirming direction is what a quiver of finite representation type is for, namely that any family of pairwise non-isomorphic finite-dimensional indecomposables is finite, so that "the indecomposables" may be counted.
Main definitions #
TauCeti.IsFiniteRepType: the quiver has finitely many finite-dimensional indecomposables up to isomorphism.
Main results #
TauCeti.not_isFiniteRepType_of_infinite: an infinite family of pairwise non-isomorphic finite-dimensional indecomposables refutes finite representation type.TauCeti.IsFiniteRepType.finite_of_pairwise_nonisomorphic: conversely, under finite representation type every such family is indexed by a finite type.
Implementation notes #
The definition does not carry [Finite Q]. The roadmap pins IsFiniteRepType with that instance
binder, to record the intended setting, but nothing in the statement consumes it and an unused
instance argument is a linter error here; a consumer that needs a finite vertex set -- Gabriel's
dichotomy does -- states it where it is used. Dropping it costs nothing: the definition reads the
same, and it stays meaningful over an infinite quiver.
The skeleton is Mathlib's CategoryTheory.Skeleton, so IsFiniteRepType is a finiteness statement
about an honest type of isomorphism classes rather than about a hand-rolled quotient. The bridge in
both proofs below is CategoryTheory.ObjectProperty.toSkeleton_eq_toSkeleton_iff_nonempty_iso,
that two objects of a full subcategory have the same class in its skeleton exactly when they are
isomorphic in the ambient category.
The two results are stated for a family M : α → QuiverRep k Q rather than for a set of
representations: a family is what the worked examples produce -- the loop quiver's family is
indexed by the base field -- and injectivity up to isomorphism is expressed directly on the index
type.
References #
This implements the "finite representation type" item of Layer 5 of
TauCetiRoadmap/RepresentationTheory/QuiverRepresentations/README.md, on which its Gabriel
dichotomy and the loop-quiver and Kronecker worked examples are stated.
Finite representation type: the quiver Q has only finitely many isomorphism classes of
finite-dimensional indecomposable representations over k. It is stated as the finiteness of the
skeleton of the full subcategory they span, so that "isomorphism class" is Mathlib's.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The elimination and introduction rule for TauCeti.IsFiniteRepType: it is the finiteness
of the skeleton of the full subcategory of finite-dimensional indecomposable representations.
An infinite family of pairwise non-isomorphic finite-dimensional indecomposables refutes finite representation type. This is how a quiver is shown to have infinite representation type: exhibit such a family.
Under finite representation type a family of pairwise non-isomorphic finite-dimensional indecomposables is finite. This is the counting form of the property: only finitely many indecomposables are available to be listed.