How statement and proof provenance work
The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.
- Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
- AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
- AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.
These labels describe origin, not correctness: citations and verification chips remain separate evidence.
Tensor Products of Modules
1 · Prerequisites
- Binary Operations, Monoids, Groups and Subgroups
- Construction of the Natural Numbers
- Construction of the Real Numbers via Cauchy Sequences
- Construction of the Real Numbers via Dedekind Cuts
- Cosets, Index and Lagrange's Theorem
- Countability and Uncountability
- Dual Spaces, Bilinear and Quadratic Forms, and Sylvester's Law of Inertia
- Finite Counting, Factorials and Binomial Coefficients
- Foundations of the Real Numbers for Analysis
- Free Modules, Exact Sequences, Projective and Injective Modules
- Group Homomorphisms and the Isomorphism Theorems
- Ideals, Quotient Rings and the Isomorphism Theorems for Rings
- Linear Independence, Bases and Dimension
- Linear Transformations, Rank-Nullity and Quotient Spaces
- Matrices, the Matrix of a Linear Map, and Change of Basis
- Modules, Submodules, Quotient Modules and the Isomorphism Theorems
- Normal Subgroups and Quotient Groups
- Order, Zorn's Lemma, and the Axiom of Choice
- Relations, Functions, and Quotients
- Rings, Subrings, Integral Domains and Fields
- Roots, Rational Powers, and Classical Inequalities
- The ZFC Axioms and the Basic Set Constructions
- Vector Spaces, Linear Subspaces, Span and Direct Sums
2 · Summary
Left and right modules over a ring, module homomorphisms with their kernels, images and cokernels, quotient modules and the universal property of a quotient are published, as are the free module on a set with its universal property, direct sums of families with theirs, Hom groups and the maps induced on them, exact and short exact sequences, the characterisations of a projective module, the fact that free modules are projective together with its choice boundary, and injective modules. Linear algebra supplies the standard basis of , dimension, the algebraic dual with its dual basis in finite dimension, and the trace of an endomorphism; ideals, ring homomorphisms and quotient groups supply the rest.
For a right module and a left module over an arbitrary ring , the page constructs by generators and relations, proves it represents the balanced maps out of , and derives uniqueness up to unique isomorphism. The universal property then yields functoriality, the criterion for a formula on elementary tensors to descend, bimodule actions, associativity, unit isomorphisms, compatibility with arbitrary direct sums, bases and dimension, internal Hom, the trace as a contraction, algebras and their tensor products, and the Hom–tensor adjunction; the module structure and symmetry over a commutative ring are given separately. Right exactness, the identification of with , flat and faithfully flat modules with the injection and ideal tests, flatness of projectives, restriction and extension of scalars with their adjunction and change-of-rings results, and the injectivity of character duals of flat modules close the page.
3 · Logical flowchart
4 · Definitions, theorems and proofs
Balanced maps from a right module and a left module, and bilinear maps over a commutative ring
Definition
Let be a unital ring, let be a right -module, let be a left -module, and let be an abelian group, written additively (Unital left and right modules over a ring; unqualified module means left module, Group and abelian group). A map is -balanced if, for all , , and ,
and
Thus a balanced map is additive in each variable and identifies the two ways in which a scalar may cross the pair. Additivity implies .
If is commutative (Commutative ring) and are -modules, a map is -bilinear if it is -linear in each variable. Equivalently, it is additive in each variable and satisfies
for all . After a commutative-ring module is regarded as a right module by , every bilinear map is balanced.
The tensor product from the additive group underlying the free -module on , elementary tensors, and finite tensor sums
Definition
Let be a unital ring, a right -module, and a left -module. Let
be the free -module on the set (The free module on a set and its standard basis, Universal property of the free module on a set), and write for its standard basis elements. The additive group of is abelian. Let be the subgroup generated (The subgroup generated by a subset, the cyclic subgroup , and cyclic groups) by all elements
and
as range over , over , and over . Since every subgroup of an abelian group is normal, the quotient group is defined (The quotient group and coset product ). The tensor product of and over is
The coset of is the elementary tensor . Every tensor is a finite sum of elementary tensors, because every element of is a finite -linear combination of basis elements and integer coefficients may be absorbed into either additive variable. The defining relations give
and
In particular . No -module structure on is part of this arbitrary-ring definition; at this stage it is an abelian group.
Universal property of the tensor product for balanced maps into abelian groups
Statement
Let be a unital ring, a right -module, a left -module, and
The map is balanced (Balanced maps from a right module and a left module, and bilinear maps over a commutative ring). For every abelian group and every balanced map , there is a unique group homomorphism
such that for all . Consequently composition with is a bijection
Facts & Assumptions
Given: A unital ring , a right -module , a left -module , an abelian group , and a balanced map .
The tensor product is , where , is generated by the additivity and balance relations, and (The tensor product from the additive group underlying the free -module on , elementary tensors, and finite tensor sums).
A balanced map is additive in each variable and satisfies (Balanced maps from a right module and a left module, and bilinear maps over a commutative ring).
Every element of has a unique finite expression with (The free module on a set and its standard basis).
Every set map into a left -module extends uniquely to an -module homomorphism taking to (Universal property of the free module on a set).
If a group homomorphism kills a normal subgroup , then it factors uniquely through a group homomorphism (A homomorphism that kills a normal subgroup factors uniquely through the quotient group).
Proof
Regard as a -module by integer multiplication in its additive group. A group homomorphism between abelian groups is automatically -linear: additivity gives for , and gives the formula for negative integers. Thus -module homomorphisms and group homomorphisms between these underlying additive groups are the same maps.
If is a group homomorphism, then is balanced because the elementary tensors satisfy all three relations in [L1].
Apply [L4] at and to extend the set map uniquely to a -linear map satisfying .
By [L2], sends each generator of to zero: the two additivity generators map respectively to and , while the balance generator maps to . Hence .
By [L5], factors uniquely through a group homomorphism , and .
The operations in steps 4.1 and 1.2 are inverse: starting from recovers its values on every pair, while starting from produces a homomorphism agreeing with on every elementary tensor, and those tensors generate . This proves both the asserted uniqueness and the displayed bijection.
No selection is made in the construction. If or , every balanced map out of is zero and the relations make every elementary tensor zero, so the same proof gives ; the zero ring is covered as well. Since a module contains its zero element, is never empty, so there is no separate empty-domain case.
Tensor products are unique up to a unique isomorphism carrying elementary tensors to elementary tensors
Statement
Let be abelian groups equipped with balanced maps and , and suppose that each pair has the universal property of Universal property of the tensor product for balanced maps into abelian groups. Then there is a unique group isomorphism such that . Its inverse is the unique map with .
Facts & Assumptions
Given: Two representing pairs and for balanced maps out of .
For any balanced map from into an abelian group, a representing pair supplies a unique group homomorphism through which that map factors (Universal property of the tensor product for balanced maps into abelian groups).
Proof
Apply [L1] for to the balanced map and for to ; this gives unique homomorphisms and with and .
Both and compose with to give , so uniqueness in [L1] gives ; similarly .
Thus is an isomorphism with inverse , and the same uniqueness clause shows that no other isomorphism carrying to exists.
Module homomorphisms induce tensor-product homomorphisms functorially
Statement
Let be a homomorphism of right -modules and a homomorphism of left -modules. There is a unique group homomorphism
such that . These maps satisfy
and
Facts & Assumptions
Given: Homomorphisms of right -modules and of left -modules.
A balanced map into an abelian group extends uniquely to a group homomorphism (Universal property of the tensor product for balanced maps into abelian groups).
A module homomorphism preserves addition and the relevant scalar action (Module homomorphism and isomorphism, kernel, image and cokernel).
Proof
The pairing is additive in both variables by [L2], and , so it is balanced.
By [L1] the pairing of step 1.1 induces a unique homomorphism with the stated formula.
The maps and agree on every elementary tensor, so uniqueness in [L1] makes them equal.
The two sides of the composition formula both send to , so uniqueness in [L1] makes them equal.
Steps 2.1, 3.1 and 3.2 prove existence, uniqueness, identity preservation, and composition preservation.
A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced
Statement
Let be an abelian group and let be a function. A prescription
extends to a group homomorphism if and only if is balanced. When it exists, the extension is unique.
The balance condition cannot be replaced by a check on tensor symbols alone. In , the prescription is not balanced and does not descend: the relation would force its value to be both and .
Facts & Assumptions
Given: A function into an abelian group.
Composition with the elementary-tensor map is a bijection from group homomorphisms to balanced maps (Universal property of the tensor product for balanced maps into abelian groups).
In the tensor product, and the elementary-tensor map is additive in each variable (The tensor product from the additive group underlying the free -module on , elementary tensors, and finite tensor sums).
The integers form a commutative unital ring (The integers form a commutative ring).
Proof
If is balanced, [L1] supplies a unique group homomorphism satisfying .
Conversely, if such a homomorphism exists, composing it with the elementary-tensor map gives ; [L2] and additivity of show that is additive in each variable and satisfies , so is balanced.
For , which is permitted by [L3], balance gives by [L2]. The function assigns to and to , so it is not balanced and no homomorphism can have the proposed elementary-tensor values.
Steps 1.1 and 1.2 prove the equivalence and uniqueness, while step 1.3 verifies the asserted failure.
-bimodules and commuting left and right scalar actions
Definition
Let and be unital rings. An -bimodule is an abelian group that is a left -module and a right -module (Unital left and right modules over a ring; unqualified module means left module) such that the two actions commute:
for every , , and . It is denoted when the rings need to be displayed.
Every ring is an -bimodule by left and right multiplication. If is commutative, every left -module becomes an -bimodule by defining .
A commuting outer scalar action descends to a tensor product
Statement
Let be a right -module and let be an -bimodule. There is a unique right -module structure on satisfying
Dually, if is an -bimodule and is a left -module, there is a unique left -module structure satisfying
If both outer actions are present, they commute, so the tensor product is an -bimodule in the evident handed situation.
Facts & Assumptions
Given: A right -module , an -bimodule , and, for the dual assertion, an -bimodule and a left -module .
In a bimodule the left and right scalar actions commute: (-bimodules and commuting left and right scalar actions).
Every balanced pairing into an abelian group induces a unique homomorphism from the tensor product (Universal property of the tensor product for balanced maps into abelian groups).
An elementary-tensor prescription descends exactly when the corresponding pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
Proof
For fixed , the pairing is additive in both variables and is balanced because by [L1].
The left-handed construction is identical: for fixed , the pairing is balanced by the commuting actions, and the induced maps satisfy the left module laws on elementary tensors.
By [L2] and [L3], step 1.1 induces an additive endomorphism of with .
On every elementary tensor one has , , and , using the right -module laws of . In each identity the two sides are additive maps that induce the same balanced pairing, so uniqueness in [L2] makes them equal on all . Thus step 2.1 defines a right -module structure.
When a left -action and a right -action are both present, on elementary tensors, so the actions commute.
The formulas determine every action map on generators, so uniqueness follows from [L2].
Associativity of tensor products for compatible bimodules
Statement
Let be a right -module, let be an -bimodule, and let be a left -module. There is a canonical group isomorphism
determined by
It respects any compatible outer module actions and is natural in .
Facts & Assumptions
Given: A right -module , an -bimodule , and a left -module .
The outer actions make a right -module and a left -module, with the stated elementary-tensor formulas (A commuting outer scalar action descends to a tensor product).
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
A formula on elementary tensors descends exactly when its underlying pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
Proof
For fixed , the pairing is -balanced: and have the same image by the outer left -action in [L1]. Thus [L2] and [L3] give a homomorphism .
For fixed , the pairing is -balanced, so [L2] and [L3] give .
The pairing from is -balanced. On generators, by [L1], and additivity extends the equality to every . Hence [L2] gives with the displayed formula.
The pairing is -balanced. It is enough to check generators , where by [L1]. Therefore [L2] gives .
The composite fixes every tensor , and fixes every tensor ; successive applications of uniqueness in [L2] show that these composites are the identity maps.
The same elementary-tensor calculation shows compatibility with any outer actions, and replacing by their images under compatible homomorphisms proves naturality because both candidate composites agree on all elementary tensors.
Thus is the asserted canonical natural isomorphism with inverse .
Over a commutative ring, is an -module with
Statement
Let be a commutative ring and let be -modules. The abelian group has a unique -module structure for which
for every , , and .
Facts & Assumptions
Given: A commutative ring and -modules , each regarded on either side by the common scalar action.
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
In a commutative ring, for all (Commutative ring).
An elementary-tensor formula descends exactly when its underlying pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
Proof
For fixed , the pairing is additive in both variables and balanced: by [L2].
By [L1] and [L3], step 1.1 induces an additive endomorphism of satisfying . The tensor balance relation also gives .
On elementary tensors, , , , and follow from the module axioms of . In each identity the two additive maps induce the same balanced pairing, so uniqueness in [L1] makes the identity hold for every tensor.
Any -module structure with the displayed formula has, for each , a scalar-multiplication endomorphism inducing the same balanced pairing as step 2.1. Uniqueness in [L1] therefore makes the structure unique.
Steps 2.1, 3.1 and 3.2 give the asserted unique -module structure and both elementary-tensor formulas.
Symmetry and associativity isomorphisms for tensor products over a commutative ring
Statement
Let be a commutative ring and let be -modules. There are natural -module isomorphisms
and
Moreover is the identity.
Facts & Assumptions
Given: A commutative ring and -modules .
The tensor product has the -module structure (Over a commutative ring, is an -module with ).
Compatible bimodules have the canonical associativity isomorphism carrying to (Associativity of tensor products for compatible bimodules).
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
Tensor maps induced by module homomorphisms preserve identities and compositions (Module homomorphisms induce tensor-product homomorphisms functorially).
Proof
The pairing is balanced because maps to , which is also the image of ; hence [L3] induces .
Regard all three modules as -bimodules. Then [L2] supplies the displayed associativity isomorphism, and [L1] shows it is -linear on elementary tensors.
Applying the same construction in the opposite order gives , and their composite fixes every ; uniqueness in [L3] makes the composite the identity, so is an isomorphism.
The map is -linear because .
Naturality of both maps follows by applying [L4]: after replacing the variables by their images under module homomorphisms, the two candidate composites agree on every elementary tensor, hence agree by [L3].
Steps 1.1 through 2.3 prove the two natural -linear isomorphisms and the involutivity of symmetry.
The regular module is a tensor unit: and
Statement
Let be a unital ring. For every left -module and every right -module , the maps
and
are group isomorphisms. Their inverses are and , respectively. The maps respect every displayed outer module structure.
Facts & Assumptions
Given: A unital ring , a left -module , and a right -module .
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
An elementary-tensor prescription descends exactly when the corresponding pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
Proof
The pairing is additive in each variable and satisfies , so it is balanced; [L1] and [L2] give .
The pairing is balanced, so [L1] and [L2] give . Define the additive map . Then , while by balance; uniqueness in [L1] makes the second composite the identity. Thus and are inverse.
The additive map , , satisfies , while by balance. Uniqueness in [L1] makes the second composite the identity, so and are inverse.
Each map commutes with any outer scalar action by associativity of that action, checked on elementary tensors. The calculations remain valid for the zero ring and for zero modules, where all displayed maps are the unique maps between zero groups.
Therefore the regular module is a left and right tensor unit with the stated natural formulas.
Finite iterated tensor products represent multilinear maps independently of parenthesization
Statement
Let be a commutative ring and let be a finite list of -modules. Any parenthesized tensor product
represents -multilinear maps from : for every -module , composition with is a bijection from to the set of multilinear maps into . Different parenthesizations are connected by the unique isomorphism preserving pure tensors.
For , take and identify zero-variable multilinear maps with chosen elements of . For , take .
Facts & Assumptions
Given: A commutative ring , a finite list of -modules, and an -module .
The binary tensor product represents balanced, hence over a commutative ring bilinear, maps (Universal property of the tensor product for balanced maps into abelian groups).
Tensor products over a commutative ring have canonical symmetry and associativity isomorphisms preserving elementary tensors (Symmetry and associativity isomorphisms for tensor products over a commutative ring).
The regular module is a tensor unit (The regular module is a tensor unit: and ).
Proof
For , an -linear map is uniquely determined by the image of , and every defines such a map by ; this is the required representation of maps from the one-point empty product.
For , the identity represents linear maps from by composition.
Assume a parenthesized product represents -linear maps. A -linear map is equivalently a bilinear map : first use the induction bijection with the last variable fixed, and then use multilinearity to see that the resulting dependence on the last variable is linear.
By [L1], the bilinear maps in step 1.3 correspond uniquely to linear maps , proving the representing property for .
By [L2], any two parenthesizations are joined by composites of elementary associativity isomorphisms preserving pure tensors. Any two such comparison maps agree on every pure tensor, so the representing uniqueness proved in step 2.1 makes them equal.
The base cases and induction step establish the representation for every finite , including the empty and singleton cases, and step 3.1 proves independence of parenthesization.
Tensor products commute with arbitrary direct sums
Statement
Let be a commutative ring, let be any family of -modules, and let be an -module. The homomorphism induced by the coordinate inclusions is a natural -module isomorphism
The same holds with the direct sum in the second variable: there is a natural -module isomorphism
The assertion includes , when both sides are zero.
Facts & Assumptions
Given: A commutative ring , a family of -modules, and an -module .
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
Tensor products over a commutative ring carry their canonical -module structure (Over a commutative ring, is an -module with ).
The direct sum consists of finite-support tuples, with coordinate inclusions , and the empty direct sum is zero (The direct sum of an indexed family of modules).
Every family of homomorphisms extends uniquely to a homomorphism whose value is the finite sum of the coordinate values (Universal property of a direct sum of modules).
Over a commutative ring there is a natural isomorphism with , and is the identity (Symmetry and associativity isomorphisms for tensor products over a commutative ring).
Proof
For each , the pairing is bilinear, so [L1] gives . By [L4], the family induces the displayed map .
Define by . This tuple has finite support by [L3], and coordinatewise calculation shows that is bilinear.
By [L1], the pairing in step 1.2 induces with .
The composite fixes every coordinate generator , so it is the identity by [L4]; the composite fixes every elementary tensor , so it is the identity by [L1].
Thus and are inverse isomorphisms. They are natural because maps and make the two composites send each coordinate generator to . When , [L3] makes both sides zero and the same construction yields the unique isomorphism.
For the second variable, put , where the middle direct sum of the symmetries is the map induced by [L4] from the family . Each is an isomorphism by [L5] and is one by step 4.1, so is an isomorphism; it is natural as a composite of natural isomorphisms. Tracing a coordinate generator gives , which is the displayed formula. At both sides are again zero.
The elementary tensors of two bases form the product basis of the tensor product
Statement
Let be a commutative ring. If is free with basis and is free with basis , then is free with basis
Equivalently, the canonical map sending the standard basis vector at to is an isomorphism. This includes an empty basis in either factor.
Facts & Assumptions
Given: A commutative ring and free modules with bases indexed by .
Tensor products commute with arbitrary direct sums in either variable: and (Tensor products commute with arbitrary direct sums).
The regular module is a tensor unit: via (The regular module is a tensor unit: and ).
A free module on is , with standard basis and unique finite coordinate expressions, including (The free module on a set and its standard basis).
A map from the basis set of a free module extends uniquely to a module homomorphism (Universal property of the free module on a set).
Proof
The chosen bases identify with and with , carrying to the corresponding standard basis vectors.
Apply [L1] in each variable and then [L2] to obtain canonical isomorphisms .
Tracing a coordinate generator through step 2.1 sends it to ; by [L3], those images therefore form a basis and every tensor has a unique finite expansion in them.
If or , then , the corresponding factor and the target free module are zero by [L3], and step 2.1 is the unique isomorphism between zero modules.
This proves the product-basis assertion in all cases.
with the product basis, and
Statement
Let be a commutative ring and . The standard finite free modules satisfy
with the tensor products of standard basis vectors corresponding to the standard basis indexed by .
If is a field and are finite-dimensional -vector spaces, then
Both assertions include a zero rank or zero-dimensional factor.
Facts & Assumptions
Given: Natural numbers , a commutative ring , and finite-dimensional vector spaces over a field .
Tensor products of free modules with bases indexed by have basis indexed by (The elementary tensors of two bases form the product basis of the tensor product).
For any unital ring , the free module has its standard basis indexed by , including the empty basis at (The free module on a set and its standard basis). For a field , this is the usual basis of (The standard list with and for is an ordered basis of ; hence , and is the zero space with basis and dimension ).
The dimension of a finite-dimensional vector space is the unique natural number equinumerous with a basis; the zero space has dimension zero (Finite-dimensional vector space, and its dimension ; infinite-dimensional means having no finite basis).
Proof
Apply [L1] to the standard bases indexed by and . Their product is indexed by , which has elements, giving the first isomorphism and its product basis.
Choose bases of and with respectively and elements. By [L1] their elementary tensors form a basis of indexed by , hence with elements.
If or , or if or , [L2] makes one basis empty and [L1] makes the product basis empty, so both sides are the zero module or have dimension zero as asserted.
By [L3], step 1.2 gives .
Steps 1.1 through 2.1 prove both formulas, including their zero boundaries.
The -module over a commutative ring
Definition
Let be a commutative ring and let be -modules. The abelian group of module homomorphisms (The abelian group and maps induced by pre- and postcomposition) becomes an -module under pointwise scalar multiplication
The function is -linear because, for ,
where the middle equality uses commutativity (Commutative ring). The remaining module laws hold pointwise in . This module is the internal Hom module and is denoted .
For finite-dimensional , the canonical map is an isomorphism
Statement
Let be a finite-dimensional vector space over a field , and let be any -vector space. The bilinear map
induces a natural isomorphism
For a basis of with dual basis , its inverse is
The empty sum gives the assertion when .
Facts & Assumptions
Given: A finite-dimensional -vector space , an -vector space , and a basis of with dual family .
Bilinear maps from induce unique homomorphisms from (Universal property of the tensor product for balanced maps into abelian groups).
Tensor products of bases have product basis (The elementary tensors of two bases form the product basis of the tensor product).
Over the commutative field , is an -module under pointwise scalar multiplication (The -module over a commutative ring).
The algebraic dual is (Linear functionals and the algebraic dual ).
The dual family of a finite basis is a basis of (The dual family of a finite basis is a basis of the dual space, with the same dimension).
Proof
The map is bilinear, so [L1] gives an -linear map .
Define . This is an -linear map because evaluation and the finite sum are linear in .
For and , one has , so is the identity.
For an elementary tensor , one has , because is the dual basis; elementary tensors generate, so is the identity.
The map is natural in : for , both routes send to the map . It is contravariantly natural in : for , both routes send to . Equality on elementary tensors gives both naturality squares.
Thus is a natural isomorphism with inverse . If , its basis and dual basis are empty, both and are zero, and the same formulas are the unique inverse maps.
Under , tensor contraction is the trace
Statement
Let be a finite-dimensional vector space over . Under the isomorphism
the contraction map , defined by , corresponds to the trace .
Facts & Assumptions
Given: A finite-dimensional -vector space and the canonical Hom-tensor isomorphism with .
The canonical isomorphism sends to the rank-one endomorphism (For finite-dimensional , the canonical map is an isomorphism).
The trace of an endomorphism is the sum of the diagonal entries of its matrix in any basis, and is zero in dimension zero (The basis-independent trace of an endomorphism of a finite-dimensional vector space).
A bilinear pairing induces a unique linear map from a tensor product (Universal property of the tensor product for balanced maps into abelian groups).
Proof
Evaluation is bilinear, so [L3] induces the linear contraction map with .
Choose a basis of and write . For , the coefficient of in is .
If , the tensor product and endomorphism space are zero and both maps are the zero map by [L2].
By [L2], .
Matrix diagonal sums are linear, so trace is linear by [L2]. Therefore trace after [L1] and contraction are linear maps inducing the same bilinear pairing by step 2.1; uniqueness in [L3] proves that they agree everywhere.
Thus tensor contraction is precisely trace under the canonical isomorphism, including the zero-dimensional case.
Algebras over a commutative ring, central structure maps, and algebra homomorphisms
Definition
Let be a commutative ring (Commutative ring). An -algebra is a unital ring together with a unital ring homomorphism
(Ring homomorphism: additive, multiplicative, and required to send to ) whose image is central: for every and . The induced scalar action is , making an -module and multiplication bilinear.
An -algebra homomorphism is a unital ring homomorphism satisfying . Such a map is automatically -linear. An -algebra is commutative when its underlying ring is commutative.
The tensor product of -algebras has multiplication
Statement
Let be a commutative ring and let be -algebras. The -module has a unique -algebra structure satisfying
and
If and are commutative, then is commutative.
Facts & Assumptions
Given: A commutative ring and central unital -algebras .
In an -algebra, the structure map is central and multiplication is -bilinear (Algebras over a commutative ring, central structure maps, and algebra homomorphisms).
Finite tensor products represent multilinear maps (Finite iterated tensor products represent multilinear maps independently of parenthesization).
The tensor product is an -module with (Over a commutative ring, is an -module with ).
An elementary-tensor formula descends exactly when its underlying pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
Proof
The map is -multilinear: centrality in [L1] lets a scalar move among the four variables without changing the value.
By [L2], step 1.1 induces an -bilinear multiplication with the displayed pure-tensor formula; [L4] ensures that the formula has descended before any ring laws are used.
Both sides of associativity are trilinear in the three tensor arguments and agree on pure tensors by associativity in and ; uniqueness in [L2] makes them equal everywhere. Both distributive laws hold because the multiplication from step 2.1 is bilinear.
Left and right multiplication by are linear maps that agree with the identity on every pure tensor, so uniqueness in the two-factor case of [L2] makes them the identity maps.
If and are commutative, then . The two bilinear multiplication maps therefore induce the same four-variable multilinear map, so [L2] gives commutativity for arbitrary tensors.
The map , , is a unital ring homomorphism and its image is central, checked on pure tensors using [L1] and [L3]. Thus the resulting ring is an -algebra.
Uniqueness in [L2] forces the multiplication from its displayed pure-tensor formula, while the identity and structure map are then forced by the displayed elements. Steps 2.1 through 4.1 prove existence and all asserted properties.
Universal mapping property of the tensor product of commutative algebras
Statement
Let be commutative -algebras. For every pair of -algebra homomorphisms and , there is a unique -algebra homomorphism
such that and . It is given by
Thus , with its two canonical maps, is the coproduct of and among commutative -algebras.
Facts & Assumptions
Given: Commutative -algebras and -algebra maps , .
The tensor product algebra has multiplication and identity (The tensor product of -algebras has multiplication ).
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
An elementary-tensor formula descends exactly when the corresponding pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
Proof
The canonical maps and are -algebra homomorphisms: [L1] gives their multiplication and identity laws, and and show compatibility with the structure maps.
The pairing is -bilinear: additivity is distributivity in , and because is commutative and both maps respect .
By [L2] and [L3], step 1.2 induces a unique -linear map satisfying .
On pure tensors, [L1] gives , where commutativity of permits the middle factors to switch; hence is multiplicative.
One has , , and , so is an -algebra homomorphism with the required restrictions.
If has the same restrictions, then by [L1], so . The underlying group homomorphisms consequently induce the same balanced pairing, and uniqueness in [L2] gives .
Step 1.1 supplies the two coproduct maps, and steps 2.1 through 4.1 prove the asserted universal mapping property.
Hom-tensor adjunction:
Statement
Let be a commutative ring and let be -modules. There is a natural -module isomorphism
It sends to the map and sends to the homomorphism determined by .
Facts & Assumptions
Given: A commutative ring and -modules .
The internal Hom is an -module under (The -module over a commutative ring).
Bilinear maps from into correspond uniquely to homomorphisms (Universal property of the tensor product for balanced maps into abelian groups).
Scalar multiplication on the tensor product satisfies (Over a commutative ring, is an -module with ).
Proof
Given , define . For fixed this is -linear in by [L3], and the dependence on is -linear by the same formula, so is an -module homomorphism.
Given , the pairing is bilinear by -linearity of and each ; [L2] therefore induces a unique .
For every , , so uniqueness in [L2] makes the identity.
For every , , so the two Hom-valued maps are equal and is the identity.
Both assignments are -linear pointwise, and precomposition or postcomposition with homomorphisms commutes with evaluation; hence the bijection is an -module isomorphism natural in all three variables, contravariantly in the Hom source variables and covariantly in .
Steps 2.1 through 2.3 prove the natural Hom-tensor adjunction.
Tensoring is right exact
Statement
Let be a commutative ring, let
be an exact sequence of -modules, and let be an -module. Then
is exact. Thus tensoring preserves cokernels and surjections, but no injectivity at the left is asserted.
Facts & Assumptions
Given: An exact sequence and an -module over a commutative ring .
Exactness means that is surjective and (Exact sequences and short exact sequences of modules).
Module homomorphisms induce tensor homomorphisms with and functorial composition (Module homomorphisms induce tensor-product homomorphisms functorially).
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
An elementary-tensor formula descends exactly when the corresponding pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
For a submodule , the quotient module has cosets and its induced scalar action (Quotient module with scalar multiplication on additive cosets).
A module homomorphism that kills factors uniquely through (A module homomorphism vanishing on factors uniquely through ).
Tensor products over carry the scalar action (Over a commutative ring, is an -module with ).
Every tensor is a finite sum of elementary tensors (The tensor product from the additive group underlying the free -module on , elementary tensors, and finite tensor sums).
Proof
The map is surjective: by [L8], every tensor is a finite sum of elementary tensors , and [L1] supplies with , so .
Functoriality gives , so .
Let . The image is a submodule because [L2] and [L7] make -linear. Since kills it, [L6] gives .
For and , choose any with and set . If is another lift, then by [L1], so ; hence is well defined.
The function is bilinear: lifts of sums may be taken as sums of lifts, scalar multiples as scalar multiples, and the tensor relations give the required equalities. By [L3] and [L4] it induces .
For , one has , while for and a lift one has ; generators and uniqueness make and inverse.
Since is an isomorphism, the kernel of is exactly the submodule quotiented out in step 2.1, namely . Together with step 1.1, this is right exactness.
The submodule generated by products of elements of an ideal with elements of a module
Definition
Let be a commutative ring, let be an ideal (Left, right and two-sided ideals), and let be an -module (Unital left and right modules over a ring; unqualified module means left module). The product of and is
where the case is the empty sum . It is the submodule of generated by the products : sums and negatives remain of the displayed form, and with .
Thus and ; the latter follows from .
naturally
Statement
Let be a commutative ring, an ideal, and an -module. There is a natural -module isomorphism
Both sides also carry the induced -module structure, and the isomorphism is -linear. For it is the tensor-unit isomorphism, while for both sides are zero.
Facts & Assumptions
Given: A commutative ring , an ideal , and an -module .
Tensoring an exact sequence ending in zero preserves exactness at the two rightmost terms (Tensoring is right exact).
The tensor-unit isomorphism sends to (The regular module is a tensor unit: and ).
The submodule consists of finite sums of products (The submodule generated by products of elements of an ideal with elements of a module ).
The quotient module consists of cosets with the induced scalar action (Quotient module with scalar multiplication on additive cosets).
A homomorphism that kills a submodule factors uniquely through the quotient module (A module homomorphism vanishing on factors uniquely through ).
Over a commutative ring the natural symmetry , , is an isomorphism (Symmetry and associativity isomorphisms for tensor products over a commutative ring).
Proof
The sequence is exact. Tensoring on the right by and applying [L1] gives the exact sequence . The symmetry isomorphisms of [L6] carry it termwise to , and since commutes with the induced maps on elementary tensors, that sequence is exact too.
Under [L2], the image of consists exactly of finite sums , hence is by [L3].
Exactness in step 1.1 identifies with the cokernel of the first map, which by step 2.1 is ; [L5] gives the resulting isomorphism.
Tracing through the quotient gives . Multiplication by an element of acts as zero on both sides, so the map and its inverse are -linear.
If , step 4.1 is [L2]. If , then [L3] gives and , so both sides are zero.
This proves the natural -linear and -linear isomorphism in every boundary case.
Flat and faithfully flat modules and ring homomorphisms
Definition
Let be a commutative ring and let be an -module. The module is flat if the functor preserves exact sequences: whenever is exact, so is
Since tensoring is always right exact (Tensoring is right exact, Exact sequences and short exact sequences of modules), the definition asks for the remaining left-hand exactness. Its equivalent formulation as preservation of injections is proved separately rather than built into the definition.
The module is faithfully flat if a sequence of -modules is exact exactly when its tensor with is exact.
For a unital ring homomorphism (Ring homomorphism: additive, multiplicative, and required to send to ) between commutative rings, is an -module by . The map is flat, respectively faithfully flat, when this -module is flat, respectively faithfully flat.
Flatness is equivalent to preserving injections and to the ideal and finitely generated ideal tests
Statement
Let be a commutative ring and let be an -module. The following are equivalent:
- is flat.
- For every injection , the induced map is injective.
- For every ideal , the multiplication map , , is injective.
- The map in claim 3 is injective for every finitely generated ideal .
Facts & Assumptions
Given: A commutative ring and an -module .
Flatness means that preserves exact sequences (Flat and faithfully flat modules and ring homomorphisms).
Tensoring is right exact (Tensoring is right exact).
Tensor products commute with direct sums and ; consequently (Tensor products commute with arbitrary direct sums, The regular module is a tensor unit: and ).
A finite list in a module determines a homomorphism taking to ; if the list generates , this map is surjective. Also (The free module on a set and its standard basis, Universal property of the free module on a set).
In a commutative ring, every submodule of the regular module is an ideal (Left, right and two-sided ideals).
A tensor product is the quotient of the free -module on pairs by the subgroup generated by the additive and balance relations (The tensor product from the additive group underlying the free -module on , elementary tensors, and finite tensor sums). Consequently, a tensor expression and a derivation that it is zero involve only finitely many generators and defining relations.
Proof
Claim 1 implies claim 2 by applying [L1] to ; claim 2 implies claim 3 by taking the inclusion and using ; claim 3 implies claim 4 by restriction to finitely generated ideals.
Assume claim 4. Claim 3 follows: an element is represented by finitely many coefficients from , hence comes from for the finitely generated ideal they generate; if its image in is zero, injectivity for makes its representative zero, so .
Let be injective and let map to zero. By [L6], the tensor uses finitely many elements of , and a derivation of its zero image uses only finitely many generators and defining relations in . Hence there are finitely generated submodules and with for which comes from an element killed in .
Under claim 3, for every submodule , the map is injective, by induction on . For it is the unique map ; for it is claim 3 by [L5].
Present using [L4], and let be the inverse image of . Right exactness identifies with and with the quotient of by .
For the induction step, let , put , and let be the image of in . Tensoring the exact rows and gives right-exact rows by [L2]; the left and right vertical maps are injective by the case and the induction hypothesis from step 2.1.
If maps to zero in , its image in maps to zero in and hence is zero by the right vertical injection in step 3.1. Right exactness of the top row lifts from some . The image of in maps to the zero image of in ; the left vertical injection in step 3.1 makes , and therefore . This completes the induction of step 2.1.
By step 4.1, both and inject into . Therefore the induced map of the quotients in step 2.2 is injective, so . This proves claim 2 from claim 4.
Finally claim 2 and right exactness [L2] imply claim 1: for an exact sequence, replace the left map by the injection of its image into the middle term; tensoring preserves that injection by claim 2 and preserves the remaining image and cokernel statements by right exactness.
Steps 1.1 through 6.1 prove the cycle of equivalences. The zero ideal, , and zero module cases occur explicitly in steps 1.2 and 2.1; no choice is made, and both directions of every equivalence have been supplied.
Every projective module over a commutative ring is flat
Statement
Every projective module over a commutative ring is flat. This implication requires no form of the Axiom of Choice.
Facts & Assumptions
Given: A commutative ring and a projective -module .
A module is flat exactly when tensoring with it preserves injections (Flatness is equivalent to preserving injections and to the ideal and finitely generated ideal tests).
Tensor products commute with arbitrary direct sums in either variable; in particular (Tensor products commute with arbitrary direct sums).
The regular module is a tensor unit (The regular module is a tensor unit: and ).
Every projective module is, without choice, a direct summand of its canonical free cover (Equivalent characterizations of projective modules).
Tensor maps preserve identities and compositions (Module homomorphisms induce tensor-product homomorphisms functorially).
Proof
Every free module is flat: for an injection , [L2] and [L3] identify with the direct sum of copies of , which is injective coordinatewise; now apply [L1].
By [L4], there are a free module and homomorphisms , with .
Let be injective and suppose satisfies . Functoriality [L5] gives .
The map is injective by step 1.1, so ; applying and using gives .
Thus preserves every injection, and [L1] makes flat. The proof used the canonical splitting supplied by projectivity and made no family of choices.
Under the stated choice boundary, free modules are projective and hence flat
Statement
Let be a commutative ring and let be a free -module with basis indexed by .
- Assuming the Axiom of Choice, is projective and therefore flat.
- If is finite, only finite choice is needed for projectivity; if , no choice is needed.
- Regardless of choice, is flat, because tensoring with is a direct sum of copies of the identity tensor functor.
Facts & Assumptions
Given: A commutative ring and a free -module with basis indexed by .
Under AC every free module is projective; a finite basis requires only finite choice, and an empty basis requires none (Free modules are projective, with the exact choice boundary).
Every projective module over a commutative ring is flat without choice (Every projective module over a commutative ring is flat).
Tensor products commute with arbitrary direct sums in either variable; in particular (Tensor products commute with arbitrary direct sums).
The regular module is a tensor unit (The regular module is a tensor unit: and ).
Proof
Under AC, [L1] makes projective and [L2] then makes it flat. The refined finite and empty-basis choice bounds are exactly those stated in [L1].
Independently of AC, write . By [L3] and [L4], tensoring an exact sequence with gives the direct sum, over , of the original exact sequence; kernels and images are computed coordinatewise, so the result remains exact. Thus is flat without any choice principle.
Step 1.1 establishes the projective route with its precise choice boundary, while step 1.2 establishes flatness unconditionally; the two routes are logically distinct.
For flat , one has
Statement
Let be a commutative ring, let be ideals, and let be a flat -module. Then
Facts & Assumptions
Given: Ideals of a commutative ring and a flat -module .
Tensoring an exact sequence with a flat module preserves exactness (Flat and faithfully flat modules and ring homomorphisms).
There is a natural isomorphism , and similarly for ( naturally).
The submodule consists of finite sums of products (The submodule generated by products of elements of an ideal with elements of a module ).
Exactness at a module is equality of the incoming image and outgoing kernel (Exact sequences and short exact sequences of modules).
Tensor products commute with direct sums (Tensor products commute with arbitrary direct sums).
Over a commutative ring the natural symmetry , , is an isomorphism (Symmetry and associativity isomorphisms for tensor products over a commutative ring).
Proof
The sequence , whose last displayed map sends to , is exact because its kernel is exactly .
Tensor step 1.1 with the flat module . By [L1], the resulting sequence is exact and begins , using [L5].
The symmetry of [L6] identifies with and likewise for , so [L2] applies and the last map in step 2.1 is , whose kernel is .
The image of is the set of finite sums of products with , namely by [L3].
Exactness in step 2.1 identifies the image in step 3.2 with the kernel in step 3.1, proving . The calculation also covers , , , or .
A short exact sequence with flat quotient remains short exact after tensoring
Statement
Let
be a short exact sequence of modules over a commutative ring . If is flat, then for every -module the sequence
is short exact.
Facts & Assumptions
Given: A short exact sequence with flat, and an -module .
Flatness makes tensoring preserve injections (Flat and faithfully flat modules and ring homomorphisms).
Tensoring is right exact (Tensoring is right exact).
Free modules are flat, and flatness means that tensoring preserves exact sequences; hence tensoring a short exact sequence with a free module preserves short exactness (Under the stated choice boundary, free modules are projective and hence flat, Flat and faithfully flat modules and ring homomorphisms).
Every module admits a canonical surjection from a free module (Every module is a quotient of a free module).
A short exact sequence has an injective first map, a surjective second map, and image equal to kernel (Exact sequences and short exact sequences of modules).
Proof
By [L4], choose a surjection from a free module and let , so is short exact.
Tensor the sequence in step 1.1 with each of . By [L2], the three resulting columns are right exact. The map is injective by flatness of and [L1].
Tensor the given short exact sequence with . Since is flat by [L3], the middle row is short exact. Tensoring it with and gives right-exact bottom and top rows by [L2].
Let map to zero in . By right exactness of the -column, lift to . Its image maps to zero in , so right exactness of the -column gives mapping to .
The image of in maps in to the image of , which is zero because came from . The injectivity in step 2.1 therefore makes the image of in zero.
By right exactness of the bottom row in step 2.2, choose mapping to . In , the images of and of are both ; injectivity of from step 2.2 makes the image of .
The composite is zero because is zero. Hence step 5.1 gives , proving injective.
Right exactness [L2] already gives exactness at , surjectivity onto , and the terminal zero. Together with step 6.1, the tensored sequence is short exact.
Restriction of scalars and extension of scalars along a ring homomorphism
Definition
Let be a unital homomorphism of commutative rings (Ring homomorphism: additive, multiplicative, and required to send to ).
For an -module , its restriction of scalars along , denoted , is the same abelian group with -action
For an -module , its extension of scalars along is
Here is an -bimodule with left action by multiplication and right action (-bimodules and commuting left and right scalar actions). The induced outer action
makes an -module (A commuting outer scalar action descends to a tensor product).
Restriction acts on a homomorphism by leaving its underlying function unchanged. Extension sends to .
Extension of scalars is left adjoint to restriction of scalars
Statement
Let be a homomorphism of commutative rings, let be an -module, and let be an -module. There is a natural bijection
It sends to . Its inverse sends to the -linear map determined by
Facts & Assumptions
Given: A ring map , an -module , and an -module .
Restriction uses , while extension is the -module with (Restriction of scalars and extension of scalars along a ring homomorphism ).
Balanced pairings induce unique homomorphisms from tensor products (Universal property of the tensor product for balanced maps into abelian groups).
An elementary-tensor prescription descends exactly when its underlying pairing is balanced (A formula on elementary tensors defines a homomorphism exactly when its underlying pairing is balanced).
Proof
If is -linear, define . Then , so is -linear into the restriction of .
If is -linear, the pairing is -balanced because . By [L2] and [L3] it induces a homomorphism .
The map is -linear because .
For , one has .
Precomposition in and postcomposition in commute with both displayed formulas, so the inverse bijections are natural in both modules.
For , one has by -linearity, so uniqueness on elementary tensors gives .
Steps 2.2, 2.3 and 3.1 prove that extension of scalars is left adjoint to restriction of scalars.
Change of rings:
Statement
Let be a homomorphism of commutative rings, let be a right -module, and let be a left -module. There is a natural group isomorphism
given by
with inverse .
Facts & Assumptions
Given: A ring map , a right -module , and a left -module .
Restriction makes a right -module and makes an -bimodule used in (Restriction of scalars and extension of scalars along a ring homomorphism ).
Compatible bimodules have the associativity isomorphism (Associativity of tensor products for compatible bimodules).
The tensor-unit isomorphism identifies with by (The regular module is a tensor unit: and ).
Proof
Apply [L2] to , the -bimodule , and , then use [L3] on the left factor to obtain and hence .
Tracing elementary tensors through step 1.1 gives ; tracing the inverse gives .
The two formulas are mutually inverse on elementary tensors. The inverse after the forward map sends to . In the other direction, balance over gives . The universal properties therefore make the composites identities.
Each construction commutes with homomorphisms in and because its elementary-tensor formula does, so the isomorphism is natural.
This proves the change-of-rings isomorphism and its stated inverse.
Extension of scalars carries flat modules to flat modules
Statement
Let be a homomorphism of commutative rings. If is a flat -module, then its extension of scalars is a flat -module.
Facts & Assumptions
Given: A ring map and a flat -module .
For every right -module , there is a natural isomorphism after restriction of scalars (Change of rings: ).
A module is flat exactly when tensoring with it preserves injections (Flatness is equivalent to preserving injections and to the ideal and finitely generated ideal tests).
Restriction of scalars leaves the underlying abelian group and function of a module map unchanged (Restriction of scalars and extension of scalars along a ring homomorphism ).
Proof
Let be an injection of right -modules. By [L3], it is still injective after restriction to -modules.
Since is flat, [L2] makes injective.
Under the natural isomorphisms [L1], the map in step 2.1 is precisely . Hence tensoring over with preserves every injection.
By [L2] over the ring , the extended module is flat.
Flatness is transitive under a flat change of rings
Statement
Let be a flat homomorphism of commutative rings. If is a flat -module, then , restricted to an -module, is flat over . Consequently a composite of flat ring homomorphisms is flat.
The same assertions hold with "faithfully flat" throughout.
Facts & Assumptions
Given: A flat ring map and a flat -module ; for the faithful assertion, assume both are faithfully flat.
A ring map is flat or faithfully flat exactly when its target has that property as a module over its source (Flat and faithfully flat modules and ring homomorphisms).
For every right -module , change of rings gives (Change of rings: ).
Restriction of scalars leaves the underlying groups and maps unchanged (Restriction of scalars and extension of scalars along a ring homomorphism ).
Proof
Let be an exact sequence of -modules. Flatness of over makes exact as a sequence of -modules.
Flatness of over preserves the exactness of step 1.1 after tensoring over .
By [L2], the sequence in step 2.1 is naturally isomorphic to , so the restricted -module is flat.
If both functors are faithful on exactness, the implications in steps 1.1 and 2.1 may be read backwards as well; [L2] then shows that tensoring with over reflects exactness, so the restricted module is faithfully flat.
Taking to be the target ring of a second flat, respectively faithfully flat, ring map and using [L1] proves the corresponding composition statement.
The character dual of a flat module is injective
Statement
Let be a commutative ring, let be a flat -module, and let be an injective -module. Define the character dual
with -action
Then is an injective -module.
Facts & Assumptions
Given: A commutative ring , a flat -module , and an injective -module .
Flatness makes injective whenever is an injection of -modules (Flat and faithfully flat modules and ring homomorphisms).
Balanced maps into an abelian group correspond uniquely to group homomorphisms from a tensor product (Universal property of the tensor product for balanced maps into abelian groups).
An injective module has the extension property along every injective module homomorphism (Injective modules and the extension property).
is an abelian group under pointwise addition (The abelian group and maps induced by pre- and postcomposition).
Proof
The displayed formula makes an -module: and agree at every , and all other module laws hold pointwise in the abelian group [L4].
Let be an injective -module homomorphism and let be -linear. By [L1], is injective.
For every -module , an -linear map determines the balanced map , and [L2] gives a group homomorphism . Conversely a group homomorphism gives ; the balance relation makes this map -linear. These constructions are inverse.
A group homomorphism between abelian groups is automatically -linear because additivity gives compatibility with positive integer multiples and with negatives. Thus the group homomorphisms in step 2.1 are precisely the -module homomorphisms to which injectivity of applies.
Transpose by step 2.1 to . By step 3.1 and injectivity [L3] at the ring , extend it along to a homomorphism .
Transpose back by step 2.1 to an -linear map . Naturality of the evaluation formulas gives .
Every -linear map into therefore extends along every injection, so [L3] makes injective.
5 · Examples, counterexamples and false statements
None yet.
Sources
Standard references
Recommended treatments; not extraction sources.
- Stacks Project, Section 10.12: Tensor products
- C. Dennis, Week 1 recap on tensor products
- H. Miller, Lectures on Algebraic Topology I, Sections 20-21
- C. Dennis, Week 4 on tensor products and flatness
- W. Li, Commutative Algebra, Lectures 9-10
- M. Barr, Acyclic Models, Chapter 2
- Stacks Project, Section 10.39: Flat modules and flat ring maps
- Stacks Project, Lemma 10.39.5
- Stacks Project, Lemma 10.39.2
- Stacks Project, Lemma 10.39.12
- MIT 18.721 Algebraic Geometry notes, Section 2.1
- MIT 18.721 Algebraic Geometry notes, Lemma 2.1.35
- Stacks Project, Lemma 10.39.7
- Stacks Project, Lemma 10.39.4
- M. Barr, Acyclic Models, Proposition 5.17