Instructor’s Solutions Manual to accompany Discrete Mathematics, 3e By Edgar G. Goodaire and Michael Parmenter Both of Memorial University of Newfoundland‘This manual contains complete solutions to all exercises in Discrete Mathematics with Graph Theory, Third Edition, by Edgar G. Goodaire and Michael M. Parmenter. It is intended sclely for the use of instructors whom, we trust, will not make it available to students. ‘The solutions here have been read many times. We think that the num- ber of errors is small, but are always grateful for help in improving accuracy. ‘Our mailing address is below and our email addresses are edgar@math .mun.ca and michael 1¢math mun.ca, Please feel free to contact either of us at any time about any aspect of our book or this solutions manual, both of which have been improved substantially by comments received since the first edi- tion of this book appeared in 1998. Edgar G. Goodaire and Michael M. Parmenter Department of Mathematics and Statistics, Memorial University of Newfoundland St. John’s, NL Canada AIC 587 August 2005Solutions to Exercises Exercises 0.1 1, (@) [BB] True (b) [BB] Not valid (c) [BB] False (0? isnot positive.) (4) Not valid (©) True (Not valid (@) False 2. (a) [BB] True, because both 4 = 2+ 2 and 7 < 50 are true statements. (©) False, because one of the two statements is false. (©) (BB) False, because 5 is not even. (@ True, because 16-4 =} (©) [BB] Truc, since 9 = 3? is true (or because 3.14 < x). (D True, because (—4)? = 16 is true. (@) [BB] Truc, because both hypothesis and conclusion are true. (h) False, because the hypothesis is true but the conclusion is false. ( [BB] Nota valid mathematical statement. ( True, because both statements are true (K) True, because this is an implication with false hypothesis. () False, because one of the statements is false while the other is true, (m) False, because the hypothesis is true but the conclusion is false. (a) [BB] False, because the area ofa circle of radius r is not 2rr and its circumference is not mr? (0) False, because the hypothesis ofthis implication is true, but the conclusion is false. (p) (BB] This is true: The hypothesis is true only when a > b and b > a, that is, when a = b, and then the conclusion is also true. (q) This implication is true because the hypothesis is always false. 3. (@) (BB) fx > 0, then 2 > 0. (©) Ifa and b are rational numbers, then abs a rational number. (6) If f isa differentiable function, then f is continuous (@) (BB) 1fG is a graph, then the sum of the degrees of the vertices of G is even. (e) [BB] If A is a matrix and A # 0, then A is invertible. (®) If P is a parallelogram, then the diagonals of P bisect each other. © Hnisaneven integer, then n <0. (h) Iftwo vectors are orthogonal, then their det product is 0. (@ If nis an integer, then <2 is not an integer. () mis a natural number, then n +3 > 2. 4, (a) [BB] True (the hypothesis is false). (b) True (hypothesis and conclusion are both true).Solutions to Exercises (©) [BB] True (the hypothesis is false). (d) False (hypothesis is true, conclusion is false). (©) [BB] False (hypothesis is true, conclusion is false: /Z = 2). (True (g)(BB]True ¢) True) [BB) True (the hypothesis is false: Vz? = |x|) @ True (k) [BB] False (1) True (@) (BB) a? < 0 and ais areal number (more simply, a = 0). (b) zis not real or 2? + 1 4 0 (more simply, «is any number, complex or real). (©) (BB) z 4 1andz # ~1, (d) There exists an integer which is not divisibie by a prime. (©) [BB] There exists areal number 2 such than < z for every integer n. ( (ab)c = a(be) for all a, b,c. (@) [BB] Every planar graph can be colored with at most four colors. (8) Some Canadian is a fan of neither the Toronto Maple Leafs nor the Montreal Canadiens. (@ There exists x > 0 and some y such that x’ + y? < 0, @ 22 202<-2. (&) [BB] There exist integers a and 6 such that for all integers q and r, b # ga +r. (©) (BB) For any infinite set, some proper Subset if not finite. (am) For every real number , there exists an integer n such that 2 ora>yora>z. (p) There exists a vector inthe plane and there exists a normal tothe plane such thatthe vector isnot ‘orthogonal to the normal. 5. (@) [BB] Converse: If 2 is an integer, then $ and $ are also integers. Contrapositive: If # is not an integer, then $ (b) Converse: Asa? =1, Contrapositive: © # Land 2 # —1 +2? 41. (©) Converse: If = 1+ V5 orz=1— VB, then 2? = 241 Contrapositive: If # 1+ V5 and x #1— V5, then 2? 2+. (@) Converse: If n? + n.—2 is an even integer, then n is an odd integer. ‘Contrapositive: If n? + 2 ~ 2 is an odd integer, then n is an even integer. (©) (BB) Converse: A connected graph is Euletan: CContrapositive: Ifa graph is not connected, then itis not Eulerian, (Converse: a = 0 orb =0— ab =0. Contrapositive: a # 0 and b #0 ab #0. (@) Converse: A four-sided figure is a square. Contrapositive: Ia figure does not have four sides, then it isnot a square (8) [BB] Converse: Ifa? = 6? +c then ABAC is aright triangle, Contrapositive: If a? # 6? + c?, then ABAC is not a right triangle. snot an integer of # isnot an integer.Section 0.1 3a (i) Converse: If p(:z) is a polynomial with at least one real root, then p(:r) has odd degree. Contrapositive: If p(2) is a polynomial with no real roots, then p(zr) has even degree. (@ Converse: A set of at most n vectors is linearly independent. Contrapositive: A set of more than n vectors is not linearly independent. (&) Converse: I fis not one-to-one, then, for all real numbers z and y, x # y and a? + zy + y?-+ arty Contrapositive: I f is one-to-one, then there exist real numbers z and y such that 2 = y or Piaytytety £0. (1) [BB] Converse: If f is not one-to-one, then there exist real numbers x and y with # y and etaytyt+rty=0. ‘Contrapositive: If f is one-to-one, then for all real numbers x and y either x = y or 2? + zy + prot 0. 7. (a) [BB] There exists a continuous function which is not differentiable. (©) % > O forall real numbers 2. (c) [BB] For every real number a. (@ Forevery set of primes p1, Pay... +P there exists a prime not inthis st (© [BB] For every positive integer n, there exist primes p1,p25....Pe Such that (f) For every real number x > 0, there exists areal number a such that a? = z. (g) [BB] For every integer n, there exists an integer m such that m 0, there exists areal number y > 0 such that y < 2. @ (BB) There exists a polynomial p such that for every real number 2, p( 2) # 0. @) For every pair of real numbers x and y with z < y, there exists a rational number a such that e 0. (0) For any integer n,n i not both even and ocd. (©) Forall integers a,b,c, a +88 #0. 8. Ifa given implication “A —+ 3" is false, then Ais true and 3 is false, ‘The converse, “B + Ais then true because its hypothesis, B,is false. It is not possible for both an implication and its converse to be false. 9. First we remember that x and y is te i and y are both true and false otherwise. Now p + q means p—qand q— p. Also = Pipa Pe A. p> gis true if pis false or if pis true and g is true. B. q— pis tue if q is false or if qis true and pis true. Ifp and q are both false, both statements A and B are tue, so p “+ q is true. Similarly if both p and q are true, then statements A and B are again true, so p + q is true. Thus p <> q is true if p and g have the same truth values. Suppose p and q have different truth values. To be specific say p is true and q is false. If pis true and q is false, we see that statement A is false, so A and B is false. Similarly if p is false and q is true, then statement B is false, so A and B is false. This verifies statement (+).4 Solutions to Exercises Exercises 0.2 1 . (@) [BB] z= -2 (0). (@) (BB] Hypothesis: a and b are postive numbers. Conclusion: a + bis positive. (©) Hypotesis is argh ange angle wi typtmse of enh ad te ter sides of eg and Conclusion: a? + i? = 2, (©) (BB) Hypothesis: pis prime, Conclusion: pis even (@) Hypothesis: n > Lis an integer. Conclusion: n is the product of prime numbers. (©) Hypothesis: A graph is planar. Conclusion: The chromatic number is 3. (@) [BB] a and b are positive is sufficient for a +6 to be postive; a +b is positive is necessary for a and b tobe positive. (©) A right angled triangle has sides of lengths a,5,c,¢ the hypotenuse, is sufficient for a? + 6® = 2 a? +H = cis necessary for a right angled triangle to have sides of lengths a,b,c, ¢ the hypotenuse, (© (BB) pis a prime is sufficient for p to be even; pis even is necessary for p to be prime. (4) n> Lan integer is sufficient forn tobe the product of primes; n a product of primes is necessary for nto be an integer bigger than 1. (©) A graph being planar is sufficient for its shromatic number to be 3. Chromatic number 3 is necessary fora graph tobe planar. ji ©ppple=4 (89,112 @)VZand a=, @ ©vimd 7 Or=sy . A can easly be proven false with the counterexample 0. No single counterexample can disprove a statement claiming “there exits” so we prove B directly. B is false because the square of a real number is nonnegative [BB] This statement is true. Suppose the hypothesis, x is an even integer, is true, Then x = 2k for some other integer k, Then x +2.= 2k +2 = 2(k-+ 1) is also twice an integer. So x + 2 is even, The conclusion is also true. ‘The converse is “x + 2 is an even integer —+ zis an even integer.” This is true for suppose that the hhypothesis, x + 2 isan even integer, is tue. Then 2 +2 = 2k for some integer k, so x = 2k — 2(e~ 1) is also twice an integer. The conclusion is also true. ‘This is tue. Let A be the statement “zc is an even integer” and let B be the statement «+ 2 is an even integer". In Exercise 5, we showed that A —+ Bis true and, in Exercise 6, that the converse B —> Ais also true. Thus A ++ 3 is also true. .. (a) Ais false: n = 0 is a counterexample,Section 0.2 5 (©) Converse: If; is not an integer, then mis an integer. This is false: 4 is a counterexample 2 = 1), aH Contrapositive: If ;!; isan integer, then mis not an integer. This is false: n — 0 is a counterex- ample. "Negation: There exists an integer n such that = is an integer. This is true: Take n = 0. 9. (a) m prime + 2” — 1 prime. (b) n prime is sufficient for 2° — 1 to be prime. (©) Alis false, For example, n = 11 is prime, but 2!” n= 11is counterexample to A. 2047 23(89) is not. The integer (@ 2" 1 prime + n prime, (©) The converse of Ais true, To show this, weestablish the contrapositive. Thus, we assume n is not prime. Then there exists a pair of integers ¢ and b such that a > 1, > 1, and n = ab. Using the hint, we can factor 2° — 1 as ma 1= (29) (2% — 1)[(2%)P? + (22)P°? + + 2" + 1], Since a > Landb > 1, we have 2°— 1 > Land (2°)*-! + (28)2-4.+-42°-41> 1,502" isthe product of two integers both of whick exceed one. Hence, 2" — 1 is not prime. 1 10, {BB} The converse is the statement, “A continuous 7 function is differentiable.” This is false. ‘The abso- Tue value function whose graph is shown to the right 2 is continuous, but not differentiable at x = 0. 1. @ %n=1) Hn (©) An > Anmty Anat Anode voor Aa > diy Aa + An: 12, [BB] A is true. It expresses the fact that every real number lies between two consecutive integers. Statement B is most definitely false. It asserts that there is a remarkable integer n with the property that every real number lies in the unit interval between n and +1. 13, Ais false; B is true. There can be no y with the property described since y is not bigger than y + 1; 2 = y + 1 provides a counterexample. To prove B, we note that for every real number x, we have 2+1>candsoz+1isasuitable y. 1 <= is rational but nq is not an integer. 14, (@) Thisis false. Suppose such an n exists. Then q = —— (b) This is true, Given a rational number g there exist integers m and n, n # 0, such that g = ™ ‘Then ng = mis an integer. " 15. (a) Since nis even, ‘2k for some integer k Thus n? +3n = 4k? + Gk = 2(2k? + 3k) is even too. (b) The converse is the statement n?-+3n even —» neven. This is false and n = 1 is counterexample. 16. (a) [BB] Case 1: a is even. In this case, we have one of the desired conclusions. Case 2: a is odd. In this case, a = 2m + 1 for some integer m, 50 a + is even, another desired result. (©) [BB] n?® +n = n(n +1) is the product of consecutive integers one of which must be even; so n? + nis even. 2m+2 (m +1)Solutions to Exercises 17. n? =n +5 = n(n—1) +5. Now either n — 1 or n is even, since these integers are consecutive. So ‘n(n — 1) iseven, Since the sum of an even integer and the odd integer 5 is odd, the result follows. 18, [BB] 22? — 42 + 3 = 2(c? — 22) +3 = 2[(2 - 1)? —1] +3 = 2(2— 1)? + Lis the sum of Landa nonnegative number. So itis atleast 1 and hence positive. 19. For a? — b? to be odd, itis necessary and sufficient for one of a or b to be even while the other is odd, Here's why. Case is a, beven. In this case, a = 2n and b = 2m for some integers m and n, so a? ~ b? = 4n? — 4m? = 4( Case it: a,b odd. In this case, a = 2n + 1 and b = 2m +1 for some integers m and n, so a? — 0? = (4n® + 4n +1) ~ (Am? + 4m-+ 1) = 4(n? +n —m? — m) is even. Case ii: a even, b odd. In this case, a = 2n and b = 2m+1 for some inlegers n and m, soa? —b? = 4n® —(4m?-+4m-+1) = 4(n? — m? — m) — Lis odd. Case iv: a odd, beven. ‘This is similar to Case if, and the result follows. im) 20, [BB] (—+) To prove this direction, we establish the contrapositive, that is, we prove that n odd implies n? odd, For this, if n is odd, then n = 2m ++1 for some integer m. ‘Thus n? = 4m? + 4m +1 = 2(2m? + 2m) + 1 is odd. (—) Here we assume that n is even. Therefore, n 4m? = 2(2m) which is even, as required, ‘2m for some integer m. Son? = (2m)? = 21 Weasottats+ 1 22itntoayte>0 1 Proof, (—») We offer a proof by contradiction. Suppose # + + > 2 but x > 0 is note; thus 2 < 0.12 =0, Lis not defined so < 0. Inti case, however, 2+ > <0, contain (Conversely, assume that x > 0. Note tha: (2~ 1)? > O implies 2? — 2 +1 > 0, which in turn 2.as required. . implies 2? +1 > 2x. Division by the positive number gives 2 + — 22. [BB] Since nis odd, m= 2k-+ 1 for some integer k. Case 1: kiseven. Tm this case k = 2m for some integer m, son = 2(2m) +1 = drm +1. ‘Case 2: is odd. Inthis case, k = 2m +1 for some integer m, son = 2(2m +1) +1=4m+3. Since each case leas to one ofthe desired con:Iusions, the result follows.Section 0.2 7 23. By Bxercise 22, there exists an integer k such that n = 4k-+ Lorn = 4k-+ 3. Case 1: n= 4k +1, If k is even, there exists an integer m such that k = 2m, son = 4(2m) +1 = 8m-+1, and the desired conclusion is true. If k is odd, there exists an integer m such that k = 2m +1, so 4(2m +1) +1 = 8m-+5, and the desired conclusion is true. Case 2: n= 4k +3. I k is even, there exists an integer m such that k = 2m, som = 4(2m) +3 the desired conclusion is true. If k is odd, there exists an integer m such that k 4(2m +1) + 3.= 8m + 7, and the desired conclusion is true. In all cases, the desired conclusion is te, 24, [BB] If the statement is false, then there does exist a smallest positive real number r. Since ris positive and smaller than r, we have reached an absurdity. So the statement must be true. 25. We give a proof by contradiction. If the result is false, then both a > \/ni and b > s/n. (Note that the negation of an “or” statement is an “and” statement.) But then n = ab > Tay =m, which isn’t true 26. [BB] Since 0 is an eigenvalue of A, there is a nonzero vector x such that Ax = 0, Now suppose that A is invertible. Then A~(Ax) = A~"0 = 0, sox = 0, a contradiction, 27. (a) The given equation is equivalent to (b — 5)V2 = 3 —a. If b 4 5, then v2 ‘number. This is false, Thus b = 5, 80 a () Note that (a+ V2)? = (a? +26) + 2aby/2. Thus, if (a-+ bV3)? = 3453, then a? +20? = 3 and 2ab = 5, by part (a). The second equation says a # O and b 0. Since a and b are integers, it follows that a? > 1 and 6? > 1 and a? +24? > 3 with equality if and only a = 1 = But then Dab #5, 28, [BB] Observe that (1+a)(1+6) is 1, There are two possibilities: in which case a =b Ef isa rational +atbbab 4a=1+b= |. Thus 1+a and 1+b are integers whose product in which case a = 6 =0,orl+a=1+6=-1, 28, Westy cotason, Sipe ist inonal Ten itsatoml theeextinee mands £0, mehiht! = 2 Size! 40, neta ao ham £0, Now ! = 2 ins = 3 is rational number, a contradiction 30. We give a proof by contradiction. Assume that eis rational, bis irational and a +b s rational. Then a-+5= 2 for integers m and n,n #0. Since ais rational, a = § for integers k and €,€ 4 0. Thus m_k_ml—kn n nt ne b is the quotient of integers with nonzero denominator. This contradicts the fact that b is not rational. 31. (BB] We begin by assuming the negation of the desired conclusion; in other words, we assume that there exist real numbers «, y, z which simultaneously satisfy each of these three equations. Subtracting the second equation from the first we see that + 5y ~ 4z = —2, Since the third equation we were given says x + 5y —4z = 0, we have x + 5y ~ 4z equal to both 0 and to 2. Thus, the original assumption has led us to a contradiction.Solutions to Exercises 32. () [BB] False: « = y = Ois a counterexample. (6) False: a = 6 isa counterexample. © (BB) Fase: 2 (© Fatse: a = v3.6 (© (BB) Fase: a counterexample -V2is a counterexample V2is a counterexample. be VO Wee 165 — sac > 0, VF Tae (® The roots of the polynomial az? + bx +c are a =o Bias ial nd ot 040th fra roti wo itn el mame = =

1,s0n® (n) 2 1. 33. The result false, A square and a rectangle (which is not a square) have equal angles but not pairwise proportional sides. 34. (@) [BB] Since n? +1 is even, n? is odd, som must also be odd. Writing n = 2k+1, then n?+1 = 2m says 4k? 4 dk +2 = 2m, som = 2k? + 2k +1 = (k-+ 1)?+ K? is the sum of two squares as required, (b) (BB) We are given that n?-+1 solution to (a) shows that 1m for n = 4373 and m = 9561565. Since n = 2(2186)-+1, our K+ (b+ 1)? where & = 2186, Thus, 9561565 = 2186? + 2187. 35. (a) 242 4-1 = 4(2!) 41 = 4(2")! +1. Applying the given identity with 2 = 2°, we get 2B ey = (2.2 4.2" 4 (2.224 4 1) (pam 4 aya — amt 42), With n = 4, we get 298 + 1.= (29 +28 + 1)(29 — 25 +1) = 545(481).. (6) 2% — 1 = (2! — 1)(2* +1) = (2° — 1) 2” + 1)(545)(481) (using the result of part (a)) 511(618)(545)(481) 36. If the result is false, then f(n) = ao + ayn +++» + aun! for some t > 1. Since f(0) = a9 = p is prime, f(n) = p-+ng(n) for g(n) = a1 ~ azn +--+ ayn'“, Replacing n by pn, we have (pn) = p+ npg(pn). The right hand side is divisible by the prime p, hence f(pn) is divisible by p. But f(pn) is prime, by hypothesis, so f(pn) =p. This means g(pn) = 0, contradicting the fact that a polynomial has only finitely many roots. 37. We offer a proof by contradiction. Suppose all the digits occur just a finite number of times. Then there is a number m, which has the property that after n digits in the decimal expansion of x, the digit 1 no longer occurs. Similarly, there is a number 1g such that after ng digits, the digit 2no longer ‘occurs, and s0 on. In general, for each k = 1,2,...,9, there is a number ny such that after my digits, the digit k no longer occurs. Let N be the largest of the numbers n,ra,... ry. Then after IV digits in the decimal expansion of x, the only digit waich can appear is 0. This contradicts the fact that the decimal expansion of x does not terminate.Chapter 0 9 38. We have proven in the text that V2 is irrational. Thus, if V2”” is rational, we are done (with a = » = VB), On the other hand, if V2"” is irrational, then let a = V/2”” and b = V2 in which case ab = v2" = 2is rational. Chapter 0 Review 1, (@) This implication is true because the hypothesis is always false: a — b > 0 and b — a > O give ‘a> band b > a, which never holds. (b) This implication is false: When a the hypotheses are true while the conclusion is false. 2. (a) wis areal number and x <5. (©) For every real number ©, there exists an intogern such that n <=. (©) There exis positive integers 2, y, = such that 2° + y® = 2° (@) There exists a graph with n vertices and n + 1 edges whose chromatic number is more than 3. (©) There exists an integer n such that for any rational number a, a # n. ® a# 00rd 0. 3. (@) Converse: If ab isan integer, then a and b are integers. CContrapositive: If ab isnot an integer, then either a or bis not an integer. Negation: There exist integers a and b such that ab isnot an integer. (©) Converse: If 2? is an even integer, then 2 is an even integer. ‘Contrapositive: If 2? is an odd integer, then «is an odd integer. Negation: There exists an even integer x such that is odd. (© Converse: Every graph which can be colored with at mast four colors is planar. CContrapositive: Every graph which cannot be colored with at most four colors is not planar. Negation: There exists a planar graph which cannot be colored with at most four colors. (© Converse: A matrix which equals its transpose is symmetric, ‘Contrapositive: Ifa matrix does not equa its transpose, then itis not symmetric Negation: There exists a symmetric matrix which is not equal to its transpose. (© Converse: A set of at least n vectors is a spanning set. ‘Contrapositve: A set of less than n vectors is nota spanning set. Negation: There exists a spanning set contsning less than n vectors. (® Converse: Ifz > ~2and x <1, then 2? +2—2<0. Contrapositive: If < —2 or x > 1, then 2? +2 —2>0. Negation: There exists an x < —2 or > I such that2® +2—2<0. 4. (@) Ais false: a 1, v = =1 provides a counterexample. (b) Converse: Given four integers a, b, u,v with u # 0, # 0, ifa = b = 0, then au + bu = 0. Negation: There exist integers a,b, u,v, u ¥ 0, v #0, with au + bv = Oand a # 0 orb #0. Contrapositive: Given four integers a, 6,1, v with u # O and v # 0, if a # 0 or b # 0, then aut bv 40.10 10, U1. 12, 13. 4. The desired formula is ab = Solutions to Review Exercises: (©) The converse is certainly true since Ou + Ov = (@) The negation is true: Take The contrapositive of A is false since A is false. = Landy = =1 (@) There exists a countable set which is infinite (b) For all positive integers m, 1 —1 and x < 2,902 +1 > Oand « ~2 <0, But then 2 (+ 1)(@—2) <0, acontradiction. 2 ‘Suppose, to the contrary, that is the largest negative rational number, where «and y are positive v = aoe integers. Then == is rational and — < =, 50 — 2y By Sy By ‘we have a contradiction. > ~5. Since — is a negative rational number, v by ‘We use a proof by contradiction which mimics that proof ofthe irrationality of v2 given in Problem 8. ‘Thus, we suppose that V3 = ¢ is rational and hence the quotient of integers a and b which have no factors in common. Squaring gives a = 34? and so a = 3k is a multiple of 3. But then 9K? = 36%, so 3k? = b?, This says that b is also a multiple of 3, contradicting our assumption that a and b have no factors in common.Section 1.1 " 16. a 18, 19, Let the rational numbers be ¢ and §. We may assume that a, b,c,d are positive integers and that 4 < §. Thus ad < be. The hint suggests that £24 is between and §, and this is the case: ¢ < ¢45 is equivalent to a(b + d) < b(a +c) and $5 <§ is equivalent to (a + c)d < (b-+ d)e, both of which are true because ad < be. (Ona standard checker board, there are 32 squares of one color and 32 of another. Since squares in ‘opposite comers have the same color, the hint shows that our defective board has 32 squares of one color and 30 of the other. Since each domino covers one square of each color, the result follows. (@) We leave the primality checking of f(1),..., £(39) to the reader, but note that /(40) = 412. (b) f(H? + 40) = 40? + 80K? + kt + 40-4 7 + 41 = AA + SIA? + 41? = (1? 4 41)? - P= (8 +41 + k)(K? +41 — i). ‘The answer is no, since 333333331 = 19607843 x 17. Exercises 1.1 1. @ (BB) [Ta [=a Ca) VP | PA (Ca) ve) rirlel 3 T rie|r| 3 r e\rlF] F F lele|r| 7 rE ©) (pT ToT Ga) a Tena TAD VCH) — a) tirlr| rt |r T rlelr| r |e r ri[r|r| r |r r Fle|r| r |r FE © [pT 4 [ave [pAGV») [= ACV) [= CAVE) oP r[rp cr] 7 F F rie} r | or F F r[r}r] oF r F elel el] F r FE ) BB) Pp Ta Tr [=a pv Ca) [Ove [P| Cove [COV CAC HVa) ti[rirfel| 3 r fri ¢ F rle|r|r| 3 r |r| r F Fi[rirlr| F r |r| r iz Flelrir| 3 ro of[r| r F rirjrle| 3 ro olrl| F F rlele|r| 3 rofel F F Fi[rilrle| F r |rl| r r Flelelr| 7 r_f[rl 3 EF12 Solutions to Exercises © pl a[ + [ar [p> Gr) [pag] @Ad Vr | GF) > (Aa VF) r(ri[r] 7 T Tr T T rlrir| 3 T FP Tr Tr Fl[r|r| 7 r F T Tr FlFir| 7 Tr P Tr Tr r\r|r| F F Tr Tr Tr rlFlr| 3 r Fr F P Fl[rir| F Tr P F F Flejel 3 r FE FE FE 2. (@) If p ~+ qs false, then necessarily pis te and q is false. (This is the only situation in which 1 qis false.) We construct the relevant row of the truth table for (p A (=)) V ((-p) ~4)- Pa [-a PACH [Oa tlelr|_ r [rl + RC Va= 9 T (©) [BB] There are thre situations in which p — q is true. The question then is whether or not the twuth value of (p/\(=)) V ((-p) —* ) isthe same in each of these cases. We construct a partial teuth able, Pla [a [PACA [= [ Ge) 4 TPAC) V (Co) = 9) t(rle| f lel 3 T Flelr| fF |r| Fr F ‘As shown, (p A (=g)) V ((-p) —+ a) has different truth values on two occasions where p —» q is tue, s0 it is not possible to answer the question in this case, 3. (BBI [yTo Tr [= [7 [eACH) p> AC [Cae ririr|r[e| F F Fl fr relies) Va [p> GACY V Fo (Cs) va} T T 4. (DTaTF Te [= a a) pS AC = [Ce Flelelelr] F r rl 4 reid vd [P= GAG Fe (Gs) Va] F Tr 5. @) (BBI[p [a [pag] Va] Ag) > pVa) tir] r| 7 T Tle) F | 7 T F\r| re | or Tr FiF| FILE r ‘The final column shows that (p A q) — (pV 4) is true for all values of p and g, so this statement isa tautology.Section 1.1 13 (b) [BB] Ba [=e [eaAa | =a [pV Ea) | (CAD ABV ED) tirlF| Ff fFl P tlrle| er |r] ¢ PF F\rir| ro fel F F Fle|r|_r |r| 7 FP ‘The final cofumn shows that (=p) Aa) A (PV (al) Is false forall values ofp and g, 80 this statement isa contradiction. 6 © Glass o>0 since = (pte or ince q+ (p —» 4) is true r[r| Tr T rle| F T all values of p and q, this ele 8 statement isa tautology. FiF T © [pfa [oral] a] (VCH ADA VCA tit| tr lFlFl F P rlelelelr) 3 F F[r| re frlr| oo F rle| e [rir] 7 F Since (pa) A (CP) V (oa) is fle Forel values of p and g, this statement is a contradiction. 7. (a) [BB] Pla]? p>a]a—r[ >) AG—> Por tirlr} rt |r T T rle\r| r | or F r Fl\r|r| 7 | or T r Flelr| r | or T r rirlr| 7 |e F F rlelr| Fe |r F r Fi[rle| 7 | F F r Flele| 7 |r T r = OAG= =P 7 = Since [(p + 4) A (a —> 7] —» [pis tre forall Z values of p,q and , this statement isa tautology T T T T T T (©) (BBY Ifp implies q which, in turn, implies r, then certainly p implies r. 8, We must show that the given “or” statement can be both true and false. We construct truth tables for ‘each part of the “or” and show that certain identical values for the variables make both parts T’(so that the “or” is true) and other certain identical values for the variables make both parts F (so that the “or” is false). Pll l=) = ba pyle = Call tir|r| ele Tr Tr Fle|r|r rE FE4 Solutions to Exercises Platt [elt] =] Civele= (oval rir{r[rirlel Tr rlrle|ririF| F F Sa Ca) =F Te = (COVAIVICHD +71 Fl oT T rl_F F 9. We are given that Ais false for any values ofits variables. 10. ML. (a) [BB] An implication p > q is false only if pis true and qis false. Since A is always false, A —> B is always true. So a tautology. (b) An implication p — q is false only if p is tre and q is false. Since Ais false and the tautology B is true for any values of the variables they contain, B —+ is always false. So itis a contradiction. (a) The tables below show that when all three variables p, q and r are false, p —+ (q — r) is true, whereas (p — q) —» r is false, Thus these statements have different truth tables and hence are not logically equivalent. plalrlaor p> @> Pla] [p>a[ eo gr Flele Tr r Flelel ¢ rE (b) The compound statement is false. ‘The truth table shows that (P¥.a) + (V4) is true for all values of p and g, 80 itis a ‘Columns three and five are the same. So the truth values of p V q and 7 (p+ q) are the same for all values of pand q. Thus these statements are @ (BB) Ta [pve tlr| rlr| Fi[r| 7 ele F ©) [pT a [=e | Ca) Xa Tov (Ce) 49) [OV Ca) Aa va tilrlr| F Tr Tr tlr|r| F Tr Tr F[ri|r| 7 Tr T rle[r| F F F © (BB) (> Ta [pva] eva] @¥a) va) tir) e lr Tr tle} 7 |r Tr F[r| r | 7 Tr Flel ele Tr. tautology. © (pTaTevaTpoa [= beg rir} re |r F rlrFl 7 | F Tr F[r| rc | F Tr Flele |r Fr logically equivalent,Section 1.2 15 Exercises 1.2 1. (BB) (ldempotence) The truth tables at the right show that p Vp <=> pand pipve| [plpar pAp => p. T| 7 rir rl Fr FIFE 2. (Commutativity) The truth tables show that pVq <=» qVpand pAg <=> Ap. 2a [evalavp pa [pray ane rir| rf[r ri[r| r |r rir] 7 |r tTlrl Fe | F F[r|r |r F\r| re | F Flr| re fr Flr[r|er 3. [BB] (Associativity) The equality of the fith and seventh columns in the truth table shows that (@va) Vr) <= @V@vr)). @vavr la py @vry SASS SASS SSSA sasssss| RSA [EI SSNS SS massssss [el ANAS SSS 2 Eg ‘The equality of the fi (PA@Ar)). seventh columns in the truth table shows that ((p Aq) Ar) <=> PAG @ADAT a. PA@AT) asa g pass aS sores iS Sas £ ass ‘e Tr P F F r F F F if 4, (Distributivity) The equality of the fi and eighth columns in te truth table shows that pV (q A 1) = (PVaA® = ") oss | BVWAN IP ls sassssssle @VaAGVT) sass |! sassssssle ASS S8h massa l nasssssh SASS TSS16 ‘The equality of the fifth and eighth columns in the truth table shows that p A (q V r) (pA) V (pAr)). Solutions to Exercises FINCEDNED ADV PAT nassssssle SASS ssS55 AAs ASS | ass |S SSS a4 nas |S 5. [BB] (Double negation) The equality of. the first and third columns inthe truth table shows that p <=» — (-p). P [I= Ca) tlr| Tr Flr| F os (The Laws of De Morgan) The equality ofthe fourth and seventh columns of the truth table shows, that =(p Vg) => ((>p) A (~4)). P E @ =S55F SS55F =a SOV a) | =P [=a CHA Cal eo flrlel F ri lje|r| oF eo \TlF Fr r_|rjr|_ + ‘The equality of the fourth ((>p) V (m9). |. [BB] The two truth tables show, respectively, that p V1 <=> 1 and PAL = p. ‘The two truth tables show, respectively, that p VO <=> pand PAO => 0. 1. [BB] The two truth tables show, respectively, that (V(-p)) => Land OAD) = 0. ). The truth tables show, respectively, that (1) => Oand (-0) <=> 1. and seventh columns of the truth table shows that -(p Aq) <=> pa [pha] =@Ag) [9 [aT CoV Ca) rir] r| Fr |Fle Fr tlF| F| vr lFir T F[r| e| ro lr}e T Fle| rel r {[r|[r T Plijevi] [p]i [ear r|r| 7 rir| r Flr| r Fl[r| F p[O[pvo] [p[OTpao rir| rt | {rle| F Flr| F | [elel e& Pl levee [2] [e [=p [PAC tlr| vr |r| |rjrl F Firt| r [|r| |Fl[r] F =iTo] foy-oyi rir el[rlrSection 1.2 7 1, (BB) The third and sixth pla le—al-a| [Cd > Ce) ans et sow ee sat T\F| F T\|F F 9) = (Ca) Co). rlr| rt jrir| Felt irr] 12. The third and sixth pP[a[poa[ pala] @>aAG—p) ‘columns of the truth table , a a a a Show tat (peo) a ‘a r\rl |r |e F (@>a)A@—?)). rlel 3 a P a 13, [BB] The third and fifth columns of pa{e=al | Cova the truth table show that a a . a oa = (CPva) tll ple) Fr rll t |r| © Oo epvienal i Ba [PAG [WAT [VEDA for all values of p and q, r\T) tT | F T this atment in tlrle | 2 r ‘uology F\F| F T T (b) By one of the laws of DeMorgan, the negation is (~p) A (p Aq). By associativity this is logically ‘equivalent to [(=p) Ap] Aq <= 0Aq <=> 0. So the negation is a contradiction. 3, (@) (BB) Using one of the laws of De Morgan and one distributive property, we obtain (Aa) ¥ (Ce) va) = (Aa) ¥ PAC-9))] + PAGY(-a)] — (AX) — p. (©) The given compound statement is ofthe form x + (y —+ 2) which is equivalent to 2 —> ((-y) V 2) <> (-a) V(-y) V 2. (By associativity, no further parentheses are required here.) So the ‘given statement is equivalent to (~(p V r}) V ((-g) Ar) V (p Vr). By commutativity, this is (-@Vr))V (PVr)V (4g) Ar) <=> 1V((>gAr) <> 1. The given statement is a tautology! (©) Using associativity to avoid extra parentheses, te left side ofthe given statement is (>a V@— ni] = (Pp) Vav (>a) Vr <= (p)VIVr ‘The given statement is equivalent to [1 A (r —+ s)] which is logically equivalent tor —+ 4. @BB] [pTa[pAa[pv@Ad] © [p[a [Val pA@va, rir| 3 T rir] T Tlr| F r r\F| 7 T Fi[t| F P Fir] 7 F rle| e FE Fle| F F 5. (a) [BB] Distributivity gives [(pVa)A(-»)] => [(eA(=P)) V(QA(~P))] => [OV((-P)Ag)] () 44.18 Solutions to Exercises (b) We have (p— (qr) => (p> (al Vr) => (-p) V (>a) Vr > (OP) VEY (9) = WAH) V (79) > (WAH) > (9). © (60%) > (>) A> P) = (CP) Va) A((-9) VP) = (A (>a) V4) = (PA (>9)) Va) 4 ((PA (-9)) V (=) => ((pV.g) A ((-9) V9) A (PY (=) A (>a) ¥ (-P))) = (PVD) AYALA (9) Vp) = (PVa) A((-a) VP) = (C9) V (>a) Aa) = (P> (a) A((-9) +P) (PO (9). (@) (BB) a[(p + 9) V (PA (-9))] = [HP 9) AA (-9))] €=> [lp (-a)) A ((-p) V a). using Exercise 5(c). (©) This is an immediate application of absorption law 4(b) with p A (ng) in place of p and g A (=r) inplace of a. (Using property 12 and associativity, lp fa Vr)] <=> [(-v) Var] => ACM) Ye (by De Morgan) => [p A (na) + +] using 12 again. (@) ~(V a) V (-p) Aa] => [(-7) Aa] V [(-P) Ag] DeMorgan) <=> (-p) A [(-9) Vv 4) Gistributvity) => (—p) V1 => - 6. (PA (-a)) +g] => (Aa) Va] => [(P) Va) Va) = [-P) va). [A (a) > = (OPAC) VCP) = [P) av (>P)] = [P) val. So these are both logically equivalent to (—p) V4 7. (a) We must show that A v @ and BV @ have the same truth tables, given that A and B have the same truth tables. This requires four rows of a truth table. ABC [Ave|Bve rir|[r| rt |r rilr|r| vr |r Fle([t| v |r Flrle| Fe | F ‘The last two columns establish our claim. (b) We must show that A.A @ and B A € have the same truth tables, given that A and ‘B have the same ‘uth tables. This requires four rows of a truth table. E[AACT BAC SANS ass |s| fn iP Tr F ‘The last two columns establish our claim,Section 1.2 19 p [pvp 1. [BB] The truth table shows that idempotence fails. Tl F Fle Pla [pvalavy tir| re | 2. Commutativity sil holds.| 7] | 7 | 7 Fl[r| r |r Flelel| er pa]? [eve] @Vavr [avr [ov@ve) t[rlr| & Tr F Tr rlrlr| 7 F Tr P elrir| 5 F F F 3. [BB] Associativity holds, rlrlr| F r r r r|r|r| F P Tr F rlrlr| 7 Tr F T elrle| 7 Tr Tr Tr Flelrl & F F F 4, Just one distributive law holds. The table which follows shows that p V (q A r) is not logically equivalent to (pV 9) A(PY.r)y Pa | [ant [ow Gar) [pve pwr | @Va)A@YT) tlelrt re] 7 [rler F On the other hand, (p A (q V.r)) <=>" (pq) V (p Ar) as shown. avr [pAGYT |p: DAT [ADV AP ASST SASS sssS fs SS sSs88 |: SASS SSS sass S naa [S| AAA 6 na laws of De Morgan fail. The table shows that -(p Vg) is not logically equivalent to (p) A (~4). Pla [pva| Ova | P| 9] CAC) tlrle |r [r[rl - ‘The next table shows that “(p A q) is not logically equivalent to (=p) V (4). P14 [eng] Ghd |p [-a| CoV Ca) rlele |r [r[r FE pi jeyi rir| F 7. [BB] The truth table shows that pV 1 is no longer 1:Solutions to Exercises pO [pyo 8. The truth table shows thatp VO <=> pT FT TF FlF| F > [eleva 9. [BB] The truth table shows that (PV (-p)) = 1: [rte] or |r Flr| 7 |r 12, (BB) This isnolongertrue: |2{2{2—-¢| -# | Co) Ya Fl[r[ 7 |r] fF Finally, neither law of absorption holds: p v (pq) is not logically equivalent to p and pA(p V.q) is not logically equivalent top. pl a[pAcTpY@Ad| [ela [pvalpa@ya) rirlr FE rlir| F (@ (BV.a) A((-B) v (-9)) is notin disjunctive normal form because the terms are not joined by V. (©) BBB] (PA) V ((-p) A (-9)) isin disjunctive normal form, © [BB] pv ((-p) A 4) is notin disjunctive normal form: not all variables are included in the first term. © (PAg)V (Ce) A (4) Ar) is not in disjunctive normal form: not all variables are included in the first term, ©) (PAG Ar) V (=p) A (79) A (ar) is in disjunctive normal form, (@) (BB] This is already in disjunctive normal form! The definition permits just a single midterm, (b) (BB] One of the laws of De Morgan gives immediately that (p A.q) V (-((-p) Vg) <=> (Aa)v @ACa)). © 9) > (PVA) > CH) ADV(CH) ACD) VARY ACP): © (Va) ACP) V a) > [A (HY (9) Va (CB) V (-a))] > (ACB) V A(CAIV IGA) VAN (a)] > [AO VP) 69) © (> DAGAN] => K—VOAGAT] > (oP) AgAr)V(GAGAr)] => (Gp) Aa Ar)v (PAGAT)V (CP) AGAr))] => [((-P) AGAT) V (PAQAT)], omitting the repeated minterm at the final tp. © (Pvlertev-n))l) => (pvilanr)viaA(-r)]) = (pv (an) Vian) —> ADV ACA)V@AQV(aA-r)) => (PAV PACA) VaAr)) = (nan yvienan(-m)vi@n(~a)Ar)v(pA(—a)AC-r))vipnav(-w))V((~B)AGA(-w))) => (OAaAMV PAGACH)VEACA)ArIV PACA) A(-¥))¥(>P) AGA (-n)),omiting the repeated minterm atthe final step. ‘August De Morgan was born in Inia in 1806, ded in England in 1871 and lived his life without the sight of his right eye, which was damaged at bith, He was the fist Professor of Mathematics at University College, London, founded the London Mathematical Society and was its first president. He was apparently quite a man of principle. He refused to study for the MA degree because of a required ‘theological exam and twice resigned his chair at University College, on matters of principle. He wasSection 1.3 2 never a Fellow of the Royal Society because he refused to let his name be put forward and, similarly, refused an honorary degree from the University of Edinburgh. ‘De Morgan's mathematical contributions include the definition and introduction of “mathematical duction”, the most important method of proof in mathematics today. See Section 5.1. His definition ‘of a limit was the first attempt to define the idea in precise mathematical terms. Nonetheless, itis the area of mathematical logic with which De Morgan's name is most closely associated. The “Laws of De ‘Morgan’ introduced in this section, tether with their set theoretical analogues—(AUB)* = ANB, (An BY = A°U BE (see Section 2.2)—are used extensively and of fundamental importance. De Morgan also developed a system of notation for symbolic logic that could denote converses and con- tradictions. Exercises 1.3 1. (@) [BB] Since [p+ (q+ r)] => [p> ((-9) Vr)] > [(-P) V (7a) Vrs the given argument ccan be rewritten (Ce) vr] v a) a Covr which s valid by disjuntive syllogism, (b) [BB] We analyze with a plat [paar | -a[ rsa) truth table In ow one, tiz|t| 7 | rlFl F |e the premises are tue but tlelr] e |r ]r| 3 the conclusion isnot. F[rir] r |r fr] Ff fs ‘The argument isnot Flelr] r |r fr] 7 |e valid t\tlr| r |r fel or |e rlrlr| re |r irl or F\rle| 7 |r fel 3 + Flele| 7 |e |r| + (©) We analyze with a truth able Pla[tle=a[r=a] =P In row thre, the premises are t[r{r] tr po | or d+ true but the conclusion isnot. tlelr| e | F |r The argument isnot valid. Fi[rir| 7 |r| F {+ Flr[r| 7 | Fr | F r|rle| r | 7 |r |e tlele| e | or |r Flrle| 7 | 7 |r |. Flele| 7 |r |r |e (@) This can be solved using a truth table with 16 rows. Alternatively, we can proceed as follows. ‘Assume that the argument is not valid, This means that we can find truth values for p, q, r and s such that the premises are true but the conclusion is false. Since s —» (r V q) is false, we must have s true and r V q false. But this means both r and q are false. Since p + q is true and q is false, p must be false. But then q V (“r) is true and p A s is false, contradicting the truth of (av (~r)) ~+ (pA). Hence we have a contradiction, so the argument is valid.22 Solutions to Exercises (©) This argument is valid. The second and third premises give q by Modus Ponens, Together with the first prem (double negation and dijunctive syllogism), we get ~p. () This argument is valid, The first and third premises give +p by modus tollens. Together with the second premise (double negation and a secoad application of modus tollens), we get r 2. (@) Ifpis true and p implies 4, then q is true. (b) This was Pause 5. (©) If pis true or qis true and p is not true, then q must be true. (@) We do this by contradiction, So assume the premises are true but the conclusion is false. Since pris false, we must have p true and r false. Since q —r is true and r is false, q must be false. Then, since p —+ q is true, p must be false, giving a contradiction, (©) We prove this by contradiction. Assume p V q is false. This means both p and q are false. Since pV ris true and p is false, r must be true. Since q'V (=r) is true and q is false, r must be false, siving a contradiction, 3. (@) [BB] We analyze with a truth table. ‘There are five rows when the premises are all tue and in each ‘case the conclusion is also true. The argument is valid. (b) We analyze with a partial truth table showing the nine situations in which both the premises are true. In every case, the conclusion is true. ‘The argument is valid. pla [pva[p=r[a=r evar tirit| r [ir [3 tT |« rle|r| 7} or |r T |. r|r|r] 7] or |r rls ele|r| e& | 7 |r role r|r|r| r | Fr | F P tlr|eF| 7] Fr | or P F[rje| 7] 7 | F FP Flele|e | 7 |r ri. Pra] * [pAa[rAs| @AD GAD tlr{r|r| r [7 T r\r|r|r) e | 7 T t\r\F|r| e& | & r Fi[r|r|r| F | 7 Tr Fl[r\rl[r| Fe | 7 Tr Fl[r|rlrl| re | F Tr Flr|rir| F | F Tr Flrlr[r| Fe | F Tr Flelele| Fe |e Tr (©) The second and third premises are p —» r and r —> s which together imply p —+ s by the chain rule, Thus the argument becomes pva pos avsSection 1.3, 23 which we check with a truth table, pl a[s[pVa[p>s [ave tir|r|] tr] r |r |« TlF|T|) 7 | rv | 7 l« F\r|r) 7 | r | 7 l* FlF|T]) F | ro) 7 r\rjeFl) r |) er) or T\F\F| 7 | PF | FP F\rjF) 7 | 7 | 7 |« FilrifFlrF| 7 fle ‘There are four rows when the premises are true and, in each case, the conclusion is also true. The argument is valid. @ Since ((-9) Ar) => (gV (ar) and ~(p As) => [(-p) V (-8)] <=> (p> (-5)), the premises become (av (or) sp p(s) so the chain rule gives (qV (=r) + (~8), which is logically equivalent to (=(9 V (=r))) V (=8), which is the desired conclusion by a law of De Morgan. 4, (@) [BB] Since [(-r) V (-a)] = [a+ (orl the first two premises give p+ (sr) by the chain rule, Now -p follows by modus tollens. (b) This argument is not valid. Ifq and r ae false, p and s are true, and ¢ takes on any truth value, then all premises are tue, yet the conclusion is false. (©) [BB] This argument is valid. Since p ++ (t's), we can replace t V with p so thatthe premises become pv (9), p+ (PVT), (>r) Vp. The fist ofthese is logically equivalent to q ~» p while the third is r sp. Using Exercise 3(a), we get q Vr +p, so certainly q Vr — pV (© This argument is valid To se this, note tha the second premise is r > (EV 8), so the chain rule gives r — p. Also, the fist premise is q —>p. Exercise (a) tells us that (@ V'r) ~» p, so certainly (avr) > (pvr). (©) This argument is not valid. Ip and r are fale while q and sare tru, all premises are tre, yet the conclusion is false. (© (BB) This is vali. The first premise is gV (-p) Vs] while the second is -9)Vr. Now (—p)VsVr follows by resolution, and this i the conclusion (@) This argument is valid. The first premise isp V {(>9) Vr], which is (-g) V p Vr. Since q Vr, resolution gives pV r. Using r ~» p and the result of part (), we get pV p, which i p (t) This argument is valid. Using (g) twice, we have p —» [s V (~p)] which is (~p) V 8 V (~p), which is (-p) Vs, which is the conclusion, » 7 ps Ista up late at night ae a een q: Tam tired in the morning. pa The given arguments __p a ‘This is valid by modus ponens.24 Solutions to Exercises atements ?* Tsay uplate at night ade is q: — Tamtired in the morning. poa ‘The given arguments __@ P Papa ‘This is not valid, asthe truth table shows. tir] 7 |« Inrow three, the two premises are true tlr| F but the conclusion is false Fi[r| r |« elel 3 I stay up late at night (©) Letpandg be the statement ee niin ne neening, Poa ‘The given argument is _ > P This is valid by modus tollens. ‘an p: stay up late at night @ Letpandg be the statements eee ein ihe nmin. Poa ‘The given argument is __—p = This is not valid, asthe truth table efapewa{ elma shows. In row three, the two rir] t Fle premises are true but the conclusion Fa aa ae ri[r| 7 |r| els rle| r [rir|s pi Lwearared ie (© (BB) Let pand qbe the statements Fe pva ‘The given argument is _- P This is valid by disjunctive syllogism. statements P: Twearared tic (© Letpand be the statements Fee pva ‘The given argument is __@ =P This is not valid, asthe truth table je pezetoe) shows. In row one, the two T/T |Fl« premises are tre but the conclusion Se is false. Tyr rie Fle |r (g) [BB] Let p,q, andr be the statements g: Team lots of money rr Ipay high taxes.Section 1.3 Poa agate Top ‘The given argument is ‘This is not valid, as the truth table shows. In row five, the two premises are true but the ‘conclusion is false. Pt (h) Let p,q, andr be the statements q: poa ‘The given argument is. _g—r- por ‘This is valid by the chain rule. Pt Gi) Letp, q, and r be the statements pra ‘The given argument is =po or aor. 25 ssasssaall ssasssssll ssaseasall aa ssssl| Tass sss k| BASS BSS work hard eam lots of money pay high taxes. work hard Team lots of money pay high taxes. ‘The conclusion is logically equivalent to r -» p, so this is the same as Exercise 5(g), hence not valid. P: @ Letp,q, andr be the statements q: Poa ~a evr argument is like mathematics Istudy ike football. ‘The first two premises give +p by modus tollens, so, since pV r is true, the conclusion follows by disjunctive syllogism, P: (k) Let p,q, and r be the statements q: avr ‘The given argumentis _r—p (a) > P avr ‘This is the same as ave like mathematics Tstudy like football t) V p_and hence valid by resolution,Solutions to Exercises : Tlike mathematics a: Istudy (9 GBB) Letp, q.and. be the statements 7 T pass mathematics 3: Tegraduate, pa (ra) vr 13) + Gr) —>0 poa aor ros s>9 ‘which is certainly not valid, as the following partial truth table shows. ‘The given argument is ‘This is the same as Pla[*|*[p>a]a>r[r>s[eaa Flele[7> er [rir [er ike mathematics Istudy ess mathematics st T graduate, pa avr (28) > (r) pos pa qor ros pos ‘which is valid by two applications of the chain rule. (in) Let p, g, and r be the statements ‘The given argument is, which is the same as 6. [BB] r-V qs logically equivalent to [>(-r) Va] <= [(-r) 4] so, with p > -r, we get p + aby the chain rule, 7. We will prove by contradiction that no such conclusion is possible. Say to the contrary that there is such a conclusion ©. Since € is nota tautology, some set of truth values for p and q must make € false. But ifr is true, then both the premises (=p) —* r and r V q are true regardless of the values of p and q. This contradicts € being a valid conclusion for this argument. 8. (@) [BB] p Aq is true precisely when p and q are both true (b) By 8(@), we can replace p Aq by the two premises p and . Using modus ponens, p and p —> r lead to the conclusion r. Using modus tollens, q and s + (=g)) lead to the conclusion -s. Finally, 8(a) says we can replace “3 and r with (=8} Ar. 9. By Exercise 8(@), the final premise is equivalentto the list of premises ¢1,2,.--sdn- Now fpr (a1 9 r)) > br > (an) Vn) > [(02) V (0) Vi)Chapter 1 a ‘Together with q1, disjunctive syllogism gives (~p1) V ri which is logically equivalent to py + ri. ‘Thus the given premises imply Pon Pion Pata which, again using 8(a), are logically equivalent to the single premise (p> 71) A (P2172) Av A Pn > Ta): 10. {BB} In Latin, modus ponens means “method of affirming” and modus tollens means “method of denying”. This isa reflection of the fact that modus tollens has a negative “p as its conclusion, while ‘modus ponens affiems the truth of a statement q. Chapter 1 Review 1 1 a> Cr) [pAG@> Cr) (a) vr CT " Sse SRSA SSAS SSS SSS Saas sls RSENS AS RARAAS aA BASASAS AG > Gry) = (Cad) V4) SANASSSS 2. We have ~r is F, 50 (=r) A sis F. Since qis 7, this means q + ((nr) A's) is F. But pis T, so DY (a> ((-r) As)) is. Alsor A tis T. Hence the entire statement has truth value 7. 3. (@) A truth table shows this is a contradiction, PI 4 [=a] =p [PACH | Cr) Va TPA CIIATCR) Val ritlrlr| F T F rlr|r|Fl| ¢ F F F\r|r\|r| F Tr F Flr[r|r| F Tr FSolutions to Review Exercises (b) A truth table shows that this is neither a tautology nor a contradiction. pla [pa pVval (p49) > Va) tir) cfr T Tle) F | oT Tr F\r| cr |r Tr Fle| rie F (©) A truth table shows this is a tautology. pla[r [-a[ PAC) [PACA + [pV [A Ca) > r] TIT|T|F| F T T rir|r|r| ¢ Tr Tr Flr|r|F| F T Tr Fle|r|r| F T Tr rir|Fl|F| F T Tr rlelF|r| o¢ F F Fi[r|Fl|F| F T Tr Flelel|r| F&F Tr Tr (@) This is neither a tautology nor a contradiction. Pl a|+ [=a] Ca) Ar [pve TIT|T|F] F TT rie|r|r| cr fr Filrir|rF| fF |r Fle|r|r| cof Fe rlrjr|r| F |r Tleje|r| fF |r Firjr|e| rf |r Fleje|r| rf [Fr Tea) ArI> Va) [ PV) > 169) ArT Va) © Ga) Ar] T F F Tr Tr T Tr F F P Tr F Tr F F T FE F T F F F T F 4, Assume that some set of truth values on the variables makes A true. If B were false, this would make AB false and B —+ A true, contradicting logical equivalence. So ‘B must be true also. Similarly, if Bis true, then A must also be true. We conclude that Ais true if and only if B is true. This means A and B are logically equivalent. 5. (a) Since A <=> B, we know that A is true precisely when ‘B is tue. Since B <—> €, B is tue precisely when @ is true. Hence Ais true ifand only if @ is tue, that is, A <=> €. ()) Property 12 says (p + q) <=> ((-p) Vg). Clearly, ((-p) Va) => (9 (-p)), So part (a) tells us (p > g) <= (qv (~p)). But Property 12 also says ((-g) > (=p) => ((>-9) VP), and clearly (=(-g) V (=p)) => (@V (-p)). So we have (p +g) <=> (qv (~p)) and alsoChapter 1 29 (av (=p) => (>a) — (>p)). Hence part (a) again gives (p > q) += ((-4) > (=P), which is Property 11. 6 (94) 97) > (ORV) 41) > (MP)VO)VF) > (WACQ)) VE) (vr) ACV) > (BV ACGACH)))- ©) PVA) > OPV QVr) > ~pVaVr(associatviyy > F(A -¥)]V F@eMorgan) <=> [pA(-a)] >. 7. @ (DVDAV (DVO ACD) > ((PAr)V GAP) V ((OA CP) ¥ (GA (2), + (PAV GAN) VOV GACH) $+ (PAN)VG@AN)V GAC) > DAFADVPATACA)V(@Ar Az) V (GAT ACP) V(gA (-p) Ar) PAGAN VDAC) Ar) V (CH) Aa Ar) V((-p) Ag A (=r) () PV GA (=r) Aa@Ar) > PVGACNIAICD Vn) (Pv @Aa-ny] Aa) v (Pv GA -r))] 4-7) (pA (~9)) V (a. (9) A (>a) VPA Cr) VGA (PA (>9) Ar) V (PA (=9) A (7) VOV (PAGA (=r) VPA (24) Ar) V RAGA (7) V (=P) AGA (>) (A (>a) Ar) V PA (9) Ar) V (PAGA (>1)) V@AGA (>7) (=P) Ags (1) 8. (a) This is not valid. If is false and q is true, the premises are true but the conclusion is not, (b) The first hypothesis can be rewritten as pV (-g), which is the same as q —> p. The second hypothesis is (=p) v (-r), which is p = (-r). The third hypothesis is (-r) — s. So the given argument is = = aCr) = = q7pP p>(7r) (rss “Gs ‘Two applications of the chain rule tell us this is valid. (©) This argument is not valid. If p, r and ¢ are true while q is false (and s takes on either truth value), the hypotheses are true while the conclusion is false. 9. This argument is valid. The hypotheses can never both be true at the same time, so there can be no ‘case when the hypotheses are true while the conclusion is false. 10. (a) This argument, an example of resolution, is valid. (b) This argument is not valid. We can write it as shown. pq When s and r are true and p is false, the hypotheses avr are true, while the conclusion is false. (or) > (28) 3p.30 Solutions to Exercises Exercises 2.1 (a) (BB) {-V5, v5} () (1,3,5,15, -1,-3, —5, -15} (©) [BB] {0, ~$} (Although +? are solutions tothe equation, they are not rational.) (@ {-1,0,1,2,3) (©) This i the empty set. There are no numbers ess than —4 and bigger than +4 (@) (BB) For example, 1 +i, 1+ 24,1 4 3%, ~8 — Bi and 17 — 434. (b) For example, {1 ~ 2V2, 1 — 52, 1 — 7/2, 316 — 2V2 and 394 — 73}. le 5 and x/y = 0. x = 1,y = &V TA, and 2/y = £1/V2 = 4V/F/24 Iz = 2, y = +V2l and z/y = +2V21/21. Five elements of the given set are 0, 24/24, = V2 /24, 2V2i/21 and —2V 7/21. © (2,3,5,6,8} (9) (BB) {1,2}, {1,2,3}, {1,2,4), {1,2,3,4) ) 0, {1}, (2), {1,2} © 0,0}, 2}, (8), 4}, 1,3}, (1,4), (2,3) 2.4}, (8,4), (3,4), (2,3,4} © {3}. {4}. {1,3} (1,4), 2,3), 12,4), (8,4). (1,2,3), (152.4), 1,3,4}, (2,3, 4), (1,2,3,4) (©) (1,23), (1,24), {1,2,3,4} © 0, {1}, {2}. [BB] Only (c) is true. The set A contains one element, {a, 6} (2) [BB] True, 3 belongs to the set {1,3,5}. (b) False. {3} is subset of (1,3,5} but not a member ofthis set. (©) True. {3} is a proper subset of {1,3,5} (@ (BB) False. {3,5} isa subset of {1,3,5}. (©) False. Although {1,3,5} is a subset of itself, it is not a proper subset. (® False. If a+ 2bis in the given set, ais even, so a + 2b is even and can't equal 1. (@ False. Ifa + bv3 = 0 and b # 0, then V3 = ~§ isthe quotient of rational numbers and, hence, rational, But his is not tue, (a) (BB) {0}; (b) {0, {0}}: © {0,{0, {0}}, {0}, { {0} 3}. (2) (BB) True, The empty set isa subset of every set. (b) True, The empty set is a subset of every set. (©) False. The empty set does not contain any elements. (@) True, {0} isa set containing one element, mmely, 0. (©) (BB] False, {1,2} is a subset of {1,2,3, (1.2, 3}}.. (O False. {1,2} isnot an element of {1,2,3, (2,2,3}). (@) True, {1,2} is a proper subset of {1,2,{{1,2}}}- (h) (BB) False, (1,2} is not an element of {1,2,{(1,2}}}Section 2.1 a” 10, 13, True. {{1,2}} contains just one element, {1,2}, and this is an element of {1,2, {1,2} [BB] Yes itis: for example, let x = {1} and A = {1, {1}. @ i. {abed} ii, [BB] {a,b,c}, {a,6,d}, {a,c,d}, {6, c,d} iit, {4,8}, {0c}, {ad}, {0,0}, {bd}, {ed} iv. {a}, (0), {ch {@) vo © 16 (@) IFA = 0, then P(A) = {0} is aset containing one element, so its power set contains two elements. () P(A) contains two elements; P(P(A)) has four elements (@ (BB}4; ——(&) (BB). (©) [BB] There are 2” subsets of a set of n elements. (See Exercise 15 in Section 5.1 fora proof.) (@) (BB) False. Let A = {2}, B = {{2}}, C = {{{2}}}. Then A is an clement of B (that is, A € B) and B is an element of C (B € C), but A isnot an element of C (since B is C’s only element). (b) True, fx € A, then x € B since AC B. Butsince « € B, then x € C since B C C. (©) True. As in the previous part, we know that A CC. To prove A # C,, we note that there is some 2 € C such that ¢ B (since BS C). Then, since x ¢ B, x ¢ A. Therefore, x is an element of C which is notin A, proving A # C. (4) [BB] True. A € B means that A belongs tothe set B. Since B is a subset of C, any element of B also belongs to C. Hence, A € C. (©) False. For example, let A = {1}, B = {{1},2} and C= {{1},2,3}. Then A € B, BCC, but AgC. ( False. Let A= {1}, B = {1,2}, = {(1,2},3}. Then AC Band BEC, but A ¢C. (@) False. Same example as 12(f) where AZ C. (a) This is false. As a counter-example, consider A = {1}, B = {2}. Then Ais not a subset of B and B is nota proper subset of A. (b) The converse of the implication in (a) isthe implication B ¢ A —+ AZ B. Thisis true. Since BG A, there exists some element a € A which isnot in B. Thus A is not a subset of B. (@) [BB] True. (—) If C € P(A), then by definition of “power set;” C is a subset of A; that is, CCA. (IEC C Athen C is a subset of A and so, again by definition of “power set." C € P(A). (b) True. (—+) Suppose A C B. We prove P(4) C P(B). For this, let X € P(A). Therefore, X is ‘a subset of A; that is, every element of X is an element of B. Since AC B, every element of X must be an element of B. So X C B; hence, X € P(B). (—) Conversely, assume P(A) C P(B). We must prove A CB. For any set A, we know that AC Aand, hence, A € P(A). Here, with P(A) C P(B), we have, therefore, A € P(B); that is, AC B, as desired, (©) The double implication here is false because the implication —> is false. If A = 0, then P(A) = {0} and {0} #0.32 Solutions to Exercises Exercises 2.2 1 = (@) (BB) 1,2, 3,4, 5,6}, B = {-1,0,1,2,3, 4,5}, C = {0,2, —2}. ,0,1,2,8,4,5,6}, BC = (0,2), -1,1,3,4,5}, A @ B = {-1,6,0}, Cx (BNC) = {(0,0), (0,2), (2,0), (2,2), (—2,0), (2, 2)}, (AN B)\C = {6}, AN (BC) = {2,6}, (BU0)N {0} =6. (©) 8 = {(1,—1), 2,0), (8,1)s (4,2), (6,3), (6,4)}: T= {(1,2), (2,2)} (@) [BB] ST = {V2,25}, SUT = {2,5, V2, 25, $.46, 3}, T x (SOT) = {(4, V2), (4,25), (25, v2), (25,25), (v2, v2), (V2, 25), (6, V2), (6,25), (3, V2); ($5 25)) (b) (BB) ZUS = {V2,x, $,0,1,-1,2,-2,...}; ZS = {25,25}; ZuT ={V2,3,0,1,—1, }s ZT = (4,25, 6}. (© Z(SUT) = {2,5,25,4,6} = (ZS) U(ZAT). The two sets are equal @ ZU(SAT) = {V,0,1,-1,2,-2,...} = ZU {V3} = (ZUS)(ZUT). ‘The two sets are equal . @) (BB) {1,9,0,6,7 (6) {46,5}; © {0,1}. A= {(1,1)5 (1,2), (1,3). (2,2), (2,8), (3,8)} end B= (43,41, 42}, (a) [BB] ie {a,b}; ) {0}; ©A @h (© (BB) 6; © {4}. (@ (BB) A= (-2,1; AP = (-00,-3]U (400); (A (@) YNZ = {3,4,5}, so X @ (¥ NZ) = {1,2,5}. (b) (XeUY)e = XN¥e=X\¥ = {i}. (@) [BB] The subsets of A containing {1,2} are obtained by taking the union of {1,2} with a subset of {3,4,5,....7}. Their number is the number of subsets of {3,4,5,...,n1} which is 2"-2, (See Exercise 11 of Section 2.1.) (©) The subsets B which have the property that BN{1,2} = @ are exactly the subsets of (3, 4,5,...,n} and these number 2°-2, (©) The subsets B which have the property that B U {1,2} = A are precisely those subsets which contain {3,4,5,....,n} and these correspond, as in (a), to the subsets of (1,2). There are four. [BB] (a,8)° = (—20,a]U[b, 00), [a,)° = (~20,@) U [b,00), (a,00)° = (—o0,a), (00,8) = (6,00). (@) [BB] CST; () (BB] MN P= OMEP; @CS\TCP; © (Mucs)nP CT. (@) Negation: CS ZT; Converse: TC CS. (b) Negation: M9 P # 0; Converse: PC M* or P.M =0,Section 2.2 33 2. 13. 4 15, 16. 1. 18. 19. 20. (©) Negation: M C P; No converse since the statement is not an implication, (@ Negation: (MU CS) NP ¢ T*; Converse: T? ¢ (MUCS) NP. (©) Negation: (MU CS) POT #0; Converse T° C (MUCS) 1 P. (@) [BBJENP 40 (be) OEZ\N @ NcZ @ ZEN © (PENCE 2eEnP @ EnP=(2} (@) [BB] Since A_s C Ay, AgU A_g = As, (b) Since A_g © As, As A-3 = As. (©) As (A-a)* = {a €Z| -3 n}. (@) Let € B. Certainly xis also in A or in A®. This suggests cases. Case 1: If r € A, then € ANB, soe C. Case 2: If ¢ A, then € APB, soz EC, Incither case, € C, 90 BCC. (©) [BB] Yes. Given AN B = ANC and A®n B= A°MC, certainly we have ANB C C and A®A.B C C0, from (@), we have that B CC. Reversing the roles of B and C’in (a), we can also conclude that C'C B; hence, B= C. (@) The Venn diagram shown in Fig. 2.1 suggests the following counterexample: Let A = {1,2,3,4}, B = {3,4,5,6} and C = {2,3,5,7}. Then AU (BMC) = AU {3,5} = {1,2,3,4,5} whereas (AUB)NC = {1,2,3,4,5,6} NC = {2, 2,5}.34 a 23, m4, 26. 21, Solutions to Exercises (b) First, we prove AU (BNC) S (AUB) (AUC). Soletz € AU(BNC). Then x € Aora € BNC. Iz € A,thenz € AUBand x € AUC, 50 2 €(AUB)N(AUC). Ifz € BNC, then € Band z € C soz € AUB and x € AUC; that is, 2 € (AUB) (AUC). Inither case, x € (AUB) (AUC) giving the desired inclusion. Second, we prove (AUB) (AUC) C AU (BNC). Solet x € (AUB)N(AUC). Thus, z € AUB andz € AUC. fz € A, then € AU(BNC). If ¢ A, then we must have x € B and 2 € C; thatis,x € BOC, sox € AU(BMC), Ineither cease, x € AU (BNC) giving the desired inclusion and equality. ‘We use the fact that (X°)® = X for any set X. Let X = A® and ¥ = B°. Then A = X° and B= ¥°,50 (AN B)* = [X°NY*]* = [(X UY)" by the first law of De Morgan) = X UY = A®U B®, as required, [BB] Using the fact that X SY = X NY, wehave (ANB)\C AN BNC = AN(BNC AN(BUC) = AX (BUC). ‘We use the laws of De Morgan and the fact that (X*)° = X and Xn X* [(AUB)en(A°UC)|°S De = [(A°NB)N(ANC®)|*\ DE = OFS DF UnD=D. for any set X. We have UND? =Un(D = A\ (BX C) = AX (BNC) = AN (BNC = AN(BEUC) = (ANB) U(ANC) = (ABU (AN (0%) = (AN B)U (ANC). (@) [BB] (AU BUC)’ = [AU (BUC) = 4° (BUCY? = Aen (BEC) = AN BEN’. (AN BNC) = [AN (BAO) = APU(BNC)’ = APU (BEUC*) = ASU BUCS © (AN(BXO)fNA = (ANBNCENA=(ASUBSUC)NA (An (APU ES) U(ANC) = (ANA?) U (ANB?) U(ANC) OU(ANB)U(ANC) = (ANB) U(ANC) (AN B)U(ANC) (©) {BB} Looking athe Venn diagram athe ight, A @ B consists 4 a ofthe points in regions Land 3. To have AB — A, we must CQ3) have bth regions? and 3 empty: tha, B= 8. On he oar hand, since A @ 0 = A, this condition is necessary and suficent (b) Looking atthe Venn diagram, A” B isthe se of points in region 2 while AU Bis the se of pins imregion 12 and 3. Hence, AM B= AUB if and only if regions 1 and 3 are bth empty: that, is and ony it = B. (8) [BB] This does not imply B = C. For example, let A = {1,2}, B = {1}, C = {2}. Then AUB = AUC, bu BC. (b) This does not imply B = C. For example, let A = {1}, B = {1,2}, C = {1,3}. Then ANB=ANnC =A, bu B#C. (©) This does imply B = C,, and here is a proof. First let 6 € B. Then, in addition, either b € A or bea. Case 1: bg ASection 2.2 35 28, 29. |. (a) False, For example, let A In this case, b € A® B, $06 € AGC and since b ¢ A it follows that b € C. Case 2: b € A. Here we have bE BMA and, hence, b ¢ A@ B,sob ¢ ABC. Since b € A, we must have b € C (otherwise, b€ ANC. CA@C). In either case, we obtain } € C. It follows that 8 C C. A similar argument shows CC B and, hence, C (@ This is false since for A= 0, Ax B= Ax C =O regardless of B and C. (@) True, Let (a,b) € Ax B. Since a € Aand AC C, we havea € C. Since b € Band BC D, b€ D. Thus, (a,8) €C x Dand Ax BCC xD. (b) False: Consider A = {1}, B = {2,3}, C ={1,2,3}. (©) False. Let A = {1}, B = 0, = {2}, D = {3}. Then Ax B= 0 {(2,3)} ALC. (@) False since, by (b), the implication — is false. (© (BB) Tue. Letz € A. Then x € AU B, soz ¢ ANB and, in particular, = € B. Thus, AC B. Similarly, we have B.C A,so.A= B. Cx D, but Let (z,) € (ANB) x C. This means x € AN Band y € C. Hence, x € A, x € B,y € C. Thus, (x,y) € Ax C and (x,y) € B x C;ie,, (2,y) € (A x C) M(B x C). Therefore, (AM B) x CC (Ax 0) (Bx 0). Now let (x,y) € (A x C) 9 (B x C). This means that (x,y) € A x C and (2,y) € B x C); that is, x € Az € B,y € C, so € ANB andy € C. Hence, (x,y) € (ANB) x C. Therefore, (Ax C) (Bx C) (ANB) x C and we have equi 1} and C= {2}. Then A\(BUC) = (1,2}\{1,2} = 0, but (AS B)U (ANC) 1,2}. (b) True, Let (x, y) € (A \ B) x C. This means that 2 € A \ B and y € Cs thatis, x € A,x ¢ B, y € C. Hence, (z,y) € A x C, but (2, y) ¢ Bx C, so (2,y) € (A x C)\ (Bx C). Therefore, (ANB) x CC (Ax0)\ (Bx). Now let (x,y) € (AC) \ (Bx C). This means that (1, y) € Ax C, but (2,4) ¢ Bx C. Since (2,y) € Ax C, we have « € A, y € C. Since y € C and (x,y) ¢ B x C, we must have x ¢ B; thatis, 2 € ANB, y € C, 80 (2,y) € (ANB) xC. Therefore, (Ax0)\(BxC) ¢ (A\B)xC and we have equality as claimed. (©) (BB) True, Let (2, y) € (A@ B) x C. Thismeans that x € A@ B and y € C; thatis, x € AUB, 2¢ANB,y eC. Ize A, thenz ¢ B, s0 (x,y) € (Ax C)\ (Bx C). Itz € B, then 2 ¢ A, 0 (2,y) € (B x C) \ (Ax C) In cither case, (z,y) € (A x C) @ (Bx C). So (A@B)xCC(AXxC)@(BXC). Now, let (z,y) € (A x C) (B x ©). This means that (z,y) € (A x C) U(B x C), but (eu) ¢ (Ax C)N (Bx C). If (x,y) € AX C, then (2,y) ¢ BxC,sore AYEC and, therefore, « ¢ B. If (x,y) € Bx C, then (2,y) ¢.A x C, soz Byy € Cand, therefore, x ¢ A. In either case, x € A@® Band y € C, s0 (z,y) € (A @B) x C. Therefore, (Ax C) (Bx C) C (A@B) x C and we have equality, as claimed. @ False, Let A = {1}, B = 2}, C = {8}, D = {4}. Then 1 € AUB, 4 € CUD, 80 (1,4) € (AUB) (CUD). But (1,4) ¢ AxC and (1,4) ¢ Bx D, s0 (1,4) ¢ (AxC)U(BxD). (©) False, Let A = {1,2}, B = {2}, C = {3}, D = {4}. Then, since 3 ¢ D, (2,3) € (Ax C)\ (B x D). However, because 2 € B,2.¢ A\ B, so (2,3) ¢ (AX B) x (CX D).36 Solutions to Exercises 31. George Boole (1815-1864) was one of the greatest mathematicians of the nineteenth century. He was the first Professor of Mathematics at University College Cork (then called Queen's College) and is best known today as the inventor of a subject called mathematical logic. Indeed he introduced much of the symbolic language and notation we use today. Like Charles Babbage and Alan Turing, Boole also had a great impact in computer science, long before the computer was even a dream. He invented ‘an algebra of logic known as Boolean Algebra, which is used widely today and forms the basis of ‘much of the internal logic of computers. His books, “The Mathematical Analysis of Logic” and “An Investigation of the Laws of Thought” form the basis of present-day computer science, Exercises 2.3 1. [BB] $ x B is the set of ordered pairs (s,8), where s is a student and b is a book; thus, S x B represents al possible pars of students and books. One sensible example of a binary relations {(s,8) | ‘shas used book 6}. 2. A x Bis the set ofall ordered pairs (a, 6) where a isa street and bis a person. One binary relation would be {(a,6) | b lives on street a}. 3. (@) [BB] not reflexive, not symmetric, not transitive. () (in most cases) reflexive, in somewhat fewer cases) symmetric, certainly not transitive! (©) [BB] not reflexive, not symmetric, but it istransitive (@ reflexive, symmetric, transitive (©) not reflexive, not symmetric, not transitive 4. (a) (BB) @ 5. (@) (BB) {(1, 1), (1,2), (2,3) () {(1,1), (2,2), (8,3), (1,2), (23): (©) {(1,2); (2,3), (2,1), (3, 2)}; @ {(2,2), 2,3), (2,3) © {(1), (2,2), (3.3), (2,3), (3,2)}; 2, ¢ (1,2), (2,2), ( @® (1,1); (2,2), (3,3)s (1,2), (2,3)5(1,3)}; (BB) {(1,2), (2,1), (1,1), (2,2): (h) {(1, 1), (2,2), (8,3), (1,2)s (2,3), (1,3); (2,1), (8,2), (8, D)}- 6. The answer is yes and the only such binary relations are subsets of the equality binary relation. To see why, let ® be a binary relation on a set A which is both symmetric and antisymmetric. Let (a,) € R. ‘Then (b,a) € R by symmetry, so a = b by antisymmetry.Section 2.3 a7 7. [BB] The argument assumes that for a € R there exists a b such that (a, 6) € R. This need not be the cease: See Exercise 5(g). 8. (a) [BB] Reflexive: Every word has atleast one letter in common with itself. ‘Symmetric: Ifa and b have at least one letter in common, then so do 6 and a. Not antisymmetric: (cat, dot) and (dot, cat) are both in the relation but dot. cat!! Not transitive: (cat, dot) and (dot, mouse) are both in the relation but (cat, mouse) is not. (b) Reflexive: Let a be a person. Ifa is not enrolled at Miskatonic University then (a,a) € R. On the other hand, if ais enrolled at MU, then ais taking atleast one course with himself, so again (a,a) €R. ‘Symmetric: If (a,b) € 7 then either itis the case that neither a nor bis enrolled at MU (so neither is 6 ora, hence, (b,a) € 7) or itis the case that a and b are both enrolled and are taking at least one course together (in which case b and a are enrolled and taking a common course, s0 (bya) €R). Inany case, if (a,b) € Ry then (bya) € R Not antisymmetric: Ifa and b are two different students in the same class at Miskatonic Univer- sity, then (a,6) € Rand (b,a) € Re, but a #6. At most universities, this is not a transitive relation. Let a,b and c be three students enrolled at MU such that a and b are enrolled in some course together and b and c are enrolled in some (other) ‘course together, but a and c are taking no courses together. Then (a,6) and (b,c) are in R. but (a,0) ¢ R. 9. (a) Not reflexive: (1,1) ¢ R. Not symmetric: (1,2) € but (2,1) ¢ R. 0. Thus, (a,a) € R. Symmetric: If (a,b) € R, then ab > 0, s0 ba > Oand hence, (b,a) € R. Not antisymmetric: (5,2) € R because 5(2) = 10 > 0 and similarly (2, 5) € R, but 5 #2. Not transitive: (5,00) € 7R because 5(0) = 0 > O and similarly (0, —6) € 2; however, (5, -6) ¢ R because 5(—6) #0. (@ Reflexive: For any a € R, a? = a, s0 (a,4) € R. ‘Symmetric: If (a,6) € 7 then a? = b2, so? = a? which says that (4,4) € R. Not antisymmetric: (1, ~1) € R and (—1,1) € R but 1 # =1 ‘Transitive: If (a,6) and (b,c) are both in R, then a? = 6? and b? = c, soa? (a,c) ER. (©) Reflexive: For any a € R, a —a = 0 < 3 and so (a,a) € R. Not symmetrie: For example, (0,7) € R because 0-7 = —7 < 3, but (7,0) ¢ R because 7-0=723.Solutions to Exercises Not antisymmetric: (2,1) € R because 2—1 = 1 < 3 and (1,2) € R because 1 —2 butt £2, Not transitive: (5,3) € because 5 ~ 3 = 2 < 3 and (3,1) € R because 3—1=2 <3, but (5,1) ¢ R because 5-1=4 £3. -183, () Reflexive: For any (a,b) € A,a—a = b~b; thus, ((a,b), (a,b) € R. Symmetrie: If ((a,2), (c,d) € R, then a—c = bd, soca = d—band, hence, ((c,d),(a,8)) € R. Not antisymmetric: ((5, 2), (15,12)) € R-because 5—15 = 2-12 and similarly, ((15, 12), (5,2)) € Rs however, (15,12) # (5,2). If ((a,8), (c,d) € R and ((c,d),(e, f)) € Rthen a —c = b—dand c—e =d— f. Thus, a-e=(a~0) + (ce) = (b= d) +(d~ f) =~ f and so ((a,b),(e,f)) ER. (@) Not reflexive: Ifn € N, then n # nis not tue. Symmetric: Ifmy # m2, then m2 # m1 Not antisymmetric: 1 4 2 and 2 4 1 so beth (1,2) and (2,1) are in R, yet 1 #2. Not transitive: 1 72,2 #1, but 1 = 1 Not reflexive: (2,2) ¢ R because 2 +2 #10. Symmetric: If (x,y) € R, then x + y = 10,50 y +2 = 10, and hence, (y,) € R. Not antisymmetric: (6,4) € R because 6 +4 = 10 and similarly, (4,6) € 7, but 6 # 4. Not transitive: (6,4) € R because 6 + 4 = 10 and similarly, (4,6) € R, but (6,6) ¢ R because 6+6 410. (@ (BB) Reflexive: If (x,y) € R?, then 2 +y Sa+y, 80 ((z,y),(z,y)) € R. Not symmetric: ((1,2),(3,4)) € Rsince 1+2<3 +4, but (8,4), (1,2)) ¢ Rsince 3+4¢142. Not antisymmetric: ((1,2),(0,3)) € Rsince1+2<0+3 and ((0,3),(1,2)) € R since 0+ 3 <1 +2, but (1,2) # (0,3). ‘Transitive: If ((a,), (¢,d)) and ((cyd),(e,)) are both in R, then a+b < c+dandc+d < e+f, soa+b its own price and length > its own length, so (a,a) € R for any book a. Not symmetric: (¥, Z) € R because the price of ¥ is greater than the price of Z and the length of Y is greater than the length of Z, but for these same reasons, (Z,Y’) ¢ R. Antisymmetric: If (a,8) and (6,2) are bota in 7, then a and b must have the same price and length. This is not the case here unless a = t ‘Transitive: If (a,b) and (b,c) are in, then the price ofa is > the price of band the price of bis > the price ofc, so the price ofa is > the price ofc. Also the length of a is > the length of band the length of b is > the length ofc, so the length ofa is > the length of c. Hence, (a,c) € () Reflexive: For any book a, the price of ais > the price of a so (a, a) € R. (One could also use a similar argument concerning length.) Not symmetric: As in part (a), (¥, Z) € R, but (Z,Y) ¢ R. Not antisymmetric: (W,X) € R because the price of W is greater than or equal to the price of X, and (X,W) € R because the length of W is greater than or equal to the length of X, but WeX. Not transitive: (Z, 7) € R because the lengih of Z is > the length of U and (U, Y) € R because the price of U is > the price of Y, bt (Z,Y) ¢ R because neither is the price of Z > the price of Y¥ noris the length of Z > the length of Y. a} #0€ P(S).40 Solutions to Exercises 13. Now the second binary relation would have an extra term, (Mike, 120}, and the third would have the extra term, {Pippy Park, 120}. But, in addition, the entry {Pippy Park, 74} would be deleted. So Mike is now clearly identified as the one who shot 120, and Pippy Park is where that occurred. Hence, ‘Mike's round of 74 was at Clovelly. Since Edgar has only one entry in binary relation two, he must have shot 72 at both courses. Finally, Bruce's 74 must have been at Clovelly and hence his 72 was at Pippy Park. All information has been retrieved in this case. Exercises 2.4 1. Reflexive: For any citizen a of New York City, either a does not own a cell phone (in which case ‘a ~ a) ora has acell phone and a's exchange isthe same as a’s exchange (in which case again a ~ a). ‘Symmetric: If a ~ b and a does not have a cell phone, then neither does b, so b ~ a; on the other hhand, if a does have a cell phone, then so does b and their exchanges are the same, so again, b ~ a. ‘Transitive: Suppose a ~ b and b ~ c. If a does not have a cell phone, then neither does b and, since b ~ , neither does c, s0 2 ~ c. On the other hand, if @ does have a cell phone then so does b and ‘a’s and b’s exchanges are the same. Since b ~ c, ¢ has a cell phone with the same exchange as b. It follows that a and c have the same exchange and so, in this case as well, a ~ ¢. ‘There is one equivalence class consisting of all residents of New York who do not own a cell phone and one equivalence class for each New York City exchange consisting of all residents who have cell phones in that exchange. 2. (@) [BB] This is not reflexive: (2,2) ¢ R. (b) This is not symmetric: (2,3) € R but (3,2) ¢ R. It would also be acceptable to not that ® is not transitive. (3,1) € R and (1,2) € R, but (3,2) ¢R. (©) This is not symmetric: (1,3) is in the relation but (3, 1) is not ‘It would also be acceptable to note that this relation is not transitive: (2,1) € R, (1,3) € R, but (2,3) ¢R. 3, [BB] Equality! The equivalence classes specify that x ~ y if and only if x = y. 4, (a) Reflexive: Ifa € S, then a and a have the same number of elements, soa ~ a. ‘Symmetric: If a ~ 6, then a and b have the same number of elements, $0 b and a have the same number of elements. Thus b ~ a. ‘Transitive: Ifa ~ b and b ~ c, then a and have the same numberof elements, and b and c have the same numberof elements, so a and chave the same number of elements, Thus a ~ ¢. (&) There are seven equivalence clases, represented by 8, {1}, {1,2}. {1,2,3}.(1,2,3,4}, (1,2,3,4,5}, {1,2,3,4,5,6}. 5. (@) [BB] Reflexive: ‘a € RX {0}, then a ~a because €Q ‘Symmetric: If a ~ b, then $ € Q and this fraction is not zero (because 0 ¢ A). So it can be inverted and we see that ® = 1/§ € Q too. Therefore, b ~ a. ‘Transitive: Ifa ~ band 6 ~ ¢, then § € Qand £ € Q. Since the product of rational numbers is rational, £ = ¢4 isin Q, soa~c.Section 2.4 a ©) (BBIT = faJa~1} = fa] $€Q} = {a]aeQ}=Qs {0} (©) (BB) B= 98 = 2 Q,s0.V3~ Vid end hence V3 = VID. 6. Reflexive: For any a € N,a ~ a since a? +a = a(a+ 1) is even, as the product of consecutive natural numbers ‘Symmetric: If a ~ 6, then a? + b is even. It follows that either a and b are both even or both are odd. I they are both even, 6? + a is the sum of even numbers, hence, even. If they are both odd, #? +a is the sum of odd numbers and, hence, again, even, In both cases 6? + a is even, sob ~ a. ‘Transitive: Ifa ~ band b ~ c, then a? +b and U? + are even, s0 (a2 +6) + (B2 +c) is even; in other words, (a? + c) + (6% + 8) is even. Since 6? + bis even, a? + cis even too; therefore, a ~ ‘The quotient set is the set of equivalence classes. Now ‘evens if ais even @= {x | 2? +aiseven} = {oz iteineda So A/~= {22,22 + 1}. 7. (a) [BB] Reflexive: For any a € R, a ~ a because a—a =0€ Z. ‘Symmetric: Ifa ~ 6, then a—b € Z, s0b—a € Z (because b—a = ~(a—b)) and, hence, b~ a ‘Transitive: Ifa ~ b and 6 ~ c, then both a ~ 6 and 6 ~ c are integers; hence, so is their sum, (a—)+(6—c) =a—c. Thus, ave. (b) [BB] The equivalence class of 5 is £=5€Zimplies x € Z. = 5 € Z} = 2, because = {elz-5}eZ} {a|2=5} +k, forsome k € Z} {2|a=5+k+ 4, forsomek €Z} {|2=n+, forsomen Z} (©) [BB] For each a € R, 0 < a < 1, there is one equivalence class, [2 R| z= a+ n for some integer n}. ‘The quotient setis {4 |0< a <1}. 8. [BB] Reflexive: For any a € Z, a~ a because 2a + 3a = 5a, Symmetric: Ifa ~ b, then 2a + 3b = 5 for some integer n. So 25 + 3a = (Sa + 5b) — (2a + 3b) = 5(a +6) — Sn = 5(a + 6—n). Since a + —mis an integer, b ~ a. ‘Transitive: Ifa ~ band b ~ c, then 2a-+ 3b = Sn and 26-+ 3c = 5m for integers n and m. Therefore, (2a +36) + (26+ 3c) = 5(n-+m) and 2a + 3c = 5(n +m) — 5b = 5(n+m —8). Since n+m——b is an integer, a ~ c 9. (a) Reflexive: For any a € Z, 34+ a= dais a multiple of 4, 0 a ~ a. Symmetric: If a ~ b, then 3a-+-b = 4k forsome integer k. Since (3a +b) + (3)-+a) = 4(a+8), wwe see that 3b +a = 4(a +b) — 4k is a multiple of 4, 30 6 ~ a,a2 10, IL 12, Solutions to Exercises ‘Transitive: Ifa ~ band b ~ c, then 3a+5 integer ¢. Since 4(k-+-€) = (3a+b) + (36-+0) is a multiple of 4 and, hence, that a ~ c. (6) D= {xe Z|x~0} = {x | 3x = 4k for some integer k}. Now if 3x = 4k, k must be a multiple of 3, So 3x = 120 for some £ € Z and x = 46.0 = 4Z. ©2 = {2 eZ | z~ 2 = {x | 3x42 = Akforsomeintegerk} = {x | 3x = 4k- 2 for some integer k} Now if Sr = 4k — 2, then 3x = 3k + k — 2 and so k — 2 is a multiple of 3, Therefore, k = 3¢ + 2 for some integer 3x = 4(3¢ + 2) — S02 = 42 +2 (@) The quotient set is {42,42 + 1,42 +2, 42+ 3). 4k for some integer k and 3b + © (Ba++e)+4b, we see that 30+ 4 for some A(k+6)—4b = 126+ Gand ¢ = 4¢ +2. () Reflexive: For any a € Z, a ~ a because 30 + 4a = 7a and a isan integer. Symmetric: If a,b € Z and a ~ 6, then 3a + 4b = 7n for some integer n. ‘Then 3b + 4a = (a+ 7) — (a+ 48) = 7a-+7—7n = 1(a+b—n) and a+b —n isan integer. Thus b~ a ‘Transitive: Suppose a,b,c € Z with a ~ band 6 ~ c. Then 3a + 45 = 7n and 3b + 4e = 7m for some integers n and m. Then 7n + 7m = (8a + 4b) + (36 + 4c) = (Ba + 4c) + 7b, 80 3a+de= 7 + 7m — 7b = 7(n + m —6)and n-+m — bis an integer. Thus a ~ c. () D= {2 Z| ~0} = {x €Z | 3x = 7 forsome integer n}. Now if 3x = 7n, n must be a ‘multiple of 3. So 3x = 21K for some k € Zand x = 7k. We conclude that 0 = 72. (a) (BB] Reflexive: If a € Z\ {0}, then aa =a? > 0,s0a~ a. Symmetric: Ifa ~ 6, then ab > 0, Soba > 0 and b ~ a. ‘Transitive: Ifa ~ b and b ~ c, then ab > Cand be > 0. Also b? > 0 since b #0. Hence, ac = (ait _ (at)(be) . since ab > 0, be > 0. Hence, a~ . (BB) 5 = {eEZ\ {0} |2~5}={e]52>0} ={2|2>0} =B = {2€Z\ {0} |2~-5} = {2| 52> 0} ={2|2 <0} (©) {BB} This equivalence relation partitions Z ~ {0} into the positive and the negative integers. (@) (BB) Reflexive: For any a € Z, a? — a? = is divisible by 3, soa ~ a. Symmetric: Ifa ~ 6, then a? ~ 6? is divisible by 8, so? — a? is divisible by 3, s0 ~ a. ‘Transitive: Ifa ~ band b ~ c, then a? — 6 is divisible by 3 and b? — cis divisible by 3, 90 a? 2 = (a? 02) + (8 — 2) isdivisibe by 3 @ 0 = {2eZ|2~0} {2 €Z| 2° is divisible by 3} {2 €Z| «is divisible by 3} = 32 {reZ|z~1} {| 23 — 1 is divisible by 3} {| («—1)(¢ +1) is divisible by 3} {x | z—Lora + 1is divisible by 3} SZ+1U3Z42 (©) This equivalence relation partitions the integers into the two disjoint sets 3Z and ( 3Z+1)U(3Z-+ 2).Section 2.4 43 13. (@) (BB) Yes, this is an equivalence relation. Reflexive: Note that if ais any triangle, a ~ a because a is congruent to itself. Symmetrie: Assume a ~ 6, Then a and tare congruent. Therefore, b and a are congruent, so bra ‘Transitive: Ifa ~ b and b ~ o, then a andb are congruent and b and c are congruent, so a and ¢ are congruent. Thus, a ~ c () Yes, this is an equivalence relation. Reflexive: Ifa is circle, then a ~ a because a has the same center as itself. Symmetric: Assume a ~ b. Then a and b have the same center. Thus, b and a have the same center, $06 ~ a ‘Transitive: Assume a ~ band b ~ c. Then a and b have the same center and b and c have the same center, 80 a and c have the same center. Thus, a ~ ¢. (© Yes, this is an equivalence relation. Reflexive: If a is a line, then a i parallel to itself, so a ~ a. ‘Symmetric: Ifa ~ b, then « is parallel to b. Thus, bis parallel to a. Hence, 6 ~ a ‘Transitive: Ifa ~ b and b ~ c, then a is parallel to b and b is parallel toc, s0 a is parallel to Thus, a~c, (@) No, this is not an equivalence relation, Tke reflexive property does not hold because no line is perpendicular to itself. Neither is this relation transitive; if &, is perpendicular to & and &2 is perpendicular tof, then é, and £5 are parallel, not perpendicular to one another. .(@) [BB] R= {(1,1), (1,2), (2, 1)5(2,2)s(8,8)s (8,4), (8,5); (4,4), (4,5), (5,5), (4,3), (5,3), (.4)} (b) R= {((1,1), (2,2), (3,3), (3,4), (4,3), (4,4), (555)} (©) R= {(1,1) (1,2) (1,3)s (154), (14 5)s (21) 5(2,2)s (2,3), (254), (255), (8,1), (8,2); (8,3), (8,4), (3,5)s(4,1)s (4,2) (4,3), (4,4), (4,5); (5s 1)5(5,2)s (5,3); (5,4) (5,8)} 15. (@) As suggested in the text, we list the partitions of {a}. There is only one; namely, {a} (b) As suggested in the text, we list the parttions of {a,b}. There are two; namely, {a,b} and {a}, {0}, (©) [BB] As suggested in the text, a good way to list the equivalence relations on {a, b,c} isto list the partitions ofthis set. Here they are: { {a}, {6}. fc} { {abc}: { {a,b}, {c} }5 { {asc}, {6} Ji { {bc}, {ah } ‘There are five in all. (@) As suggested in the text, we ist the patitions of (a5, c,d} { {a}, {0}, fe}, {a} } {{a,b,e.d) } { {a,b}, {4} }: { {aye}, {bd} }: { {0,4}, {b,0} } { {a,b}, Le}, {4} 3s { {are}, {0}, £4} Js (Lard), (0), fe} Js { {0,2}, fa), fa} { {bd}, {a}, {ch}: {Led}, {a}, {0} { {asb eh, {4} fs { {5b d}, fe} { {aed}, {0} 3: { {ed}, {a} } ‘There are 15 in all.“ Solutions to Exercises 16, (a) {BB] The given statement isan implication which concludes “x — y = x ~ y" whereas what is required i a logical argument which concludes “so ~ i reflexive.” Acorrect argument is this: For any (x, y) € R?, x — y = x — y; thus, (x,y) ~ (z,y). Therefore, ~ is reflexive, (b) There is confusion between the elements of binary relation on a set A (which are ordered pairs) and the elements of A which are themselves ordered pairs inthis situation, The given statement is corect provided each of z andy is understood to be an ordered pair of eal numbers, and we understand R = {(x,y) | x ~ y} but this is very misleading. Much better is to state symmetry like this: it ey) ~ (uy2)sthon (u,0) ~ (20) (©) The first statement assets the implication “z — y = u—v —+ (z,y) ~ (1nv)" which isthe converse of what should have been sid. Here isthe correct argument: 1F (eu) ~ (w0)s then 2 y= w~v,s0u~v = 2—yand, hence, (1,0) ~ (2,)- (d) This suggested answer is utterly confusing. Logical arguments consist ofa sequence of implica- tions but here i isnot clear where these implications start. Certainly the fist sentence isnot an implication. If (x,y) ~ (u,v) and (u,v) ~ (w, 2) then —y = u—vandu— «© —y=w-—zand, hence, (x,y) ~ (w, 2). (©) ~ defines an equivalence relation on R? because it isa reflexive, symmetric and transitive binary relation on R?, w—2. So (® The equivalence class of (0,0) is {(a.y) | (tu) ~ (0,0)} = {(@y) | z—y = 0-0} = {(z¥) |v = 2} Which i a straight line of slope 1 in the Certesian plane passing through the origin. The equiva- lence class of (2,3) is {(z,v) | (@u) ~ (2,3)} = (2,9) |z- y= 2-3 = 1} = {(z,y) [y= 24 1h ‘which isa straight line of slope 1 passing through the point (2,3). 17. [BB] Reflexive: If (z,y) € R2, then 2? — y? =z? —y?, so (x,y) ~ (z,). ‘Symmetric: If (x, y) ~ (u,v), then 2? — y? =u? — v?, sou? — a? —y? and (u,v) ~ (2,9). ‘Transitive: If (¢,y) ~ (u,v) and (u,v) ~ (w,2)sthen 22 — y® = u? — v? and w? — v? = w? — 22, $0.2? — y? =u? =v? = w = 29, 2? — y? =v? — 2? and (2,y) ~ (w,2). Pore ‘Thus, the equivalence class of (0,0) isthe pair of lines with equations y = 1,0) = {(2,u) | (24) ~ (1,0)} = {(2,y) | 2? -9? =? -P = 1). ‘Thus, the equivalence class of (1,0) is the hyperbola whose equation is 2? (zu) | (eu) ~ ,0)} = (zu) | 2? 18, (@) This is an equivalence relation. Reflexive: If (a,8) € R®, then a +2b= a+ 2b, s0(a,8) ~ (a,b). Symmetric: If (a,b) ~ (c,d), then a + 2b= c+ 2d, so c+ 2 a+ 2b and (c,d) ~ (4,8).Section 2.4 45 ‘Transitive: If (a,b) ~ (c,d) and (c,d) ~ (e, f), then a +2b= c+ 2dandc+2d=e + 2f,s0 a+2b=e + 2f and (a,b) ~(e,f). ‘The quotient set isthe set of equivalence classes. We have (ab) = {(x,) | (zy) ~(@,8)} = {(2,y) | 2+ 2y = a+ 2b} ={(z,y) |y—b=-3(2-a)} Which describes the line through (a,8) with slope —}. The quotient set isthe set of lines with slope —. (b) This is an equivalence relation, Reflexive: If (a,8) € R?, then ab = ab, s0 (a,8) ~ (2,6). Symmetric: If (a,8) ~ (c,d), then ab = et so ed = ab and (c,d) ~ (2,6). ‘Transitive: If (a,b) ~ (c,d) and (c,d) ~ (e, f), then ab = ed and ed = ef, s0 ab = ab = ef and (2,8) ~ (e, f)- ‘The quotient set is the st of equivalence classes. We have (2,5) = {(zy) | (zu) ~ (a,6)} = {(@,y) | zy = ab} and consider two cases. If either a. or b = 0, then (a,6) = {(x,y) | cy = 0}; that is, {(x,v) | z = Oory = O}. Hence, (a,6) is the union of the z-axis and the y-axis. On the other hand, ifa #0 and b # 0, then TD = {(z,y) | ay = ab) ={ewvly= since 2 # 0 in this case. This time, (a, ) isthe hyperbola whose equation is y = ab /x. (©) This is not an equivalence relation. We have (0,2) ~ (1,1) because 0? +2 = 2 = 141% however, (1,1) 7 (0,2) because 1? +1=2 4 4 = 0-+ 2°. The relation is not symmetric, (@) Reflexive: For any (a, 6) € R?, a = a, s0 (2,6) ~ (a,b). Symmetric: If (a,b) ~ (¢,d), then a = ¢, 50 ¢= a and, hence, (c,d) ~ (a,2). ‘Transitive: If (a,b) ~ (c,d) and (c,d) ~ (e,f), then a = cand ¢ = e; hence, a = ¢ and 50 (a,b) ~ (e, f). Since the elation is reflexive, symmetric ard transitive, its an equivalence relation. The quotient sets the set of equivalence classes. The equivalence class of (a, 6) is {(@u) ER? | (2) ~ (a8) Geometrically, this set is the vertical line with equation x = a. The quotient set is the set of (2,9) €R? | z= a} vertical lines (©) This is not an equivalence relation. For example, itis not reflexive: (1,2) % (1,2) because 12) =241=1 19. (a) “Ifanb =O, thena 4 5” (b) The converse is true. ITB = then a €abuta ¢ Bo 5, 20. Remembering that zis just the set of elements equivalent to x, we are given that a ~ b, ¢~ dand d~b, By Proposition 2.43, a = 6,46 Solutions to Exercises 21. (@) [BB] The ordered pairs defined by ~ are (1, 1), (1,4), (1,9), (2,2). (2,8). (3,3), (4,1), (4,4). (4,9), (5,5), (6,6), (7,7), (8,2), (8,8), (9:1), (9,4), (9,9). (®) [BB] T = {1,4,9} j.2 = {2,8} =3,3 = (3};5 = (5): 6 = {6};7 = {7}. (©) [BB] Since the sets {1,4,9}, {2,8}, {3}, {5}. {6} and {7} partition A, they determine an equiv- ‘lence relation, namely, that equivalence relation in which a ~ b if and only if a and b belong to the same one of these sets. This is the given relation. 22, [BB] Reflexive: Ifa € A, then a? is a perfect square, so a ~ a. ‘Symmetric: If a ~ b, then ab is a perfect square. Since ba = ab, ba is also a perfect square, so b ~ a. ‘Transitive: If a ~ b and b ~ c, then ab and bare each perfect squares. Thus ab = 2? and bc = y? forineger andy, Now aie = 247, wae = 2" = (2). Resa ac ian nee, 0 ao 23. (a) The order pairs of ~ are (1,1), (2,2), (3,8), (4,4), (5,5), (6,6). (7,7). (1,2) (2,1), (1,4). (4,1), (2,4), (4,2), (8,6), (6,3). (6) The sets {1,2,4}, {3,6}, {5}, {7} patition A, so the given relation is an equivalence relation. 24, Reflexive: Ifa € A, then © = 1= 2° is apower of 2, soa ~ a. Symmetric: fa ~ 6 then $ = 2,50 2 = 2, Since —t is an integer, © is also a power of 2, so bra. Transitive: a ~ band b~c,then & = 2 and 2 = 2 for integers and. Thus © showing that a ~ c 25. We have to prove that the given sets are disjoint and have union S. For the later, we note that since Ris reflexive for any a € S, (a,a) € Rand so a and a are elements of the same set Si; that is, @ € S; for some i. To prove that the sets are disjoint, suppose there is some « € Sy Se. Since Se Z Usa Su there exists y CS such that y ¢ S; for any j 7 k. Similarly, there exists = © Se such that 2 ¢ 5) if j # €. Now if y,2 € Sj, then (y,2) € Rand 2, z € Se implies (2,2) € R. By transitivity, (y, 2) € R, hence, y and z belong to the same set. But the only set to which y belongs is Sx. Since z does not belong to Se, we have a contradiction: Nox € Sk. S¢ exists. Exercises 2.5 1. (@) {BB} This defines a partial order. Reflexive: For any a € R,a >a. Antisymmetrie: Ifa, € R, a > band b > a, then a ‘Transitive: If a,b,c € Ra > band b > c,then a > c ‘This partial order is a total order because for any a, € R, either a > bor b > a. (©) [BB} This is nota partial order because the relation is not reflexive; for example, 1 <1 is not true.Section 2.5 a7 (©) This is not a partial order because the relation is not antisymmetric; for example, ~3 ~ 3 because (-3)? <3? and 3 = -3 because 3? < (—3)? but -3 4 3. (@) This is not a partial order because the relation is not antisymmetric; for example, (1,4) ~ (1,8) because 1 < 1 and similarly, (1,8) ~ (1,4), but (1,4) # (1,8). (©) This is a partial order Reflexive: For any (a,8) € N x N, (a,6) = (a,6) because a < a.and 6 > 6. Antisymmetrie: If (2,8), (c,d) € N x N, {a,6) = (¢,d) and (c,d) ~ (a,8), thena < 0,6 > d, ¢Saandd > b. Soa=c, b= d and, hence, (a,6) ‘Transitive: If (a,), (cd), (e, f) € NX N, (a,6) + (c,d) and (c,d) = (¢, f), then a < c,6 > d, eS eandd > f. Soa < e (because a < ¢ f (because b > d > f) and, therefore, (a,b) = (e, f). ‘This is not a total order; for example, (1,4) and (2,5) are incomparable. (® This is reflexive and transitive but not antisrmmetric and, hence, not a partial order. For example, cat < dog and dog ~ cat but dog # cat. =( 2. (@) [BB] 1, 10, 100, 1000, 1001, 101, 1010, 11, 110, 111 (b) 1, 11, 111, 110, 10, 101, 1010, 100, 1001, 1000 3. (a) [BB] (a,b), (a,c), (a,d), (b,¢),(b,d),(c.2); (6) [BB] (a, 6), (c,d); ©) (a,b), (a,d), (ed); @ (a,8),(a,2), (a,d); (©) (ad), (aye), (65), (b,€)s (bf), (f)s (se); (D(a, f)s (b,¢)s (4s )s (ds €)s (ds), (ds), (€5€),(€,4)s (95 F)5 (B44). 4. (a) [BB] ais minimal and minimum; d is maximal and maximum. (b) [BB] a and c are minimal; b and d are maximal; there are no minimum nor maximum elements. (©) @ and c are minimal; b and d are maximal; there are no minimum nor maximum elements. (@) ais minimal and minimum; 6, c and d and maximal; there is no maximum element. (©) aand b are minimal; ¢ and f and maximal; there are no minimum nor maximum elements. (0 a,d, ¢ and g are minimal; c, f and i are maximal; there are no minimum nor maximum elements. 5. @) (b) {a,b,c,d} {a,b,c} {0,5} {a,c} No {ed} {a} 6. (@) Lis minimal and minimum; 6 is maximal and maximum. (b) {a} and {c,d} are minimal; there is no minimum. ‘The set {a,6, ¢,d} is maximal and maximum. 7. (BB] A B and the set B contains exactly one more element than A.48 Solutions to Exercises 8. Helmut Hasse (1898-1979) was one of the more important mathematicians of the twentieth century. He grew up in Berlin and was a member of Germany's navy during the first World War. He received his PRD from the University of Gottingen in 1921 for a thesis in number theory, which was to be the subject of his life’s work. He is known for his research with Richard Brauer and Emmy Noether on simple algebras, his proof of the Riemann Hypothesis (one of today's most famous open problems) for zeta functions on elliptic curves, and his work on the arithmetical properties of abelian number fields. Hasse’s career started at Kiel and continued at Halle and Marburg. When the Nazis came to power in 1933, all Jewish mathematicians, including eighteen at the University of Gottingen, were summarily dismissed from their jobs. tis hard to know the degree of ambivalence Hasse may have had ‘when he received an offer of employment at Gétingen around this time, but he accepted the position, While some of Hasse’s closest research collaborators were Jewish, he nonetheless made no secret of his support for Hitler’ policies. In 1945, he wes dismissed by the British, lost his right to teach and eventually moved to Berlin. In May 1949, he was appointed professor at the Humboldt University in East Berlin but he moved to Hamburg the next year and worked there until his retirement in 1966. 9. (a) [BB] Let (A, x) be a finite poset and let a < A. If a is not maximal, there is an element a, such that ay > a. Ha; is not maximal, there is an element aa such that a2 > a). Continue, Since A is finite, eventually this process must stop, end it stops at a maximal element. A similar argument shows that (A, *) must also contain minimal elements, (b) The result of (a) is not true in general. For example, (R, <) is a poset without maximal elements ‘and without minimal elements. 10, (@) Reflexive: For any a = (a1,02) € A,a 2 (since {a} and {b} are not comparable if a 4 }), But 0 is a minimum because 0 is a subset of any set and the set $ itself is a maximum because any of its subsets is contained in i. (©) (Z,S) oF (R, <) are obvious examples. 17, Suppose a is maximal in a totally ordered set (A, ~) and let b be any other element of A. Since A is totaly ordered, either a ~ b or b < a. In the frst case, a = b because a is maximal so in either case, ba. Thus, ais a maximum, 18. (@) [BB] We have to prove that if b ~ a, then b = a. So suppose b = a. Since a is minimum, we have also a ~ 6, By antisymmetry, b = «. (b) Let b be a minimal element. We claim b = a. To see why, note that a minimum implies a = b. ‘Then minimality of b says a50 Solutions to Review Exercises Chapter 2 Review 1 . PA) = Since A = {1,2,3,4,5,6} and B = {3,4,5,6,7}, we have A @ B = {1,2,7} and (A@ B)\ C= {1,7}. (@ A={-1,0,1,2}, B = {-5,-3,-1, 1}, C = {-2,—2,0,41, 42,43, 43}; (©) ANB = {41}, $0(ANB) x B= {(1,—8)s(14—B)s (0,1), (L41)s (1-8). (1 3), (A=), (DK © B\C={-5,-3; — @) A@C={-3,-3,-2, 4}, 44). (a) True, (—+) Suppose AN B = A and leta € A. Thena € AN B,soa € B. Thus AC B. (—) Suppose A C B. To prove AN B = A, we prove each of the two sets, A B and A, is a subset ofthe othe, Lets © ANB. By defnition of 1 isin both A and B, in particular, x © A. ‘Thus AN BC A. Conversely, let x € A. Since A C B, x € B. Since z is in both A and B, 2€ ANB. Thus AC ANB. (©) This is false. If A = 0, (AM B)UC = C while AN (BUC) = 0, sony C # 0, any B, and A= 0 provides a counterexample. (©) False, Take A = B= 0, 1. Let b € B and let a be any element of A. Then (a,8) € A x B, so (a,b) € Ax C. Thus b € C. This shows tht BC C and a similar argument shows C'C B, so B = C. (a) Region 2: (ANC)\B_ Region3: ANBNC__ Region 4: (AN B)\ Region 5:(BNC)\ A Region 6: BX (AUC) Region 7: C \ (AUB) (b) Region 2, 3,4,5,7 is (AM B) UC; region 2,3,4 is AN (BUC) © BS (CA) consists of egons 3,4, and 6, (B.C) x A consists of region 6. “2 ‘ SE) ° ) i, (AUB)NC = {2,3,8,9} ii, AX (BX C) = {2,3,7,9) iii, A® B = {2,6,7,8,9} iv. (ANB) x (BNC) = {(2,3),(2,8),(7,3), (7,8), (9,8), (9,8)} 0, A}, s0 P(P(A)) = {0, {0}, {4}, {0,4}. (a) Take A = B = C = {3}. Then BX C = (50 A@(B\C) = A. On the other hand, A® B= 0 80(A®B)\C=0F A, (b) Let (a,8) € Ax B. Thena € Aandb € B. Since a € Aand AC C,a€ C. Since b € Band BCD,bE D. So(a,t) €C x D. Hence Ax BCC xD. ‘Take B= hand, AZ C. |, A = {1} = D. Then Ax B= = Cx D,s0 Ax BCC x D, Onthe otherChapter 2 5 10, 2. 13, 44 15, 16. ‘This follows quickly from one of the laws of De Morgan and the identity X \ ¥ = XY*, AX (BNC) = AN (BNC) = An (BUC) £ (AN BU (ANC) = (AN B)U(ANC), using (3), p. 62 atthe spot marked withthe arrow. (@) A binary relation on Aisa subset of Ax A. (b) If A has 10 elements, A x A has 100 elements, so there are 2!° binary relations on A. Reflexive: For any a with |a| < 1, we have a? = |a?| Symmetric by definition Not antisymmetric because (4, 4) is in ® (3)? < 4 and (3)? < }) but} # 4 Not transitive. We have (J, 3) ¢ because (J)? < § and (J)? < J and (§, 4) © R because (3)? $ }and (5)? < J, but (5, §) is notin R because (J)? £ §- {al|a < Jal, thus (a,a) € R. () Reflexive: For any natural number a, we have a < 2a, soa ~ a. Not symmetric: 2 ~ 5 because 2 < 2(6), but 5 7 2 because 5 £ 2(2). Not antisymmetric: Let a = 1 and b = 2, Then a ~ b because 1 < 2(2) and also b ~ a because 2< 21). Not transitive: Let a = 3, b = 2,¢ = 1. Then a ~ b because 3 < 2(2) and b ~ c because 2<2(1). However, a # cbecause 3 £ 2(1) Since the relation isnot transitive, its not an equivalence relation and its nota partial order. (b) Not reflexive: (1,2) ¢4 (1,2) because 1 <2 and 2 £1. Not symmetric: (1,2) ~ (4,3) because 1 < 2 and 3 <4, but (4,3) of (1,2) because 4 £3. Not antisymmetric: (1,1) ~ (2,2) because 1 < 1 and 2 < 2, and (2,2) ~ (1, 1) forthe same reason, but (1,1) # (2,2). ‘Transitive: if (a,8) ~ (c,d) and (c,d) ~ (¢, f).thena < bd < a implies a = 6. Since the hypothesis is always false, this implication is true. The relation is antisymmetric. Reflexive: For any a € Z, 4a +a = 5a is a multiple of 5. Symmetric: If @Rb, then 4a + bis a multiple of 5, so 46 +a = 5(a-+b) — (4a +b) is also a multiple of 5 that is, BR. ‘Transitive: If aRb and bRe, then both 4a + band 4b + c are multiples of 5, hence so is their sum, 4a + 5b + ¢, It follows that (4a + 56-+ c) — 56 = 4a + cis also a multiple of 5, so ae. (@) Reflexive: For any a € Z, aRa because 2a-+ 5a = 7a is a multiple of 7 ‘Symmetric: Ifa, € Z and aR, then 2a-+ 5b = 7k for some integer k, so 5a+2b = 7(a +6) — (2a-+ 5b) is the difference of multiples of 7, hence also a multiple of 7. Thus ba. ‘Transitive: If a,),¢ € Z with aRb and bRe, then 2a + 5b = 7k for some integer k and 2b + 5c = 7¢ for some integer &. Thus (2a +58) + (2b + 5c) = 2a + 7 + Se = 7(k + é) and 2a+ Se = 7(k + &) ~ 7bis the difference of multiples of 7, hence a multiple of 7. Thus aR.52 Solutions to Review Exercises (b) We have OR7 and 70, yet 0 # 7. The relation is not antisymmetric, so it is nota partial order. 17. (—) Assume x ~ a, Let t € #. Then t ~ 2 sot ~ a by transitivity. Thus ¢ € @. This proves FC a. imilarly, a C %, so = 7, (—) Assume Z =a. Since x € 7, by symmetry, x € 7. Thus 2 ~ a. 18. Since a € 5, a ~ band hence & = B by Proposition 2.43. Similarly d ¢ Bimplies @ = B, hence @ Now d ¢ Zimplies d # 2, so 2M d = 0 by Proposition 2.4.4, Since @ = d, so also = 0. 19, (@) We must show that = i reflexive, antisymmetric and transitive on A, The relation is reflexive: For any (a,B) € A, (a,b) ~ (a,8) because a Saand b 1, f(n) = n ~ 1, for example (©) the constant function f(n) = 107 forall n, for example. (@) the identify function, x, for example: u(r) = m, forall n. 5. (@) [BB] If z © X, then « is a country in th: British Commonwealth with a uniquely determined Prime Minister y who lives in that country: that is, y € Y. Ify € Y, then y isa person living in one of the countries in the British Commonwealth. Thus, the domicile of y is a uniquely determined clement x € X. (b) If-z, # zz are two different countries, then their Prime Ministers are different individuals, so Prime Minister is one-to-one. On the otherhand, Prime Minister is not onto since there are people {in a country who are not the Prime Minister. (© Itz € X, then x is a country. If y is any person living in that country, then Domicile: y -» 2, ‘so Domicile is onto. On the other hand, ify: and y2 are different people living in country z, then Domicile: yi +> x and Domicile: 1+ x, s0 Domicile is not one-to-one. 6. {BB} Parking rates, bus fares, admission prices are several common examples.10. 1 2 13. 14 15, 16. Solutions to Exercises (@) Jumps inthe graph of y = [2x ~ 3] occur whenever 2 ~ 3 is an integer, thats, at (3 | n € Z}. (b) Jumps in the graph of y = [}.2-+7] ovcur waenever 4+ 7 is an integer, that is, at {4n | n € Z}. (©) Jumps in the graph of y = [42 occur whenever 3(c + 3) is an integer, that is, Sn ~3,n € Z} t {el z= . (a) (BB) The answer is “yes.” By definition of “‘loor”, we know that kz is the unique integer satisfy- ing n —1 < ka 1. Thus, for example, 0 ¢ mg g. (6) [BB] g is not one-to-one as in (a). This time itis onto, however, because for n € N, the equation (x) has the solution = n— 1, (Note that n > O for n € N, so jn —1|=n—1,) (@) [BB] If f(e1) = f(x2), then 32 +5 = 323 +5, $0. Also f is onto, since given y € Q, y = f(}(y — 5)) with (©) This function is one-to-one, asin (a), but itis not onto because f(n) = $n +5 > 8 for any n € N and 50, for example, 1 is not inthe range off. (@) his not one-to-one since, for example, h(1) = 3 = A(—1) and # his not onto since 1 ¢ rng h: 1 € Z, but if we try to solve 2? + has no solutions in Z. (©) his one-to-one since if h(z1) = h(a), then 23 +2 = 2} + 2, 80.22 = 23 and 2, = xp, Since 2 and 22 are both in N, we must have 2 = 22. ‘his not onto; as in (a), 1 is not in the range of h.Section 3.1 55 17. (@) The function is not onto since, for example, g(z) = 0 has no integer solutions (by the quadratic formula), but itis one-to-one. To see why, suppose g(21) = (2). Then 32] + 14zy — 51 Sr} + 1dza ~ 51, so 3(z4 ~ 20)(e1 + 22) = 14(z> ~ 23). Thus 21 = 22 oF 2 + 22 =~ Since xy and are integers, x1 + 2 = —'f is impossible. Thus, 21 = 22, $09 is one-to-one. (b) Since the graph of gis a parabola, g is neither one-to-one nor onto. 18. (@) [BB] Note that f(z) = (x +7)? — 100. if f(21) = f(a) it follows that (21 + 7)? - 100 = (a2 +7)? — 100, so (x1 +7)? = (x2 + 7)? and, taking square roots, ry + 7| = laa + 7|. Since 1,2 € N, we know x1 +7 > Oand 22 —7 > 0. Thus, ¢1 +7 = 2+ 7 and 21 = za. Thus, F is one-to-one, but it is not onto: For example, 1 € B, but there is no x © N with f(x) = 1 since (2 +7)? — 100 = 1 implies (x + 7)? = 101 and this equation has no solution in the natural numbers. (b) This function is not one-to-one since, for example, f(0) = in @). (-14) 51) and itis not onto, as (© This function is not one-to-one; asin (b), f(0) = f(—14). But its onto since for any y > ~100, z= VIO — 7 isa solution o y= f(a) 19. (a) [BB] The domain of f is R. Its range is also R because every y € R can be written y = for some 2; namely, = y/®, The functin i, therefore, onto. Itis also one-to-one: If f(2 f(a), then 2} = 2} and this implies 2 =. () (b) The domain of g is R and so is the range. For the latter, note that 2 ife>0 aey={ -2 ife 0, then y = g(,/y) while if y < 0, y = g(-—/=9). Since rng g = R, g is onto. To see that g is 1-1, ether inspect the graph or assume g(r) = g(r) and consider the cases. Case 1: Both x; and x» are nonnegative. 4, 90 Va] = ay hence, [x3] Case 2: Both 2; and x2 are negative. In this case, —2} = ~23, so 2} = 23, V7] = YZ, [ail = [xal, Case 3: One of 21,222 is nonnegative and the other is negative. Here we may assume that 2 is nonnegative, so 23 = ~23. Since the lft side is at least andthe right side is less than 0, the equation cannot be true. Inall cases x; = 22, 80 gis 1-1. Intthis case, 2? |za| and 21 = 9. (©) The domain of @ is (4,00) as given. Its range is R because for any y € R, y = A(x) for x = 12" +4), Since mga = R, f isonto, Itis also one-to-one: If (1) = (a2), then aleea(32i-4) — ploes(8e3-4), 59 32, — 4 = 32q —4and 2) = Z9. (@) The domain of f is Ras given. To find the ange, we must remember that 2 > 0 for all t and thus, ‘S(2) > 8 forall x. Infact mg f = (3,00) since, for any y > 3, y = f(t) for x = 1+loga(y-3). Since mg f # R, f isnot onto. Itis one-to-one, however, for if f(z1) = f(a), then 2-243 = 28-1 4-3, 90.2811 = 29°1. Taking logy of each side, 21 — 1 = 22 —1,$0 21 = 2256 Solutions to Exercises 20. (@) (BB] The graph of f shown at the right makes it clear that f is v ‘one-to-one and onto (b) [BB] Solution 1. Since f: R > R is one-to-one by part (a), itis also one-to-one as a function with domain Z. Here, however, f is rot onto for we note that f(0) = 0, f(1) = 4 and fis increasing, 0 1 is not in the range of f. Solution 2, (This solution mimics that given in Problem 8 in our discussion of discrete functions in this section.) If f(x1) = f(x2), then 3x} +2, = 32}+22, s03(x3—x3) 2 —ay. Ify # 22, we must have 2f + 222 +2} ‘Thus, 2) = 29 and fis one-to-one, 22) and 3(1—22)(3-+21 22429) = 4, which is impossible for integers 21,3. On the other hand, f is not onto. In particular 1 is not in the range of f since 1 = f(a) for some a implies 3a° +a = 1; that is, (3a? + 1) = 1. Bat the only pairs of integers whose product is 1 are the pairs 1,1 and —1, —1, Sohhere, we'd have to have a = 3a? +1 = 1 ora = 3a? +1 = —1, neither of which is possible, 21. (@) The graph of g shown at the right makes it clear that g is ‘onto but not one-to-one. (©) Since 9(0) = 9(1), 9 isnot one-to-one. Furthermore, 9 is rot onto; for example, 2 = g(2) has no integer solution since 2 = 2° — 2+ 1 implies 2(2* — 1) 2),s02 = 22-1 = which is possible. (©) The argument given in (b) applies to this case as well, so g is not onto. ‘To determine whether or not g is one-to-one, suppose g(1) = g(z2). Then 2} ~ 21 +1 = 23 ~ 22-41, 0 (2 ~22)(e} + yt2+ 23) = a1 — a. If-zy # 22, we rust have 22 + 212 + 23 = 1, For natural numbers 71,22, this is not possible because a} + xyx2 + x3 > 3. Thus, x1 = zz and g is one-to-one, 22. (@) (BB] This is not one-to-one since, for example, f(1,3) = f(4,1) = 11 while (1,3) # (4,1). ‘The function is not onto since, for example, the equation f(z) = 1 has no solution x: = (n,m) (because 2n + 3m > 5 for every n,m € N). (©) [BB] This is not one-to-one since, for example, f(1,3) = f(4,1) = 11 while (1,3) # (4,1). The function is onto, however, since for k € Z, the equation k = f(z) has the solution x = (—k,k) (©) This is not one-to-one since, for example, f(12, 1) = f(1,8) (= 190) but (12, 1) # (1,8). Its not conto because (n,m) is even for any (n, mm) so, for example, the equation f(n, m) = 1 cannot be solved. (@) This is not one-to-one since, for example, f(246,0) = f(0,89). It's onto since (—17)(246) + 47(89) = 1 implies that for any k € Z, k = (—17k)(246) + (47k)(89) = f(47k, -17h). () This is not one-to-one because, for example, f(7,8) = f(8,7). I's not onto because, for example, the equation f(r) = 4 has no solution = (n,m): n? +m?-+1 # 4 forn,m € Z. £(3,2) (O This is not one-to-one because, for example, f(2,2) £(1,2) and, for n > 2, we have n = f(n = 1,1). ). Ibis onto, however, becauseSection 3.1 87 23, (+) Suppose fis one-to-one and 21 # 2a. We must prove that 23-4 2122-4 23-+a(2y + 22) +b #0. Since f is one-to-one and 21 % 2, we know that f(21) # (xa); thus, 2} + az} + ba +e 7 3 + ax} + bra + c; equivalently, (ef = 2) + (29 — 28) + 8(01 — 2) £0. Since x; ~ x2 # 0, we may divide this last inequality by x1 — 2. Since x} ~ 23 = (x1 — 22)(2}+ 2122 + 23) and 2} — 23 = (xy —22)(21 + 22), we obtain (23 + 2122 + 23) + a(x +22) +040 as required, (Suppose that ry 2 implies 2} + xy: +23 + a(y + a2) +b # 0. We must prove that fis ‘one-to-one. Thus, we assume f(z) = f(e2) and prove 2; = 22. The equation f(z1) = f(x2) says i+ aa} +bry +e= xf + axf + bee + ¢; or equivalently, as before, (9 = 28) + (a - 2) + 6(¢1 - 22) =0. If, # 2, then we divide by x ~ 22 as before obtaining 2} + 22 +23 + a(t +42) += Oa contradiction, Thus 21 = 2. 24. (@) IBD] We rege = #8 ap we ake A (2 R] 2 3}. Teng = {yy #0} because if v#0y= f(a) forz=3+ 1 dom (b) We require 1 — 2 > 0, so we take A= ee {u | y > 0} because for any y > 0.y = f(a) forz =1- 4 €domf. 25. (@) The function is not one-to-one because (a,2) and (d,<) are both in f, but a # d. Restricting the domain to {a,b,c}, for instance, we obtain the one-to-one function {(a, a), (b, 8), (c,7)}: (b) [BB] Note that f(2) = ~(2x — 3). This function is not one-to-one since, for example, f(1) = £(2) (= =1). Restrict the domain to {z | x > 3/2} (or to {z | x < }}). (©) The sine function is not one-to-one since, for example, sin = sin. The largest interval on which a one-to-one function can be defined has length x; for example, {z | —x/2 X. The one-to-one functions from X — ¥ are {(a,1), (b,2)} {(@,1), (6,3)} {(@,2),(6,)} {(a,2),(b,3)} {(@,3),(6,0)} {(@,3),(0,2)}. (©) [BB] There are no onto functions X + ¥. The onto functions ¥ —+ X are {(1,4), (2,4), (3,8)} {(1,a), (2,6), (3,4)} {(1,a), (2,0), (3,6)} {(1,b),(2,a), (3,4)} {(1,6), (2,4), (3,0)} {(1,4); (2,6), (3,4)}. 29. (a) There are no one-to-one functions Y —+ X, Here are the 24 one-to-one functions X —> Y. {(a,1),(b,2),(¢,3)} {(a,1), (6,2), (c,4)} {(a, 1), (0,3), (€,2)} (@).0,3.64)} {(0,0.0,44(6.2)} {0.1} (6.4) (63)) {(a, 2), (b,3), (c,I)} {02.9 (64) {(a,2), (0,4), (6 1)} ; } {(a,2), (0,4), (¢,3)} {(a,2), (2 {(a,2),(b,1), (c,4)} {(a,3), (0,1), (€,2)} {(a,3),(b,1),(c,4)} {(a, 4), (6,1), (c,2)} {(a,4),(0,1),(3)} {(@,3),(b,2),(1)} {(a,3), (b,2), (c4)} {(@,4); (0,2), (c1)} {(a,4);(0,2),(c,3)} {(a,3), (0,4), (,1)} {(2,3), (0,4); (6,2)} {(4,4),(6,3),(,1)} {(a,4), (6,3), (€.2)} (b) There are no onto functions X —+ Y. Here are the 36 onto functions ¥ —> X. {(1,4), (2,4), (3,8), (4,e)} {(1,a), (2,4), (3,¢),(4,5)} {(1, 6), (2,8), (3,4), (4,¢)} {(1,5),(2,b),(3,¢),(4,a)} {(1,¢),!2,¢), (3,),(4,6)} {(1,0), (2,¢), (3,6), (4,a)} {(1, a), (2,0), (3,4), (4,c)} {(1,@), (25), (3,a),(4,0)} {(1,), (2,4), (3,6), (4,¢)} {(1)s (2,0), (3,8), ia a)} {(1,e), (2,4), (3,¢),(4,b)} {(1,¢), (2,8); (3,0), (4,4)} {(1,a), (2,0), (3,¢),(4,a)} {(1,4), (2,¢), (3,6), (4,a)} {(1,8), (2,4), (8,¢),(4,0)} {(1,0),(2,0),(B,a),(4,6)} {(1se)s(2.a),(B,8),(4,e)} {C0}, (2.8) (ya), (4,e)} {(1,b); (254), (35), (4,€)} {(1s€)s (25a), (3,4), (4,)} {1 a), (2,0), (3,0), (4,0)} {(1¢)5 (256), (3,8), (4,a)} {(1,a)s(2,¢),(3,),(4,6)} {(1,0), (250), (3,¢), (4,0)} {(1, 4), (2,4), (3,¢),(4,a)} {(1,€)s 12,0), (3,8), (4,a)} {(1,a@), (2,0), (3,¢),(4,6)} {(1¢)5 (256), (8,4), (4,0)} {(1s4)5(2,e), (3,6), (4se)} (140), (250), (3,4), (4,0)} {(1,4)s (2s), (3,),(4,a)} {(1,€)s 12,6), (8,4),(4,a)} {(1,a), (2,0), (3,0), (4,6)} {(1,¢); (254), (3:5),(4,0)} {(154)5 (2,0), (8,¢)s(4e)} {(1,4), (2,4), (3,0), (4,e)}Section 3.2 59 ‘We guess that the number of functions X > Y isn. 2 Et 31. (@) Let A = {a1,43,... 44m}. If f: A+ B were one-to-one, then f(a), contradicting the fact that B has only m < () Let B = {b1,bay...,0m}- Ef: A B were onto, in A such that f(a,) = i, distinct by definition of has n < m elements. ld exist elements a),.02)..- 4m 32. (@) (BB (—)] Suppose A and B each contain nelements. Assume that f: A —+ B is one-to-one F(a) | a € A}. Since f(a1) # f(x) if a # az, C is a subset of B containing n ‘elements; s0 C= B. Therefore, fis onto. Conversely, suppose f is onto. Then {f(a) | a € A} = B and so the set on the left here contains nistinct elements. Since A contains only n elements, we cannot have f(ax) = f(aa) for distinct us, fis one-to-one. 1a forall a € N. This does not contradict (a) because N is not finite (©) f(1) =1, f(n) =n—1 for n > 2, Again there is no contradiction since N is not finite, 33. (@) (BB) (#) (BB] The domain of f is R. The range is the set ‘ofall real numbers ofthe form 2n.++ a, where n € Z and 0 < a < 1. To prove this algebraically, we first rote that any y = 2n + ais f(a) for 2 =n-+a. On tke other hand, suppose y = f(z) for some. Writing [z].0 0,thus y= + “1(a), (c). Then z = f(y) = y?, soy = YZ. Since y € dom f,y < 0, soy = Vz =60 Solutions to Exercises 4, Lety = f-¥(2). Then 2 = f(y) = —y.50y = =F). 5. Since g(z) isan integer, fo g(z) = g(a). Similarly, 90 f(z) = f(x). 6, (a) (BB) f-?: R— Ris defined by f-*(z) = }(a—5). (0) J74: Ro Ris defined by fz) = ( +2)" (© B74: R—+ (4,00) is given by B-"(z) = 3(2" +4). 7 1 VE ifz>0 ARO Ris Aa) = (@) gut: R— Ris defined by 9~*(2) { Ve iteco. ow @ Nik 7. (@ (BBIIE f(21) = faa), then 1+ WG = 14 a and 1 —4= 42-4, a ‘Thus 2 = 22 and f is one-to-one. Next vemgfoy=f(2) forsomere A © there is an z €-A such that y = 1+ = ‘e+ there is an x € A such that y— 1 za © there is an x € A such that (y—1)(¢—4) =1 wyel Thus mg f = B= {ye Rv #1) and f as an inverse B+ A, To find formula for *), lety = f-\(2), 2 € B. Then z = f(y) =1+ wa-i= Th @-Dy-#=1 y and, since x #1,y—4= and foe) =y 2 16) te aite 1 (©) Suppose f(21) = (@2). Then 5— 4m ta +a IyaSection 3.2 6 and x1 = 22. Thus f is one-to-one, Next vemgfoy=f(z) forsomere A there is an x € A such that y = ie © there is an € A such that y— = +2 theres an x € A such that (y ~5)(1-+2) = —1 wuss = {y€R|y #5} and fhas an inverse B+ A. To find a formula for f(z), 1 l Tay @~ Gta) = 1 3e1 302 (© Suppose f(2i) = f(a). Then ay = govt y, 80 Oxize + xy = Gryty + Srp and 2 = 2p. Thus f is one-to-one. Now vemgfoy=f(z) forsomere A 3c. Bert there is an ¢ € A such that 2zy + y = 32 “+ there is an ¢ € A such that 2(2y — 3) + there is an ¢ € A such that y = 3 outs ‘Thus mg, ={y€R]y # 9) and J has an inverse B + A. To find a formula for f"(2), lety = f-"(a),2 € B. Then z = f(y) =, so 22y +2 = By and y(2z~3) y+ (x), -z, Since z-3 BS 8 2-8 co nyry4301—3r2-9 = m22-Br1 +3229, H43 nn F3 ‘2. Thus f is one-to-one. Next 2 € B,we know 2x —3 40; thus, (@) Suppose f(21) = f(v2). Then 62) = 6z9 and 2 = yemgfoy= f(z) forsomere A is a3 ‘+ there is an = € A such that y = 2—5 ‘+ there is an s € A such that y(x + 3) ‘+ there is an x € A such that 2 — ‘++ there is an x € A such that 2(1 — y) syAl ‘Thus mg f = Then x = f(y) = tél fl10. 12, 13, 4. Solutions to Exercises (a) First we show that f is one-to-one, Suppose then that f(1) = f(2). If xy and zp are both negative, then 2/14] = 2|ra|. Since || = —a for « < 0, we have ~2r, — —2z2 and hence, 1 = 22. If ay and zp are both > 0, wehave 2x +1 = 2a + 1, s0 2xy = 2p and again, ‘21 = 22, Finally, we note that it is impossible for f(x) = f(z2) with 21 <0 and za > 0 (or vice versa) since in this case, one of f(2). f(z2) is an even integer while the other is oda. Next we show that fis onto. Suppose thatn € N. If n = 2k is even, then n = 2|— k| = f(A) since —k <0, while if n = 2k-+ 1 is odd, then n = (k) since k > 0. Since f is one-to-one and onto, it has an inverse. (b) Let f-(2586) = a. We must solve the equation 2586 = f(a). Since 2586 is even, we sce that 2586 = f(—2586/2) = f(—1298). (@) [BB] maternal grandmother (b) patemal grandmother __(¢) maternal grandfather (@) mother-in-law © father (9 fatherintaw (you (h) maternal grandmother (i) maternal grandmother (@) [BB] fog = {(1,1), (3,8), (2,1), (4,9), 6. )}s {9 f is not defined because mg f = {1,2,3,8,9} £ domg = {1,2,3,4,5)- 0 F is not defined because rng f = {1,2.3,8,9} Z dom f = {1,2,3,4,5}. 90g = {(1,2), (2,2), (3,2), (4,1), (5,2)}. (®) f is one-to-one and onto; 17 is not one-to-one: It contains both (1,2) and (2,2); is not onto: 4 isnot in the range of g (©) Since f is one-to-one and onto, it hasan inverse: f-? = {(8, 1), (9,3), (3,4), (1,2), (2,5)}. (@ Since g is not one-to-one (or since g is not onto), g does not have an inverse. (@) [BB] 910 fog ={(1,3),(21,(3,2(4.4) 0) fog tog © go fog = {(1,2),(2,4),(3,3), 4D}: @ gogtof (© ft ogo fog = {(1,1); (24), (3,2), (4,3)}, 1 1 1 _142(a? +41) _ 2a? +3 (BoM) = Gyo yi ~ apa e el IO) gee aT hogof(e) =3:9eho f(z) = 9(8) = Since 1 0 f(z) = 2, we have go ft o f(z) (2) Since f-4(2) = 2 — 2, we have 1 Pegs) =S'(Gaa =fe+ aha go f(z) hogof(e) [BBI(g f)(z) = a(f(2)) = F(@) — c. Thus the graph of g 0 f is the graph of f translated vertically Ccunits down if c > 0 and —c units up if ¢ < 0. The graphs are identical if ¢ = 0.Section 3.2 63 15. (fog)(z) = f(g(2)) = Fe —c). Thus the graph of f 0 is the graph of f, but translated horizontally units to the right if ¢ > O and —c units to the left if ¢ < 0. The graphs are identical if ¢ = 0. ince <|e| {®t 20 cole) = f(cle L202) fe>0 16. (a) [BB] Since wife Tee) wehave Fogle) =f m= {45 a So the graph of f 0 g is the same as the graph of f to the left of the y-axis (where « < 0) while to the right of the y-axis, the graph of f og is the reflection (mirror image) of the left half of the ‘graph of fin the y-axis. We call fog an even function since it is symmetric with respect tothe aris: fo9(-2) = fo9(2) ; =f(e) itf(2) 20 b) Since go f(z) = =| f(@)| = the graph of of isthe same a the, (©) Since go f(2) = -I4(@)| ia styl) co meeaPhof aos raph of f wherever the graph of f is below the z-axis (y < 0) while it is the reflection (mirror image) of the graph of f in the z-axis wherever the graph of f is above the x-axis. (In particular, the ‘graph of go f lies entirely on or below the z-axis.) 17. (@) feg(z) = flo(z)) =F (1-2) =a). gor(z) = 9(r(z)) = 9( >) eis 1-2=s(z), ee fakrs wpe fag hres tipo eshr aloe frsh Allhr se fg rir shoes slsnrr fg . function || e| f] 9 |r| r|s by All these functions have in o re nee ‘verse ela t ft ete ts inction fi| fo| fa | fa | fo 0 Att totimtneineres ol Lad al Le 19. (@) (BB) f og = {(144)s (2,8) (8,2) (451), (5,5): 90 F = {(145)s (2,3), (8,2) (444), (5, 1)} Clearly, fog #90 f. (b) f-* = {(1,2), (2,1), (8,5), (4,3), (5,4) } 9 = {(1,3)5 (24), (3,1), (4,5), (5,2)} Functions f and g have inverses because they are one-to-one and onto while h does not have an inverse because it is not one-to-one (equally because it is not onto).a1. 2 24, Solutions to Exercises (©) (Fog) = {(1,8)s (2,8) (8,2) (4, 1)6(5,8)) go f-* = {(1,4),(2,3)5(3,2)s (451), (5,5)} = (Fog). $10.97? = {(1,5), (2,3); (3,2); (4,4), (5,)} # (Fog)-* a) —We-1) 2 2n. 2 = tg 7 Feats (@) (BB) Forz € B, (fo9)(2) = (2 =n (b) [BB] For z € A, (go f)(z) = g(¢ Proposition 3.2.7, f and g are inverses. (a) Suppose 0 € A and let a = f(0). Then 0 = f-*(a) ‘number f(a). (b) Since f and f* each have domain A, we have only to prove that f*(a) = a for all a € A. Let then a be some element of A. Let a; = f(a), a2 = f(a) = f2(a), a3 = f(aa) = f(a) and 1 = 77 which is not possible for any real Fo Pos y a4 = Has) = F(a). We mus show th ag = a, From a = f(a), we have a= f¥(an) = Fleas 802 f(a,) = =. From a3 = f(a2), we obtain az = f-"(a3) = Flea) and so 1 t= aas desired 7a (@) [BB] Suppose 941) (ba) for by, bo € B. Since some a;,a2 € A. Thus, 9(f(ai)) = g(f(a2)); that one-to-one, ai = az. Therefore, bi = f(ai) = f(a) (B) Given b € B, we must find a € A such tha f(a) onto, there is some a € A with go f(a) = 9 F(a) = b, so we have the desired clement a by = f(a) and be = J(a2) for . 9 © f(a1) = go flag). Since go f is ba proving that g i one-to-one. Consider g(b) € C. Since go f: A> Cis is, a(f(a)) = 9(0). But g one-to-one implies (@) (BB) Suppose f: A + B and g: B + C are one-to-one. We prove that go f: A ~» Cis one-o- one. For this, suppose (90 f)(a1) = (9 fi(a2) for some ay, an € A. Then 9((a)) = 9(F(e2)) {an equation of the form g(b1) = 9(b2)]. Since g is one-to-one, we conclude that f(a) = f(a), and then, since f is one-to-one, that a; = a2. (b) Let f = {(1, 1), (2,2)} and g = {(—1,5),(1,5), (2, 10)}. Then g is not 1-1, but the composition 9° f = {(1,5), (2, 10)} is. (© Let f and g be functions, f: A + B,g: B-» C and suppose go f: A —+ C'is 1-1. Then f must be 1-1 since f(as) = f(@2) implies g(f(2x)) = g(f(02)) thats, g 2 f(ax) = 9 © fan). from which we conclude that a = a2, because 79 f is I=L (@) Suppose f: A+ B and g: B + C are onto. We prove that go f: A — C is onto, Let then ‘© C. We must find a € A such that go (a) = ¢; that is, @(f(a)) = c [an equation of the form ‘9(t) = cl. Since gis onto, there is b € B such that 9(6) = c. Also, since fis onto, we know there isa € A such that b = f(a). Thus, 9(f(a)) = 9(b) = cas desired. (&) [BB] With A = {1,2}, B= {-1, 1,2}, C= {5,10}, f = {(1,1),(2,2)} and 9 = {(-1,5), (1,5), (2,10)}, we have g of = {(1,5),(2,10}. Then go f: A —> C'is onto but f isnot. (©) Suppose f: A+ B, 9: BC and g¢ f: A + Cis onto. Then g must be onto because if © € C, there exists a € A with g o f(a) = ¢(g 0 f is onto), so g(f(a)) = c. Thus, there exists be Bwith g(0) =e = f(a).