mathlib documentation

algebra.module.basic

Modules over a ring #

In this file we define

Implementation notes #

In typical mathematical usage, our definition of module corresponds to "semimodule", and the word "module" is reserved for module R M where R is a ring and M an add_comm_group. If R is a field and M an add_comm_group, M would be called an R-vector space. Since those assumptions can be made by changing the typeclasses applied to R and M, without changing the axioms in module, mathlib calls everything a module.

In older versions of mathlib, we had separate semimodule and vector_space abbreviations. This caused inference issues in some cases, while not providing any real advantages, so we decided to use a canonical module typeclass throughout.

Tags #

semimodule, module, vector space

@[class]
structure module (R : Type u) (M : Type v) [semiring R] [add_comm_monoid M] :
Type (max u v)

A module is a generalization of vector spaces to a scalar semiring. It consists of a scalar semiring R and an additive monoid of "vectors" M, connected by a "scalar multiplication" operation r • x : M (where r : R and x : M) with some natural associativity and distributivity axioms similar to those on a ring.

Instances
@[instance]
def module.to_mul_action_with_zero {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] :

A module over a semiring automatically inherits a mul_action_with_zero structure.

Equations
theorem add_smul {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] (r s : R) (x : M) :
(r + s) x = r x + s x
theorem two_smul (R : Type u) {M : Type w} [semiring R] [add_comm_monoid M] [module R M] (x : M) :
2 x = x + x
theorem two_smul' (R : Type u) {M : Type w} [semiring R] [add_comm_monoid M] [module R M] (x : M) :
2 x = bit0 x
def function.injective.module (R : Type u) {M : Type w} {M₂ : Type x} [semiring R] [add_comm_monoid M] [module R M] [add_comm_monoid M₂] [has_scalar R M₂] (f : M₂ →+ M) (hf : function.injective f) (smul : ∀ (c : R) (x : M₂), f (c x) = c f x) :
module R M₂

Pullback a module structure along an injective additive monoid homomorphism. See note [reducible non-instances].

Equations
def function.surjective.module (R : Type u) {M : Type w} {M₂ : Type x} [semiring R] [add_comm_monoid M] [module R M] [add_comm_monoid M₂] [has_scalar R M₂] (f : M →+ M₂) (hf : function.surjective f) (smul : ∀ (c : R) (x : M), f (c x) = c f x) :
module R M₂

Pushforward a module structure along a surjective additive monoid homomorphism.

Equations
def module.comp_hom {R : Type u} {S : Type v} (M : Type w) [semiring R] [add_comm_monoid M] [module R M] [semiring S] (f : S →+* R) :
module S M

Compose a module with a ring_hom, with action f s • m.

See note [reducible non-instances].

Equations
@[simp]
theorem module.to_add_monoid_End_apply_apply (R : Type u) (M : Type w) [semiring R] [add_comm_monoid M] [module R M] (ᾰ : R) (ᾰ_1 : M) :
((module.to_add_monoid_End R M) ᾰ) ᾰ_1 = ᾰ_1
def module.to_add_monoid_End (R : Type u) (M : Type w) [semiring R] [add_comm_monoid M] [module R M] :

(•) as an add_monoid_hom.

This is a stronger version of distrib_mul_action.to_add_monoid_End

Equations
def smul_add_hom (R : Type u) (M : Type w) [semiring R] [add_comm_monoid M] [module R M] :
R →+ M →+ M

A convenience alias for module.to_add_monoid_End as an add_monoid_hom, usually to allow the use of add_monoid_hom.flip.

Equations
@[simp]
theorem smul_add_hom_apply {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] (r : R) (x : M) :
((smul_add_hom R M) r) x = r x
theorem module.eq_zero_of_zero_eq_one {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] (x : M) (zero_eq_one : 0 = 1) :
x = 0
theorem list.sum_smul {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] {l : list R} {x : M} :
l.sum x = (list.map (λ (r : R), r x) l).sum
theorem multiset.sum_smul {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] {l : multiset R} {x : M} :
l.sum x = (multiset.map (λ (r : R), r x) l).sum
theorem finset.sum_smul {R : Type u} {M : Type w} {ι : Type z} [semiring R] [add_comm_monoid M] [module R M] {f : ι → R} {s : finset ι} {x : M} :
(∑ (i : ι) in s, f i) x = ∑ (i : ι) in s, f i x
def module.add_comm_monoid_to_add_comm_group (R : Type u) {M : Type w} [ring R] [add_comm_monoid M] [module R M] :

An add_comm_monoid that is a module over a ring carries a natural add_comm_group structure. See note [reducible non-instances].

Equations
@[nolint]
structure module.core (R : Type u) (M : Type w) [semiring R] [add_comm_group M] :
Type (max u w)
  • to_has_scalar : has_scalar R M
  • smul_add : ∀ (r : R) (x y : M), r (x + y) = r x + r y
  • add_smul : ∀ (r s : R) (x : M), (r + s) x = r x + s x
  • mul_smul : ∀ (r s : R) (x : M), (r * s) x = r s x
  • one_smul : ∀ (x : M), 1 x = x

A structure containing most informations as in a module, except the fields zero_smul and smul_zero. As these fields can be deduced from the other ones when M is an add_comm_group, this provides a way to construct a module structure by checking less properties, in module.of_core.

def module.of_core {R : Type u} {M : Type w} [semiring R] [add_comm_group M] (H : module.core R M) :
module R M

Define module without proving zero_smul and smul_zero by using an auxiliary structure module.core, when the underlying space is an add_comm_group.

Equations
@[ext]
theorem module_ext {R : Type u_1} [semiring R] {M : Type u_2} [add_comm_monoid M] (P Q : module R M) (w : ∀ (r : R) (m : M), r m = r m) :
P = Q

To prove two module structures on a fixed add_comm_monoid agree, it suffices to check the scalar multiplications agree.

@[simp]
theorem neg_smul {R : Type u} {M : Type w} [ring R] [add_comm_group M] [module R M] (r : R) (x : M) :
-r x = -(r x)
@[simp]
theorem neg_smul_neg {R : Type u} {M : Type w} [ring R] [add_comm_group M] [module R M] (r : R) (x : M) :
-r -x = r x
@[simp]
theorem units.neg_smul {R : Type u} {M : Type w} [ring R] [add_comm_group M] [module R M] (u : units R) (x : M) :
-u x = -(u x)
theorem neg_one_smul (R : Type u) {M : Type w} [ring R] [add_comm_group M] [module R M] (x : M) :
(-1) x = -x
theorem sub_smul {R : Type u} {M : Type w} [ring R] [add_comm_group M] [module R M] (r s : R) (y : M) :
(r - s) y = r y - s y
theorem module.subsingleton (R : Type u_1) (M : Type u_2) [semiring R] [subsingleton R] [add_comm_monoid M] [module R M] :

A module over a subsingleton semiring is a subsingleton. We cannot register this as an instance because Lean has no way to guess R.

def ring_hom.to_module {R : Type u} {S : Type v} [semiring R] [semiring S] (f : R →+* S) :
module R S

A ring homomorphism f : R →+* M defines a module structure by r • x = f r * x.

Equations
@[instance]

The tautological action by R →+* R on R.

This generalizes function.End.apply_mul_action.

Equations
@[simp]
theorem ring_hom.smul_def {R : Type u} [semiring R] (f : R →+* R) (a : R) :
f a = f a
theorem nsmul_eq_smul_cast (R : Type u) {M : Type w} [semiring R] [add_comm_monoid M] [module R M] (n : ) (b : M) :
n b = n b

nsmul is equal to any other module structure via a cast.

theorem nat_smul_eq_nsmul {M : Type w} [add_comm_monoid M] (h : module M) (n : ) (x : M) :
n x = n x

Convert back any exotic -smul to the canonical instance. This should not be needed since in mathlib all add_comm_monoids should normally have exactly one -module structure by design.

All -module structures are equal. Not an instance since in mathlib all add_comm_monoid should normally have exactly one -module structure by design.

Equations
@[instance]
def add_comm_monoid.nat_is_scalar_tower {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] :
@[instance]
def add_comm_monoid.nat_smul_comm_class {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] :
@[instance]
def add_comm_monoid.nat_smul_comm_class' {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] :
theorem gsmul_eq_smul_cast (R : Type u) {M : Type w} [ring R] [add_comm_group M] [module R M] (n : ) (b : M) :
n b = n b

gsmul is equal to any other module structure via a cast.

theorem int_smul_eq_gsmul {M : Type w} [add_comm_group M] (h : module M) (n : ) (x : M) :
n x = n x

Convert back any exotic -smul to the canonical instance. This should not be needed since in mathlib all add_comm_groups should normally have exactly one -module structure by design.

All -module structures are equal. Not an instance since in mathlib all add_comm_group should normally have exactly one -module structure by design.

Equations
@[instance]
def add_comm_group.int_is_scalar_tower {R : Type u} {M : Type w} [ring R] [add_comm_group M] [module R M] :
@[instance]
def add_comm_group.int_smul_comm_class {S : Type v} {M : Type w} [semiring S] [add_comm_group M] [module S M] :
@[instance]
def add_comm_group.int_smul_comm_class' {S : Type v} {M : Type w} [semiring S] [add_comm_group M] [module S M] :
theorem add_monoid_hom.map_nat_module_smul {M : Type w} {M₂ : Type x} [add_comm_monoid M] [add_comm_monoid M₂] (f : M →+ M₂) (x : ) (a : M) :
f (x a) = x f a
theorem add_monoid_hom.map_int_module_smul {M : Type w} {M₂ : Type x} [add_comm_group M] [add_comm_group M₂] (f : M →+ M₂) (x : ) (a : M) :
f (x a) = x f a
theorem add_monoid_hom.map_int_cast_smul {M : Type w} {M₂ : Type x} [add_comm_group M] [add_comm_group M₂] (f : M →+ M₂) (R : Type u_1) (S : Type u_2) [ring R] [ring S] [module R M] [module S M₂] (x : ) (a : M) :
f (x a) = x f a
theorem add_monoid_hom.map_nat_cast_smul {M : Type w} {M₂ : Type x} [add_comm_monoid M] [add_comm_monoid M₂] (f : M →+ M₂) (R : Type u_1) (S : Type u_2) [semiring R] [semiring S] [module R M] [module S M₂] (x : ) (a : M) :
f (x a) = x f a
theorem add_monoid_hom.map_inv_int_cast_smul {E : Type u_1} {F : Type u_2} [add_comm_group E] [add_comm_group F] (f : E →+ F) (R : Type u_3) (S : Type u_4) [division_ring R] [division_ring S] [module R E] [module S F] (n : ) (x : E) :
f ((n)⁻¹ x) = (n)⁻¹ f x
theorem add_monoid_hom.map_inv_nat_cast_smul {E : Type u_1} {F : Type u_2} [add_comm_group E] [add_comm_group F] (f : E →+ F) (R : Type u_3) (S : Type u_4) [division_ring R] [division_ring S] [module R E] [module S F] (n : ) (x : E) :
f ((n)⁻¹ x) = (n)⁻¹ f x
theorem add_monoid_hom.map_rat_cast_smul {E : Type u_1} {F : Type u_2} [add_comm_group E] [add_comm_group F] (f : E →+ F) (R : Type u_3) (S : Type u_4) [division_ring R] [division_ring S] [module R E] [module S F] (c : ) (x : E) :
f (c x) = c f x
theorem add_monoid_hom.map_rat_module_smul {E : Type u_1} [add_comm_group E] [module E] {F : Type u_2} [add_comm_group F] [module F] (f : E →+ F) (c : ) (x : E) :
f (c x) = c f x
theorem subsingleton_rat_module (E : Type u_1) [add_comm_group E] :

There can be at most one module ℚ E structure on an additive commutative group. This is not an instance because simp becomes very slow if we have many subsingleton instances, see [gh-6025].

theorem inv_int_cast_smul_eq {E : Type u_1} (R : Type u_2) (S : Type u_3) [add_comm_group E] [division_ring R] [division_ring S] [module R E] [module S E] (n : ) (x : E) :

If E is a vector space over two division rings R and S, then scalar multiplications agree on inverses of integer numbers in R and S.

theorem inv_nat_cast_smul_eq {E : Type u_1} (R : Type u_2) (S : Type u_3) [add_comm_group E] [division_ring R] [division_ring S] [module R E] [module S E] (n : ) (x : E) :

If E is a vector space over two division rings R and S, then scalar multiplications agree on inverses of natural numbers in R and S.

theorem rat_cast_smul_eq {E : Type u_1} (R : Type u_2) (S : Type u_3) [add_comm_group E] [division_ring R] [division_ring S] [module R E] [module S E] (r : ) (x : E) :
r x = r x

If E is a vector space over two division rings R and S, then scalar multiplications agree on rational numbers in R and S.

no_zero_smul_divisors #

This section defines the no_zero_smul_divisors class, and includes some tests for the vanishing of elements (especially in modules over division rings).

@[class]
structure no_zero_smul_divisors (R : Type u_1) (M : Type u_2) [has_zero R] [has_zero M] [has_scalar R M] :
Prop
  • eq_zero_or_eq_zero_of_smul_eq_zero : ∀ {c : R} {x : M}, c x = 0c = 0 x = 0

no_zero_smul_divisors R M states that a scalar multiple is 0 only if either argument is 0. This a version of saying that M is torsion free, without assuming R is zero-divisor free.

The main application of no_zero_smul_divisors R M, when M is a module, is the result smul_eq_zero: a scalar multiple is 0 iff either argument is 0.

It is a generalization of the no_zero_divisors class to heterogeneous multiplication.

Instances
@[simp]
theorem smul_eq_zero {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] [no_zero_smul_divisors R M] {c : R} {x : M} :
c x = 0 c = 0 x = 0
theorem smul_ne_zero {R : Type u} {M : Type w} [semiring R] [add_comm_monoid M] [module R M] [no_zero_smul_divisors R M] {c : R} {x : M} :
c x 0 c 0 x 0
theorem eq_zero_of_smul_two_eq_zero (R : Type u) {M : Type w} [semiring R] [add_comm_monoid M] [module R M] [no_zero_smul_divisors R M] [char_zero R] {v : M} (hv : 2 v = 0) :
v = 0
theorem smul_right_injective {R : Type u} (M : Type w) [semiring R] [add_comm_group M] [module R M] [no_zero_smul_divisors R M] {c : R} (hc : c 0) :
function.injective (λ (x : M), c x)
theorem eq_zero_of_eq_neg (R : Type u) {M : Type w} [semiring R] [add_comm_group M] [module R M] [no_zero_smul_divisors R M] [char_zero R] {v : M} (hv : v = -v) :
v = 0
theorem smul_left_injective (R : Type u) {M : Type w} [ring R] [add_comm_group M] [module R M] [no_zero_smul_divisors R M] {x : M} (hx : x 0) :
function.injective (λ (c : R), c x)
theorem ne_neg_of_ne_zero {R : Type u} [ring R] [char_zero R] [no_zero_divisors R] {v : R} (hv : v 0) :
v -v
@[instance]
@[simp]
theorem nat.smul_one_eq_coe {R : Type u_1} [semiring R] (m : ) :
m 1 = m
@[simp]
theorem int.smul_one_eq_coe {R : Type u_1} [ring R] (m : ) :
m 1 = m