3.1. Definitions.
In group theory, we started with a set equipped with a bilinear
operation which mapped to itself.
The operation was required to be associative, and there needed to
be inverses and an identity element.
In ring theory, we went further to assume was not only an
abelian group, we placed the group operation with and then defined a multiplication which is was associative and left- and right-
distributive.
Finally, we reach module theory, which considers again an abelian
group with operation but lets a ring
act on , whose addition agrees with the one
which acts on but whose multiplication acts on and .
Note how abelian groups and rings are special cases of modules.
This will be more clear once we introduce the axioms.
Let be a ring with identity, and an abelian group
equipped with . Then is an
left -module if we equip with
multiplication and for all and
\begin{enumerate}
\item
\item
\item
\item where is the identity of .
\end{enumerate}
Alternatively, an abelian group is a right -module if we equip
with multiplication and for all and
\begin{enumerate}
\item
\item
\item
\item where is the identity of .
\end{enumerate}
Notice that we can think of these products as a group action, or sort of a
"ring action" acting on . That is, an -module is just an abelian
group that a ring can act on. If you have an abelian group
that simply cannot act on and satisfy the above axioms, then
is not an -module.
\
For convenience, we will develop the theory of -modules by
solely working with left -modules, since all proofs and
statements will be equivalent up to a swap of variables for right
-modules.
\
Examples.\
- 1. \textcolor{NavyBlue}{Note that if is commutative, then a left
-module coincides with a right -module. To see this, let
be a left -module. Then construct the right -module by
defining the multiplication as
Then we see that for all , ,
\begin{enumerate}
* \checkmark
* \checkmark
* \checkmark
* . \checkmark
\end{enumerate}
Note that in part is where we used the fact that is
commutative. So whenever is commutative, the existence of a
left -module automatically implies that existence of a
right -module, and vice versa.
}
* 2. Let be a ring. Then if we substitute in the above
definition, and let the multiplication $
\cdot$ be the multiplication on then is a left and a right
-module. This is because is an abelian group which is
associative and left- and right-distributive. Hence, it satisfies
all of the above axioms.
So keep in mind that a ring is just a left- and right-
module that acts on .
Here's another example which shows that abelian groups are simply
modules.
Let be an abelian group. Then is a left and right
-module.
Let act on as follows. Define
and
Then with this definition of multiplication, it is easy to
show that the axioms (a)-(d) are satisfied.
- 3. If is a ring and is a left (right) ideal of
, then is a left (right) -module.
- 4. Let be a vector space defined over a field .
Then is an -module. (Now it is clear why there are a
million axioms in the definition of a vector space!)
With -modules introduced and understood, we can jump right into
homomorphisms.
Let be a ring and and be -modules. We define
to be an -module homomorphism if
- 1. for any
- 2. for all and .
If is a bijective -module homomorphism, then we say
that is an isomorphism and that .
Thus we see that -module homomorphisms must not only be linear
over the elements of , but they must also pull out scalar
multiplication by elements of .
Recall earlier that we said a vector space over a field is
an -module. Now if is another vector space and
is a linear transformation, then we see that is also an
-module homomorphism!
In the language of linear algebra, a
linear transformation is usually defined as a function
such that for any and
we have that
- 1.
- 2. $T(\alpha\bf{v}) = $ .
As we will see, linear algebra is basically a special case of
module theory.
Let be a ring and and a pair of -modules. Then
is the set of all -module homomorphisms from
to .
\textcolor{MidnightBlue}{It turns out we can turn
into an abelian group, and under special circumstances it can
actually be an -module itself. It will be the case
that will actually be an important functor, but that is
for later.}
\
\indent To turn this into an abelian group, we define addition of
the elements to be
for all . We let the identity be the
zero map, and realize associativity and closedness are a given to
conclude that this is in fact an abelian group.
\
Suppose we want to make , our ring, act on in
order for it to be an -module. Then we define scalar
multiplication to be ; a pretty reasonable
definition for scalar multiplication.
\textcolor{Red}{This issue with this is that will
not be closed under scalar multiplication of elements of
unless is a commutative ring.
}
We'll demonstrate this as follows. Let and
. Then the second property of an -module
homomorphism tells us that for all . Now suppose we try to use our definition of scalar
multiplication, and consider where . Then if we try
to see if will pass the second criterion for being an
-module homomorphism, we see that
That is, we see that isn't an -module homomorphism because
(which is required for an -module homomorphism); rather, Now if
is a commutative ring, then
so we can then say that , in which case
passes the test for being an -module homomorphism.
This proves the following propsition, which will be useful for
reference for later.
Let and be -modules. Then is an
abelian group. Furthermore, it is an
-module if and only if is a commutative ring.
Next, we make the following definitions for completeness.
Let be a ring and and be -modules. If is an -module homomorphism, then
- 1. The set is
the kernal of
- 2. The set is the
image of .