Decomposing a module along a complete orthogonal system of idempotents #
A complete orthogonal system of idempotents (eᵢ) in a commutative ring R decomposes R as a
finite product of rings R ⧸ (1 - eᵢ) — that is mathlib's
CompleteOrthogonalIdempotents.bijective_pi. This file supplies the module analogue: the same
system decomposes every R-module M as ∏ᵢ M ⧸ (1 - eᵢ)M, compatibly with the ring
decomposition.
Applying this with mathlib's QuotSMulTop.equivQuotTensor rewrites the factors as
(R ⧸ (1 - eᵢ)) ⊗[R] M, which is the form in which the pieces are modules over the factor rings.
Main results #
CompleteOrthogonalIdempotents.bijective_piQuotSMulTop: the comparison mapM → ∏ᵢ M ⧸ (1 - eᵢ)Mis bijective.CompleteOrthogonalIdempotents.linearEquivPiQuotSMulTop: that map packaged as a linear equivalence.CompleteOrthogonalIdempotents.linearEquivPiQuotientTensor: the same, with the factors written as(R ⧸ (1 - eᵢ)) ⊗[R] M.
Implementation notes #
This is a statement mathlib is missing, not a competing formulation of one it has: mathlib's
CompleteOrthogonalIdempotents.bijective_pi is exactly the case M = R, and the proof here is
the same two-line computation carried out with scalars acting on M. Both directions are
elementary — injectivity because x = (∑ᵢ eᵢ) • x and each eᵢ • x vanishes, surjectivity by
assembling ∑ᵢ eᵢ • xᵢ from lifts of the components.
The factors are QuotSMulTop (1 - eᵢ) M, mathlib's abbreviation for M ⧸ (1 - eᵢ) • ⊤, rather
than the tensor products (R ⧸ (1 - eᵢ)) ⊗[R] M, because the quotient form is what the proof
manipulates. linearEquivPiQuotientTensor converts.
The comparison map M → ∏ᵢ M ⧸ (1 - eᵢ)M attached to a family of elements of R.
Equations
- piQuotSMulTop M e = LinearMap.pi fun (i : I) => ((1 - e i) • ⊤).mkQ
Instances For
A complete orthogonal system of idempotents decomposes every module. The module analogue
of CompleteOrthogonalIdempotents.bijective_pi, which is the case M = R.
The decomposition of M along a complete orthogonal system of idempotents,
M ≃ ∏ᵢ M ⧸ (1 - eᵢ)M.
Equations
Instances For
The decomposition of M with the factors written as base changes,
M ≃ ∏ᵢ (R ⧸ (1 - eᵢ)) ⊗[R] M. This is the form in which each factor is a module over the
corresponding factor ring of R.
Equations
- One or more equations did not get rendered due to their size.
Instances For
The tensor-product module decomposition sends an element to its pure-tensor image in each idempotent factor.