So far we have studied algebraic structures with only one operation. Now we study rings which are sets with two (compatible) binary operations. Unlike groups, these two operations are usually denoted by + and · . One can, of course, go for general notations for these operations. However, that generalization doesn’t seem to pay much, but complicates matters. We stick to the conventions.
A ring (R, +, ·) (or R in short) is a set R together with two binary operations + and · on R such that the following conditions are satisfied. As in the case of multiplicative groups we write ab for a · b.
|
Notice that it is more conventional to define a ring as an algebraic structure (R, +, ·) that satisfies conditions (1), (2) and (5) only. A ring (by the conventional definition) is called a commutative ring (resp. a ring with identity), if it (additionally) satisfies condition (3) (resp. (4)). As per our definition, a ring is always a commutative ring with identity. Rings that are not commutative or that do not contain the identity element are not used in the rest of the book. So let us be happy with our unconventional definition of a ring.[3]
[3] Cool! But what’s circular in a ring? Historically, such algebraic structures were introduced by Hilbert to designate a Zahlring (a number ring, see Section 2.13). If α is an algebraic integer (Definition 2.95) and we take a Zahlring of the form and consider the powers α, α2, α3, . . . , we eventually get an αd which can be expressed as a linear combination of the previous (that is, smaller) powers of α. This is perhaps the reason that prompted Hilbert to call such structures “rings”. Also see Footnote 1.
We do not rule out the possibility that 0 = 1 in R. In that case, for any , we have a = a · 1 = a · 0 = 0 (See Proposition 2.6), that is to say, the set R consists of the single element 0. In this case, R is called the zero ring and is denoted (by an abuse of notation) by 0.
Finally, note that R is, in general, not a group under multiplication. This is because we do not expect a ring R to contain the multiplicative inverse of every element of R. Indeed the multiplicative inverse of the element 0 exists if and only if R = 0.
|
Let R be a ring. For all a, , we have:
Proof
|
|
Let R be a ring.
|
A field is an integral domain. Proof Recall from Definition 2.13 that an element in a ring cannot be simultaneously a unit and a zero-divisor. |
Let R be a non-zero ring. The characteristic of R, denoted char R, is the smallest positive integer n such that 1 + 1 + · · · + 1 (n times) = 0. If no such integer exists, then we take char R = 0. |
, , and are rings of characteristic zero. If R is a non-zero finite ring, then the elements 1, 1 + 1, 1 + 1 + 1, · · · cannot be all distinct. This shows that there are positive integers m and n, m < n, such that 1+1+· · · + 1 (n times) = 1 + 1 + · · · + 1 (m times). But then 1 + 1 + · · · + 1 (n – m times) = 0. Thus any non-zero finite ring has positive (that is, non-zero) characteristic. If char R = t is finite, then for any one has .
In what follows, we will often denote by n the element 1 + 1 + · · · + 1 (n times) of any ring. One should not confuse this with the integer n. One can similarly identify a negative integer –n with the ring element –(1 + 1 + · · · + 1)(n times) = (–1) + (–1) + · · · + (–1)(n times).
Let R be an integral domain of positive characteristic p. Then p is a prime. Proof If p is composite, then we can write p = mn with 1 < m < p and 1 < n < p. But then p = mn = 0 (in R). Since R is an integral domain, we must have m = 0 or n = 0 (in R). This contradicts the minimality of p. |
Just as we studied subgroups of groups, it is now time to study subrings of rings. It, however, turns out that subrings are not that important for the study of rings as the subsets called ideals are. In fact, ideals (and not subrings) help us construct quotient rings. This does not mean that ideals are “normal” subrings! In fact, ideals are, in general, not subrings at all, and conversely. The formal definitions are waiting!
is a subring of , and , whereas and are field extensions.
We demand that a ring always contains the multiplicative identity (Definition 2.12). This implies that if S is a subring of R, then for all integers n, the elements are also in S (though they need not be pairwise distinct). Similarly, if R and S are fields, then S contains all the elements of the form mn–1 for m, , (cf. Exercise 2.26). Thus , the set of all even integers, is not a subring of , though it is a subgroup of (, +) (Example 2.2).
Let R be a ring. A subset of R is called an ideal of R, if is an additive subgroup of (R, +) and if for all and .[4]
|
In this book, we will use Gothic letters (usually lower case) like , , , , to denote ideals.[5]
[5] Mathematicians always run out of symbols. Many believe if it is Gothic, it is just ideal!
The condition for being an ideal is in one sense more stringent than that for being a subring, that is, an ideal has to be closed under multiplication by any element of the entire ring. On the other hand, we do not demand an ideal to necessarily contain the identity element 1. In fact, is an ideal of . Conversely, is a subring of but not an ideal. Subrings and ideals are different things.
|
The only ideals of a field are the zero ideal and the unit ideal. Proof By definition, every non-zero element of a field is a unit. |
Let R be a ring and ai, , a family of elements of R. The ideal generated by ai, , is defined to be the sum of the principal ideals Rai. We denote this as . In this case, we also say that is generated by ai, . If I is finite, then we say that is finitely generated. In particular, if #I = 1, then is a principal ideal (See Example 2.7). An integral domain every ideal of which is principal is called a principal ideal domain or PID in short. A ring every ideal of which is finitely generated is called Noetherian. Thus principal ideal domains are Noetherian. |
Note that an ideal may have different generating sets of varying cardinalities. For example, the unit ideal in any ring is principal, since it is generated by 1. The integers 2 and 3 generate the unit ideal of , since . However, neither 2 nor 3 individually generates the unit ideal of . Indeed, using Bézout’s relation (Proposition 2.16) one can show that for every there is a (minimal) generating set of the unit ideal of , that contains exactly n integers. Interested readers may try to construct such generating sets as an (easy) exercise.
A very similar argument proves the following theorem. The details are left to the reader. Also see Exercise 2.31.
We now prove a very important theorem:
Two particular types of ideals are very important in algebra.
Let R be a ring.
|
Prime and maximal ideals can be characterized by some nice equivalent criteria. See Proposition 2.9.
Let R be a ring and an ideal of R. Then is a subgroup of the group (R, +). Since (R, +) is Abelian, is a normal subgroup (Definition 2.6). Thus the cosets , , form an additive Abelian group. We define multiplication on these cosets as . It is easy to check that this multiplication is well-defined. Furthermore, the set of these cosets, denoted , becomes a ring under this addition and multiplication. The ring is called the quotient ring of R with respect to . We say that two elements a, are congruent modulo an ideal (of R) and write a ≡ b (mod ), if . Thus a ≡ b (mod ) if and only if a and b lie in the same coset of , that is, . |
|
The last proposition in conjunction with Corollary 2.1 indicates:
Maximal ideals are prime. |
For every , the quotient ring is a field. In particular, is a maximal ideal of . Proof Since is a prime ideal of , is an integral domain. But is finite, so by Exercise 2.25 is a field. |
Recall how we have defined homomorphisms of groups. In a similar manner, we define homomorphisms of rings. A ring homomorphism is a map from one ring to another, which respects addition, multiplication and the identity element. More precisely:
|
Let f : R → S be a ring homomorphism.
Proof
|
The ideal of the above proposition is called the contraction of and is often denoted by . If R ⊆ S and f is the inclusion homomorphism, then .
Let f : R → S be a ring homomorphism. The set is called the kernel of f and is denoted by Ker f. The set is called the image of f and is denoted by f(R) or Im f. |
With the notations of the last definition, Ker f is an ideal of R, Im f is a subring of S and R/ Ker f ≅ Im f. Proof Consider the map that takes a + Ker f ↦ f(a). It is easy to verify that is a well-defined ring homomorphism and is bijective. The details are left to the reader. Also see Theorem 2.3. |
Two ideals and of a ring R are called relatively prime or coprime if , that is, if there exist and with a + b = 1. |
In Section 2.5, we will see an interesting application of this theorem. Notice that the injectivity of in the last proof does not require the coprimality of ; the surjectivity of requires this condition.
Now we introduce the concept of divisibility in a ring. We also discuss about an important type of rings known as unique factorization domains. This study is a natural generalization of that of the rings and K[X], K a field.
Note that for the concepts of prime and irreducible elements are the same. This is indeed true for any PID (Proposition 2.12). Thus our conventional definition of a prime integer p > 0 as one which has only 1 and p as (positive) divisors tallies with the definition of irreducible elements above. For the ring K[X], on the other hand, it is more customary to talk about irreducible polynomials instead of prime polynomials; they are the same thing anyway.
Let R be an integral domain and a prime. Then p is irreducible. Proof Let p = ab. Then p|(ab), so that by hypothesis p|a or p|b. If p|a, then a = up for some . Hence p = ab = upb, that is, (1 – ub)p = 0. Since R is an integral domain and p ≠ 0, we have 1 – ub = 0, that is, ub = 1, that is, b is a unit. Similarly, p|b implies a is a unit. |
Let R be a PID. An element is prime if and only if p is irreducible. Proof [if] Let p be irreducible, but not prime. Then there are a, such that a ∉ 〈p〉 and b ∉ 〈p〉, but . Consider the ideal . Since , we have p = cα for some . By hypothesis, p is irreducible, so that either c or α is a unit. If c is a unit, 〈p〉 = 〈α〉 = 〈p〉 + 〈a〉, that is, , a contradiction. So α is a unit. Then 〈p〉 + 〈a〉 = R which implies that there are elements u, such that up + va = 1. Similarly, there are elements u′, such that u′p + v′b = 1. Multiplying these two equations gives (uu′p + uv′b + u′va)p + (vv′)ab = 1. Now , so that ab = wp for some . But then (uu′p + uv′b + u′va + vv′w)p = 1, which shows that p is a unit, a contradiction. [only if] Immediate from Proposition 2.11. |
Let R be a UFD. An element is prime if and only if p is irreducible. Proof The only if part is immediate from Proposition 2.11. For proving the if part, let p = up1 · · · pr ( and pi primes in R) be irreducible. If r = 0, p is a unit, a contradiction. If r > 1, then p can be written as the product of two non-units up1 · · · pr–1 and pr, again a contradiction. So r = 1. |
A classical example of an integral domain that is not a UFD is . In this ring, we have two essentially different factorizations of 6 into irreducible elements. The failure of irreducible elements to be primes in such rings is a serious thing to patch up!
A PID is a UFD Proof Let R be a PID and . We show that a has a factorization of the form a = up1 · · · pr, where u is a unit and p1, . . . , pr are prime elements of R. If a is a unit, we are done. So assume that a =: a0 is a non-unit and let . Since , there is a maximal ideal containing (Exercise 2.23). Then p1 is a prime that divides a0. Let a0 = a1p1. We have . If is the unit ideal, we are done. Otherwise we choose as before a prime p2 dividing a1 and with a1 = a2p2 get the ideal properly containing . Repeating this process we can generate a strictly ascending chain of ideals of R. Since R is a PID and hence Noetherian, this process must stop after finitely many steps (Exercise 2.33). |
The converse of the above theorem is not necessarily true. For example, the polynomial ring K[X1, . . . , Xn] over a field K is a UFD for every , but not a PID for n ≥ 2.
Divisibility in a UFD can be rephrased in terms of prime factorizations. Let R be a UFD and let the non-zero elements a, have the prime factorizations and with units u, u′, pairwise non-associate primes p1, . . . , pr and with αi ≥ 0 and βi ≥ 0. Then a|b if and only if αi ≤ βi for all i = 1, . . . , r. This notion leads to the following definitions.
It is clear that these definitions of gcd and lcm can be readily generalized for any arbitrary finite number of elements.
Let R be a UFD and a, not both zero. Then gcd(a, b) · lcm(a, b) is an associate of ab. Proof Immediate from the definitions. |
Let R be a UFD and a, b, with a|bc. If gcd(a, c) = 1, then a|b. Proof Consider the prime factorizations of a, b and c. |
For a PID, the gcd and lcm have equivalent characterizations.
A direct corollary to the last proposition is the following.
This completes our short survey of factorization in rings. Note that and K[X] (for a field K) are PID and hence UFD. Thus all the results we have proved in this section apply equally well to both these rings. It is because of this (and not of a mere coincidence) that these two rings enjoy many common properties. Thus our abstract treatment saves us from the duplicate effort of proving the same results once for integers (Section 2.5) and once more for polynomials (Section 2.6).
2.21 | For a non-zero ring R, prove the following assertions:
Let K be a field. What are the units in the polynomial ring K[X]? In K[X1, . . . , Xn]? In the ring K(X) of rational functions? In K(X1, . . . , Xn)? |
2.22 | Binomial theorem Let R be a ring, a, and . Show that where
are the binomial coefficients. |
2.23 | Show that every non-zero ring has a maximal (and hence prime) ideal. More generally, show that every non-unit ideal of a non-zero ring is contained in a maximal ideal. [H] |
2.24 | Let R be a ring.
|
2.25 | Show that a finite integral domain R is a field. [H] |
2.26 | Let R be a ring of characteristic 0. Show that:
|
2.27 | Let f : R → S be a ring-homomorphism and let and be ideals in R and S respectively. Find examples to corroborate the following statements.
|
2.28 | Let K be a field.
|
2.29 |
|
2.30 | Let R be a ring and let and be ideals of R with . Show that is an ideal of and that . [H] |
2.31 | An integral domain R is called a Euclidean domain (ED) if there is a map satisfying the following two conditions:
Show that:
|
2.32 | Let R be a ring and an ideal. Consider the set
Show that is an ideal of R. It is called the radical or root of . If , then is called a radical or a root ideal. For arbitrary ideals and of R, prove the following assertions.
|
2.33 | Let R be a ring. An ascending chain of ideals is a sequence . The ascending chain is called stationary, if there is some such that for all n ≥ n0. Show that the following conditions are equivalent. [H]
|
2.34 |
|
18.191.176.194