be a commutative ring. (Recall that we always assume ring to
be unital associative.)
Amazingly enough(!?), any element
in
is central.
As we have seen in the Schur's Lemma, for any “finite dimensional”
irreducible representation
of
,
should be a
scalar. Thus we see that any irreducible “finite dimensional”
irreducible representation of
should be one dimensional.
Though this argument does not make sense when
has no restriction
such as “
is finitely generated over a field”,
we may begin by considering a one-dimensional representation of
.
That means, a ring homomorphism
is a field.
One knows that
is an integral domain. That means, it has no
zero-divisor other than zero. (In this sense,
is said to
be a prime ideal of
.)
is decomposed in the following way.
is the field of fractions of a ring
.
With a suitable definition of “equivalence" of such representations,
we may identify equivalence class of representation with
the kernel
.
In other words, we are interested in prime ideals.
be a commutative ring.
Then we define the set
of spectrum of
as the set of
prime ideals of
.
We note that for any
, we have a ring homomorphism
(“representation associated to
”)
defined by
is an inclusion, we may say, by abuse of language,
that the value of an element
under the representation
is equal to
. We note further that
Let us now define a topology on
.
be a commutative ring.
For any
, we define a subset
of
defined by
be a commutative ring.
Then we have
.
.
Thus we may introduce a topology on
whose open sets are
unions of various
.

.
In Part II, we always equip
with the Zariski topology.
Thus for any commutative ring
, we may always associate a topological
space
.