Physics 5040 Spring 2009 Problem Set 8 Solutions 1. (a) For the set (Z, −) we have 0 as the identity element, so each integer is its own inverse. However, this is not a group because n − 0 6= 0 − n, and also note that subtraction is not associative since m − (n − q) 6= (m − n) − q. (b) The set G = (Q − {0}, ÷) is not a group. A potential inverse is 1, but 1 ÷ x 6= x ÷ 1. Furthermore, division is not associative. (c) This is just the cyclic group of order 7, C7 . The identity is a0 and the inverses are (a0 )−1 = a0 and (ai )−1 = a7−i for 1 ≤ i ≤ 6. (d) This forms a group with identity element 1 and inverse given by (2m 3n )−1 = 2−m 3−n . 2. Given f ∈ F [R], define 0 ∈ F [R] by 0(x) = 0 and −f ∈ F [R] by (−f )(x) = −f (x). Then we make F [R] into a group by defining the identity element to be e = 0 ∈ F [R] and for every f ∈ F [R] we let f −1 = −f ∈ F [R]. To show that F [R] satisfies the group axioms, we have for every f, g, h ∈ F [R] (i) (f + g)(x) = f (x) + g(x) so that f + g ∈ F [R]. (ii) [(f + g) + h](x) = (f + g)(x) + h(x) = f (x) + g(x) + h(x) = f (x) + (g + h)(x) = [f + (g + h)](x) so that (f + g) + h = f + (g + h). (iii) Clearly (f + 0)(x) = f (x) + 0 = f (x) = 0 + f (x) = (0 + f )(x) so that f + 0 = 0 + f = f . (iv) [f + (−f )](x) = f (x) − f (x) = 0 = −f (x) + f (x) = [(−f ) + f ](x) so that f + (−f ) = (−f ) + f = 0. 3. (a) If A, B are groups, and if (a1 , b1 ), (a2 , b2 ) ∈ A × B with (a1 , b1 )(a2 , b2 ) := (a1 a2 , b1 b2 ) ∈ A × B, then this satisfies the closure requirement for a group. (Since a1 a2 ∈ A and b1 b2 ∈ B.) As to associativity, we have [(a1 , b1 )(a2 , b2 )](a3 , b3 ) = (a1 a2 , b1 b2 )(a3 , b3 ) = (a1 a2 a3 , b1 b2 b3 ) = (a1 , b1 )(a2 a3 , b2 b3 ) = (a1 , b1 )[(a2 , b2 )(a3 , b3 )] . 1 Since A, B are groups, there are identity elements eA ∈ A and eB ∈ B. Then (eA , eB )(a, b) = (eA a, eB b) = (a, b) = (aeA , beB ) = (a, b)(eA , eB ) so that the identity in A × B is just (eA , eB ). As to the inverse, again we use the fact that A and B are groups so every a ∈ A and b ∈ B has an inverse. Then (a−1 , b−1 )(a, b) = (a−1 a, b−1 b) = (eA , eB ) = (aa−1 , bb−1 ) = (a, b)(a−1 , b−1 ) and hence (a, b)−1 = (a−1 , b−1 ). (b) If G is a group, consider the set D = {(x, x) ∈ G × G}. With group operations defined as in part (a), we have for the group axioms: (i) (x, x)(y, y) = (xy, xy) ∈ D (since xy ∈ G) so D is closed. (ii) [(x, x)(y, y)](z, z) = (xy, xy)(z, z) = (xyz, xyz) = (x, x)(yz, yz) = (x, x)[(y, y)(z, z)] so we have associativity. (iii) If e is the identity in G, then (e, e) ∈ D and (e, e)(x, x) = (ex, ex) = (x, x) = (xe, xe) = (x, x)(e, e) so that (e, e) is the identity in D. (iv) Every x ∈ G has an inverse x−1 ∈ G so that (x−1 , x−1 ) ∈ D and (x, x)(x−1 , x−1 ) = (xx−1 , xx−1 ) = (e, e) = (x−1 x, x−1 x) = (x−1 , x−1 )(x, x) so that (x, x)−1 = (x−1 , x−1 ) ∈ D. 4. (a) (123)(45)(16789)(15) = (123)(45)(156789) = (123)(1456789) = (145678923) . (b) (12)(123)(12) = (12)(13) = (132) . 5. (a) Let ψ : A → B and φ : B → C be homomorphisms. Then φ ◦ ψ : A → C is a homomorphism because (φ ◦ ψ)(a1 a2 ) = φ(ψ(a1 a2 )) = φ(ψ(a1 )ψ(a2 )) = φ(ψ(a1 ))φ(ψ(a2 )) = (φ ◦ ψ)(a1 )(φ ◦ ψ)(a2 ) . 2 (b) Let G denote the set of all groups, and let Φ be the set of all group isomorphisms. We say that G1 , G2 ∈ G are equivalent if G2 = φ(G1 ) for some φ ∈ Φ, and we write this as G1 ∼ G2 . To show that this defines an equivalence relation on G , we have (i) Given G ∈ G , let id be the identity map on G. This is clearly an isomorphism (since id(ab) = ab = id(a)id(b)) so that id ∈ Φ. Then G = id(G) so that G ∼ G. (ii) If G1 ∼ G2 , then φ(G1 ) = G2 for some φ ∈ Φ. But φ−1 ∈ Φ so we have G1 = φ−1 (G2 ) so that G2 ∼ G1 . (iii) If G1 ∼ G2 and G2 ∼ G3 , then G2 = φ(G1 ) and G3 = ψ(G2 ) for some φ, ψ ∈ Φ. By part (a) we know that ψ ◦ φ is a homomorphism and hence also an isomorphism, and (ψ ◦ φ)(G1 ) = ψ(φ(G1 )) = ψ(G2 ) = G3 so that G1 ∼ G3 (since ψ ◦ φ ∈ Φ). 6. (a) Let G = (R − {0}, ·) and let ϕ : G → G be defined by ϕ(x) = x2 . We first note that the identity on G is 1 ∈ R, and that ϕ(1) = 1 as it must for a homomorphism. Then we also see that ϕ(xy) = (xy)2 = x2 y 2 = ϕ(x)ϕ(y) and hence ϕ is a homomorphism. If ϕ(x) = x2 = 1, then we must have x = ±1 so that Ker ϕ = {1, −1}. Note this also shows that ϕ is not an isomorphism. (b) Let G = (R, +) and let ϕ : G → G be defined by ϕ(x) = 1 + x. Since ϕ(x + y) = 1 + (x + y) 6= (1 + x) + (1 + y) we see that ϕ is not a homomorphism. We could also observe that ϕ(0) = 1 which is not the identity on G (which is 0). (c) Let G = (R, +) and let ϕ : G → G be defined by ϕ(x) = kx for some fixed k ∈ R. (Assume that k 6= 0 or ϕ is just the zero map.) Now we have ϕ(0) = 0 as it must for a homomorphism. Furthermore, we have ϕ(x + y) = k(x + y) = kx + ky = ϕ(x) + ϕ(y) so that ϕ is a homomorphism. Since ϕ(x) = kx = 0 implies that x = 0, we see that Ker ϕ = {0} and ϕ is an isomorphism. 7. (a) Let ϕ : G → G′ be a homomorphism with kernel Kϕ . We first show that Kϕ is a subgroup. If k1 , k2 ∈ Kϕ , then ϕ(k1 k2 ) = ϕ(k1 )ϕ(k2 ) = e′ e′ = e′ so that k1 k2 ∈ Kϕ . Next, if k ∈ Kϕ , then ϕ(k) = ϕ(ke) = ϕ(k)ϕ(e) implies that ϕ(e) = e′ so that e ∈ Kϕ . Finally, for any k ∈ Kϕ we 3 have e′ = ϕ(e) = ϕ(kk −1 ) = ϕ(k)ϕ(k −1 ) = e′ ϕ(k −1 ) = ϕ(k −1 ) so that k −1 ∈ Kϕ . This shows that Kϕ is a subgroup of G. Now, to show that Kϕ is invariant (i.e., normal), simply note that if g ∈ G and k ∈ Kϕ , then ϕ(gkg −1 ) = ϕ(g)ϕ(k)ϕ(g −1 ) = ϕ(g)e′ ϕ(g −1 ) = ϕ(gg −1 ) = ϕ(e) = e′ so that gkg −1 ∈ Kϕ . Hence Kϕ is a normal subgroup of G. (b) Define the center of G by Z = {z ∈ G : zg = gz for all g ∈ G}. To show that this is a subgroup, we note that if z1 , z2 ∈ Z, then for any g ∈ G we have z1 z2 g = z1 gz2 = gz1 z2 which proves closure; associativity follows from associativity in G; Z contains the identity since eg = ge; and if z ∈ Z, then zg = gz so multiplying from the right and left by z −1 yields gz −1 = z −1 g so that z −1 ∈ Z. Therefore Z is a subgroup of G. Finally, it is easy to see that Z is normal because gzg −1 = zgg −1 = z ∈ Z. 8. (a) The group D2 consists of the set {e, a, b, c} where a and b represent rotations by π about the two symmetry axes of the rectangle, and c represents a rotation by π about the perpendicular axis. Clearly a2 = b2 = c2 = e. Now consider the product ab. Labeling the corners of the rectangle for ease of identification, we see that starting from 3 2 b 4 1 a we first apply b to obtain 4 1 b 2 3 a and then a to obtain 4 1 4 b 2 3 a which has the same net effect as simply c, and hence we see that ab = c. To evaluate ac, we apply a to the last figure above (which was just the result of c applied alone) to obtain 4 1 b 2 3 a which is the same as only b, so that we easily find e a e e a a a e b b c c c b ac = b. Continuing in this manner b b c e a c c b . a e However, you can save yourself some work if you realize that each row and column must contain all of the group elements (this is just the rearrangement theorem), so if you find three of them, then you know the fourth. (b) Let us write D2 = {e, a, b, c} = {g1 , g2 , g3 , g4 }. Multiplying from the left by e doesn’t change anything, so gi = egi = gei corresponds to the identity permutation pe = (1)(2)(3)(4) ∈ S4 . Next, multiplying by a yields {a, e, c, b} = {g2 , g1 , g4 , g3 } so that (a1 , a2 , a3 , a4 ) = (2, 1, 4, 3) which is the permutation pa = (12)(34) ∈ S4 . Multiplying by b we have {b, c, e, a} = {g3 , g4 , g1 , g2 } so (b1 , b2 , b3 , b4 ) = (3, 4, 1, 2) which corresponds to the permutation (13)(24) ∈ S4 . Finally, multiplying by c we have {c, b, a, e} = {g4 , g3 , g2 , g1 } so (c1 , c2 , c3 , c4 ) = (4, 3, 2, 1) and the permutation is (14)(23) ∈ S4 . This shows that D2 corresponds to the elements {(1)(2)(3)(4), (12)(34), (13)(24), (14)(23)} ∈ S4 . We must still show that this is a subgroup. It contains the identity, and using the fact that disjoint cycles commute together with the fact that 5 2-cycles are their own inverses, you can easily show that this is indeed a subgroup of S4 . For example, (12)(34) · (13)(24) = (14)(23) which corresponds to ab = c. 9. If D is reducible, then there is an invariant subspace and a similarity transformation R that takes the matrix representation of D to block triangular form T = R−1 DR. Solving this for D we have D = RT R−1 . Then if U = SDS −1 , we have U = S(RT R−1)S −1 = (SR)T (SR)−1 so that T = (SR)−1 U (SR). This shows that the transformation SR puts U into block triangular form, and hence U is reducible. 10. Let xi be a representative element of Ci , so that Ci contains all elements of the form gxi g −1 for all g ∈ G. Similarly, Cj contains all elements of the form hxj h−1 . If Ci Cj contains e, then for some g, h we must have e = (gxi g −1 )(hxj h−1 ). This is the same as g −1 h = xi g −1 hxj or xj = −1 (g −1 h)−1 x−1 h) ∈ Ci′ . Thus, if j 6= i′ , Ci Cj can not contain e = [e] = C1 . i (g Probably an easier way to say the same thing is let xi ∈ Ci , xj ∈ Cj and suppose that xi xj = e. Then xj = xi−1 ∈ Ci′ . But classes are either disjoint or identical, so this would imply that Cj = Ci′ . 11. It might be helpful to look at this diagram: φ G′ G ψ G/Kφ (a) Let the homomorphism φ : G → G′ have kernel Kφ , and define ψ : G/Kφ → G′ by ψ(X) = φ(g) where X = Kφ g. To show that ψ is welldefined, suppose that Kφ g = Kφ h. Then g = kh for some k ∈ Kφ , and hence φ(g) = φ(kh) = φ(k)φ(h) = e′ φ(h) = φ(h) so that ψ is well-defined. To show that ψ is a homomorphism, let X = Kφ g ∈ G/Kφ , Y = Kφ h ∈ G/Kφ and simply calculate using the fact that Kφ is a normal subgroup and φ is a homomorphism: ψ(XY ) = ψ(Kφ gKφ h) = ψ(Kφ Kφ gh) = ψ(Kφ gh) = φ(gh) = φ(g)φ(h) = ψ(X)ψ(Y ) . To show that ψ is surjective, let g ′ ∈ G′ be arbitrary. Since φ is surjective, there exists g ∈ G such that g ′ = φ(g). But then ψ(Kφ g) = φ(g) = g ′ so that ψ is surjective. Lastly, to show that Ker ψ = Kφ (the identity in G/Kφ ), suppose that ψ(Kφ g) = e′ . Then φ(g) = ψ(Kφ g) = e′ so that 6 g ∈ Kφ , and hence Kφ g = Kφ . Thus Ker ψ = Kφ which is the zero element (or the identity) in G/Kφ . This completes the proof that ψ is an isomorphism of G/Kφ onto G′ . (b) To say U (G) is degenerate means that the homomorphism ϕ : G → U (G) has a non-trivial kernel Kϕ , which from Problem 7a we know is a normal subgroup. But then part (a) of this problem shows that G/Kϕ is isomorphic to U (G), which thus defines a faithful representation of G/Kϕ . 12. From Example 16 we know that S3 has three classes C1 = {e}, C2 = {(12), (23), (31)} and C3 = {(123), (321)}, and hence Theorem 15 tells us there are three inequivalent irreps. Let D1 (S3 ) denote the one-dimensional identity representation p → 1 for every p ∈ S3 . Clearly the corresponding three characters (one for each class) are (1, 1, 1). From Example 21 we have a second one-dimensional induced representation D2 (g) = +1 for g ∈ C1 , C3 , and D2 (g) = −1 for g ∈ C2 . This then gives Pus the characters for each class as (1, −1, 1). From equation (11) we have µ nµ 2 = nG , and since nG = 6 for S3 , the only solution to this is n1 = 1, n2 = 1 and n3 = 2, and therefore there is a two-dimensional irrep. We need the characters χ3 (Ci ). Since the representation of e ∈ S3 is the two-dimensional identity matrix, we know that χ3 (C1 ) = 2. The character table up to this point is D1 D2 D3 C1 1 1 2 3C2 2C3 1 1 . −1 1 To fill in the remaining two entries we use the fact that the columns are orthogonal in the sense of equation (12), which in this (real) case becomes X χµ (Ck )χµ (Ci ) = µ 6 δki . nk If we let k = i = 2, then nk = 3 and this equation says 12 +(−1)2 +[χ3 (C2 )]2 = 2 so that we must have χ3 (C2 ) = 0. Alternatively, we could let k = 1 and i = 2 to write 1 ·1 + 1 ·(−1)+ 2 ·χ3(C2 ) = 0 to again conclude that χ3 (C2 ) = 0. Finally, the same argument applied to the first and third columns shows that χ3 (C3 ) = −1, and we have D1 D2 D3 C1 1 1 2 7 3C2 2C3 1 1 . −1 1 0 −1