<< стр. 8(всего 14)СОДЕРЖАНИЕ >>
Every nonzero p-adic integer has the form x = pn u where n в‰Ґ 0 and u is a p-adic unit.
Consequently, Оёp is an integral domain. Furthermore, Оёp has only one prime element p,
and every x в€€ Оёp is a power of p, up to multiplication by a unit.
Proof. The series representation for x has a nonzero term an pn of lowest degree n, where
an can be taken between 1 and p в€’ 1. Factor out pn to obtain x = pn u, where u is a unit
by (7.9.2). в™Ј

7.9.4 Deп¬Ѓnitions and Comments
The quotient п¬Ѓeld Qp of Оёp is called the п¬Ѓeld of p-adic numbers. By (7.9.3), each О± в€€ Qp
has the form pm u, where m is an integer (possibly negative) and u is a unit in Оёp . Thus
О± has a вЂњLaurent expansionвЂќ
aв€’r aв€’1
+ В·В·В· + + a0 + a1 p + В· В· В· .
r
p p
Another representation is О± = x/pr , where x is a p-adic integer and r в‰Ґ 0. This version
is convenient for doing addition and multiplication in Qp .
The rationals Q are a subп¬Ѓeld of Qp . To see this, let a/b be a rational number in
lowest terms (a and b relatively prime). If p does not divide b, then by (7.9.2), b is a unit
of Оёp . Since a в€€ Z вЉ† Оёp , we have a/b в€€ Оёp . If b = pt b where p does not divide b , we
can factor out pt and reduce to the previous case. Thus a/b always belongs to Qp , and
a/b в€€ Оёp iп¬Ђ p does not divide b. Rational numbers belonging to Оёp are sometimes called
p-integers.
We now outline a procedure for constructing the p-adic numbers formally.

7.9.5 Deп¬Ѓnitions and Comments
The p-adic valuation on Qp is deп¬Ѓned by

vp (pm u) = m.

In general, a valuation v on a п¬Ѓeld F is a real-valued function on F \ {0} satisfying:
(a) v(xy) = v(x) + v(y);
(b) v(x + y) в‰Ґ min(v(x), v(y)).
By convention, we take v(0) = + в€ћ.
The representation x = pm u shows that vp is indeed a valuation on Qp . If c is any
real number greater than 1, then the valuation v induces an absolute value on F , namely,

|x| = cв€’v(x) .

When v = vp , the constant c is usually taken to be p, and we obtain the p-adic absolute
value

|x|p = pв€’vp (x) .
32 CHAPTER 7. INTRODUCING ALGEBRAIC NUMBER THEORY

Thus the p-adic absolute value of pn is pв€’n , which approaches 0 exponentially as n ap-
proaches inп¬Ѓnity.
In general, an absolute value on a п¬Ѓeld F is a real-valued function on F such that:

(i) |x| в‰Ґ 0, with equality if and only if x = 0;
(ii) |xy| = |x||y|;
(iii) |x + y| в‰¤ |x| + |y|.

By (b), an absolute value induced by a valuation satisп¬Ѓes a property that is stronger
than (iii):

(iv) |x + y| в‰¤ max(|x|, |y|).

An absolute value satisfying (iv) is said to be nonarchimedian.

7.9.6 Proposition
Let F be the quotient п¬Ѓeld of an integral domain R. The absolute value | | on F is
nonarchimedian if and only if |n| в‰¤ 1 for every integer n = 1 В± В· В· В· В± 1 в€€ R.

Proof. Assume a nonarchimedian absolute value. By property (ii) of (7.9.5), | В± 1| = 1. If
|n| в‰¤ 1, then by property (iv), |nВ±1| в‰¤ 1, and the desired conclusion follows by induction.
Conversely, assume that the absolute value of every integer is at most 1. To prove (iv), it
suп¬ѓces to show that |x + 1| в‰¤ max(|x|, 1) for every x в€€ F . [If y = 0 in (iv), divide by |y|.]
By the binomial theorem,
n n
nr n
|x + 1| = | x |в‰¤ | ||x|r .
n
r r
r=0 r=0

n
has absolute value at most 1. If |x| > 1, then |x|r в‰¤ |x|n
By hypothesis, the integer
r
for all r = 0, 1, . . . , n. If |x| в‰¤ 1, then |x|r в‰¤ 1. Consequently,

|x + 1|n в‰¤ (n + 1) max(|x|n , 1).

Take nth roots and let n в†’ в€ћ to get |x + 1| в‰¤ max(|x|, 1). в™Ј

The next result may seem innocuous, but it leads to a remarkable property of nonar-
chimedian absolute values.

7.9.7 Proposition
If | | is a nonarchimedian absolute value, then

|x| = |y| implies |x + y| = max(|x|, |y|).
7.9. P-ADIC NUMBERS 33

Proof. First note that | в€’ y| = |(в€’1)y| = | в€’ 1||y| = |y|. We can assume without loss of
generality that |x| > |y|. Using property (iv) of (7.9.5), we have

|x| = |x + y в€’ y| в‰¤ max(|x + y|, |y|) = |x + y|.

[Otherwise, max(|x + y|, |y|) = |y|, hence |x| в‰¤ |y| < |x|, a contradiction.] Since |x + y| в‰¤
max(|x|, |y|) = |x|, the result follows. в™Ј
Any absolute value determines a metric via d(x, y) = |x в€’ y|. This distance function
can be used to measure the length of the sides of a triangle.

7.9.8 Corollary
With respect to the metric induced by a nonarchimedian absolute value, all triangles are
isosceles.
Proof. Let the vertices of the triangle be x, y and z. Then

|x в€’ y| = |(x в€’ z) + (z в€’ y)|.

If |x в€’ z| = |z в€’ y|, then two side lengths are equal. If |x в€’ z| = ||z в€’ y|, then by (7.9.7),
|x в€’ y| = max(|x в€’ z|, |z в€’ y|), and again two side lengths are equal. в™Ј
We now look at the p-adic numbers from the viewpoint of valuation theory.

7.9.9 Deп¬Ѓnitions and Comments
Let | | be a nonarchimedian absolute value on the п¬Ѓeld F . The valuation ring of | | is

Оё = {x в€€ F : |x| в‰¤ 1}.

In the p-adic case, Оё = {x в€€ Qp : vp (x) в‰Ґ 0} = Оёp . By properties (ii) and (iv) of (7.9.5), Оё
is a subring of F .
The valuation ideal of | | is

ОІ = {x в€€ F : |x| < 1}.

In the p-adic case, ОІ = {x в€€ Qp : vp (x) в‰Ґ 1} = pОёp , those p-adic integers whose series
representation has no constant term. To verify that ОІ is an ideal of Оё, note that if x, y в€€ ОІ
and r в€€ Оё, then |rx| = |r||x| в‰¤ |x| < 1 and |x + y| в‰¤ max(|x|, |y|) < 1.
Now if x в€€ Оё \ ОІ, then |x| = 1, hence |xв€’1 | = 1/|x| = 1, so xв€’1 в€€ Оё and x is a
unit of Оё. On the other hand, if x в€€ ОІ, then x cannot be a unit of Оё. [If xy = 1, then
1 = |x||y| в‰¤ |x| < 1, a contradiction.] Thus the ideal ОІ is the set of all nonunits of Оё. No
proper ideal I of Оё can contain a unit, so I вЉ† ОІ. It follows that ОІ is the unique maximal
ideal of Оё. A ring with a unique maximal ideal is called a local ring. We will meet such
rings again when we examine the localization process in Section 8.5.
To construct the p-adic numbers, we start with the p-adic valuation on the integers,
and extend it to the rationals in the natural way: vp (a/b) = vp (a) в€’ vp (b). The p-adic
valuation then determines the p-adic absolute value, which induces a metric d on Q.
34 CHAPTER 7. INTRODUCING ALGEBRAIC NUMBER THEORY

[Because d comes from a nonarchimedian absolute value, it will satisfy the ultrametric
inequality d(x, y) в‰¤ max(d(x, z), d(z, y)), which is stronger than the triangle inequality.]
The process of constructing the real numbers by completing the rationals using equivalence
classes of Cauchy sequences is familiar. The same process can be carried out using the
p-adic absolute value rather than the usual absolute value on Q. The result is a complete
metric space, the п¬Ѓeld of p-adic numbers, in which Q is dense.
OstrowskiвЂ™s theorem says that the usual absolute value | |в€ћ on Q, along with the p-adic
absolute values | |p for all primes p, and the trivial absolute value (|0| = 0; |x| = 1 for
x = 0), essentially exhaust all possibilities. To be more precise, two absolute values on a
п¬Ѓeld F are equivalent if the corresponding metrics on F induce the same topology. Any
nontrivial absolute value on Q is equivalent to | |в€ћ or to one of the | |p .

Problems For Section 7.9
1. Take p = 3, and compute the standard representation of (2 + p + p2 )(2 + p2 ) in two
ways, using (1) and (2) of (7.9.1). Check the result by computing the product using
ordinary multiplication of two integers, and then expanding in base p = 3.
2. Express the p-adic integer -1 as an inп¬Ѓnite series of the form (1), using the standard
representation.
3. Show that every absolute value on a п¬Ѓnite п¬Ѓeld is trivial.
4. Show that an absolute value is archimedian iп¬Ђ the set S = {|n| : n в€€ Z} is unbounded.
5. Show that a п¬Ѓeld that has an archimedian absolute value must have characteristic 0.
zn of p-adic numbers converges if and only if zn в†’ 0 as
6. Show that an inп¬Ѓnite series
n в†’ в€ћ.
7. Show that the sequence an = n! of p-adic integers converges to 0.
8. Does the sequence an = n converge in Qp ?
Chapter 8

Introducing Algebraic
Geometry

(Commutative Algebra 2)

We will develop enough geometry to allow an appreciation of the Hilbert Nullstellensatz,
and look at some techniques of commutative algebra that have geometric signiп¬Ѓcance. As
in Chapter 7, unless otherwise speciп¬Ѓed, all rings will be assumed commutative.

8.1 Varieties
8.1.1 Deп¬Ѓnitions and Comments
We will be working in k[X1 , . . . , Xn ], the ring of polynomials in n variables over the п¬Ѓeld
k. (Any application of the Nullstellensatz requires that k be algebraically closed, but we
will not make this assumption until it becomes necessary.) The set An = An (k) of all
n-tuples with components in k is called aп¬ѓne n-space. If S is a set of polynomials in
k[X1 , . . . , Xn ], then the zero-set of S, that is, the set V = V (S) of all x в€€ An such that
f (x) = 0 for every f в€€ S, is called a variety. (The term вЂњaп¬ѓne varietyвЂќ is more precise,
but we will use the short form because we will not be discussing projective varieties.)
Thus a variety is the solution set of simultaneous polynomial equations.
If I is the ideal generated by S, then I consists of all п¬Ѓnite linear combinations gi fi
with gi в€€ k[X1 , . . . , Xn ] and fi в€€ S. It follows that V (S) = V (I), so every variety is the
variety of some ideal. We now prove that we can make An into a topological space by
taking varieties as the closed sets.

8.1.2 Proposition
(1) If VО± = V (IО± ) for all О± в€€ T , then VО± = V ( IО± ). Thus an arbitrary intersection of
varieties is a variety.

1
2 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

r
(2) If Vj = V (Ij ), j = 1, . . . , r, then j=1 Vj = V ({f1 В· В· В· fr : fj в€€ Ij , 1 в‰¤ j в‰¤ r}). Thus a
п¬Ѓnite union of varieties is a variety.
(3) An = V (0) and в€… = V (1), so the entire space and the empty set are varieties.

Consequently, there is a topology on An , called the Zariski topology, such that the
closed sets and the varieties coincide.

Proof. (1) If x в€€ An , then x в€€ VО± iп¬Ђ every polynomial in every IО± vanishes at x iп¬Ђ
x в€€ V ( IО± ).
r
(2) x в€€ j=1 Vj iп¬Ђ for some j, every fj в€€ Ij vanishes at x iп¬Ђ x в€€ V ({f1 В· В· В· fr : fj в€€ Ij
for all j}).
(3) The zero polynomial vanishes everywhere and a nonzero constant polynomial van-
ishes nowhere. в™Ј

Note that condition (2) can also be expressed as
пЈ« пЈ¶
r
=V пЈ­ Ij пЈё = V в€©r Ij .
в€Єr Vj
j=1 j=1
j=1

[See (7.6.1) for the deп¬Ѓnition of a product of ideals.]
We have seen that every subset of k[X1 , . . . , Xn ], in particular every ideal, determines
a variety. We can reverse this process as follows.

8.1.3 Deп¬Ѓnitions and Comments
If X is an arbitrary subset of An , we deп¬Ѓne the ideal of X as I(X) = {f в€€ k[X1 , . . . , Xn ] :
f vanishes on X}. By deп¬Ѓnition we have:

(4) If X вЉ† Y then I(X) вЉ‡ I(Y ); if S вЉ† T then V (S) вЉ‡ V (T ).

Now if S is any set of polynomials, deп¬Ѓne IV (S) as I(V (S)), the ideal of the zero-set
of S; we are simply omitting parentheses for convenience. Similarly, if X is any subset
of An , we can deп¬Ѓne V I(X), IV I(X), V IV (S), and so on. From the deп¬Ѓnitions we
have:

(5) IV (S) вЉ‡ S; V I(X) вЉ‡ X.

[If f в€€ S then f vanishes on V (S), hence f в€€ IV (S). If x в€€ X then every polynomial in
I(X) vanishes at x, so x belongs to the zero-set of I(X).]
If we keep applying V вЂ™s and IвЂ™s alternately, the sequence stabilizes very quickly:

(6) V IV (S) = V (S); IV I(X) = I(X).

[In each case, apply (4) and (5) to show that the left side is a subset of the right side. If
x в€€ V (S) and f в€€ IV (S) then f (x) = 0, so x в€€ V IV (S). If f в€€ I(X) and x в€€ V I(X)
then x belongs to the zero-set of I(X), so f (x) = 0. Thus f vanishes on V I(X), so
f в€€ IV I(X).]
Since every polynomial vanishes on the empty set (vacuously), we have:
8.1. VARIETIES 3

(7) I(в€…) = k[X1 , . . . , Xn ].

The next two properties require a bit more eп¬Ђort.

(8) If k is an inп¬Ѓnite п¬Ѓeld, then I(An ) = {0};
(9) If x = (a1 , . . . , an ) в€€ An , then I({x}) = (X1 в€’ a1 , . . . , Xn в€’ an ).

Property (8) holds for n = 1 since a nonconstant polynomial in one variable has only
п¬Ѓnitely many zeros. Thus f = 0 implies that f в€€ I(A1 ). If n > 1, let f = ar X1 + В· В· В· +
r
/
a1 X1 + a0 where the ai are polynomials in X2 , . . . , Xn and ar = 0. By the induction
hypothesis, there is a point (x2 , . . . , xn ) at which ar does not vanish. Fixing this point,
we can regard f as a polynomial in X1 , which cannot possibly vanish at all x1 в€€ k. Thus
f в€€ I(An ).
/
To prove (9), note that the right side is contained in the left side because Xi в€’ ai is 0
when Xi = ai . Also, the result holds for n = 1 by the remainder theorem (2.5.2). Thus
assume n > 1 and let f = br X1 + В· В· В· + b1 X1 + b0 в€€ I({x}), where the bi are polynomials
r

in X2 , . . . , Xn and br = 0. By the division algorithm (2.5.1), we have

f = (X1 в€’ a1 )g(X1 , . . . , Xn ) + h(X2 , . . . , Xn )

and h must vanish at (a2 , . . . , an ). By the induction hypothesis, h в€€ (X2 в€’a2 , . . . , Xn в€’an ),
hence f в€€ (X1 в€’ a1 , X2 в€’ a2 , . . . , Xn в€’ an ).

Problems For Section 8.1
A variety is said to be reducible if it can be expressed as the union of two proper subva-
rieties; otherwise the variety is irreducible. In Problems 1вЂ“4, we are going to show that a
variety V is irreducible if and only if I(V ) is a prime ideal.

1. Assume that I(V ) is not prime, and let f1 f2 в€€ I(V ) with f1 , f2 в€€ I(V ). If x в€€ V ,
/
show that x в€€ V (f1 ) implies x в€€ V (f2 ) (and similarly, x в€€ V (f2 ) implies x в€€ V (f1 )).
/ /
2. Show that V is reducible.
3. Show that if V and W are varieties with V вЉ‚ W , then I(V ) вЉѓ I(W ).
4. Now assume that V = V1 V2 , with V1 , V2 вЉ‚ V . By Problem 3, we can choose
fi в€€ I(Vi ) with fi в€€ I(V ). Show that f1 f2 в€€ I(V ), so I(V ) is not a prime ideal.
/
5. Show that any variety is the union of п¬Ѓnitely many irreducible subvarieties.
6. Show that the decomposition of Problem 5 is unique, assuming that we discard any
subvariety that is contained in another one.
7. Assume that k is algebraically closed. Suppose that An is covered by open sets An \
V (Ii ) in the Zariski topology. Let I is the ideal generated by the Ii , so that I = Ii ,
the set of all п¬Ѓnite sums xi1 + В· В· В· xir with xij в€€ Iij . Show that 1 в€€ I. (You may appeal
to the weak Nullstellensatz, to be proved in Section 8.4.)
8. Show that An is compact in the Zariski topology.
4 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

8.2 The Hilbert Basis Theorem
If S is a set of polynomials in k[X1 , . . . , Xn ], we have deп¬Ѓned the variety V (S) as the
zero-set of S, and we know that V (S) = V (I), where I is the ideal generated by S. Thus
any set of simultaneous polynomial equations deп¬Ѓnes a variety. In general, inп¬Ѓnitely many
equations may be involved, but as Hilbert proved, an inп¬Ѓnite collection of equations can
always be replaced by a п¬Ѓnite collection. The reason is that every ideal of k[X1 , . . . , Xn ]
has a п¬Ѓnite set of generators, in other words, k[X1 , . . . , Xn ] is a Noetherian ring. The
п¬Ѓeld k is, in particular, a PID, so k is Noetherian. The key step is to show that if R is a
Noetherian ring, then the polynomial ring in n variables over R is also Noetherian.

8.2.1 Hilbert Basis Theorem
If R is a Noetherian ring, then R[X1 , . . . , Xn ] is also Noetherian.

Proof. By induction, we can assume n = 1. Let I be an ideal of R[X], and let J be the ideal
of all leading coeп¬ѓcients of polynomials in I. (The leading coeп¬ѓcient of 5X 2 в€’ 3X + 17
is 5; the leading coeп¬ѓcient of the zero polynomial is 0.) By hypothesis, J is п¬Ѓnitely
generated, say by a1 , . . . , an . Let fi be a polynomial in I whose leading coeп¬ѓcient is ai ,
and let di be the degree of fi . Let I в€— consist of all polynomials in I of degree at most
d = max{di : 1 в‰¤ i в‰¤ n}. Then I в€— is an R-submodule of the free R-module M of all
polynomials b0 + b1 X + В· В· В· + bd X d , bi в€€ R. Now a п¬Ѓnitely generated free R-module is a
п¬Ѓnite direct sum of copies of R, hence M , and therefore I в€— , is Noetherian. Thus I в€— can
be generated by п¬Ѓnitely many polynomials g1 , . . . , gm . Take I0 to be the ideal of R[X]
generated by f1 , . . . , fn , g1 , . . . , gm . We will show that I0 = I, proving that I is п¬Ѓnitely
generated.
First observe that fi в€€ I and gj в€€ I в€— вЉ† I, so I0 вЉ† I. Thus we must show that each
h в€€ I belongs to I0 .
Case 1 : deg h в‰¤ d
Then h в€€ I в€— , so h is a linear combination of the gj (with coeп¬ѓcients in R вЉ† R[X]), so
h в€€ I0 .
Case 2 : deg h = r > d
n
Let a be the leading coeп¬ѓcient of h. Since a в€€ J, we have a = i=1 ci ai with the
ci в€€ R. Let
n
q =hв€’ ci X rв€’di fi в€€ I.
i=1

The coeп¬ѓcient of X r in q is
n
aв€’ ci ai = 0
i=1

so that deg q < r. We can iterate this degree-reduction process until the resulting poly-
nomial has degree d or less, and therefore belongs to I0 . But then h is a п¬Ѓnite linear
combination of the fi and gj . в™Ј
8.2. THE HILBERT BASIS THEOREM 5

8.2.2 Corollary
Every variety is the intersection of п¬Ѓnitely many hypersurfaces (zero-sets of single poly-
nomials).
Proof. Let V = V (I) be a variety. By (8.2.1), I has п¬Ѓnitely many generators f1 , . . . , fr .
r
But then V = i=1 V (fi ). в™Ј

8.2.3 Formal Power Series
The argument used to prove the Hilbert basis theorem can be adapted to show that if R is
Noetherian, then the ring R[[X]] of formal power series is Noetherian. We cannot simply
reproduce the proof because an inп¬Ѓnite series has no term of highest degree, but we can
look at the lowest degree term. If f = ar X r + ar+1 X r+1 + В· В· В· , where r is a nonnegative
integer and ar = 0, let us say that f has degree r and leading coeп¬ѓcient ar . (If f = 0,
take the degree to be inп¬Ѓnite and the leading coeп¬ѓcient to be 0.)
If I is an ideal of R[[X]], we must show that I is п¬Ѓnitely generated. We will inductively
construct a sequence of elements fi в€€ R[[X]] as follows. Let f1 have minimal degree among
elements of I. Suppose that we have chosen f1 , . . . , fi , where fi has degree di and leading
coeп¬ѓcient ai . We then select fi+1 satisfying the following three requirements:
1. fi+1 belongs to I;
2. ai+1 does not belong to (a1 , . . . , ai ), the ideal of R generated by the aj , j = 1, . . . , i;
3. Among all elements satisfying the п¬Ѓrst two conditions, fi+1 has minimal degree.
The second condition forces the procedure to terminate in a п¬Ѓnite number of steps;
otherwise there would be an inп¬Ѓnite ascending chain (a1 ) вЉ‚ (a1 , a2 ) вЉ‚ (a1 , a2 , a3 ) вЉ‚ В· В· В· .
If stabilization occurs at step k, we will show that I is generated by f1 , . . . , fk .
Let g = aX d + В· В· В· be an element of I of degree d and leading coeп¬ѓcient a. Then
a в€€ (a1 , . . . , ak ) (Problem 1).
Case 1 : d в‰Ґ dk . Since di в‰¤ di+1 for all i (Problem 2), we have d в‰Ґ di for i = 1, . . . , k.
k
Now a = i=1 ci0 ai with the ci0 в€€ R. Deп¬Ѓne
k
ci0 X dв€’di fi
g0 =
i=1

so that g0 has degree d and leading coeп¬ѓcient a, and consequently g в€’ g0 has degree
r
greater than d. Having deп¬Ѓned g0 , . . . , gr в€€ (f1 , . . . , fk ) such that g в€’ i=0 gi has degree
greater than d + r, say
r
gв€’ gi = bX d+r+1 + . . . .
i=0

(The argument is the same if the degree is greater than d + r + 1.) Now b в€€ (a1 , . . . , ak )
(Problem 1 again), so
k
b= ci,r+1 ai
i=1
6 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

with ci,r+1 в€€ R. We deп¬Ѓne
k
ci,r+1 X d+r+1в€’di fi
gr+1 =
i=1

r+1
so that g в€’ i=0 gi has degree greater than d + r + 1. Thus
в€ћ в€ћ k
cir X d+rв€’di fi
g= gr =
r=0 r=0 i=1

and it follows upon reversing the order of summation that g в€€ (f1 , . . . , fk ). (The reversal
is legal because the inner summation is п¬Ѓnite. For a given nonnegative integer j, there
are only п¬Ѓnitely many terms of the form bX j .)
Case 2 : d < dk . As above, a в€€ (a1 , . . . , ak ), so there is a smallest m between 1 and
k such that a в€€ (a1 , . . . , am ). It follows that d в‰Ґ dm (Problem 3). As in case 1 we have
m
a = i=1 ci ai with ci в€€ R. Deп¬Ѓne
m
ci X dв€’di fi в€€ (f1 , . . . , fk ) вЉ† I.
h=
i=1

The leading coeп¬ѓcient of h is a, so the degree of g в€’ h is greater than d. We replace g by
g в€’ h and repeat the procedure. After at most dk в€’ d iterations, we produce an element
g в€’ hi in I of degree at least dk , with all hi в€€ (f1 , . . . , fk ). By the analysis of case 1,
g в€€ (f1 , . . . , fk ).

Problems For Section 8.2
1. Justify the step a в€€ (a1 , . . . , ak ) in (8.2.3).
2. Justify the step di в‰¤ di+1 in (8.2.3).
3. Justify the step d в‰Ґ dm in (8.2.3).
4. Let R be a subring of the ring S, and assume that S is п¬Ѓnitely generated as an algebra
over R. In other words, there are п¬Ѓnitely many elements x1 , . . . , xn в€€ S such that the
smallest subring of S containing the xi and all elements of R is S itself. Show that S
is a homomorhic image of the polynomial ring R[X1 , . . . , Xn ].
5. Continuing Problem 4, show that if R is Noetherian, then S is also Noetherian.

8.3 The Nullstellensatz: Preliminaries
We have observed that every variety V deп¬Ѓnes an ideal I(V ) and every ideal I deп¬Ѓnes a
variety V (I). Moreover, if I(V1 ) = I(V2 ), then V1 = V2 by (6) of (8.1.3). But it is entirely
possible for many ideals to deп¬Ѓne the same variety. For example, the ideals (f ) and (f m )
need not coincide, but their zero-sets are identical. Appearances to the contrary, the two
statements in part (6) of (8.1.3) are not symmetrical. A variety V is, by deп¬Ѓnition, always
expressible as V (S) for some collection S of polynomials, but an ideal I need not be of the
8.3. THE NULLSTELLENSATZ: PRELIMINARIES 7

special form I(X). HilbertвЂ™s Nullstellensatz says that if two ideals deп¬Ѓne the same variety,
then, informally, the ideals are the same вЂњup to powersвЂќ. More precisely, if g belongs to
one of the ideals, then g r belongs to the other ideal for some positive integer r. Thus the
only factor preventing a one-to-one correspondence between ideals and varieties is that
a polynomial can be raised to a power without aп¬Ђecting its zero-set. In this section we
collect some results needed for the proof of the Nullstellensatz. We begin by showing that
each point of An determines a maximal ideal.

8.3.1 Lemma
If a = (a1 , . . . , an ) в€€ An , then I = (X1 в€’ a1 , . . . , Xn в€’ an ) is a maximal ideal.

Proof. Suppose that I is properly contained in the ideal J, with f в€€ J \ I. Apply the
division algorithm n times to get

f = A1 (X1 в€’ a1 ) + A2 (X2 в€’ a2 ) + В· В· В· + An (Xn в€’ an ) + b

where A1 в€€ k[X1 , . . . , Xn ], A2 в€€ k[X2 , . . . , Xn ], . . . , An в€€ k[Xn ], b в€€ k. Note that b
cannot be 0 since f в€€ I. But f в€€ J, so by solving the above equation for b we have b в€€ J,
/
hence 1 = (1/b)b в€€ J. Consequently, J = k[X1 , . . . , Xn ]. в™Ј

The following deп¬Ѓnition will allow a precise statement of the Nullstellensatz.

8.3.2 Deп¬Ѓnition
The radical of an ideal I (in any commutative ring R) is the set of all elements f в€€ R
such that f r в€€ I for some positive integer r. в€љ
A popular notation for the radical of I is I. If f r в€љ g s belong to I, then by the
and
в€€ I, and it follows that I is an ideal.
r+sв€’1
binomial theorem, (f + g)

8.3.3 Lemma
в€љ
I вЉ† IV (I).
If I is any ideal of k[X1 , . . . , Xn ], then
в€љ
Proof. If f в€€ I, then f r в€€ I for some positive integer r. But then f r vanishes on V (I),
hence so does f . Therefore f в€€ IV (I). в™Ј
в€љ
The Nullstellensatz states that IV (I) = I, and the hard part is to prove that
в€љ
IV (I) вЉ† I. The technique is known as the вЂњRabinowitsch trickвЂќ, and it is indeed
very clever. Assume that f в€€ IV (I). We introduce a new variable Y and work in
k[X1 , . . . , Xn , Y ]. If I is an ideal of k[X1 , . . . , Xn ], then by the Hilbert basis theorem,
I is п¬Ѓnitely generated, say by f1 , . . . , fm . Let I в€— be the ideal of k[X1 , . . . , Xn , Y ] gener-
ated by f1 , . . . , fm , 1 в€’ Y f . [There is a slight ambiguity: by fi (X1 , . . . , Xn , Y ) we mean
fi (X1 , . . . , Xn ), and similarly for f .] At an appropriate moment we will essentially set Y
equal to 1/f and come back to the original problem.
8 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

8.3.4 Lemma
If (a1 , . . . , an , an+1 ) is any point in An+1 and (a1 , . . . , an ) в€€ V (I) (in other words, the fi ,
i = 1, . . . , m, vanish at (a1 , . . . , an )), then (a1 , . . . , an , an+1 ) в€€ V (I в€— ).
/
Proof. We are assuming that f в€€ IV (I), so that f vanishes on the zero-set of {f1 , . . . , fm }.
In particular, f (a1 , . . . , an ) = 0. The value of 1 в€’ Y f at (a1 , . . . , an , an+1 ) is therefore
1 в€’ an+1 f (a1 , . . . , an ) = 1 в€’ an+1 (0) = 1 = 0. But 1 в€’ Y f в€€ I в€— , so (a1 , . . . , an , an+1 ) does
not belong to the zero-set of I в€— . в™Ј

8.3.5 Lemma
If (a1 , . . . , an , an+1 ) is any point in An+1 and (a1 , . . . , an ) в€€ V (I), then (a1 , . . . , an , an+1 ) в€€
/ /
в€— в€—
V (I ). Consequently, by (8.3.4), V (I ) = в€….
Proof. By hypothesis, fi (a1 , . . . , an , an+1 ) = 0 for some i, and since fi в€€ I в€— , (a1 , . . . , an+1 )
cannot belong to the zero-set of I в€— . в™Ј
At this point we are going to assume what is called the weak Nullstellensatz, namely
that if I is a proper ideal of k[X1 , . . . , Xn ], then V (I) is not empty.

8.3.6 Lemma
There are polynomials g1 , . . . , gm , h в€€ k[X1 , . . . , Xn , Y ] such that
m
gi fi + h(1 в€’ Y f ).
1= (1)
i=1

This equation also holds in the rational function п¬Ѓeld k(X1 , . . . , Xn , Y ) consisting of quo-
tients of polynomials in k[X1 , . . . , Xn , Y ].
Proof. By (8.3.4) and (8.3.5), V (I в€— ) = в€…, so by the weak Nullstellensatz, I в€— = k[X1 , . . . ,
Xn , Y ]. In particular, 1 в€€ I в€— , and since I в€— is generated by f1 , . . . , fm , 1 в€’ Y f , there is an
equation of the speciп¬Ѓed form. The equation holds in the rational function п¬Ѓeld because
a polynomial is a rational function. в™Ј

8.3.7 The Rabinowitsch Trick
The idea is to set Y = 1/f , so that (1) becomes
m
1= gi (X1 , . . . , Xn , 1/f (X1 , . . . , Xn ))fi (X1 , . . . , Xn ). (2)
i=1
в€љ
Is this legal? First of all, if f is the zero polynomial, then certainly f в€€ I, so we
can assume f = 0. To justify replacing Y by 1/f , consider the ring homomorphism
from k[X1 , . . . , Xn , Y ] to k(X1 , . . . , Xn ) determined by Xi в†’ Xi , i = 1, . . . , n, Y в†’
1/f (X1 , . . . , Xn ). Applying this mapping to (1), we get (2). Now the right side of (2) is a
sum of rational functions whose denominators are various powers of f . If f r is the highest
8.3. THE NULLSTELLENSATZ: PRELIMINARIES 9

power that appears, we can absorb all denominators by multiplying (2) by f r . The result
is an equation of the form
m
r
f= hi (X1 , . . . , Xn )fi (X1 , . . . , Xn )
i=1

where the hi are polynomials in k[X1 , . . . , Xn ]. Consequently, f r в€€ I. в™Ј
The п¬Ѓnal ingredient is a major result in its own right.

8.3.8 Noether Normalization Lemma
Let A be a п¬Ѓnitely generated k-algebra, where k is a п¬Ѓeld. In other words, there are
п¬Ѓnitely many elements x1 , . . . , xn in A that generate A over k in the sense that every
element of A is a polynomial in the xi . Equivalently, A is a homomorphic image of the
polynomial ring k[X1 , . . . , Xn ] via the map determined by Xi в†’ xi , i = 1, . . . , n.
There exists a subset {y1 , . . . , yr } of A such that the yi are algebraically independent
over k and A is integral over k[y1 , . . . , yr ].
Proof. Let {x1 , . . . , xr } be a maximal algebraically independent subset of {x1 , . . . , xn }.
If n = r we are п¬Ѓnished, since we can take yi = xi for all i. Thus assume n > r, in which
case x1 , . . . , xn are algebraically dependent over k. Thus there is a nonzero polynomial
f в€€ k[X1 , . . . , Xn ] such that f (x1 , . . . , xn ) = 0. We can assume n > 1, for if n = 1 and
r = 0, then A = k[x1 ] and we can take {y1 , . . . , yr } to be the empty set.
We п¬Ѓrst assume that k is inп¬Ѓnite and give a proof by induction on n. (It is possible to
go directly to the general case, but the argument is not as intricate for an inп¬Ѓnite п¬Ѓeld.)
Decompose f into its homogeneous components (sums of monomials of the same degree).
Say that g is the homogeneous component of maximum degree d. Then, regarding g as a
polynomial in Xn whose coeп¬ѓcients are polynomials in the other Xi , we have, relabeling
variables if necessary, g(X1 , . . . , Xnв€’1 , 1) = 0. Since k is inп¬Ѓnite, it follows from (8.1.3)
part (8) that there are elements a1 , . . . , anв€’1 в€€ k such that g(a1 , . . . , anв€’1 , 1) = 0. Set
zi = xi в€’ ai xn , i = 1, . . . , n в€’ 1, and plug into f (x1 , . . . , xn ) = 0 to get an equation of the
form

g(a1 , . . . , anв€’1 , 1)xd + terms of degree less than d in xn = 0.
n

A concrete example may clarify the idea. If f (x1 , x2 ) = g(x1 , x2 ) = x2 x3 and x1 =
12
z1 + a1 x2 , then the substitution yields

(z1 + 2a1 z1 x2 + a2 x2 )x3
2
12 2

which indeed is g(a1 , 1)x5 plus terms of degree less than 5 in x2 . Divide by
2
g(a1 , . . . , anв€’1 , 1) = 0 to conclude that xn is integral over B = k[z1 , . . . , znв€’1 ]. By
the induction hypothesis, there are elements y1 , . . . , yr algebraically independent over
k such that B is integral over k[y1 , . . . , yr ]. But the xi , i < n, are integral over B since
xi = zi + ai xn . By transitivity (see (7.1.4)), x1 , . . . , xn are integral over k[y1 , . . . , yr ].
Thus (see (7.1.5)) A is integral over k[y1 , . . . , yr ].
Now we consider arbitrary k. As before, we produce a nonzero polynomial f such
that f (x1 , . . . , xn ) = 0. We assign a weight wi = snв€’i to the variable Xi , where s is a
10 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

large positive integer. (It suп¬ѓces to take s greater than the total degree of f , that is,
the sum of the degrees of all monomials in f .) If h = О»X1 1 В· В· В· Xnn is a monomial of f ,
a a
n
we deп¬Ѓne the weight of h as w(h) = i=1 ai wi . The point is that if h = ВµX1 1 В· В· В· Xnn ,
b b

then w(h) > w(h ) iп¬Ђ h > h in the lexicographic ordering, that is, for some m we have
ai = bi for i в‰¤ m, and am+1 > bm+1 . We take h to be the monomial of maximum weight.
(If two monomials diп¬Ђer in the lexicographic ordering, they must have diп¬Ђerent weights.)
Set zi = xi в€’ xwi , 1 в‰¤ i в‰¤ n в€’ 1, and plug into f (x1 , . . . , xn ) = 0 to get
n

cxw(h) + terms of lower degree in xn = 0.
n

For example, if f (x1 , x2 ) = h(x1 , x2 ) = x3 x2 , then x1 = z1 + xw1 gives
12 2

(z1 + 3z1 xw1 + 3z1 x2w1 + x3w1 )x2
3 2
2
2 2 2

and w(h) = 3w1 + 2w2 = 3w1 + 2 since snв€’2 = s0 = 1. Thus xn is integral over
B = k[z1 , . . . , znв€’1 ], and an induction argument п¬Ѓnishes the proof as in the п¬Ѓrst case. в™Ј

8.3.9 Corollary
Let B be a п¬Ѓnitely generated k-algebra, where k is a п¬Ѓeld. If I is a maximal ideal of B,
then B/I is a п¬Ѓnite extension of k.

Proof. The п¬Ѓeld k can be embedded in B/I via c в†’ c + I, c в€€ k. [If c в€€ I,
c = 0, then cв€’1 c = 1 в€€ I, a contradiction.] Since A = B/I is also a п¬Ѓnitely gener-
ated k-algebra, it follows from (8.3.8) that there is a subset {y1 , . . . , yr } of A with the
yi algebraically independent over k and A integral over k[y1 , . . . , yr ]. Now A is a п¬Ѓeld
(because I is a maximal ideal), and therefore so is k[y1 , . . . , yr ] (see the Problems in Sec-
tion 7.1). But this will lead to a contradiction if r в‰Ґ 1, because 1/y1 в€€ k[y1 , . . . , yr ]. /
(If 1/y1 = g(y1 , . . . , yr ) в€€ k[y1 , . . . , yr ], then y1 g(y1 , . . . , yr ) = 1, contradicting algebraic
independence.) Thus r must be 0, so A is integral, hence algebraic, over the п¬Ѓeld k.
Therefore A is generated over k by п¬Ѓnitely many algebraic elements, so by (3.3.3), A is a
п¬Ѓnite extension of k. в™Ј

8.3.10 Corollary
Let A be a п¬Ѓnitely generated k-algebra, where k is a п¬Ѓeld. If A is itself a п¬Ѓeld, then A is
a п¬Ѓnite extension of k.

Proof. As in (8.3.9), with B/I replaced by A. в™Ј

Problems For Section 8.3
1. Let S be a multiplicative subset of the ring R (see (2.8.1)). If I is an ideal that is
disjoint from S, then by ZornвЂ™s lemma, there is an ideal J that is maximal among
ideals disjoint from S. Show that J must be prime.
8.4. THE NULLSTELLENSATZ: EQUIVALENT VERSIONS AND PROOF 11

2. Show that the radical of the ideal I is the intersection of all prime ideals containing I.
в€љ
[If f r в€€ I вЉ† P , P prime, then f в€€ P . Conversely, assume f в€€ I. With a clever
/
choice of multiplicative set S, show that for some prime ideal P containing I, we have
f в€€ P .]
/
3. An algebraic curve is a variety deп¬Ѓned by a nonconstant polynomial in two variables.
Show (using the Nullstellensatz) that the polynomials f and g deп¬Ѓne the same algebraic
curve iп¬Ђ f divides some power of g and g divides some power of f . Equivalently, f and
g have the same irreducible factors.
4. Show that the variety V deп¬Ѓned over the complex numbers by the two polynomials
Y 2 в€’ XZ and Z 2 в€’ X 2 Y is the union of the line L given by Y = Z = 0, X arbitrary,
and the set W of all (t3 , t4 , t5 ), t в€€ C.
5. The twisted cubic is the variety V deп¬Ѓned over the complex numbers by Y в€’ X 2 and
Z в€’ X 3 . In parametric form, V = {(t, t2 , t3 ) : t в€€ C}. Show that V is irreducible. [The
same argument works for any variety that can be parametrized over an inп¬Ѓnite п¬Ѓeld.]
6. Find parametrizations of the following algebraic curves over the complex numbers. (It
is permissible for your parametrizations to fail to cover п¬Ѓnitely many points of the
curve.)
(a) The unit circle x2 + y 2 = 1;
(b) The cuspidal cubic y 2 = x3 ;
(c) The nodal cubic y 2 = x2 + x3 .
7. Let f be an irreducible polynomial, and g an arbitrary polynomial, in k[x, y]. If f does
not divide g, show that the system of simultaneous equations f (x, y) = g(x, y) = 0 has
only п¬Ѓnitely many solutions.

8.4 The Nullstellensatz: Equivalent Versions
And Proof
We are now in position to establish the equivalence of several versions of the Nullstellen-
satz.

8.4.1 Theorem
For any п¬Ѓeld k and any positive integer n, the following statements are equivalent.
(1) Maximal Ideal Theorem The maximal ideals of k[X1 , . . . , Xn ] are the ideals of
the form (X1 в€’ a1 , . . . , Xn в€’ an ), a1 , . . . , an в€€ k. Thus maximal ideals correspond to
points.
(2) Weak Nullstellensatz If I is an ideal of k[X1 , . . . , Xn ] and V (I) = в€…, then I =
k[X1 , . . . , Xn ]. Equivalently, if I is a proper ideal, then V (I) is not empty.
(3) Nullstellensatz If I is an ideal of k[X1 , . . . , Xn ], then
в€љ
IV (I) = I.
12 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

(4) k is algebraically closed.

Proof. (1) implies (2). Let I be a proper ideal, and let J be a maximal ideal containing I.
By (8.1.3), part (4), V (J) вЉ† V (I), so it suп¬ѓces to show that V (J) is not empty. By (1),
J has the form (X1 в€’ a1 , . . . , Xn в€’ an ). But then a = (a1 , . . . , an ) в€€ V (J). [In fact
V (J) = {a}.]
(2) implies (3). This was done in Section 8.3.
(3) implies (2). We use the fact that the radical of an ideal I is the intersection
of all prime ideals containing I; see Section 8.3, Problem 2. Let I be a proper ideal of
k[X1 , . . .в€љXn ]. Then I is contained in a maximal, hence prime, ideal P . By the result just
, в€љ
quoted, I is also contained in P , hence I is a proper ideal. By (3), IV (I) is a proper
ideal. But if V (I) = в€…, then by (8.1.3) part (7), IV (I) = k[X1 , . . . , Xn ], a contradiction.
(2) implies (1). If I is a maximal ideal, then by (2) there is a point a = (a1 , . . . , an ) в€€
V (I). Thus every f в€€ I vanishes at a, in other words, I вЉ† I({a}). But (X1 в€’ a1 , . . . ,
Xn в€’ an ) = I({a}); to see this, decompose f в€€ I({a}) as in the proof of (8.3.1). Therefore
the maximal ideal I is contained in the maximal ideal (X1 в€’ a1 , . . . , Xn в€’ an ), and it
follows that I = (X1 в€’ a1 , . . . , Xn в€’ an ).
(4) implies (1). Let I be a maximal ideal of k[X1 , . . . , Xn ], and let K = k[X1 , . . . ,
Xn ]/I, a п¬Ѓeld containing an isomorphic copy of k via c в†’ c + I, c в€€ k. By (8.3.9), K
is a п¬Ѓnite extension of k, so by (4), K = k. But then Xi + I = ai + I for some ai в€€ k,
i = 1, . . . , n. Therefore Xi в€’ ai is zero in k[X1 , . . . , Xn ]/I, in other words, Xi в€’ ai в€€ I.
Consequently, I вЉ‡ (X1 в€’ a1 , . . . , Xn в€’ an ), and we must have equality by (8.3.1).
(1) implies (4). Let f be a nonconstant polynomial in k[X1 ] with no root in k. We
can regard f is a polynomial in n variables with no root in An . Let I be a maximal ideal
containing the proper ideal (f ). By (1), I is of the form (X1 в€’ a1 , . . . , Xn в€’ an ) = I({a})
for some a = (a1 , . . . , an ) в€€ An . Therefore f vanishes at a, a contradiction. в™Ј

8.4.2 Corollary
If the ideals I and J deп¬Ѓne the same variety and a polynomial g belongs to one of the
ideals, then some power of g belongs to the other ideal.

в€љ в€љ в€љ
J. If g в€€ I вЉ† I, then g r в€€ J
Proof. If V (I) = V (J), then by the Nullstellensatz, I=
for some positive integer r. в™Ј

8.4.3 Corollary
The maps V в†’ I(V ) and I в†’ V (I) set up a one-to-one correspondence between varieties
в€љ
and radical ideals (deп¬Ѓned by I = I).

в€љ
Proof. By (8.1.3) part 6, V I(V ) = V . By the Nullstellensatz, IV (I) = I = I for radical
ideals. It remains to prove that for any variety V , I(V ) is a radical ideal. If f r в€€ I(V ),
then f r , hence f , vanishes on V , so f в€€ I(V ). в™Ј
8.5. LOCALIZATION 13

8.4.4 Corollary
Let f1 , . . . , fr , g в€€ k[X1 , . . . , Xn ], and assume that g vanishes wherever the fi all vanish.
Then there are polynomials h1 , . . . , hr в€€ k[X1 , . . . , Xn ] and a positive integer s such that
g s = h1 f1 + В· В· В· + hr fr .
Proof. Let I be the ideal generated by f1 , . . . , fr . Then V (I) is the set of points at which
all fi vanish, so that IV (I) is the set of polynomials that vanish wherever all fi vanish.
в€љ
Thus g belongs to IV (I), which is I by the Nullstellensatz. Consequently, for some
positive integer s, we have g s в€€ I, and the result follows. в™Ј

Problems For Section 8.4
1. Let f be a polynomial in k[X1 , . . . , Xn ], and assume that the factorization of f into
irreducibles is f = f1 1 В· В· В· fr r . Show that the decomposition of the variety V (f ) into
n n

irreducible subvarieties (Section 8.1, Problems 5 and 6) is given by V (f ) = в€Єr V (fi ).
i=1
2. Under the hypothesis of Problem 1, show that IV (f ) = (f1 В· В· В· fr ).
3. Show that there is a one-to-one correspondence between irreducible polynomials in
k[X1 , . . . , Xn ] and irreducible hypersurfaces (see (8.2.2))in An (k), if polynomials that
diп¬Ђer by a nonzero multiplicative constant are identiп¬Ѓed.
4. For any collection of subsets Xi of An , show that I(в€Єi Xi ) = в€©i I(Xi ).
5. Show that every radical ideal I of k[X1 , . . . , Xn ] is the intersection of п¬Ѓnitely many
prime ideals.
6. In Problem 5, show that the decomposition is unique, subject to the condition that
the prime ideals P are minimal, that is, there is no prime ideal Q with I вЉ† Q вЉ‚ P .
7. Suppose that X is a variety in A2 , deп¬Ѓned by equations f1 (x, y) = В· В· В· = fm (x, y) = 0,
m в‰Ґ 2. Let g be the greatest common divisor of the fi . If g is constant, show that X
is a п¬Ѓnite set (possibly empty).
8. Show that every variety in A2 except for A2 itself is the union of a п¬Ѓnite set and an
algebraic curve. 9. Give an example of two distinct irreducible polynomials in k[X, Y ]
with the same zero-set, and explain why this cannot happen if k is algebraically closed.
10. Give an explicit example of the failure of a version of the Nullstellensatz in a
non-algebraically closed п¬Ѓeld.

8.5 Localization
8.5.1 Geometric Motivation
Suppose that V is an irreducible variety, so that I(V ) is a prime ideal. A polynomial g
will belong to I(V ) if and only if it vanishes on V . If we are studying rational functions
f /g in the neighborhood of a point x в€€ V , we must have g(x) = 0. It is very convenient
to have every polynomial g в€€ I(V ) available as a legal object, even though g may vanish
/
at some points of V . The technical device that makes this possible is the construction of
the ring of fractions S в€’1 R, the localization of R by S, where R = k[X1 , . . . , Xn ] and S is
the multiplicative set R \ I(V ). We will now study the localization process in general.
14 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

8.5.2 Notation
Recalling the setup of Section 2.8, let S be a multiplicative subset of the ring R, and
S в€’1 R the ring of fractions of R by S. Let h be the natural homomorphism of R into
S в€’1 R, given by h(a) = a/1. If X is any subset of R, deп¬Ѓne S в€’1 X = {x/s : x в€€ X, s в€€ S}.
We will be especially interested in such a set when X is an ideal.
If I and J are ideals of R, the product of I and J, denoted by IJ, is deп¬Ѓned (as in
(7.6.1)) as the set of all п¬Ѓnite sums i xi yi , xi в€€ I, yi в€€ J. It follows from the deп¬Ѓnition
that IJ is an ideal. The sum of two ideals has already been deп¬Ѓned in (2.2.8).

8.5.3 Lemma
If I is an ideal of R, then S в€’1 I is an ideal of S в€’1 R. If J is another ideal of R, the

(i) S в€’1 (I + J) = S в€’1 I + S в€’1 J;
(ii) S в€’1 (IJ) = (S в€’1 I)(S в€’1 J);
(iii) S в€’1 (I в€© J) = S в€’1 I в€© S в€’1 J;
(iv) S в€’1 I is a proper ideal iп¬Ђ S в€© I = в€….

Proof. The deп¬Ѓnition of addition and multiplication in S в€’1 R implies that S в€’1 I is an
ideal, and that in (i), (ii) and (iii), the left side is contained in the right side. The reverse
inclusions in (i) and (ii) follow from

ab at + bs ab ab
+= , =.
s t st st st

To prove (iii), let a/s = b/t where a в€€ I, b в€€ J, s, t в€€ S. There exists u в€€ S such that
u(at в€’ bs) = 0. Then a/s = uat/ust = ubs/ust в€€ S в€’1 (I в€© J).
Finally, if s в€€ S в€© I then 1/1 = s/s в€€ S в€’1 I, so S в€’1 I = S в€’1 R. Conversely, if
S в€’1 I = S в€’1 R, then 1/1 = a/s for some a в€€ I, s в€€ S. There exists t в€€ S such that
t(s в€’ a) = 0, so at = st в€€ S в€© I. в™Ј

Ideals in S в€’1 R must be of a special form.

8.5.4 Lemma
If J is an ideal of S в€’1 R and I = hв€’1 (J), then I is an ideal of R and S в€’1 I = J.

Proof. I is an ideal by the basic properties of preimages of sets. Let a/s в€€ S в€’1 I, with
a в€€ I and s в€€ S. Then a/1 в€€ J, so a/s = (a/1)(1/s) в€€ J. Conversely, let a/s в€€ J, with
a в€€ R, s в€€ S. Then h(a) = a/1 = (a/s)(s/1) в€€ J, so a в€€ I and a/s в€€ S в€’1 I. в™Ј

Prime ideals yield sharper results.
8.5. LOCALIZATION 15

8.5.5 Lemma
If I is any ideal of R, then I вЉ† hв€’1 (S в€’1 I), with equality if I is prime and disjoint from
S.

Proof. If a в€€ I, then h(a) = a/1 в€€ S в€’1 I. Thus assume that I is prime and disjoint from
S, and let a в€€ hв€’1 (S в€’1 I). Then h(a) = a/1 в€€ S в€’1 I, so a/1 = b/s for some b в€€ I, s в€€ S.
There exists t в€€ S such that t(as в€’ b) = 0. Thus ast = bt в€€ I, with st в€€ I since S в€© I = в€….
/
Since I is prime, we have a в€€ I. в™Ј

8.5.6 Lemma
If I is a prime ideal of R disjoint from S, then S в€’1 I is a prime ideal of S в€’1 R.

Proof. By (8.5.3), part (iv), S в€’1 I is a proper ideal. Let (a/s)(b/t) = ab/st в€€ S в€’1 I, with
a, b в€€ R, s, t в€€ S. Then ab/st = c/u for some c в€€ I, u в€€ S. There exists v в€€ S such that
v(abu в€’ cst) = 0. Thus abuv = cstv в€€ I, and uv в€€ I because S в€© I = в€…. Since I is prime,
/
ab в€€ I, hence a в€€ I or b в€€ I. Therefore either a/s or b/t belongs to S в€’1 I. в™Ј

The sequence of lemmas can be assembled to give a precise conclusion.

8.5.7 Theorem
There is a one-to-one correspondence between prime ideals P of R that are disjoint from
S and prime ideals Q of S в€’1 R, given by

P в†’ S в€’1 P and Q в†’ hв€’1 (Q).

Proof. By (8.5.4), S в€’1 (hв€’1 (Q)) = Q, and by (8.5.5), hв€’1 (S в€’1 P ) = P . By (8.5.6), S в€’1 P
is a prime ideal, and hв€’1 (Q) is a prime ideal by the basic properties of preimages of sets.
If hв€’1 (Q) meets S, then by (8.5.3) part (iv), Q = S в€’1 (hв€’1 (Q)) = S в€’1 R, a contradiction.
Thus the maps P в†’ S в€’1 P and Q в†’ hв€’1 (Q) are inverses of each other, and the result
follows. в™Ј

8.5.8 Deп¬Ѓnitions and Comments
If P is a prime ideal of R, then S = R \ P is a multiplicative set. In this case, we write
R(P ) for S в€’1 R, and call it the localization of R at P . (The usual notation is RP , but itвЂ™s
easier to read without subscripts.) If I is an ideal of R, we write I(P ) for S в€’1 I. We are
going to show that R(P ) is a local ring, that is, a ring with a unique maximal ideal. First
we give some conditions equivalent to the deп¬Ѓnition of a local ring.

8.5.9 Proposition
For a ring R, the following conditions are equivalent.

(i) R is a local ring;
(ii) There is a proper ideal I of R that contains all nonunits of R;
16 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

(iii) The set of nonunits of R is an ideal.

Proof. (i) implies (ii). If a is a nonunit, then (a) is a proper ideal, hence is contained in
the unique maximal ideal I.
(ii) implies (iii). If a and b are nonunits, so are a + b and ra. If not, then I contains
a unit, so I = R, a contradiction.
(iii) implies (i). If I is the ideal of nonunits, then I is maximal, because any larger
ideal J would have to contain a unit, so that J = R. If H is any proper ideal, then H
cannot contain a unit, so H вЉ† I. Therefore I is the unique maximal ideal. в™Ј

8.5.10 Theorem
R(P ) is a local ring.

Proof. Let Q be a maximal ideal of R(P ). Then Q is prime, so by (8.5.7), Q = I(P ) for
some prime ideal I of R that is disjoint from S, in other words, contained in P . Thus
Q = I(P ) вЉ† P (P ). If P (P ) = R(P ), then by (8.5.3) part (iv), P is not disjoint from
S = R\P , which is impossible. Therefore P (P ) is a proper ideal containing every maximal
ideal, so it must be the unique maximal ideal. в™Ј

If R is an integral domain and S is the set of all nonzero elements of R, then S в€’1 R is
the quotient п¬Ѓeld of R. In this case, S в€’1 R is a local ring, because any п¬Ѓeld is a local ring.
({0} is the unique maximal ideal.) Alternatively, we can appeal to (8.5.10) with P = {0}.

8.5.11 Localization of Modules
If M is an R-module and S a multiplicative subset of R, we can essentially repeat the
construction of Section 2.8 to form the localization S в€’1 M of M by S, and thereby divide
elements of M by elements of S. If x, y в€€ M and s, t в€€ S, we call (x, s) and (y, t)
equivalent if for some u в€€ S, u(tx в€’ sy) = 0. The equivalence class of (x, s) is denoted by
x/s, and addition is deп¬Ѓned by

xy tx + sy
+= .
s t st

If a/s в€€ S в€’1 R and x/t в€€ S в€’1 M , we deп¬Ѓne
ax ax
= .
st st

In this way, S в€’1 M becomes an S в€’1 R-module. Exactly as in (8.5.3), if M and N are
submodules of a module L, then

S в€’1 (M + N ) = S в€’1 M + S в€’1 N and S в€’1 (M в€© N ) = S в€’1 M в€© S в€’1 N.

Further properties will be given in the exercises.
8.6. PRIMARY DECOMPOSITION 17

Problems For Section 8.5
1. Let M be a maximal ideal of R, and assume that for every x в€€ M , 1 + x is a unit.
Show that R is a local ring (with maximal ideal M ). [Show that if x в€€ M , then x is a
/
unit, and apply (8.5.9).]
2. Show that if p is prime and n is a positive integer, then Zpn is a local ring with maximal
ideal (p).
3. Let R be the ring of all n by n matrices with coeп¬ѓcients in a п¬Ѓeld F . If A is a nonzero
element of R and 1 is the identity matrix, is {1, A, A2 , . . . } always a multiplicative set?
Let S be a multiplicative subset of the ring R. We are going to construct a mapping
from R-modules to S в€’1 R-modules, and another mapping from R-module homomorphisms
to S в€’1 R-module homomorphisms, as follows. If M is an R-module, we let M в†’ S в€’1 M .
If f : M в†’ N is an R-module homomorphism, we deп¬Ѓne S в€’1 f : S в€’1 M в†’ S в€’1 N by
x f (x)
в†’ .
s s
Since f is a homomorphism, so is S в€’1 f .
4. If g : N в†’ L and composition of functions is written as a product, show that S в€’1 (gf ) =
S в€’1 (g)S в€’1 (f ), and if 1M is the identity mapping on M , then S в€’1 (1M ) = 1S в€’1 M . We
say that S в€’1 is a functor from the category of R-modules to the category of S в€’1 R-
modules. This terminology will be explained in great detail in Chapter 10.
5. If
f g
в†’ в†’
M N L
is an exact sequence, show that
S в€’1 f S в€’1 g
S в€’1 M S в€’1 N S в€’1 L
в†’ в†’

is exact. We say that S в€’1 is an exact functor. Again, we will study this idea in
Chapter 10.
6. Let R be the ring of rational functions f /g with f, g в€€ k[X1 , . . . , Xn ] and g(a) = 0,
where a = (a1 , . . . , an ) is a п¬Ѓxed point in An . Show that R is a local ring, and identify
the unique maximal ideal.
7. If M is an R-module and S is a multiplicative subset of R, denote S в€’1 M by MS . If
N is a submodule of M , show that (M/N )S в€ј MS /NS .
=

8.6 Primary Decomposition
We have seen that every radical ideal in k[X1 , . . . , Xn ] can be expressed as an intersection
of п¬Ѓnitely many prime ideals (Section 8.4, Problem 5). A natural question is whether a
similar result holds for arbitrary ideals. The answer is yes if we generalize from prime to
primary ideals.
18 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

8.6.1 Deп¬Ѓnitions and Comments
The ideal Q in the ring R is primary if Q is proper and whenever a product ab belongs to
Q, either a в€€ Q or bn в€€ Q for some positive integer n. [The condition on b is equivalent
в€љ
to b в€€ Q.] An equivalent statement is that R/Q = 0 and whenever (a + Q)(b + Q) = 0
in R/Q, either a + Q = 0 or (b + Q)n = 0 for some positive integer n. This says that if
b + Q is a zero-divisor in R/Q, then it is nilpotent, that is, some power of b + Q is 0.
It в€љ
follows from the deп¬Ѓnition that every prime ideal is primary. Also, if Q is primary,
в€љ
then Q is the smallest prime ideal containing Q. [Since Q is the intersection of all
в€љ
prime ideals containing Q (Section 8.3, Problem 2), it suп¬ѓces to show that Q is prime.
в€љ if a b в€€ Q, then a в€€ Q or b в€€ Q for some m, so either a or b must belong to
nn n nm
But
Q. Note also that since Q is proper, it is contained in a maximal, hence prime, ideal,
в€љ
so Q is also proper.] в€љ
If Q is primary and Q = P , we say that Q is P -primary.

8.6.2 Examples
1. In Z, the primary ideals are {0} and (pr ), where p is prime. In Z6 , 2 and 3 are
zero-divisors that are not nilpotent, and a similar situation will occur in Zm whenever
more than one prime appears in the factorization of m.
2. Let R = k[X, Y ] where k is any п¬Ѓeld, and take Q = (X, Y 3 ), the ideal generated by
X and Y 3 . This is a nice example of analysis in quotient rings. We are essentially setting
X and Y 3 equal to zero, and this collapses the ring R down to polynomials a0 +a1 Y +a2 Y 2 ,
with the ai в€€ k and arithmetic mod Y 3 . Formally, R/Q is isomorphic to k[Y ]/(Y 3 ). The
zero-divisors in R/Q are of the form cY + dY 2 , c в€€ k, and they are nilpotent. Thus Q is
primary. If f в€€ R, then the only way for f not to belong to the radical of Q is for the
в€љ
constant term of f to be nonzero. Thus Q = (X, Y ), a maximal ideal by (8.3.1).
Now we claim that Q cannot be a power of a prime ideal; this will be a consequence
of the next result.

8.6.3 Lemma
в€љ
If P is a prime ideal, then for every positive integer n, P n = P .
в€љ
Proof. Since P is a prime ideal containing P n , P n вЉ† P . If x в€€ P , then xn в€€ P n , so
в€љ
x в€€ P n. в™Ј

Returning to Example 2 of (8.6.2), if Q = (X, Y 3 ) is a prime power P n , then its radical
is P , so P must be (X, Y ). But X в€€ Q and X в€€ P n , n в‰Ґ 2; since Y belongs to P but
/
not Q, we have reached a contradiction.
After a preliminary deп¬Ѓnition, we will give a convenient suп¬ѓcient condition for an
ideal to be primary.

8.6.4 Deп¬Ѓnition
The nilradical N (R) of a ring R is the set of nilpotent elements of R, that is, {x в€€ R :
xn = 0 for some positive integer n}. Thus N (R) is the radical of the zero ideal, which is
8.6. PRIMARY DECOMPOSITION 19

the intersection of all prime ideals of R.

8.6.5 Proposition
If the radical of the ideal Q is maximal, then Q is primary.
в€љ
Proof. Since Q is maximal, it must be the only prime ideal containing Q. By the
correspondence theorem and the fact that the preimage of a prime ideal is a prime ideal
(cf. (8.5.7)), R/Q has exactly one prime ideal, which must coincide with N (R/Q). Any
element of R/Q that is not a unit generates a proper ideal, which is contained in a
maximal ideal, which again must be N (R/Q). Thus every element of R/Q is either a unit
or nilpotent. Since a zero-divisor cannot be a unit, every zero-divisor of R/Q is nilpotent,
so Q is primary. в™Ј

8.6.6 Corollary
If M is a maximal ideal, then M n is M -primary for all n = 1, 2, . . . .

Proof. By (8.6.3), the radical of M n is M , and the result follows from (8.6.5). в™Ј

Here is another useful property.

8.6.7 Proposition
If Q is a п¬Ѓnite intersection of P -primary ideals Qi , i = 1, . . . , n, then Q is P -primary.

Proof. First note that the radical of a п¬Ѓnite intersection of ideals is the intersection of the
radicals (see Problem 1). It follows that the radical of Q is P , and it remains to show
that Q is primary. If ab в€€ Q but a в€€ Q, then for some i we have a в€€ Qi . Since Qi is
/ /
в€љ
P -primary, b belongs to P = Qi . But then some power of b belongs to Q. в™Ј

We are going to show that in a Noetherian ring, every proper ideal I has a primary
decomposition, that is, I can be expressed as a п¬Ѓnite intersection of primary ideals.

8.6.8 Lemma
Call an ideal I irreducible if for any ideals J and K, I = J в€© K implies that I = J or
I = K. If R is Noetherian, then every ideal of R is a п¬Ѓnite intersection of irreducible
ideals.

Proof. Suppose that the collection S of all ideals that cannot be so expressed is nonempty.
Since R is Noetherian, S has a maximal element I, necessarily reducible. Let I = J в€© K,
where I is properly contained in both J and K. By maximality of I, the ideals J and K are
п¬Ѓnite intersections of irreducible ideals, and consequently so is I, contradicting I в€€ S. в™Ј
20 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

If we can show that every irreducible proper ideal is primary, we then have the desired
primary decomposition. Let us focus on the chain of reasoning we will follow. If I is
an irreducible proper ideal of R, then by the correspondence theorem, 0 is an irreducible
ideal of the Noetherian ring R/I. If we can show that 0 is primary in R/I, then again by
the correspondence theorem, I is primary in R.

8.6.9 Primary Decomposition Theorem
Every proper ideal in a Noetherian ring R has a primary decomposition. (We can drop
the word вЂњproperвЂќ if we regard R as the intersection of the empty collection of primary
ideals.)

Proof. By the above discussion, it suп¬ѓces to show that if 0 is an irreducible ideal of R,
then it is primary. Let ab = 0 with a = 0. Since R is Noetherian, the sequence of
annihilators
ann b вЉ† ann b2 вЉ† ann b3 вЉ† В· В· В·
stabilizes, so ann bn = ann bn+1 for some n. If we can show that

(a) в€© (bn ) = 0

we are п¬Ѓnished, because a = 0 and the zero ideal is irreducible (by hypothesis). Thus let
x = ca = dbn for some c, d в€€ R. Then bx = cab = dbn+1 = 0 (because ab = 0), so d
annihilates bn+1 , hence d annihilates bn . Thus x = dbn = 0. в™Ј

Problems For Section 8.6
1. If I1 , . . . , In are arbitrary ideals, show that

в€љ
n n
Ii = I i.
i=1 i=1

2. Let I be the ideal (XY в€’Z 2 ) in k[X, Y, Z], where k is any п¬Ѓeld, and let R = k[X, Y, Z]/I.
If P is the ideal (X + I, Z + I), show that P is prime.
3. Continuing Problem 2, show that P 2 , whose radical is prime by (8.6.3) and which is a
power of a prime, is nevertheless not primary.
4. Let R = k[X, Y ], and let P1 = (X), P2 = (X, Y ), Q = (X 2 , Y ). Show that P1 is prime
2
and P2 and Q are P2 -primary.
5. Continuing Problem 4, let I = (X 2 , XY ). Show that P1 в€© P2 and P1 в€© Q are both
2

primary decompositions of I.

Notice that the radicals of the components of the primary decomposition (referred to
в€љ
as the в€љ
primes associated with I) are P1 and P2 in both cases. [P1 is prime, so P 1 = P1 ;
в€љ
P2 вЉ† Q and P2 is maximal, so P2 = Q;] Uniqueness questions involving primary
decompositions are treated in detail in textbooks on commutative algebra.
8.7. TENSOR PRODUCT OF MODULES OVER A COMMUTATIVE RING 21

6. We have seen in Problem 5 of Section 8.4 that every radical ideal in R = k[X1 , . . . , Xn ]
is the intersection of п¬Ѓnitely many prime ideals. Show that this result holds in an
arbitrary Noetherian ring R.
7. Let R = k[X, Y ] and let In be the ideal (X 3 , XY, Y n ). Show that for every positive
integer n, In is a primary ideal of R.

8.7 Tensor Product of Modules Over a Commutative
Ring

8.7.1 Motivation
In many areas of algebra and its applications, it is useful to multiply, in a sensible way, an
element x of an R-module M by an element y of an R-module N . In group representation
theory, M and N are free modules, in fact п¬Ѓnite-dimensional vector spaces, with bases
{xi } and {yj }. Thus if we specify that multiplication is linear in each variable, then we
need only specify products of xi and yj . We require that the these products, to be denoted
by xi вЉ— yj , form a basis for a new R-module T .
If f : R в†’ S is a ring homomorphism and M is an S-module, then M becomes an
R-module via rx = f (r)x, r в€€ R, x в€€ M . This is known as restriction of scalars.
In algebraic topology and algebraic number theory, it is often desirable to reverse this
process. If M is an R-module, we want to extend the given multiplication rx, r в€€ R,
x в€€ M , to multiplication of an arbitrary s в€€ S by x в€€ M . This is known as extension of
scalars, and it becomes possible with the aid of the tensor product construction.
The tensor product arises in algebraic geometry in the following way. Let M be the
coordinate ring of a variety V in aп¬ѓne space Am , in other words, M is the set of all
polynomial functions from V to the base п¬Ѓeld k. Let N be the coordinate ring of the
variety W in An . Then the cartesian product V Г— W is a variety in Am+n , and its
coordinate ring turns out to be the tensor product of M and N .
LetвЂ™s return to the п¬Ѓrst example above, where M and N are free modules with bases
{xi } and {yj }. Suppose that f is a bilinear map from M Г—N to an R-module P . (In other
words, f is R-linear in each variable.) Information about f can be completely encoded
into a function g of one variable, where g is an R-module homomorphism from T to P .
We take g(xi вЉ— yj ) = f (xi , yj ) and extend by linearity. Thus f is the composition of
the bilinear map h from M Г— N to T speciп¬Ѓed by (xi , yj ) в†’ xi вЉ— yj , followed by g. To
summarize:

Every bilinear mapping on M Г— N can be factored through T .

The R-module T is called the tensor product of M and N , and we write T = M вЉ—R N .
We are going to construct a tensor product of arbitrary modules over a commutative ring,
and sketch the generalization to noncommutative rings.
22 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

8.7.2 Deп¬Ѓnitions and Comments
Let M and N be arbitrary R-modules, and let F be a free R-module with basis M Г— N .
Let G be the submodule of F generated by the вЂњrelationsвЂќ

(x + x , y) в€’ (x, y) в€’ (x , y); (x, y + y ) в€’ (x, y) в€’ (x, y );
(rx, y) в€’ r(x, y); (x, ry) в€’ r(x, y)

where x, x в€€ M , y, y в€€ N , r в€€ R. Deп¬Ѓne the tensor product of M and N (over R) as

T = M вЉ—R N = F/G

and denote the element (x, y) + G of T by x вЉ— y. Thus the general element of T is a п¬Ѓnite
sum of the form

xi вЉ— yi
t= (1)
i

with xi в€€ M and yi в€€ N . It is important to note that the representation (1) is not
necessarily unique.
The relations force x вЉ— y to be linear in each variable, so that

x вЉ— (y + y ) = x вЉ— y + x вЉ— y , (x + x ) вЉ— y = x вЉ— y + x вЉ— y, (2)
r(x вЉ— y) = rx вЉ— y = x вЉ— ry. (3)

See Problem 1 for details. Now if f is a bilinear mapping from M Г— N to the R-module
P , then f extends uniquely to a homomorphism from F to P , also called f . Bilinearity
means that the kernel of f contains G, so by the factor theorem, there is a unique R-
homomorphism g : T в†’ P such that g(x вЉ— y) = f (x, y) for all x в€€ M , y в€€ N . As in
(8.7.1), if we compose the bilinear map h : (x, y) в†’ x вЉ— y with g, the result is f . Again,
we say that

Every bilinear mapping on M Г— N can be factored through T .

We have emphasized this sentence, known as a universal mapping property (abbrevi-
ated UMP), because along with equations (1), (2) and (3), it indicates how the tensor
product is applied in practice. The detailed construction we have just gone through can
now be forgotten. In fact any two R-modules that satisfy the universal mapping property
are isomorphic. The precise statement and proof of this result will be developed in the
exercises.
In a similar fashion, using multilinear rather than bilinear maps, we can deп¬Ѓne the
tensor product of any п¬Ѓnite number of R-modules. [In physics and diп¬Ђerential geometry,
a tensor is a multilinear map on a product M1 Г— В· В· В· Г— Mr , where each Mi is either a п¬Ѓnite-
dimensional vector space V or its dual space V в€— . This suggests where the terminology
вЂњtensor productвЂќ comes from.]
In the discussion to follow, M , N and P are R-modules. The ring R is assumed п¬Ѓxed,
and we will usually write вЉ— rather than вЉ—R .
8.7. TENSOR PRODUCT OF MODULES OVER A COMMUTATIVE RING 23

8.7.3 Proposition
M вЉ— N в€ј N вЉ— M.
=
Proof. Deп¬Ѓne a bilinear mapping f : M Г— N в†’ N вЉ— M by f (x, y) = y вЉ— x. By the UMP,
there is a homomorphism g : M вЉ— N в†’ N вЉ— M such that g(x вЉ— y) = y вЉ— x. Similarly,
there is a homomorphism g : N вЉ— M в†’ M вЉ— N with g (y вЉ— x) = x вЉ— y. Thus g is an
isomorphism (with inverse g ). в™Ј

8.7.4 Proposition
M вЉ— (N вЉ— P ) в€ј (M вЉ— N ) вЉ— P .
=
Proof. Deп¬Ѓne f : M Г— N Г— P в†’ (M вЉ— N ) вЉ— P by f (x, y, z) = (x вЉ— y) вЉ— z. The UMP
produces g : M Г— (N вЉ— P ) в†’ (M вЉ— N ) вЉ— P with g((x, (y вЉ— z))) = (x вЉ— y) вЉ— z. [We are
applying the UMP for each п¬Ѓxed x в€€ M , and assembling the maps to produce g.] Since g
is bilinear (by Equations (2) and (3)), the UMP yields h : M вЉ— (N вЉ— P ) в†’ (M вЉ— N ) вЉ— P
with h(x вЉ— (y вЉ— z)) = (x вЉ— y) вЉ— z. Exactly as in (8.7.3), we can construct the inverse of
h, so h is the desired isomorphism. в™Ј

8.7.5 Proposition
M вЉ— (N вЉ• P ) в€ј (M вЉ— N ) вЉ• (M вЉ— P ).
=
Proof. Let f be an arbitrary bilinear mapping from M Г— (N вЉ• P ) to Q. If x в€€ M ,
y в€€ N , z в€€ P , then f (x, y + z) = f (x, y) + f (x, z). The UMP gives homomorphisms
g1 : M вЉ—N в†’ Q and g2 : M вЉ—P в†’ Q such that g1 (xвЉ—y) = f (x, y) and g2 (xвЉ—z) = f (x, z).
The maps g1 and g2 combine to give g : (M вЉ— N ) вЉ• (M вЉ— P ) в†’ Q such that

g((x вЉ— y) + (x вЉ— z)) = g1 (x вЉ— y) + g2 (x вЉ— z).

In particular, with x = x,

g((x вЉ— y) + (x вЉ— z)) = f (x, y + z),

so if h : M Г— (N вЉ• P ) в†’ M вЉ— (N вЉ• P ) is deп¬Ѓned by

h(x, y + z) = (x вЉ— y) + (x вЉ— z),

then f = gh. Thus (M вЉ— N ) вЉ• (M вЉ— P ) satisп¬Ѓes the universal mapping property, hence
must be isomorphic to the tensor product. в™Ј

8.7.6 Proposition
Regarding R as a module over itself, R вЉ—R M в€ј M .
=
Proof. The map (r, x) в†’ rx of R Г— M в†’ M is bilinear, so there is a homomorphism
g : R вЉ— M в†’ M such that g(r вЉ— x) = rx. Deп¬Ѓne h : M в†’ R вЉ— M by h(x) = 1 вЉ— x. Then
h(rx) = 1 вЉ— rx = r1 вЉ— x = r вЉ— x. Thus g is an isomorphism (with inverse h). в™Ј
24 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

8.7.7 Corollary
Let Rm be the direct sum of m copies of R, and M m the direct sum of m copies of M .
Then Rm вЉ— M в€ј M m .
=
Proof. By (8.7.5), Rm вЉ— M is isomorphic to the direct sum of m copies of R вЉ— M , which
is isomorphic to M m by (8.7.6). в™Ј

8.7.8 Proposition
Rm вЉ— Rn в€ј Rmn . Moreover, if {x1 , . . . , xm } is a basis for Rm and {y1 , . . . , yn } is a basis
=
for R , then {xi вЉ— yj , i = 1, . . . , m, j = 1, . . . , n} is a basis for Rmn .
n

Proof. This follows from the discussion in (8.7.1). [The п¬Ѓrst assertion can also be proved
by taking M = Rn in (8.7.7).] в™Ј

8.7.9 Tensor Product of Homomorphisms
Let f1 : M1 в†’ N1 and f2 : M2 в†’ N2 be R-module homomorphisms. The map (x1 , x2 ) в†’
f1 (x1 ) вЉ— f2 (x2 ) of M1 Г— M2 into N1 вЉ— N2 is bilinear, and induces a unique f : M1 вЉ— M2 в†’
N1 вЉ— N2 such that
f (x1 вЉ— x2 ) = f1 (x1 ) вЉ— f2 (x2 ), x1 в€€ M1 , x2 в€€ M2 .
We write f = f1 вЉ—f2 , and call it the tensor product of f1 and f2 . Similarly, if g1 : N1 в†’ P1
and g2 : N2 в†’ P2 , then we can compose g1 вЉ— g2 with f1 вЉ— f2 , and
(g1 вЉ— g2 )(f1 вЉ— f2 )(x1 вЉ— x2 ) = g1 f1 (x1 ) вЉ— g2 f2 (x2 ),
hence
(g1 вЉ— g2 ) в—¦ (f1 вЉ— f2 ) = (g1 в—¦ f1 ) вЉ— (g2 в—¦ f2 ).
When M1 = N1 = V , a free R-module of rank m, and M2 = N2 = W , a free R-
module of rank n, there is a very concrete interpretation of the tensor product of the
endomorphisms f : V в†’ V and g : W в†’ W . If f is represented by the matrix A and g by
the matrix B, then the action of f and g on basis elements is given by

f (vj ) = aij vi , g(wl ) = bkl wk
i k

where i and j range from 1 to m, and k and l range from 1 to n. Thus

(f вЉ— g)(vj вЉ— wl ) = f (vj ) вЉ— g(wl ) = aij bkl (vi вЉ— wk ).
i,k

The mn by mn matrix representing the endomorphism f вЉ— g : V вЉ— W в†’ V вЉ— W is denoted
by A вЉ— B and called the tensor product or Kronecker product of A and B. It is given by
пЈ® пЈ№
a11 B a12 B В· В· В· a1m B
пЈЇ. . пЈє.
AвЉ—B =пЈ° . .пЈ»
. .
В·В·В·
am1 B am2 B amm B
8.8. GENERAL TENSOR PRODUCTS 25

The ordering of the basis of V вЉ— W is

v1 вЉ— w1 , . . . , v1 вЉ— wn , . . . , vm вЉ— w1 , . . . , vm вЉ— wn .

To determine the column of A вЉ— B corresponding to vj вЉ— wl , locate the aij B block
(i = 1, . . . , m; j п¬Ѓxed) and proceed to column l of B. As we move down this column, the
indices i and k vary according to the above ordering of basis elements. If this road map
is not clear, perhaps writing out the entire matrix for m = 2 and n = 3 will help.

Problems For Section 8.7
1. Verify Equations (2) and (3) of (8.7.2).
2. If m and n are relatively prime, show that Zm вЉ—Z Zn = 0.
3. If A is a п¬Ѓnite abelian group and Q is the additive group of rationals, show that
A вЉ—Z Q = 0. Generalize to a wider class of abelian groups A.
4. The deп¬Ѓnition of M вЉ—R N via a universal mapping property is as follows. The tensor
product is an R-module T along with a bilinear map h : M Г— N в†’ T such that given
any bilinear map f : M Г— N в†’ P , there is a unique R-homomorphism g : T в†’ P such
that f = gh. See the diagram below.

GT
h
M Г—N
qq
qq
qq
f qqq
g
5
P

Now suppose that another R-module T , along with a bilinear mapping h : M Г— N в†’
T , satisп¬Ѓes the universal mapping property. Using the above diagram with P = T
and f replaced by h , we get a unique homomorphism g : T в†’ T such that h = gh.
Reversing the roles of T and T , we get g : T в†’ T such that h = g h .
Show that T and T are isomorphic.
5. Consider the element nвЉ—x in ZвЉ—Zn , where x is any element of Zn and we are tensoring
over Z, i.e., R = Z. Show that n вЉ— x = 0.
6. Continuing Problem 5, take x = 0 and regard n вЉ— x as an element of nZ вЉ— Zn rather
than Z вЉ— Zn . Show that n вЉ— x = 0.
7. Let M, N, M , N be arbitrary R-modules, where R is a commutative ring. Show that
the tensor product of homomorphisms induces a linear map from HomR (M, M ) вЉ—R
HomR (N, N ) to HomR (M вЉ—R N, M вЉ—R N ).
8. Let M be a free R-module of rank m, and N a free R-module of rank n. Show that
there is an R-module isomorphism of EndR (M ) вЉ—R EndR (N ) and EndR (M вЉ— N ).

8.8 General Tensor Products
We now consider tensor products of modules over noncommutative rings. A natural
question is вЂњWhy not simply repeat the construction of (8.7.2) for an arbitrary ring R?вЂќ.
26 CHAPTER 8. INTRODUCING ALGEBRAIC GEOMETRY

 << стр. 8(всего 14)СОДЕРЖАНИЕ >>