Homework 9 Solutions

0.0(0)
Studied by 0 people
call kaiCall Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/23

encourage image

There's no tags or description

Looks like no tags are added yet.

24 Terms

1
New cards

Exercise 1(1). Prove or disprove: Every subring of Z\mathbb Z is generated by one element.

True. Let S⊆ZS\subseteq\mathbb Z be a subring. Since SS is closed under addition and additive inverses, (S,+)(S,+) is an additive subgroup of (Z,+)(\mathbb Z,+). By the classification of subgroups of Z\mathbb Z, every such subgroup has the form nZn\mathbb Z for some n≥0n\geq 0. Thus S=nZ=kn:k∈ZS=n\mathbb Z={kn:k\in\mathbb Z}. This set is generated additively by the single element nn. It is also the subring generated by nn, since it is closed under multiplication: (an)(bn)=abn2∈nZ(an)(bn)=abn^2\in n\mathbb Z. Therefore every subring of Z\mathbb Z is generated by one element. If subrings are required to contain the identity, the only possibility is Z\mathbb Z itself, generated by 11.

2
New cards

Exercise 1(2). Prove or disprove: In a ring with identity without zero-divisors, the only elements rr such that r2=1r^2=1 are 11 and −1-1.

True. Let RR be a ring with identity and no zero-divisors, and suppose r2=1r^2=1. Subtracting 11 gives r2−1=0r^2-1=0. Since 11 commutes with every element, we can factor this expression even when RR is noncommutative: (r−1)(r+1)=r2−1=0(r-1)(r+1)=r^2-1=0. By the assumption that RR has no zero-divisors, a product can equal zero only if at least one factor is zero. Therefore r−1=0r-1=0 or r+1=0r+1=0, giving r=1r=1 or r=−1r=-1. Conversely, 12=11^2=1 and (−1)2=1(-1)^2=1. Hence the solutions are exactly 11 and −1-1, which may coincide in characteristic 22.

3
New cards

Exercise 1(3). Prove or disprove: The set of rational numbers with odd denominator (when written in lowest terms) is a subring of Q\mathbb Q.

True. Define S=a/b∈Q:a,b∈Z, b≠0, b oddS={a/b\in\mathbb Q:a,b\in\mathbb Z,\ b\neq 0,\ b\text{ odd}}. First, 0=0/1∈S0=0/1\in S, so SS is nonempty. Let a/b,c/d∈Sa/b,c/d\in S, where b,db,d are odd. Their difference is a/b−c/d=(ad−bc)/(bd)a/b-c/d=(ad-bc)/(bd). Since the product of two odd integers is odd, bdbd is odd. Similarly, their product is (a/b)(c/d)=ac/(bd)(a/b)(c/d)=ac/(bd), again with odd denominator. Reducing either fraction to lowest terms only divides its denominator by an integer, so the resulting denominator remains odd. Thus SS is closed under subtraction and multiplication. By the subring test, SS is a subring of Q\mathbb Q.

4
New cards

Exercise 1(4). Prove or disprove: An element rr of a ring with identity RR with two distinct right inverses (distinct s,t∈Rs,t\in R such that rs=rt=1rs=rt=1) is a left zero-divisor.

True. Suppose r∈Rr\in R has two distinct right inverses s,t∈Rs,t\in R, meaning rs=1rs=1 and rt=1rt=1, with s≠ts\neq t. Subtract these equations: rs−rt=1−1=0rs-rt=1-1=0. By right distributivity, r(s−t)=0r(s-t)=0. Because s≠ts\neq t, their difference s−ts-t is nonzero. Also, r≠0r\neq 0, since rs=1rs=1 and the identity of the ring is nonzero under the stated hypothesis that two distinct right inverses exist. Thus we have found nonzero elements rr and s−ts-t whose product is zero, with rr on the left. By definition, rr is a left zero-divisor.

5
New cards

Exercise 2(1). Let (G,+)(G,+) be an abelian group, R=group homomorphisms G→GR={\text{group homomorphisms }G\to G}, and define (φ+ψ)(g)=φ(g)+ψ(g)(\varphi+\psi)(g)=\varphi(g)+\psi(g) for every g∈Gg\in G. Let ∘\circ denote composition. Show that (R,+,∘)(R,+,\circ) is a ring with identity.

First show closure: for a,b∈Ga,b\in G, (φ+ψ)(a+b)=φ(a)+φ(b)+ψ(a)+ψ(b)=(φ+ψ)(a)+(φ+ψ)(b)(\varphi+\psi)(a+b)=\varphi(a)+\varphi(b)+\psi(a)+\psi(b)=(\varphi+\psi)(a)+(\varphi+\psi)(b), using commutativity of GG. Pointwise associativity and commutativity follow from GG. The zero homomorphism is the additive identity, and (−φ)(g)=−φ(g)(-\varphi)(g)=-\varphi(g) gives additive inverses. Composition is closed because compositions of homomorphisms are homomorphisms, and it is associative because function composition is associative. Right distributivity follows from ((φ+ψ)∘θ)(g)=φ(θ(g))+ψ(θ(g))((\varphi+\psi)\circ\theta)(g)=\varphi(\theta(g))+\psi(\theta(g)). Left distributivity follows from (φ∘(ψ+θ))(g)=φ(ψ(g)+θ(g))=φ(ψ(g))+φ(θ(g))(\varphi\circ(\psi+\theta))(g)=\varphi(\psi(g)+\theta(g))=\varphi(\psi(g))+\varphi(\theta(g)). Finally, id⁡G∘φ=φ∘id⁡G=φ\operatorname{id}_G\circ\varphi=\varphi\circ\operatorname{id}_G=\varphi. Therefore RR is a ring with identity id⁡G\operatorname{id}_G.

6
New cards

Exercise 2(2). Let (G,+)(G,+) be an abelian group, R=group homomorphisms G→GR={\text{group homomorphisms }G\to G}, with pointwise addition (φ+ψ)(g)=φ(g)+ψ(g)(\varphi+\psi)(g)=\varphi(g)+\psi(g) and multiplication by composition. If GG were not abelian, would RR still be a ring?

No. Suppose GG is nonabelian. The identity map id⁡G\operatorname{id}_G is a group homomorphism, so it belongs to RR. If RR were closed under pointwise addition, then id⁡G+id⁡G\operatorname{id}_G+\operatorname{id}_G would also be a homomorphism. Therefore, for every a,b∈Ga,b\in G, the homomorphism property would imply (a+b)+(a+b)=(a+a)+(b+b)(a+b)+(a+b)=(a+a)+(b+b). Written without parentheses, associativity gives a+b+a+b=a+a+b+ba+b+a+b=a+a+b+b. Cancel the first aa on the left and the last bb on the right to obtain b+a=a+bb+a=a+b. Since a,ba,b were arbitrary, GG would be abelian, contradicting our assumption. Thus pointwise addition is not closed on RR, so the proposed operations do not make RR a ring.

7
New cards

Exercise 2(3). Let (G,+)(G,+) be an abelian group, R=group homomorphisms G→GR={\text{group homomorphisms }G\to G}, with pointwise addition and multiplication by composition. Find a nontrivial group GG for which RR is commutative and one for which RR is not commutative.

For a commutative example, take G=(Z,+)G=(\mathbb Z,+). Every homomorphism φ:Z→Z\varphi:\mathbb Z\to\mathbb Z is determined by φ(1)=n\varphi(1)=n, so φ(k)=nk\varphi(k)=nk. For any two endomorphisms φn,φm\varphi_n,\varphi_m, we have (φn∘φm)(k)=nmk=mnk=(φm∘φn)(k)(\varphi_n\circ\varphi_m)(k)=nmk=mnk=(\varphi_m\circ\varphi_n)(k). Hence R≅ZR\cong\mathbb Z is commutative. For a noncommutative example, take G=(Z2,+)G=(\mathbb Z^2,+). Define φ(x,y)=(y,0)\varphi(x,y)=(y,0) and ψ(x,y)=(0,x)\psi(x,y)=(0,x). Both are homomorphisms. But (φ∘ψ)(x,y)=(x,0)(\varphi\circ\psi)(x,y)=(x,0), whereas (ψ∘φ)(x,y)=(0,y)(\psi\circ\varphi)(x,y)=(0,y). These are different, for example at (1,0)(1,0). Thus composition is not commutative and R≅M2(Z)R\cong M_2(\mathbb Z) is noncommutative.

8
New cards

Exercise 2(4). Let (G,+)(G,+) be an abelian group, R=group homomorphisms G→GR={\text{group homomorphisms }G\to G}, with pointwise addition and multiplication by composition. Show that the units of RR are the group automorphisms of GG.

Recall that a unit of RR is an endomorphism φ\varphi possessing a two-sided inverse under composition. Suppose φ\varphi is a unit. Then there exists ψ∈R\psi\in R such that φ∘ψ=ψ∘φ=id⁡G\varphi\circ\psi=\psi\circ\varphi=\operatorname{id}_G. Therefore φ\varphi has a two-sided inverse as a function, so it is bijective. Since φ\varphi is already a homomorphism, it is a group automorphism. Conversely, suppose φ\varphi is an automorphism of GG. It is a bijective homomorphism, and its inverse function φ−1\varphi^{-1} is also a homomorphism: if a=φ(x)a=\varphi(x) and b=φ(y)b=\varphi(y), then φ−1(a+b)=x+y=φ−1(a)+φ−1(b)\varphi^{-1}(a+b)=x+y=\varphi^{-1}(a)+\varphi^{-1}(b). Thus φ−1∈R\varphi^{-1}\in R, and both compositions equal id⁡G\operatorname{id}_G. Hence R×=Aut⁡(G)R^\times=\operatorname{Aut}(G).

9
New cards

Exercise 3(1). Let (R,+R,⋅R)(R,+_R,\cdot_R) and (S,+S,⋅S)(S,+_S,\cdot_S) be rings. On R×SR\times S define (r1,s1)+(r2,s2)=(r1+Rr2,s1+Ss2)(r_1,s_1)+(r_2,s_2)=(r_1+_Rr_2,s_1+_Ss_2) and (r1,s1)⋅(r2,s2)=(r1⋅Rr2,s1⋅Ss2)(r_1,s_1)\cdot(r_2,s_2)=(r_1\cdot_Rr_2,s_1\cdot_Ss_2). Show that (R×S,+,⋅)(R\times S,+,\cdot) is a ring.

Verify the ring axioms coordinatewise. Closure under addition and multiplication follows because each coordinate lies in its respective ring. The additive identity is (0R,0S)(0_R,0_S), and the additive inverse of (r,s)(r,s) is (−r,−s)(-r,-s). Addition is associative and commutative because addition in RR and SS has these properties. Multiplication is associative since ((r,s)(u,v))(x,y)=((ru)x,(sv)y)=(r(ux),s(vy))=(r,s)((u,v)(x,y))((r,s)(u,v))(x,y)=((ru)x,(sv)y)=(r(ux),s(vy))=(r,s)((u,v)(x,y)). For distributivity, (r,s)((u,v)+(x,y))=(r(u+x),s(v+y))=(ru+rx,sv+sy)=(r,s)(u,v)+(r,s)(x,y)(r,s)((u,v)+(x,y))=(r(u+x),s(v+y))=(ru+rx,sv+sy)=(r,s)(u,v)+(r,s)(x,y). The other distributive law follows similarly. All ring axioms hold, so R×SR\times S is a ring.

10
New cards

Exercise 3(2). Let R,SR,S be rings and define addition and multiplication on R×SR\times S componentwise. Show that R×SR\times S is commutative if and only if both RR and SS are commutative.

Prove both directions. First suppose RR and SS are commutative. For arbitrary (r,s),(u,v)∈R×S(r,s),(u,v)\in R\times S, we have (r,s)(u,v)=(ru,sv)=(ur,vs)=(u,v)(r,s)(r,s)(u,v)=(ru,sv)=(ur,vs)=(u,v)(r,s). Hence R×SR\times S is commutative. Conversely, suppose R×SR\times S is commutative. Choose arbitrary r,u∈Rr,u\in R and consider (r,0S)(r,0_S) and (u,0S)(u,0_S). Their commutativity gives (ru,0S)=(ur,0S)(ru,0_S)=(ur,0_S), so ru=urru=ur. Thus RR is commutative. Similarly, comparing (0R,s)(0R,v)(0_R,s)(0_R,v) with (0R,v)(0R,s)(0_R,v)(0_R,s) for arbitrary s,v∈Ss,v\in S gives sv=vssv=vs. Therefore SS is commutative. Both implications hold, proving the equivalence.

11
New cards

Exercise 3(3). Let R,SR,S be rings and define addition and multiplication on R×SR\times S componentwise. Show that R×SR\times S has an identity if and only if both RR and SS have identities.

First assume RR and SS have multiplicative identities 1R1_R and 1S1_S. Then for every (r,s)∈R×S(r,s)\in R\times S, (1R,1S)(r,s)=(r,s)=(r,s)(1R,1S)(1_R,1_S)(r,s)=(r,s)=(r,s)(1_R,1_S). Thus (1R,1S)(1_R,1_S) is the identity of the product. Conversely, assume R×SR\times S has identity (eR,eS)(e_R,e_S). For arbitrary r∈Rr\in R, multiplying (r,0S)(r,0_S) on both sides by this identity gives (eRr,0S)=(r,0S)=(reR,0S)(e_Rr,0_S)=(r,0_S)=(re_R,0_S). Therefore eRr=reR=re_Rr=re_R=r for every r∈Rr\in R, so eRe_R is the identity of RR. Similarly, applying the identity to (0R,s)(0_R,s) shows eSs=seS=se_Ss=se_S=s for every s∈Ss\in S. Thus both rings have identities, completing the equivalence.

12
New cards

Exercise 3(4). Let R,SR,S be rings with identity and define multiplication on R×SR\times S componentwise. Show that (R×S)×=R××S×(R\times S)^\times=R^\times\times S^\times, where R×R^\times denotes the group of units of RR.

Suppose (r,s)∈(R×S)×(r,s)\in(R\times S)^\times. Then there exists (u,v)∈R×S(u,v)\in R\times S such that (r,s)(u,v)=(u,v)(r,s)=(1R,1S)(r,s)(u,v)=(u,v)(r,s)=(1_R,1_S). By componentwise multiplication, this means ru=ur=1Rru=ur=1_R and sv=vs=1Ssv=vs=1_S. Therefore rr is a unit of RR with inverse uu, and ss is a unit of SS with inverse vv. Thus (r,s)∈R××S×(r,s)\in R^\times\times S^\times. Conversely, suppose r∈R×r\in R^\times and s∈S×s\in S^\times. Their inverses exist, and (r,s)(r−1,s−1)=(1R,1S)=(r−1,s−1)(r,s)(r,s)(r^{-1},s^{-1})=(1_R,1_S)=(r^{-1},s^{-1})(r,s). Hence (r,s)(r,s) is a unit in R×SR\times S. The two inclusions establish (R×S)×=R××S×(R\times S)^\times=R^\times\times S^\times, with inverses computed coordinatewise.

13
New cards

Exercise 3(5). Let R,SR,S be rings and define addition and multiplication on R×SR\times S componentwise. Determine under which conditions on R,SR,S the ring R×SR\times S is a division ring.

Recall that a division ring is a nontrivial ring with identity in which every nonzero element has a multiplicative inverse. Suppose both RR and SS are nontrivial. Choose r≠0r\neq0 in RR and s≠0s\neq0 in SS. Then (r,0S)(r,0_S) and (0R,s)(0_R,s) are nonzero, but (r,0S)(0R,s)=(0R,0S)(r,0_S)(0_R,s)=(0_R,0_S). This is impossible in a division ring, since division rings have no nonzero zero-divisors. Hence at least one factor must be the zero ring. If S=0S={0}, the map R→R×SR\to R\times S given by r↦(r,0)r\mapsto(r,0) is a ring isomorphism. Therefore R×SR\times S is a division ring precisely when RR is one. The case R=0R={0} is symmetric. Thus exactly one factor must be zero and the other a division ring.

14
New cards

Exercise 4(1). Let R,SR,S be rings, and let R′⊆RR'\subseteq R and S′⊆SS'\subseteq S be subrings. Show that R′×S′R'\times S' is a subring of R×SR\times S, where the product operations are componentwise.

Apply the subring test. Because R′R' and S′S' are subrings, they contain their respective additive identities, so (0R,0S)∈R′×S′(0_R,0_S)\in R'\times S'. Thus the proposed subring is nonempty. Take arbitrary (a,b),(c,d)∈R′×S′(a,b),(c,d)\in R'\times S'. Since subrings are closed under subtraction, a−c∈R′a-c\in R' and b−d∈S′b-d\in S'. Consequently, (a,b)−(c,d)=(a−c,b−d)∈R′×S′(a,b)-(c,d)=(a-c,b-d)\in R'\times S'. Similarly, closure of R′R' and S′S' under multiplication gives ac∈R′ac\in R' and bd∈S′bd\in S', so (a,b)(c,d)=(ac,bd)∈R′×S′(a,b)(c,d)=(ac,bd)\in R'\times S'. Therefore R′×S′R'\times S' is nonempty and closed under subtraction and multiplication. By the subring test, it is a subring of R×SR\times S.

15
New cards

Exercise 4(2). Let R,SR,S be rings. Find R,SR,S such that R×SR\times S contains a subring TT not of the form R′×S′R'\times S' for subrings R′⊆RR'\subseteq R, S′⊆SS'\subseteq S. Explain why TT cannot be written that way.

Take R=S=ZR=S=\mathbb Z and define the diagonal subset T=(n,n):n∈Z⊆Z×ZT={(n,n):n\in\mathbb Z}\subseteq\mathbb Z\times\mathbb Z. First verify it is a subring. It contains (0,0)(0,0), and (a,a)−(b,b)=(a−b,a−b)∈T(a,a)-(b,b)=(a-b,a-b)\in T and (a,a)(b,b)=(ab,ab)∈T(a,a)(b,b)=(ab,ab)\in T. Now suppose for contradiction that T=R′×S′T=R'\times S' for some subrings R′,S′⊆ZR',S'\subseteq\mathbb Z. Since (1,1)∈T(1,1)\in T, we must have 1∈R′1\in R' and 1∈S′1\in S'. Also 0∈S′0\in S' because every subring contains zero. Therefore (1,0)∈R′×S′=T(1,0)\in R'\times S'=T. But every element of TT has equal coordinates, whereas 1≠01\neq0. This contradiction proves that TT cannot be expressed as a direct product of two subrings.

16
New cards

Exercise 4(3). Let R,SR,S be nontrivial rings and let r∈R∖0r\in R\setminus{0}. Show that (r,0S)(r,0_S) is a zero-divisor in R×SR\times S.

Since SS is a nontrivial ring, there exists an element s∈Ss\in S with s≠0Ss\neq0_S. Consider (0R,s)∈R×S(0_R,s)\in R\times S. This element is nonzero because its second coordinate is nonzero. By assumption, r≠0Rr\neq0_R, so (r,0S)(r,0_S) is also nonzero. Using componentwise multiplication, compute (r,0S)(0R,s)=(r0R,0Ss)=(0R,0S)(r,0_S)(0_R,s)=(r0_R,0_Ss)=(0_R,0_S). Hence (r,0S)(r,0_S) multiplies a nonzero element on its right to give the zero element of R×SR\times S. Therefore (r,0S)(r,0_S) is a left zero-divisor. In fact, the reverse product also vanishes, so it is a two-sided zero-divisor. This proves the statement.

17
New cards

Exercise 4(4). Let R,SR,S be rings. Show that R×SR\times S is an integral domain if and only if RR is an integral domain and S=0S={0}, or SS is an integral domain and R=0R={0}.

Recall that an integral domain is a nontrivial commutative ring with identity and no zero-divisors. Suppose R×SR\times S is an integral domain. If both factors were nontrivial, choose r≠0r\neq0 and s≠0s\neq0. Then (r,0S)(0R,s)=(0R,0S)(r,0_S)(0_R,s)=(0_R,0_S), contradicting the absence of zero-divisors. Thus at least one factor is zero. They cannot both be zero because an integral domain is nontrivial. Suppose S=0S={0}. The map φ:R→R×S\varphi:R\to R\times S defined by φ(r)=(r,0)\varphi(r)=(r,0) is bijective and preserves addition and multiplication, so it is a ring isomorphism. Hence RR is an integral domain. The other case is symmetric. Conversely, if RR is an integral domain and S=0S={0}, then R×S≅RR\times S\cong R is an integral domain. Similarly for the reversed factors.

18
New cards

Exercise 5(1). Let RR be a ring. Define its center by Z(R)=r∈R:rs=sr for all s∈RZ(R)={r\in R:rs=sr\text{ for all }s\in R}. Show that Z(R)Z(R) is a subring of RR.

Use the subring test. First, 0∈Z(R)0\in Z(R) because 0r=r0=00r=r0=0 for every r∈Rr\in R. Let a,b∈Z(R)a,b\in Z(R) and choose any r∈Rr\in R. Since aa and bb commute with every element of RR, we have ar=raar=ra and br=rbbr=rb. Therefore (a−b)r=ar−br=ra−rb=r(a−b)(a-b)r=ar-br=ra-rb=r(a-b), showing a−b∈Z(R)a-b\in Z(R). For multiplication, (ab)r=a(br)=a(rb)=(ar)b=(ra)b=r(ab)(ab)r=a(br)=a(rb)=(ar)b=(ra)b=r(ab). Thus ab∈Z(R)ab\in Z(R). Since the center is nonempty and closed under subtraction and multiplication, it is a subring of RR. Furthermore, any two elements of Z(R)Z(R) commute, since each commutes with every element of RR.

19
New cards

Exercise 5(2). Let RR be a division ring and define Z(R)=a∈R:ar=ra for all r∈RZ(R)={a\in R:ar=ra\text{ for all }r\in R}. Show that Z(R)Z(R) is a subfield of RR.

By part (1), Z(R)Z(R) is a subring of RR. It is commutative because any a,b∈Z(R)a,b\in Z(R) satisfy ab=baab=ba. Also 1R∈Z(R)1_R\in Z(R), since the identity commutes with every element. It remains to show that every nonzero central element has an inverse belonging to the center. Let a∈Z(R)a\in Z(R) with a≠0a\neq0. Since RR is a division ring, a−1∈Ra^{-1}\in R exists. For any r∈Rr\in R, centrality gives ar=raar=ra. Multiply this equality on the left and right by a−1a^{-1} to obtain ra−1=a−1rra^{-1}=a^{-1}r. Thus a−1a^{-1} commutes with every r∈Rr\in R, so a−1∈Z(R)a^{-1}\in Z(R). Hence Z(R)Z(R) is a commutative ring with identity in which every nonzero element is invertible, proving it is a field.

20
New cards

Exercise 5(3). Let RR be a ring with identity and n∈N+n\in\mathbb N^+. Prove Z(Mn(R))=(aij)∈Mn(R):aij=0 for i≠j, a11=⋯=ann∈Z(R)Z(M_n(R))={(a_{ij})\in M_n(R):a_{ij}=0\text{ for }i\neq j,\ a_{11}=\cdots=a_{nn}\in Z(R)}.

Write the proposed center as aIn:a∈Z(R){aI_n:a\in Z(R)} and prove both inclusions. If a∈Z(R)a\in Z(R) and B=(bij)B=(b_{ij}) is arbitrary, then (aInB)<em>ij=ab</em>ij=bija=(BaIn)<em>ij(aI_nB)<em>{ij}=ab</em>{ij}=b_{ij}a=(BaI_n)<em>{ij}, so aInaI_n is central. Conversely, let A=(a</em>ij)∈Z(Mn(R))A=(a</em>{ij})\in Z(M_n(R)) and denote the standard matrix units by EijE_{ij}. Commuting AA with every EiiE_{ii} forces all off-diagonal entries of AA to be zero, so AA is diagonal. Commuting with EijE_{ij} for i≠ji\neq j gives aii=ajja_{ii}=a_{jj}; hence A=aInA=aI_n for some a∈Ra\in R. Finally, for arbitrary r∈Rr\in R, commuting AA with rE11rE_{11} gives arE11=raE11arE_{11}=raE_{11}, hence ar=raar=ra. Thus a∈Z(R)a\in Z(R). For n=1n=1, this final step alone suffices. Both inclusions establish the equality.

21
New cards

Exercise 6. Let RR be a commutative ring and p∈R[X]∖0p\in R[X]\setminus{0}. Show that pp is a zero-divisor in R[X]R[X] if and only if there exists r∈R∖0r\in R\setminus{0} such that pr=0pr=0.

First, if pr=0pr=0 for some nonzero constant r∈Rr\in R, then rr is a nonzero polynomial, so pp is a zero-divisor. Conversely, suppose pq=0pq=0 for some nonzero q∈R[X]q\in R[X]. Choose such a qq of minimal degree mm. Write p=∑i=0naiXip=\sum_{i=0}^n a_iX^i and q=∑j=0mbjXjq=\sum_{j=0}^m b_jX^j, with bm≠0b_m\neq0. The coefficient of Xn+mX^{n+m} in pq=0pq=0 gives anbm=0a_nb_m=0. Hence deg⁡(anq)<m\deg(a_nq)<m if anq≠0a_nq\neq0. But p(anq)=anpq=0p(a_nq)=a_npq=0, so minimality forces anq=0a_nq=0. Induct downward: assuming ai+1q=⋯=anq=0a_{i+1}q=\cdots=a_nq=0, the coefficient of Xi+mX^{i+m} in pqpq gives aibm=0a_ib_m=0, because all higher-index contributions vanish. Again p(aiq)=0p(a_iq)=0 and deg⁡(aiq)<m\deg(a_iq)<m, so aiq=0a_iq=0. Thus aibm=0a_ib_m=0 for every ii, and therefore pbm=0pb_m=0. Since bm≠0b_m\neq0, take r=bmr=b_m.

22
New cards

Exercise 7(1). Let RR be a nontrivial ring and GG a nontrivial group. Let RGRG be the group ring. Show that ε:RG→R\varepsilon:RG\to R, defined by ε(∑g∈Grgg)=∑g∈Grg\varepsilon(\sum_{g\in G}r_gg)=\sum_{g\in G}r_g, is well-defined and a ring homomorphism.

Every element of the group ring RGRG has a unique representation as a finite formal sum ∑g∈Grgg\sum_{g\in G}r_gg, with only finitely many nonzero coefficients. Therefore ε\varepsilon is well-defined because the coefficient sum is finite and independent of the representation. Let a=∑grgga=\sum_g r_gg and b=∑hshhb=\sum_h s_hh. Addition is coefficientwise, so ε(a+b)=∑g(rg+sg)=ε(a)+ε(b)\varepsilon(a+b)=\sum_g(r_g+s_g)=\varepsilon(a)+\varepsilon(b). For multiplication, the group-ring rule gives ab=∑g,hrgsh(gh)ab=\sum_{g,h}r_gs_h(gh). Hence ε(ab)=∑g,hrgsh=(∑grg)(∑hsh)=ε(a)ε(b)\varepsilon(ab)=\sum_{g,h}r_gs_h=(\sum_g r_g)(\sum_h s_h)=\varepsilon(a)\varepsilon(b) by distributivity and finite summation. Thus ε\varepsilon preserves addition and multiplication and is a ring homomorphism. If RR has identity, it also preserves identity.

23
New cards

Exercise 7(2). Let RR be a nontrivial ring and GG a nontrivial group. Assume RR has identity. Show that rg∈(RG)×rg\in(RG)^\times for every r∈R×r\in R^\times and g∈Gg\in G.

Suppose r∈R×r\in R^\times and g∈Gg\in G. Since rr is a unit, there exists r−1∈Rr^{-1}\in R satisfying rr−1=r−1r=1Rrr^{-1}=r^{-1}r=1_R. Since GG is a group, gg has inverse g−1∈Gg^{-1}\in G with gg−1=g−1g=eGgg^{-1}=g^{-1}g=e_G. Consider the element r−1g−1∈RGr^{-1}g^{-1}\in RG. By the multiplication rule in the group ring, (rg)(r−1g−1)=(rr−1)(gg−1)=1ReG(rg)(r^{-1}g^{-1})=(rr^{-1})(gg^{-1})=1_Re_G. Similarly, (r−1g−1)(rg)=(r−1r)(g−1g)=1ReG(r^{-1}g^{-1})(rg)=(r^{-1}r)(g^{-1}g)=1_Re_G. The element 1ReG1_Re_G is the multiplicative identity of RGRG. Thus r−1g−1r^{-1}g^{-1} is a two-sided multiplicative inverse of rgrg. Therefore rg∈(RG)×rg\in(RG)^\times, with (rg)−1=r−1g−1(rg)^{-1}=r^{-1}g^{-1}.

24
New cards

Exercise 7(3). Let RR be a nontrivial ring and GG a nontrivial group. Assume GG has a nontrivial element of finite order. Show that the group ring RGRG has a zero-divisor.

Choose g∈Gg\in G with finite order n>1n>1, so gn=eGg^n=e_G, and choose 0≠r∈R0\neq r\in R. Define a=rg−reGa=rg-re_G and b=reG+rg+⋯+rgn−1b=re_G+rg+\cdots+rg^{n-1} in RGRG. Both are nonzero: the elements eG,g,…,gn−1e_G,g,\ldots,g^{n-1} are distinct, and the coefficients rr and −r-r are nonzero. Compute their product using distributivity: ab=r2(g−eG)(eG+g+⋯+gn−1)ab=r^2(g-e_G)(e_G+g+\cdots+g^{n-1}). The intermediate terms cancel telescopically, leaving ab=r2(gn−eG)ab=r^2(g^n-e_G). Since gn=eGg^n=e_G, this equals zero. Therefore a≠0a\neq0 and b≠0b\neq0 but ab=0ab=0, proving that RGRG contains a zero-divisor. The argument does not require RR to have an identity.