Unformatted text preview: Section 1.1 Propositional Logic 1 CHAPTER 1
The Foundations: Logic and Proofs
SECTION 1.1 Propositional Logic 2. Propositions must have clearly deﬁned truth values, so a proposition must be a declarative sentence with no
free variables.
a) This is not a proposition; it’s a command.
b) This is not a proposition; it’s a question.
c) This is a proposition that is false, as anyone who has been to Maine knows.
d) This is not a proposition; its truth value depends on the value of x .
e) This is a proposition that is false.
f) This is not a proposition; its truth value depends on the value of n .
4. a) I did not buy a lottery ticket this week.
b) Either I bought a lottery ticket this week or [in the inclusive sense] I won the million dollar jackpot on
Friday.
c) If I bought a lottery ticket this week, then I won the million dollar jackpot on Friday.
d) I bought a lottery ticket this week and I won the million dollar jackpot on Friday.
e) I bought a lottery ticket this week if and only if I won the million dollar jackpot on Friday.
f) If I did not buy a lottery ticket this week, then I did not win the million dollar jackpot on Friday.
g) I did not buy a lottery ticket this week, and I did not win the million dollar jackpot on Friday.
h) Either I did not buy a lottery ticket this week, or else I did buy one and won the million dollar jackpot on
Friday.
6. a) The election is not decided.
b) The election is decided, or the votes have been counted.
c) The election is not decided, and the votes have been counted.
d) If the votes have been counted, then the election is decided.
e) If the votes have not been counted, then the election is not decided.
f) If the election is not decided, then the votes have not been counted.
g) The election is decided if and only if the votes have been counted.
h) Either the votes have not been counted, or else the election is not decided and the votes have been counted.
Note that we were able to incorporate the parentheses by using the words either and else .
8. a) If you have the ﬂu, then you miss the ﬁnal exam.
b) You do not miss the ﬁnal exam if and only if you pass the course.
c) If you miss the ﬁnal exam, then you do not pass the course.
d) You have the ﬂu, or miss the ﬁnal exam, or pass the course.
e) It is either the case that if you have the ﬂu then you do not pass the course or the case that if you miss
the ﬁnal exam then you do not pass the course (or both, it is understood).
f) Either you have the ﬂu and miss the ﬁnal exam, or you do not miss the ﬁnal exam and do pass the course. 2 Chapter 1 10. a) r ∧ ¬q
12. a) This is
b) This is
c) This is
d) This is b) p ∧ q ∧ r c) r → p d) p ∧ ¬q ∧ r The Foundations: Logic and Proofs e) (p ∧ q ) → r f) r ↔ (q ∨ p) T ↔ T , which is true.
T ↔ F , which is false.
F ↔ F , which is true.
F ↔ T , which is false. 14. a) This is F → F , which is true.
b) This is F → F , which is true.
c) This is T → F , which is false.
d) This is T → T , which is true.
16. a) The employer making this request would be happy if the applicant knew both of these languages, so this
is clearly an inclusive or .
b) The restaurant would probably charge extra if the diner wanted both of these items, so this is an exclusive
or .
c) If a person happened to have both forms of identiﬁcation, so much the better, so this is clearly an inclusive
or .
d) This could be argued either way, but the inclusive interpretation seems more appropriate. This phrase
means that faculty members who do not publish papers in research journals are likely to be ﬁred from their
jobs during the probationary period. On the other hand, it may happen that they will be ﬁred even if they
do publish (for example, if their teaching is poor).
18. a) The necessary condition is the conclusion: If you get promoted, then you wash the boss’s car.
b) If the winds are from the south, then there will be a spring thaw.
c) The suﬃcient condition is the hypothesis: If you bought the computer less than a year ago, then the
warranty is good.
d) If Willy cheats, then he gets caught.
e) The “only if” condition is the conclusion: If you access the website, then you must pay a subscription fee.
f) If you know the right people, then you will be elected.
g) If Carol is on a boat, then she gets seasick.
20. a) If I am to remember to send you the address, then you will have to send me an email message. (This has
been slightly reworded so that the tenses make more sense.)
b) If you were born in the United States, then you are a citizen of this country.
c) If you keep your textbook, then it will be a useful reference in your future courses. (The word “then” is
understood in English, even if omitted.)
d) If their goaltender plays well, then the Red Wings will win the Stanley Cup.
e) If you get the job, then you had the best credentials.
f) If there is a storm, then the beach erodes.
g) If you log on to the server, then you have a valid password.
h) If you do not begin your climb too late, then you will reach the summit.
22. a) You will get an A in this course if and only if you learn how to solve discrete mathematics problems.
b) You will be informed if and only if you read the newspaper every day. (It sounds better in this order; it
would be logically equivalent to state this as “You read the newspaper every day if and only if you will be
informed.”) Section 1.1 Propositional Logic 3 c) It rains if and only if it is a weekend day.
d) You can see the wizard if and only if he is not in.
24. a) Converse: If I stay home, then it will snow tonight. Contrapositive: If I do not stay at home, then it will
not snow tonight. Inverse: If it does not snow tonight, then I will not stay home.
b) Converse: Whenever I go to the beach, it is a sunny summer day. Contrapositive: Whenever I do not go
to the beach, it is not a sunny summer day. Inverse: Whenever it is not a sunny day, I do not go to the beach.
c) Converse: If I sleep until noon, then I stayed up late. Contrapositive: If I do not sleep until noon, then I
did not stay up late. Inverse: If I don’t stay up late, then I don’t sleep until noon.
26. A truth table will need 2n rows if there are n variables.
a) 22 = 4
b) 23 = 8
c) 26 = 64
d) 25 = 32
28. To construct the truth table for a compound proposition, we work from the inside out. In each case, we will
show the intermediate steps. In part (d), for example, we ﬁrst construct the truth tables for p ∧ q and for
p ∨ q and combine them to get the truth table for (p ∧ q ) → (p ∨ q ) . For parts (a) and (b) we have the
following table (column three for part (a), column four for part (b)).
p
¬p
p → ¬p
p ↔ ¬p
T
F F
T F
T For parts (c) and (d) we have the following table.
pq
p∨q
p∧q
T
T
F
F T
F
T
F T
T
T
F (p ∧ q ) → (p ∨ q ) F
F
T
F T
T
T
T F
F
T
T p↔q (q → ¬p) ↔ (p ↔ q ) T
F
F
T F
F
F
T p ↔ ¬q (p ↔ q ) ⊕ (p ↔ ¬q ) F
T
T
F T
T
T
T F
T
T
T For part (f ) we have the following table.
p q ¬q
p↔q
TT
TF
FT
FF p ⊕ (p ∨ q ) T
F
F
F For part (e) we have the following table.
p q ¬p
q → ¬p
TT
TF
FT
FF F
T
F
T F
F T
F
F
T 30. For parts (a) and (b) we have the following table (column two for part (a), column four for part (b)).
p
p⊕p
¬p
p ⊕ ¬p
T
F F
F F
T T
T For parts (c) and (d) we have the following table (columns ﬁve and six).
pq
¬p
¬q
p ⊕ ¬q
¬p ⊕ ¬q
T
T
F
F T
F
T
F F
F
T
T F
T
F
T T
F
F
T F
T
T
F 4 Chapter 1 The Foundations: Logic and Proofs For parts (e) and (f ) we have the following table (columns ﬁve and six). This time we have omitted the column
explicitly showing the negation of q . Note that the ﬁrst is a tautology and the second is a contradiction (see
deﬁnitions in Section 1.2).
p p⊕q p ⊕ ¬q (p ⊕ q ) ∨ (p ⊕ ¬q ) (p ⊕ q ) ∧ (p ⊕ ¬q ) F
T
T
F q T
F
F
T T
T
T
T F
F
F
F TT
TF
FT
FF 32. For parts (a) and (b), we have
p q r T
T
T
T
F
F
F
F p∨q (p ∨ q ) ∨ r (p ∨ q ) ∧ r T
T
T
T
T
T
F
F T
T
T
T
T
T
T
F T
F
T
F
T
F
F
F p∧q (p ∧ q ) ∨ r (p ∧ q ) ∧ r T
T
F
F
F
F
F
F T
T
T
F
T
F
T
F T
F
F
F
F
F
F
F TT
TF
FT
FF
TT
TF
FT
FF For parts (c) and (d), we have
p q r T
T
T
T
F
F
F
F TT
TF
FT
FF
TT
TF
FT
FF Finally, for parts (e) and (f ) we have
p q r ¬r p∨q (p ∨ q ) ∧ ¬r p∧q (p ∧ q ) ∨ ¬r T
T
T
T
F
F
F
F TT
TF
FT
FF
TT
TF
FT
FF F
T
F
T
F
T
F
T T
T
T
T
T
T
F
F F
T
F
T
F
T
F
F T
T
F
F
F
F
F
F T
T
F
T
F
T
F
T 34. This time the truth table needs 24 = 16 rows. Section 1.1 Propositional Logic 5 q r s p→q (p → q ) → r ((p → q ) → r) → s TT
TT
TT
TT
TF
TF
TF
TF
FT
FT
FT
FT
FF
FF
FF
FF T
T
F
F
T
T
F
F
T
T
F
F
T
T
F
F T
F
T
F
T
F
T
F
T
F
T
F
T
F
T
F T
T
T
T
F
F
F
F
T
T
T
T
T
T
T
T T
T
F
F
T
T
T
T
T
T
F
F
T
T
F
F T
F
T
T
T
F
T
F
T
F
T
T
T
F
T
T p 36. a) Since the condition is true, the statement is executed, so x is incremented and now has the value 2 .
b) Since the condition is false, the statement is not executed, so x is not incremented and now still has the
value 1 .
c) Since the condition is true, the statement is executed, so x is incremented and now has the value 2 .
d) Since the condition is false, the statement is not executed, so x is not incremented and now still has the
value 1 .
e) Since the condition is true when it is encountered (since x = 1 ), the statement is executed, so x is
incremented and now has the value 2 . (It is irrelevant that the condition is now false.)
38. a) 1 1000 ∧ (0 1011 ∨ 1 1011) = 1 1000 ∧ 1 1011 = 1 1000
b) (0 1111 ∧ 1 0101) ∨ 0 1000 = 0 0101 ∨ 0 1000 = 0 1101
c) (0 1010 ⊕ 1 1011) ⊕ 0 1000 = 1 0001 ⊕ 0 1000 = 1 1001
d) (1 1011 ∨ 0 1010) ∧ (1 0001 ∨ 1 1011) = 1 1011 ∧ 1 1011 = 1 1011
40. The truth value of “Fred and John are happy” is min(0.8, 0.4) = 0.4 . The truth value of “Neither Fred nor
John is happy” is min(0.2, 0.6) = 0.2 , since this statement means “Fred is not happy, and John is not happy,”
and we computed the truth values of the two propositions in this conjunction in Exercise 35.
42. This cannot be a proposition, because it cannot have a truth value. Indeed, if it were true, then it would
be truly asserting that it is false, a contradiction; on the other hand if it were false, then its assertion that
it is false must be false, so that it would be true—again a contradiction. Thus this string of letters, while
appearing to be a proposition, is in fact meaningless.
44. No. This is a classical paradox. (We will use the male pronoun in what follows, assuming that we are talking
about males shaving their beards here, and assuming that all men have facial hair. If we restrict ourselves to
beards and allow female barbers, then the barber could be female with no contradiction.) If such a barber
existed, who would shave the barber? If the barber shaved himself, then he would be violating the rule that
he shaves only those people who do not shave themselves. On the other hand, if he does not shave himself,
then the rule says that he must shave himself. Neither is possible, so there can be no such barber.
46. a) If the explorer (a woman, so that our pronouns will not get confused here—the cannibals will be male)
encounters a truthteller, then he will honestly answer “no” to her question. If she encounters a liar, then the 6 Chapter 1 The Foundations: Logic and Proofs honest answer to her question is “yes,” so he will lie and answer “no.” Thus everybody will answer “no” to
the question, and the explorer will have no way to determine which type of cannibal she is speaking to.
b) There are several possible correct answers. One is the following question: “If I were to ask you if you
always told the truth, would you say that you did?” Then if the cannibal is a truth teller, he will answer yes
(truthfully), while if he is a liar, then, since in fact he would have said that he did tell the truth if questioned,
he will now lie and answer no.
48. a) “But” means “and”: r ∧ ¬p .
b) “Whenever” means “if”: (r ∧ p) → q .
c) Access being denied is the negation of q , so we have ¬r → ¬q .
d) The hypothesis is a conjunction: (¬p ∧ r) → q .
50. We write these symbolically: u → ¬a , a → s , ¬s → ¬u . Note that we can make all the conclusion true by
making a false, s true, and u false. Therefore if the users cannot access the ﬁle system, they can save new
ﬁles, and the system is not being upgraded, then all the conditional statements are true. Thus the system is
consistent.
52. This system is consistent. We use L , Q , N , and B to stand for the basic propositions here, “The ﬁle system
is locked,” “New messages will be queued,” “The system is functioning normally,” and “New messages will
be sent to the message buﬀer,” respectively. Then the given speciﬁcations are ¬L → Q , ¬L ↔ N , ¬Q → B ,
¬L → B , and ¬B . If we want consistency, then we had better have B false in order that ¬B be true. This
requires that both L and Q be true, by the two conditional statements that have B as their consequence. The
ﬁrst conditional statement therefore is of the form F → T , which is true. Finally, the biconditional ¬L ↔ N
can be satisﬁed by taking N to be false. Thus this set of speciﬁcations is consistent. Note that there is just
this one satisfying truth assignment.
54. This is similar to Example 17, about universities in New Mexico. To search for hiking in West Virginia, we
could enter WEST AND VIRGINIA AND HIKING. If we enter (VIRGINIA AND HIKING) NOT WEST,
then we’ll get websites about hiking in Virginia but not in West Virginia, except for sites that happen to use
the word “west” in a diﬀerent context (e.g., “Follow the stream west until you come to a clearing”).
56. If A is a knight, then his statement that both of them are knights is true, and both will be telling the truth.
But that is impossible, because B is asserting otherwise (that A is a knave). If A is a knave, then B ’s
assertion is true, so he must be a knight, and A ’s assertion is false, as it should be. Thus we conclude that A
is a knave and B is a knight.
58. We can draw no conclusions. A knight will declare himself to be a knight, telling the truth. A knave will lie
and assert that he is a knight. Since everyone will say “I am a knight,” we can determine nothing.
60. a) We look at the three possibilities of who the innocent men might be. If Smith and Jones are innocent
(and therefore telling the truth), then we get an immediate contradiction, since Smith said that Jones was a
friend of Cooper, but Jones said that he did not even know Cooper. If Jones and Williams are the innocent
truthtellers, then we again get a contradiction, since Jones says that he did not know Cooper and was out
of town, but Williams says he saw Jones with Cooper (presumably in town, and presumably if we was with
him, then he knew him). Therefore it must be the case that Smith and Williams are telling the truth. Their
statements do not contradict each other. Based on Williams’ statement, we know that Jones is lying, since he
said that he did not know Cooper when in fact he was with him. Therefore Jones is the murderer. Section 1.2 Propositional Equivalences 7 b) This is just like part (a), except that we are not told ahead of time that one of the men is guilty. Can
none of them be guilty? If so, then they are all telling the truth, but this is impossible, because as we just
saw, some of the statements are contradictory. Can more than one of them be guilty? If, for example, they
are all guilty, then their statements give us no information. So that is certainly possible.
62. This information is enough to determine the entire system. Let each letter stand for the statement that
the person whose name begins with that letter is chatting. Then the given information can be expressed
symbolically as follows: ¬K → H , R → ¬V , ¬R → V , A → R , V → K , K → V , H → A , H → K .
Note that we were able to convert all of these statements into conditional statements. In what follows we will
sometimes make use of the contrapositives of these conditional statements as well. First suppose that H is
true. Then it follows that A and K are true, whence it follows that R and V are true. But R implies that
V is false, so we get a contradiction. Therefore H must be false. From this it follows that K is true; whence
V is true, and therefore R is false, as is A . We can now check that this assignment leads to a true value for
each conditional statement. So we conclude that Kevin and Vijay are chatting but Heather, Randy, and Abby
are not.
64. Note that Diana’s statement is merely that she didn’t do it.
a) John did it. There are four cases to consider. If Alice is the sole truthteller, then Carlos did it; but this
means that John is telling the truth, a contradiction. If John is the sole truthteller, then Diana must be
lying, so she did it, but then Carlos is telling the truth, a contradiction. If Carlos is the sole truthteller, then
Diana did it, but that makes John truthful, again a contradiction. So the only possibility is that Diana is the
sole truthteller. This means that John is lying when he denied it, so he did it. Note that in this case both
Alice and Carlos are indeed lying.
b) Again there are four cases to consider. Since Carlos and Diana are making contradictory statements, the
liar must be one of them (we could have used this approach in part (a) as well). Therefore Alice is telling the
truth, so Carlos did it. Note that John and Diana are telling the truth as well here, and it is Carlos who is
lying. SECTION 1.2 Propositional Equivalences 2. There are two cases. If p is true, then ¬(¬p) is the negation of a false proposition, hence true. Similarly, if p
is false, then ¬(¬p) is also false. Therefore the two propositions are logically equivalent.
4. a) We construct the relevant truth table and note that the ﬁfth and seventh columns are identical.
p q r p ∨ q (p ∨ q ) ∨ r
q∨r
p ∨ (q ∨ r )
TT
TT
TF
TF
FT
FT
FF
FF T
F
T
F
T
F
T
F T
T
T
T
T
T
F
F T
T
T
T
T
T
T
F T
T
T
F
T
T
T
F T
T
T
T
T
T
T
F b) Again we construct the relevant truth table and note that the ﬁfth and seventh columns are identical. 8 Chapter 1 The Foundations: Logic and Proofs q r p∧q (p ∧ q ) ∧ r q∧r p ∧ (q ∧ r ) TT
TT
TF
TF
FT
FT
FF
FF T
F
T
F
T
F
T
F T
T
F
F
F
F
F
F T
F
F
F
F
F
F
F T
F
F
F
T
F
F
F T
F
F
F
F
F
F
F p 6. We see that the fourth and seventh columns are identical.
pq
p∧q
¬(p ∧ q )
¬p ¬q ¬p ∨ ¬q TT
TF
FT
FF F
T
F
T F
T
T
T T
F
F
F F
T
T
T F
F
T
T 8. We need to negate each part and swap “and” with “or.”
a) Kwame will not take a job in industry and will not go to graduate school.
b) Yoshiko does not know Java or does not know calculus.
c) James is not young, or he is not strong.
d) Rita will not move to Oregon and will not move to Washington.
10. We construct a truth table for each conditional statement and note that the relevant column contains only
T’s. For part (a) we have the following table.
pq
¬p
p∨q
¬p ∧ (p ∨ q )
[¬p ∧ (p ∨ q )] → q
T
T
F
F
For part (b) we have the
will ﬁt on the page.
pqr
T
T
T
T
F
F
F
F T
T
F
F
T
T
F
F T
F
F
F
T
T
F
T
following table. T
T
T
F
We omit the F
F
T
F
columns showing p T
T
T
T
→ q and q → r so that the table (p → q ) → (q → r) q→r [(p → q ) → (q → r)] → (p → r) T
F
T
F
T
F
T
T T
T
T
F
T
T
T
T T
T
F
T
T
F
F
T T
F
T
F
T
F
T
F For part (c) we have the following table.
pq
p→q
TT
TF
FT
FF T
F
T
T p ∧ (p → q ) [p ∧ (p → q )] → q T
F
F
F T
T
T
T For part (d) we have the following table. We have omitted some of the intermediate steps to make the table
ﬁt. Section 1.2 Propositional Equivalences
p q r T
T
T
T
F
F
F
F TT
TF
FT
FF
TT
TF
FT
FF 9 (p ∨ q ) ∧ (p → r) ∧ (p → r) [(p ∨ q ) ∧ (p → r) ∧ (p → r)] → r T
F
T
F
T
F
F
F T
T
T
T
T
T
T
T 12. We argue directly by showing that if the hypothesis is true, then so is the conclusion. An alternative approach,
which we show only for part (a), is to use the equivalences listed in the section and work symbolically.
a) Assume the hypothesis is true. Then p is false. Since p ∨ q is true, we conclude that q must be true. Here
is a more “algebraic” solution: [¬p ∧ (p ∨ q )] → q ≡ ¬[¬p ∧ (p ∨ q )] ∨ q ≡ ¬¬p ∨ ¬(p ∨ q )] ∨ q ≡ p ∨ ¬(p ∨ q ) ∨ q ≡
(p ∨ q ) ∨ ¬(p ∨ q ) ≡ T . The reasons for these logical equivalences are, respectively, Table 7, line 1; De Morgan’s
law; double negation; commutative and associative laws; negation law.
b) We want to show that if the entire hypothesis is true, then the conclusion p → r is true. To do this, we
need only show that if p is true, then r is true. Suppose p is true. Then by the ﬁrst part of the hypothesis,
we conclude that q is true. It now follows from the second part of the hypothesis that r is true, as desired.
c) Assume the hypothesis is true. Then p is true, and since the second part of the hypothesis is true, we
conclude that q is also true, as desired.
d) Assume the hypothesis is true. Since the ﬁrst part of the hypothesis is true, we know that either p or q
is true. If p is true, then the second part of the hypothesis tells us that r is true; similarly, if q is true, then
the third part of the hypothesis tells us that r is true. Thus in either case we conclude that r is true.
14. This is not a tautology. It is saying that knowing that the hypothesis of an conditional statement is false
allows us to conclude that the conclusion is also false, and we know that this is not valid reasoning. To show
that it is not a tautology, we need to ﬁnd truth assignments for p and q that make the entire proposition
false. Since this is possible only if the conclusion if false, we want to let q be true; and since we want the
hypothesis to be true, we must also let p be false. It is easy to check that if, indeed, p is false and q is true,
then the conditional statement is false. Therefore it is not a tautology.
16. The ﬁrst of these propositions is true if and only if p and q have the same truth value. The second is true if
and only if either p and q are both true, or p and q are both false. Clearly these two conditions are saying
the same thing.
18. It is easy to see from the deﬁnitions of conditional statement and negation that each of these propositions
is false in the case in which p is true and q is false, and true in the other three cases. Therefore the two
propositions are logically equivalent.
20. It is easy to see from the deﬁnitions of the logical operations involved here that each of these propositions is
true in the cases in which p and q have the same truth value, and false in the cases in which p and q have
opposite truth values. Therefore the two propositions are logically equivalent.
22. Suppose that (p → q ) ∧ (p → r) is true. We want to show that p → (q ∧ r) is true, which means that we
want to show that q ∧ r is true whenever p is true. If p is true, since we know that both p → q and p → r
are true from our assumption, we can conclude that q is true and that r is true. Therefore q ∧ r is true, as
desired. Conversely, suppose that p → (q ∧ r) is true. We need to show that p → q is true and that p → r is
true, which means that if p is true, then so are q and r . But this follows from p → (q ∧ r) . 10 Chapter 1 The Foundations: Logic and Proofs 24. We determine exactly which rows of the truth table will have T as their entries. Now (p → q ) ∨ (p → r) will
be true when either of the conditional statements is true. The conditional statement will be true if p is false,
or if q in one case or r in the other case is true, i.e., when q ∨ r is true, which is precisely when p → (q ∨ r)
is true. Since the two propositions are true in exactly the same situations, they are logically equivalent.
26. Applying the third and ﬁrst equivalences in Table 7, we have ¬p → (q → r) ≡ p ∨ (q → r) ≡ p ∨ ¬q ∨ r .
Applying the ﬁrst equivalence in Table 7 to q → (p ∨ r) shows that ¬q ∨ p ∨ r is equivalent to it. But these
are equivalent by the commutative and associative laws.
28. We know that p ↔ q is true precisely when p and q have the same truth value. But this happens precisely
when ¬p and ¬q have the same truth value, that is, ¬p ↔ ¬q .
30. The conclusion q ∨ r will be true in every case except when q and r are both false. But if q and r are both
false, then one of p ∨ q or ¬p ∨ r is false, because one of p or ¬p is false. Thus in this case the hypothesis
(p ∨ q ) ∧ (¬p ∨ r) is false. An conditional statement in which the conclusion is true or the hypothesis is false
is true, and that completes the argument.
32. We just need to ﬁnd an assignment of truth values that makes one of these propositions true and the other
false. We can let p be true and the other two variables be false. Then the ﬁrst statement will be F → F ,
which is true, but the second will be F ∧ T , which is false.
34. We apply the rules stated in the preamble.
a) p ∧ ¬q
b) p ∨ (q ∧ (r ∨ F))
c) (p ∨ ¬q ) ∧ (q ∨ T)
36. If s has any occurrences of ∧ , ∨ , T , or F , then the process of forming the dual will change it. Therefore
s∗ = s if and only if s is simply one propositional variable (like p ). A more diﬃcult question is to determine
when s∗ will be logically equivalent to s . For example, p ∨ F is logically equivalent to its dual p ∧ T , because
both are logically equivalent to p .
38. The table is in fact displayed so as to exhibit the duality. The two identity laws are duals of each other, the
two domination laws are duals of each other, etc. The only law not listed with another, the double negation
law, is its own dual, since there are no occurrences of ∧ , ∨ , T, or F to replace.
40. Following the hint, we easily see that the answer is p ∧ q ∧ ¬r .
42. The statement of the problem is really the solution. Each line of the truth table corresponds to exactly one
combination of truth values for the n atomic propositions involved. We can write down a conjunction that
is true precisely in this case, namely the conjunction of all the atomic propositions that are true and the
negations of all the atomic propositions that are false. If we do this for each line of the truth table for which
the value of the compound proposition is to be true, and take the disjunction of the resulting propositions,
then we have the desired proposition in its disjunctive normal form.
44. Given a compound proposition p , we can, by Exercise 43, write down a proposition q that is logically equivalent
to p and uses only ¬ , ∧ , and ∨ . Now by De Morgan’s law we can get rid of all the ∨’s by replacing each
occurrence of p1 ∨ p2 ∨ · · · ∨ pn with ¬(¬p1 ∧ ¬p2 ∧ · · · ∧ ¬pn ) .
46. We write down the truth table corresponding to the deﬁnition.
pq
pq
T
T
F
F T
F
T
F F
T
T
T Section 1.2 Propositional Equivalences 11 48. We write down the truth table corresponding to the deﬁnition.
p q p↓q T
T
F
F T
F
T
F F
F
F
T 50. a) From the deﬁnition (or as seen in the truth table constructed in Exercise 48), p ↓ p is false when p is true
and true when p is false, exactly as ¬p is; thus the two are logically equivalent.
b) The proposition (p ↓ q ) ↓ (p ↓ q ) is equivalent, by part (a), to ¬(p ↓ q ) , which from the deﬁnition (or
truth table or Exercise 49) is clearly equivalent to p ∨ q .
c) By Exercise 45, every compound proposition is logically equivalent to one that uses only ¬ and ∨ . But
by parts (a) and (b) of the present exercise, we can get rid of all the negations and disjunctions by using
NOR ’s . Thus every compound proposition can be converted into a logically equivalent compound proposition
involving only NOR ’s .
52. This exercise is similar to Exercise 50. First we can see from the truth tables that (p  p) ≡ (¬p) and that
((p  p)  (q  q )) ≡ (p ∨ q ) . Then we argue exactly as in part (c) of Exercise 50: by Exercise 45, every
compound proposition is logically equivalent to one that uses only ¬ and ∨ . But by our observations at the
beginning of the present exercise, we can get rid of all the negations and disjunctions by using NAND ’s . Thus
every compound proposition can be converted into a logically equivalent compound proposition involving only
NAND ’s .
54. To show that these are not logically equivalent, we need only ﬁnd one assignment of truth values to p , q , and
r for which the truth values of p  (q  r) and (p  q )  r diﬀer. One such assignment is T for p and F for q
and r . Then computing from the truth tables (or deﬁnitions), we see that p  (q  r) is false and (p  q )  r is
true.
56. To say that p and q are logically equivalent is to say that the truth tables for p and q are identical; similarly,
to say that q and r are logically equivalent is to say that the truth tables for q and r are identical. Clearly
if the truth tables for p and q are identical, and the truth tables for q and r are identical, then the truth
tables for p and r are identical (this is a fundamental axiom of the notion of equality). Therefore p and r are
logically equivalent. (We are assuming—and there is no loss of generality in doing so—that the same atomic
variables appear in all three propositions.)
58. If we want the ﬁrst two of these to be true, then p and q must have the same truth value. If q is true, then
the third and fourth expressions will be true, and if r is false, the last expression will be true. So all ﬁve of
these disjunctions will be true if we set p and q to be true, and r to be false.
60. In each case we hunt for truth assignments that make all the disjunctions true.
a) Since p occurs in four of the ﬁve disjunctions, we can make p true, and then make q false (and make r
and s anything we please). Thus this proposition is satisﬁable.
b) This is satisﬁable by, for example, setting p to be false (that takes care of the ﬁrst, second, and fourth
disjunctions), s to be false (for the third and sixth disjunctions), q to be true (for the ﬁfth disjunction), and
r to be anything.
c) It is not hard to ﬁnd a satisfying truth assignment, such as p , q , and s true, and r false. 12 Chapter 1 SECTION 1.3 The Foundations: Logic and Proofs Predicates and Quantiﬁers 2. a) This is true, since there is an a in orange .
c) This is false, since there is no a in true . b) This is false, since there is no a in lemon .
d) This is true, since there is an a in false . 4. a) Here x is still equal to 0 , since the condition is false.
b) Here x is still equal to 1 , since the condition is false.
c) This time x is equal to 1 at the end, since the condition is true, so the statement x := 1 is executed.
6. The answers given here are not unique, but care must be taken not to confuse nonequivalent sentences. Parts
(c) and (f ) are equivalent; and parts (d) and (e) are equivalent. But these two pairs are not equivalent to
each other.
a) Some student in the school has visited North Dakota. (Alternatively, there exists a student in the school
who has visited North Dakota.)
b) Every student in the school has visited North Dakota. (Alternatively, all students in the school have visited
North Dakota.)
c) This is the negation of part (a): No student in the school has visited North Dakota. (Alternatively, there
does not exist a student in the school who has visited North Dakota.)
d) Some student in the school has not visited North Dakota. (Alternatively, there exists a student in the
school who has not visited North Dakota.)
e) This is the negation of part (b): It is not true that every student in the school has visited North Dakota.
(Alternatively, not all students in the school have visited North Dakota.)
f) All students in the school have not visited North Dakota. (This is technically the correct answer, although
common English usage takes this sentence to mean—incorrectly—the answer to part (e). To be perfectly
clear, one could say that every student in this school has failed to visit North Dakota, or simply that no
student has visited North Dakota.)
8. Note that part (b) and part (c) are not the sorts of things one would normally say.
a) If an animal is a rabbit, then that animal hops. (Alternatively, every rabbit hops.)
b) Every animal is a rabbit and hops.
c) There exists an animal such that if it is a rabbit, then it hops. (Note that this is trivially true, satisﬁed,
for example, by lions, so it is not the sort of thing one would say.)
d) There exists an animal that is a rabbit and hops. (Alternatively, some rabbits hop. Alternatively, some
hopping animals are rabbits.)
10. a) We assume that this means that one student has all three animals: ∃x(C (x) ∧ D(x) ∧ F (x)) .
b) ∀x(C (x) ∨ D(x) ∨ F (x))
c) ∃x(C (x) ∧ F (x) ∧ ¬D(x))
d) This is the negation of part (a): ¬∃x(C (x) ∧ D(x) ∧ F (x)) .
e) Here the owners of these pets can be diﬀerent: (∃x C (x)) ∧ (∃x D(x)) ∧ (∃x F (x)) . There is no harm in using
the same dummy variable, but this could also be written, for example, as (∃x C (x)) ∧ (∃y D(y )) ∧ (∃z F (z )) .
12. a) Since 0 + 1 > 2 · 0 , we know that Q(0) is true.
b) Since (−1) + 1 > 2 · (−1) , we know that Q(−1) is true.
c) Since 1 + 1 > 2 · 1 , we know that Q(1) is false.
d) From part (a) we know that there is at least one x that makes Q(x) true, so ∃x Q(x) is true.
e) From part (c) we know that there is at least one x that makes Q(x) false, so ∀x Q(x) is false.
f) From part (c) we know that there is at least one x that makes Q(x) false, so ∃x ¬Q(x) is true.
g) From part (a) we know that there is at least one x that makes Q(x) true, so ∀x ¬Q(x) is false. Section 1.3 Predicates and Quantiﬁers 13 14. a) Since (−1)3 = −1 , this is true.
b) Since ( 1 )4 < ( 1 )2 , this is true.
2
2
c) Since (−x)2 = ((−1)x)2 = (−1)2 x2 = x2 , we know that ∀x((−x)2 = x2 ) is true.
d) Twice a positive number is larger than the number, but this inequality is not true for negative numbers
or 0 . Therefore ∀x(2x > x) is false.
√
√
2)
b) false ( −1 is not a real number)
c) true (the lefthand side is always at least 2 )
d) false (not true for x = 1 or x = 0) 16. a) true ( x = 18. Existential quantiﬁers are like disjunctions, and universal quantiﬁers are like conjunctions. See Examples 11
and 16.
a) We want to assert that P (x) is true for some x in the domain, so either P (−2) is true or P (−1) is true
or P (0) is true or P (1) is true or P (2) is true. Thus the answer is P (−2) ∨ P (−1) ∨ P (0) ∨ P (1) ∨ P (2) . The
other parts of this exercise are similar. Note that by De Morgan’s laws, the expression in part (c) is logically
equivalent to the expression in part (f ), and the expression in part (d) is logically equivalent to the expression
in part (e).
b) P (−2) ∧ P (−1) ∧ P (0) ∧ P (1) ∧ P (2)
c) ¬P (−2) ∨ ¬P (−1) ∨ ¬P (0) ∨ ¬P (1) ∨ ¬P (2)
d) ¬P (−2) ∧ ¬P (−1) ∧ ¬P (0) ∧ ¬P (1) ∧ ¬P (2)
e) This is just the negation of part (a): ¬(P (−2) ∨ P (−1) ∨ P (0) ∨ P (1) ∨ P (2))
f) This is just the negation of part (b): ¬(P (−2) ∧ P (−1) ∧ P (0) ∧ P (1) ∧ P (2))
20. Existential quantiﬁers are like disjunctions, and universal quantiﬁers are like conjunctions. See Examples 11
and 16.
a) We want to assert that P (x) is true for some x in the domain, so either P (−5) is true or P (−3) is true or
P (−1) is true or P (1) is true or P (3) is true or P (5) is true. Thus the answer is P (−5) ∨ P (−3) ∨ P (−1) ∨
P (1) ∨ P (3) ∨ P (5) .
b) P (−5) ∧ P (−3) ∧ P (−1) ∧ P (1) ∧ P (3) ∧ P (5)
c) The formal translation is as follows: ((−5 = 1) → P (−5)) ∧ ((−3 = 1) → P (−3)) ∧ ((−1 = 1) → P (−1)) ∧
((1 = 1) → P (1)) ∧ ((3 = 1) → P (3)) ∧ ((5 = 1) → P (5)) . However, since the hypothesis x = 1 is false when
x is 1 and true when x is anything other than 1, we have more simply P (−5) ∧ P (−3) ∧ P (−1) ∧ P (3) ∧ P (5) .
d) The formal translation is as follows: ((−5 ≥ 0) ∧ P (−5)) ∨ ((−3 ≥ 0) ∧ P (−3)) ∨ ((−1 ≥ 0) ∧ P (−1)) ∨ ((1 ≥
0) ∧ P (1)) ∨ ((3 ≥ 0) ∧ P (3)) ∨ ((5 ≥ 0) ∧ P (5)) . Since only three of the x’s in the domain meet the condition,
the answer is equivalent to P (1) ∨ P (3) ∨ P (5) .
e) For the second part we again restrict the domain: (¬P (−5) ∨ ¬P (−3) ∨ ¬P (−1) ∨ ¬P (1) ∨ ¬P (3) ∨ ¬P (5)) ∧
(P (−1) ∧ P (−3) ∧ P (−5)) . This is equivalent to (¬P (1) ∨ ¬P (3) ∨ ¬P (5)) ∧ (P (−1) ∧ P (−3) ∧ P (−5)) .
22. Many answer are possible in each case.
a) A domain consisting of a few adults in certain parts of India would make this true. If the domain were all
residents of the United States, then this is certainly false.
b) If the domain is all residents of the United States, then this is true. If the domain is the set of pupils in a
ﬁrst grade class, it is false.
c) If the domain consists of all the United States Presidents whose last name is Bush, then the statement is
true. If the domain consists of all United States Presidents, then the statement is false.
d) If the domain were all residents of the United States, then this is certainly true. If the domain consists of
all babies born in the last ﬁve minutes, one would expect the statement to be false (it’s not even clear that
these babies “know” their mothers yet). 14 Chapter 1 The Foundations: Logic and Proofs 24. In order to do the translation the second way, we let C (x) be the propositional function “ x is in your class.”
Note that for the second way, we always want to use conditional statements with universal quantiﬁers and
conjunctions with existential quantiﬁers.
a) Let P (x) be “ x has a cellular phone.” Then we have ∀x P (x) the ﬁrst way, or ∀x(C (x) → P (x)) the
second way.
b) Let F (x) be “ x has seen a foreign movie.” Then we have ∃x F (x) the ﬁrst way, or ∃x(C (x) ∧ F (x)) the
second way.
c) Let S (x) be “ x can swim.” Then we have ∃x ¬S (x) the ﬁrst way, or ∃x(C (x) ∧ ¬S (x)) the second way.
d) Let Q(x) be “ x can solve quadratic equations.” Then we have ∀x Q(x) the ﬁrst way, or ∀x(C (x) → Q(x))
the second way.
e) Let R(x) be “ x wants to be rich.” Then we have ∃x ¬R(x) the ﬁrst way, or ∃x(C (x) ∧ ¬R(x)) the second
way.
26. In all of these, we will let Y (x) be the propositional function that x is in your school or class, as appropriate.
a) If we let U (x) be “ x has visited Uzbekistan,” then we have ∃x U (x) if the domain is just your schoolmates,
or ∃x(Y (x) ∧ U (x)) if the domain is all people. If we let V (x, y ) mean that person x has visited country y ,
then we can rewrite this last one as ∃x(Y (x) ∧ V (x, Uzbekistan)) .
b) If we let C (x) and P (x) be the propositional functions asserting that x has studied calculus and C++,
respectively, then we have ∀x(C (x)∧P (x)) if the domain is just your schoolmates, or ∀x(Y (x) → (C (x)∧P (x)))
if the domain is all people. If we let S (x, y ) mean that person x has studied subject y , then we can rewrite
this last one as ∀x(Y (x) → (S (x, calculus) ∧ S (x, C++))) .
c) If we let B (x) and M (x) be the propositional functions asserting that x owns a bicycle and a motorcycle,
respectively, then we have ∀x(¬(B (x) ∧ M (x))) if the domain is just your schoolmates, or ∀x(Y (x) → ¬(B (x) ∧
M (x))) if the domain is all people. Note that “no one” became “for all . . . not.” If we let O(x, y ) mean that
person x owns item y , then we can rewrite this last one as ∀x(Y (x) → ¬(O(x, bicycle) ∧ O(x, motorcycle))) .
d) If we let H (x) be “ x is happy,” then we have ∃x ¬H (x) if the domain is just your schoolmates, or
∃x(Y (x) ∧ ¬H (x)) if the domain is all people. If we let E (x, y ) mean that person x is in mental state y , then
we can rewrite this last one as ∃x(Y (x) ∧ ¬E (x, happy)) .
e) If we let T (x) be “ x was born in the twentieth century,” then we have ∀x T (x) if the domain is just your
schoolmates, or ∀x(Y (x) → T (x)) if the domain is all people. If we let B (x, y ) mean that person x was born
in the y th century, then we can rewrite this last one as ∀x(Y (x) → B (x, 20)) .
28. Let R(x) be “ x is in the correct place”; let E (x) be “ x is in excellent condition”; let T (x) be “ x is a [or
your] tool”; and let the domain of discourse be all things.
a) There exists something not in the correct place: ∃x ¬R(x) .
b) If something is a tool, then it is in the correct place place and in excellent condition: ∀x (T (x) → (R(x) ∧
E (x))) .
c) ∀x (R(x) ∧ E (x))
d) This is saying that everything fails to satisfy the condition: ∀x ¬(R(x) ∧ E (x)) .
e) There exists a tool with this property: ∃x (T (x) ∧ ¬R(x) ∧ E (x)) .
30. a) P (1, 3) ∨ P (2, 3) ∨ P (3, 3)
c) ¬P (2, 1) ∨ ¬P (2, 2) ∨ ¬P (2, 3) b) P (1, 1) ∧ P (1, 2) ∧ P (1, 3)
d) ¬P (1, 2) ∧ ¬P (2, 2) ∧ ¬P (3, 2) 32. In each case we need to specify some propositional functions (predicates) and identify the domain of discourse.
a) Let F (x) be “ x has ﬂeas,” and let the domain of discourse be dogs. Our original statement is ∀x F (x) .
Its negation is ∃x ¬F (x) . In English this reads “There is a dog that does not have ﬂeas.” Section 1.3 Predicates and Quantiﬁers 15 b) Let H (x) be “ x can add,” where the domain of discourse is horses. Then our original statement is ∃x H (x) .
Its negation is ∀x ¬H (x) . In English this is rendered most simply as “No horse can add.”
c) Let C (x) be “ x can climb,” and let the domain of discourse be koalas. Our original statement is ∀x C (x) .
Its negation is ∃x ¬C (x) . In English this reads “There is a koala that cannot climb.”
d) Let F (x) be “ x can speak French,” and let the domain of discourse be monkeys. Our original statement
is ¬∃x F (x) or ∀x ¬F (x) . Its negation is ∃x F (x) . In English this reads “There is a monkey that can speak
French.”
e) Let S (x) be “ x can swim” and let C (x) be “ x can catch ﬁsh,” where the domain of discourse is pigs. Then
our original statement is ∃x (S (x) ∧ C (x)) . Its negation is ∀x ¬(S (x) ∧ C (x)) , which could also be written
∀x (¬S (x) ∨ ¬C (x)) by De Morgan’s law. In English this is “No pig can both swim and catch ﬁsh,” or “Every
pig either is unable to swim or is unable to catch ﬁsh.”
34. a) Let S (x) be “ x obeys the speed limit,” where the domain of discourse is drivers. The original statement
is ∃x ¬S (x) , the negation is ∀x S (x) , “All drivers obey the speed limit.”
b) Let S (x) be “ x is serious,” where the domain of discourse is Swedish movies. The original statement is
∀x S (x) , the negation is ∃x ¬S (x) , “Some Swedish movies are not serious.”
c) Let S (x) be “ x can keep a secret,” where the domain of discourse is people. The original statement is
¬∃x S (x) , the negation is ∃x S (x) , “Some people can keep a secret.”
d) Let A(x) be “ x has a good attitude,” where the domain of discourse is people in this class. The original
statement is ∃x ¬A(x) , the negation is ∀x A(x) , “Everyone in this class has a good attitude.”
36. a) Since 12 = 1 , this statement is false; x = 1 is a counterexample. So is x = 0 (these are the only two
counterexamples).
√
√
b) There are two counterexamples: x = 2 and x = − 2 .
c) There is one counterexample: x = 0.
38. a) Some system is open.
b) Every system is either malfunctioning or in a diagnostic state.
c) Some system is open, or some system is in a diagnostic state.
d) Some system is unavailable.
e) No system is working. (We could also say “Every system is not working,” as long as we understood that
this is diﬀerent from “Not every system is working.”)
40. There are many ways to write these, depending on what we use for predicates.
a) Let F (x) be “There is less than x megabytes free on the hard disk,” with the domain of discourse being
positive numbers, and let W (x) be “User x is sent a warning message.” Then we have F (30) → ∀x W (x) .
b) Let O(x) be “Directory x can be opened,” let C (x) be “File x can be closed,” and let E be the proposition
“System errors have been detected.” Then we have E → ((∀x ¬O(x)) ∧ (∀x ¬C (x))) .
c) Let B be the proposition “The ﬁle system can be backed up,” and let L(x) be “User x is currently logged
on.” Then we have (∃x L(x)) → ¬B .
d) Let D(x) be “Product x can be delivered,” and let M (x) be “There are at least x megabytes of memory available” and S (x) be “The connection speed is at least x kilobits per second,” where the domain of
discourse for the last two propositional functions are positive numbers. Then we have (M (8) ∧ S (56)) →
D(video on demand) .
42. There are many ways to write these, depending on what we use for predicates.
a) Let A(x) be “User x has access to an electronic mailbox.” Then we have ∀x A(x) .
b) Let A(x, y ) be “Group member x can access resource y ,” and let S (x, y ) be “System x is in state y .”
Then we have S (ﬁle system, locked) → ∀x A(x, system mailbox) . 16 Chapter 1 The Foundations: Logic and Proofs c) Let S (x, y ) be “System x is in state y .” Recalling that “only if” indicates a necessary condition, we have
S (ﬁrewall, diagnostic) → S (proxy server, diagnostic) .
d) Let T (x) be “The throughput is at least x kbps,” where the domain of discourse is positive numbers,
let M (x, y ) be “Resource x is in mode y ,” and let S (x, y ) be “Router x is in state y .” Then we have
(T (100) ∧ ¬T (500) ∧ ¬M (proxy server, diagnostic)) → ∃x S (x, normal) .
44. We want propositional functions P and Q that are sometimes, but not always, true (so that the second
biconditional is F ↔ F and hence true), but such that there is an x making one true and the other false. For
example, we can take P (x) to mean that x is an even number (a multiple of 2 ) and Q(x) to mean that x is
a multiple of 3 . Then an example like x = 4 or x = 9 shows that ∀x(P (x) ↔ Q(x)) is false.
46. a) There are two cases. If A is true, then (∀xP (x)) ∨ A is true, and since P (x) ∨ A is true for all x ,
∀x(P (x) ∨ A) is also true. Thus both sides of the logical equivalence are true (hence equivalent). Now suppose
that A is false. If P (x) is true for all x , then the lefthand side is true. Furthermore, the righthand side is
also true (since P (x) ∨ A is true for all x ). On the other hand, if P (x) is false for some x , then both sides
are false. Therefore again the two sides are logically equivalent.
b) There are two cases. If A is true, then (∃xP (x)) ∨ A is true, and since P (x) ∨ A is true for some (really
all) x , ∃x(P (x) ∨ A) is also true. Thus both sides of the logical equivalence are true (hence equivalent). Now
suppose that A is false. If P (x) is true for at least one x , then the lefthand side is true. Furthermore, the
righthand side is also true (since P (x) ∨ A is true for that x ). On the other hand, if P (x) is false for all x ,
then both sides are false. Therefore again the two sides are logically equivalent.
48. a) There are two cases. If A is false, then both sides of the equivalence are true, because a conditional
statement with a false hypothesis is true. If A is true, then A → P (x) is equivalent to P (x) for each x , so
the lefthand side is equivalent to ∀x P (x) , which is equivalent to the righthand side.
b) There are two cases. If A is false, then both sides of the equivalence are true, because a conditional
statement with a false hypothesis is true (and we are assuming that the domain is nonempty). If A is true,
then A → P (x) is equivalent to P (x) for each x , so the lefthand side is equivalent to ∃x P (x) , which is
equivalent to the righthand side.
50. It is enough to ﬁnd a counterexample. It is intuitively clear that the ﬁrst proposition is asserting much more
than the second. It is saying that one of the two predicates, P or Q , is universally true; whereas the second
proposition is simply saying that for every x either P (x) or Q(x) holds, but which it is may well depend
on x . As a simple counterexample, let P (x) be the statement that x is odd, and let Q(x) be the statement
that x is even. Let the domain of discourse be the positive integers. The second proposition is true, since
every positive integer is either odd or even. But the ﬁrst proposition is false, since it is neither the case that
all positive integers are odd nor the case that all of them are even.
52. a) This is false, since there are many values of x that make x > 1 true.
b) This is false, since there are two values of x that make x2 = 1 true.
c) This is true, since by algebra we see that the unique solution to the equation is x = 3.
d) This is false, since there are no values of x that make x = x + 1 true.
54. There are only three cases in which ∃x!P (x) is true, so we form the disjunction of these three cases. The
answer is thus (P (1) ∧ ¬P (2) ∧ ¬P (3)) ∨ (¬P (1) ∧ P (2) ∧ ¬P (3)) ∨ (¬P (1) ∧ ¬P (2) ∧ P (3)) .
56. A Prolog query returns a yes/no answer if there are no variables in the query, and it returns the values that
make the query true if there are. Section 1.4 Nested Quantiﬁers 17 a) None of the facts was that Kevin was enrolled in EE 222. So the response is no.
b) One of the facts was that Kiko was enrolled in Math 273. So the response is yes.
c) Prolog returns the names of the courses for which Grossman is the instructor, namely just cs301.
d) Prolog returns the names of the instructor for CS 301, namely grossman.
e) Prolog returns the names of the instructors teaching any course that Kevin is enrolled in, namely chan,
since Chan is the instructor in Math 273, the only course Kevin is enrolled in.
58. Following the idea and syntax of Example 28, we have the following rule:
grandfather(X,Y) : father(X,Z), father(Z,Y); father(X,Z), mother(Z,Y).
Note that we used the comma to mean “and” and the semicolon to mean “or.” For X to be the grandfather
of Y , X must be either Y ’s father’s father or Y ’s mother’s father.
60. a) ∀x(P (x) → Q(x))
b) ∃x(R(x) ∧ ¬Q(x))
c) ∃x(R(x) ∧ ¬P (x))
d) Yes. The unsatisfactory excuse guaranteed by part (b) cannot be a clear explanation by part (a).
c) ∀x(Q(x) → P (x))
d) ∀x(Q(x) → ¬R(x))
62. a) ∀x(P (x) → ¬S (x))
b) ∀x(R(x) → S (x))
e) Yes. If x is one of my poultry, then he is a duck (by part (c)), hence not willing to waltz (part (a)). Since
oﬃcers are always willing to waltz (part (b)), x is not an oﬃcer. SECTION 1.4 Nested Quantiﬁers 2. a) There exists a real number x such that for every real number y , xy = y . This is asserting the existence
of a multiplicative identity for the real numbers, and the statement is true, since we can take x = 1.
b) For every real number x and real number y , if x is nonnegative and y is negative, then the diﬀerence
x − y is positive. Or, more simply, a nonnegative number minus a negative number is positive (which is true).
c) For every real number x and real number y , there exists a real number z such that x = y + z . This is a
true statement, since we can take z = x − y in each case.
4. a) Some student in your class has taken some computer science course.
b) There is a student in your class who has taken every computer science course.
c) Every student in your class has taken at least one computer science course.
d) There is a computer science course that every student in your class has taken.
e) Every computer science course has been taken by at least one student in your class.
f) Every student in your class has taken every computer science course.
6. a) Randy Goldberg is enrolled in CS 252.
b) Someone is enrolled in Math 695.
c) Carol Sitea is enrolled in some course.
d) Some student is enrolled simultaneously in Math 222 and CS 252.
e) There exist two distinct people, the second of whom is enrolled in every course that the ﬁrst is enrolled in.
f) There exist two distinct people enrolled in exactly the same courses.
8. a) ∃x∃yQ(x, y )
b) This is the negation of part (a), and so could be written either ¬∃x∃yQ(x, y ) or ∀x∀y ¬Q(x, y ) .
c) We assume from the wording that the statement means that the same person appeared on both shows:
∃x(Q(x, Jeopardy) ∧ Q(x, Wheel of Fortune))
d) ∀y ∃xQ(x, y )
e) ∃x1 ∃x2 (Q(x1 , Jeopardy) ∧ Q(x2 , Jeopardy) ∧ x1 = x2 ) 18 Chapter 1 10. a) ∀xF (x, Fred) b) ∀yF (Evelyn, y ) c) ∀x∃yF (x, y ) The Foundations: Logic and Proofs d) ¬∃x∀yF (x, y ) e) ∀y ∃xF (x, y ) f) ¬∃x(F (x, Fred) ∧ F (x, Jerry))
g) ∃y1 ∃y2 (F (Nancy, y1 ) ∧ F (Nancy, y2 ) ∧ y1 = y2 ∧ ∀y (F (Nancy, y ) → (y = y1 ∨ y = y2 )))
h) ∃y (∀xF (x, y ) ∧ ∀z (∀xF (x, z ) → z = y ))
i) ¬∃xF (x, x)
j) ∃x∃y (x = y ∧ F (x, y ) ∧ ∀z ((F (x, z ) ∧ z = x) → z = y )) (We do not assume that this sentence is asserting
that this person can or cannot fool her/himself.)
12. The answers to this exercise are not unique; there are many ways of expressing the same propositions symbolically. Note that C (x, y ) and C (y, x) say the same thing.
a) ¬I (Jerry)
b) ¬C (Rachel, Chelsea)
c) ¬C (Jan, Sharon)
d) ¬∃x C (x, Bob)
f) ∃x ¬I (x)
g) ¬∀x I (x) (same as (f ))
e) ∀x(x = Joseph ↔ C (x, Sanjay))
h) ∃x∀y (x = y ↔ I (y ))
i) ∃x∀y (x = y ↔ I (y ))
j) ∀x(I (x) → ∃y (x = y ∧ C (x, y )))
k) ∃x(I (x) ∧ ∀y (x = y → ¬C (x, y )))
l) ∃x∃y (x = y ∧ ¬C (x, y ))
m) ∃x∀y C (x, y )
n) ∃x∃y (x = y ∧ ∀z ¬(C (x, z ) ∧ C (y, z ))) o) ∃x∃y (x = y ∧ ∀z (C (x, z ) ∨ C (y, z ))) 14. The answers to this exercise are not unique; there are many ways of expressing the same propositions symbolically. Our domain of discourse for persons here consists of people in this class. We need to make up a
predicate in each case.
a) Let S (x, y ) mean that person x can speak language y . Then our statement is ∃x S (x, Hindi) .
b) Let P (x, y ) mean that person x plays sport y . Then our statement is ∀x∃y P (x, y ) .
c) Let V (x, y ) mean that person x has visited state y . Then our statement is ∃x(V (x, Alaska) ∧ ¬V (x,
Hawaii)) .
d) Let L(x, y ) mean that person x has learned programming language y . Then our statement is ∀x∃y L(x, y ) .
e) Let T (x, y ) mean that person x has taken course y , and let O(y, z ) mean that course y is oﬀered by
department z . Then our statement is ∃x∃z ∀y (O(y, z ) → T (x, y )) .
f) Let G(x, y ) mean that persons x and y grew up in the same town. Then our statement is ∃x∃y (x =
y ∧ G(x, y ) ∧ ∀z (G(x, z ) → (x = y ∨ x = z ))) .
g) Let C (x, y, z ) mean that persons x and y have chatted with each other in chat group z . Then our
statement is ∀x∃y ∃z (x = y ∧ C (x, y, z )) .
16. We let P (s, c, m) be the statement that student s has class standing c and is majoring in m . The variable
s ranges over students in the class, the variable c ranges over the four class standings, and the variable m
ranges over all possible majors.
a) The proposition is ∃s∃mP (s, junior, m) . It is true from the given information.
b) The proposition is ∀s∃cP (s, c, computer science) . This is false, since there are some mathematics majors.
c) The proposition is ∃s∃c∃m P (s, c, m) ∧ (c = junior) ∧ (m = mathematics) . This is true, since there is a
sophomore majoring in computer science.
d) The proposition is ∀s ∃cP (s, c, computer science) ∨ ∃mP (s, sophomore, m) . This is false, since there is a
freshman mathematics major.
e) The proposition is ∃m∀c∃sP (s, c, m) . This is false. It cannot be that m is mathematics, since there is no
senior mathematics major, and it cannot be that m is computer science, since there is no freshman computer
science major. Nor, of course, can m be any other major.
18. a) ∀f (H (f ) → ∃c A(c)) , where A(x) means that console x is accessible, and H (x) means that fault condition
x is happening
b) (∀u∃m (A(m) ∧ S (u, m))) → ∀u R(u) , where A(x) means that the archive contains message x , S (x, y )
means that user x sent message y , and R(x) means that the email address of user x can be retrieved Section 1.4 Nested Quantiﬁers 19 c) (∀b∃m D(m, b)) ↔ ∃p ¬C (p) , where D(x, y ) means that mechanism x can detect breach y , and C (x)
means that process x has been compromised
d) ∀x∀y (x = y → ∃p∃q (p = q ∧ C (p, x, y ) ∧ C (q, x, y ))) , where C (p, x, y ) means that path p connects endpoint
x to endpoint y
e) ∀x ((∀u K (x, u)) ↔ x = SysAdm) , where K (x, y ) means that person x knows the password of user y
20. a) ∀x∀y ((x < 0) ∧ (y < 0) → (xy > 0))
b) ∀x∀y ((x > 0) ∧ (y > 0) → ((x + y )/2 > 0))
c) What does “necessarily” mean in this context? The best explanation is to assert that a certain universal
conditional statement is not true. So we have ¬∀x∀y ((x < 0) ∧ (y < 0) → (x − y < 0)) . Note that we do
not want to put the negation symbol inside (it is not true that the diﬀerence of two negative integers is never
negative), nor do we want to negate just the conclusion (it is not true that the sum is always nonnegative).
We could rewrite our solution by passing the negation inside, obtaining ∃x∃y ((x < 0) ∧ (y < 0) ∧ (x − y ≥ 0)) .
d) ∀x∀y (x + y  ≤ x + y )
22. ∃x∀a∀b∀c ((x > 0) ∧ x = a2 + b2 + c2 ) , where the domain of discourse consists of all integers
24. a) There exists an additive identity for the real numbers—a number that when added to every number does
not change its value.
b) A nonnegative number minus a negative number is positive.
c) The diﬀerence of two nonpositive numbers is not necessarily nonpositive.
d) The product of two numbers is nonzero if and only if both factors are nonzero.
26. a) This is false, since 1 + 1 = 1 − 1 . b) This is true, since 2 + 0 = 2 − 0 . c) This is false, since there are many values of y for which 1 + y = 1 − y .
d) This is false, since the equation x + 2 = x − 2 has no solution.
e) This is true, since we can take x = y = 0.
f) This is true, since we can take y = 0 for each x .
g) This is true, since we can take y = 0.
i) This is certainly false. h) This is false, since part (d) was false. 28. a) true (let y = x2 ) b) false (no such y exists if x is negative)
c) true (let x = 0)
d) false (the commutative law for addition always holds)
e) true (let y = 1/x )
f) false (the reciprocal of y depends on y —there is not one x that works for all y )
g) true (let y = 1 − x )
h) false (this system of equations is inconsistent)
i) false (this system has only one solution; if x = 0 , for example, then no y satisﬁes y = 2 ∧ −y = 1)
j) true (let z = (x + y )/2) 30. We need to use the transformations shown in Table 2 of Section 1.3, replacing ¬∀ by ∃¬ , and replacing ¬∃
by ∀¬ . In other words, we push all the negation symbols inside the quantiﬁers, changing the sense of the
quantiﬁers as we do so, because of the equivalences in Table 2 of Section 1.3. In addition, we need to use De
Morgan’s laws (Section 1.2) to change the negation of a conjunction to the disjunction of the negations and to
change the negation of a disjunction to the conjunction of the negations. We also use the fact that ¬¬p ≡ p .
a) ∀y ∀x ¬P (x, y )
b) ∃x∀y ¬P (x, y )
c) ∀y (¬Q(y ) ∨ ∃x R(x, y ))
d) ∀y (∀x ¬R(x, y ) ∧ ∃x ¬S (x, y ))
e) ∀y (∃x∀z ¬T (x, y, z ) ∧ ∀x∃z ¬U (x, y, z ))
32. As we push the negation symbol toward the inside, each quantiﬁer it passes must change its type. For logical
connectives we either use De Morgan’s laws or recall that ¬(p → q ) ≡ p ∧ ¬q (Table 7 in Section 1.2) and that
¬(p ↔ q ) ≡ ¬p ↔ q (Exercise 21 in Section 1.2). 20 Chapter 1
a) The Foundations: Logic and Proofs ¬∃z ∀y ∀x T (x, y, z ) ≡ ∀z ¬∀y ∀x T (x, y, z )
≡ ∀z ∃y ¬∀x T (x, y, z )
≡ ∀z ∃y ∃x ¬T (x, y, z ) b) ¬(∃x∃y P (x, y ) ∧ ∀x∀y Q(x, y )) ≡ ¬∃x∃y P (x, y ) ∨ ¬∀x∀y Q(x, y )
≡ ∀x¬∃y P (x, y ) ∨ ∃x¬∀y Q(x, y )
≡ ∀x∀y ¬P (x, y ) ∨ ∃x∃y ¬ Q(x, y ) c) ¬∃x∃y (Q(x, y ) ↔ Q(y, x)) ≡ ∀x¬∃y (Q(x, y ) ↔ Q(y, x))
≡ ∀x∀y ¬(Q(x, y ) ↔ Q(y, x))
≡ ∀x∀y (¬Q(x, y ) ↔ Q(y, x)) d) ¬∀y ∃x∃z (T (x, y, z ) ∨ Q(x, y )) ≡ ∃y ¬∃x∃z (T (x, y, z ) ∨ Q(x, y ))
≡ ∃y ∀x¬∃z (T (x, y, z ) ∨ Q(x, y ))
≡ ∃y ∀x∀z ¬(T (x, y, z ) ∨ Q(x, y ))
≡ ∃y ∀x∀z (¬T (x, y, z ) ∧ ¬Q(x, y )) 34. The logical expression is asserting that the domain consists of at most two members. (It is saying that
whenever you have two unequal objects, any object has to be one of those two. Note that this is vacuously
true for domains with one element.) Therefore any domain having one or two members will make it true (such
as the female members of the United States Supreme Court in 2005), and any domain with more than two
members will make it false (such as all members of the United States Supreme Court in 2005).
36. In each case we need to specify some predicates and identify the domain of discourse.
a) Let L(x, y ) mean that person x has lost y dollars playing the lottery. The original statement is then
¬∃x∃y (y > 1000 ∧ L(x, y )) . Its negation of course is ∃x∃y (y > 1000 ∧ L(x, y )) ; someone has lost more than
$1000 playing the lottery.
b) Let C (x, y ) mean that person x has chatted with person y . The given statement is ∃x∃y (y = x ∧ ∀z (z =
x → (z = y ↔ C (x, z )))) . The negation is therefore ∀x∀y (y = x → ∃z (z = x ∧ ¬(z = y ↔ C (x, z )))) . In
English, everybody in this class has either chatted with no one else or has chatted with two or more others.
c) Let E (x, y ) mean that person x has sent email to person y . The given statement is ¬∃x∃y ∃z (y = z ∧ x =
y ∧ x = z ∧ ∀w(w = x → (E (x, w) ↔ (w = y ∨ w = z )))) . The negation is obviously ∃x∃y ∃z (y = z ∧ x =
y ∧ x = z ∧ ∀w(w = x → (E (x, w) ↔ (w = y ∨ w = z )))) . In English, some student in this class has sent
email to exactly two other students in this class.
d) Let S (x, y ) mean that student x has solved exercise y . The statement is ∃x∀y S (x, y ) . The negation is
∀x∃y ¬S (x, y ) . In English, for every student in this class, there is some exercise that he or she has not solved.
(One could also interpret the given statement as asserting that for every exercise, there exists a student—
perhaps a diﬀerent one for each exercise—who has solved it. In that case the order of the quantiﬁers would
be reversed. Word order in English sometimes makes for a little ambiguity.)
e) Let S (x, y ) mean that student x has solved exercise y , and let B (y, z ) mean that exercise y is in section z
of the book. The statement is ¬∃x∀z ∃y (B (y, z ) ∧ S (x, y )) . The negation is of course ∃x∀z ∃y (B (y, z ) ∧ S (x, y )) .
In English, some student has solved at least one exercise in every section of this book.
38. a) In English, the negation is “Some student in this class does not like mathematics.” With the obvious
propositional function, this is ∃x¬L(x) .
b) In English, the negation is “Every student in this class has seen a computer.” With the obvious propositional
function, this is ∀xS (x) . Section 1.4 Nested Quantiﬁers 21 c) In English, the negation is “For every student in this class, there is a mathematics course that this student
has not taken.” With the obvious propositional function, this is ∀x∃c¬T (x, c) .
d) As in Exercise 15f, let P (z, y ) be “Room z is in building y ,” and let Q(x, z ) be “Student x has been
in room z .” Then the original statement is ∃x∀y ∃z P (z, y ) ∧ Q(x, z ) . To form the negation, we change all
the quantiﬁers and put the negation on the inside, then apply De Morgan’s law. The negation is therefore
∀x∃y ∀z ¬P (z, y ) ∨ ¬Q(x, z ) , which is also equivalent to ∀x∃y ∀z P (z, y ) → ¬Q(x, z ) . In English, this could
be read, “For every student there is a building such that for every room in that building, the student has not
been in that room.”
40. a) There are many counterexamples. If x = 2 , then there is no y among the integers such that 2 = 1/y ,
since the only solution of this equation is y = 1/2 . Even if we were working in the domain of real numbers,
x = 0 would provide a counterexample, since 0 = 1/y for no real number y .
b) We can rewrite y 2 − x < 100 as y 2 < 100 + x . Since squares can never be negative, no such y exists if x
is, say, −200 . This x provides a counterexample.
c) This is not true, since sixth powers are both squares and cubes. Trivial counterexamples would include
x = y = 0 and x = y = 1 , but we can also take something like x = 27 and y = 9 , since 272 = 36 = 93 .
42. The distributive law is just the statement that x(y + z ) = xy + xz for all real numbers. Therefore the expression
we want is ∀x∀y ∀z (x(y + z ) = xy + xz ) , where the quantiﬁers are assumed to range over (i.e., the domain of
discourse is) the real numbers.
44. We want to say that for each triple of coeﬃcients (the a , b , and c in the expression ax2 + bx + c , where we
insist that a = 0 so that this actually is quadratic), there are at most two values of x making that expression
equal to 0 . The domain here is all real numbers. We write ∀a∀b∀c(a = 0 → ∀x1 ∀x2 ∀x3 (ax2 + bx1 + c =
1
0 ∧ ax2 + bx2 + c = 0 ∧ ax2 + bx3 + c = 0) → (x1 = x2 ∨ x1 = x3 ∨ x2 = x3 )) .
2
3
46. This statement says that there is a number that is less than or equal to all squares.
a) This is false, since no matter how small a positive number x we might choose, if we let y =
x = 2y 2 , and it will not be true that x ≤ y 2 .
b) This is true, since we can take x = −1 , for example. x/2 , then c) This is true, since we can take x = −1 , for example.
48. We need to show that each of these propositions implies the other. Suppose that ∀xP (x) ∨ ∀xQ(x) is true. We
want to show that ∀x∀y (P (x) ∨ Q(y )) is true. By our hypothesis, one of two things must be true. Either P is
universally true, or Q is universally true. In the ﬁrst case, ∀x∀y (P (x) ∨ Q(y )) is true, since the ﬁrst expression
in the disjunction is true, no matter what x and y are; and in the second case, ∀x∀y (P (x) ∨ Q(y )) is also
true, since now the second expression in the disjunction is true, no matter what x and y are. Next we need
to prove the converse. So suppose that ∀x∀y (P (x) ∨ Q(y )) is true. We want to show that ∀xP (x) ∨ ∀xQ(x)
is true. If ∀xP (x) is true, then we are done. Otherwise, P (x0 ) must be false for some x0 in the domain of
discourse. For this x0 , then, the hypothesis tells us that P (x0 ) ∨ Q(y ) is true, no matter what y is. Since
P (x0 ) is false, it must be the case that Q(y ) is true for each y . In other words, ∀yQ(y ) is true, or, to change
the name of the meaningless quantiﬁed variable, ∀xQ(x) is true. This certainly implies that ∀xP (x) ∨ ∀xQ(x)
is true, as desired.
50. a) By Exercises 45 and 46b in Section 1.3, we can simply bring the existential quantiﬁer outside: ∃x(P (x) ∨
Q(x) ∨ A).
b) By Exercise 48 of the current section, the expression inside the parentheses is logically equivalent to
∀x∀y (P (x) ∨ Q(y )) . Applying the negation operation, we obtain ∃x∃y ¬(P (x) ∨ Q(y )) . 22 Chapter 1 The Foundations: Logic and Proofs c) First we rewrite this using Table 7 in Section 1.2 as ∃xQ(x) ∨ ¬∃xP (x) , which is equivalent to ∃xQ(x) ∨
∀x¬P (x) . To combine the existential and universal statements we use Exercise 49b of the current section,
obtaining ∀x∃y (¬P (x) ∨ Q(y )) , which is in prenex normal form.
52. We simply want to say that there exists an x such that P (x) holds, and that every y such that P (y ) holds
must be this same x . Thus we write ∃x P (x) ∧ ∀y (P (y ) → y = x) . Even more compactly, we can write
∃x∀y (P (y ) ↔ y = x) . SECTION 1.5 Rules of Inference 2. This is modus tollens. The ﬁrst statement is p → q , where p is “George does not have eight legs” and q is
“George is not an insect.” The second statement is ¬q . The third is ¬p . Modus tollens is valid. We can
therefore conclude that the conclusion of the argument (third statement) is true, given that the hypotheses
(the ﬁrst two statements) are true.
4. a) We have taken the conjunction of two propositions and asserted one of them. This is, according to Table 1,
simpliﬁcation.
b) We have taken the disjunction of two propositions and the negation of one of them, and asserted the other.
This is, according to Table 1, disjunctive syllogism. See Table 1 for the other parts of this exercise as well.
c) modus ponens
d) addition
e) hypothetical syllogism
6. Let r be the proposition “It rains,” let f be the proposition “It is foggy,” let s be the proposition “The
sailing race will be held,” let l be the proposition “The life saving demonstration will go on,” and let t be the
proposition “The trophy will be awarded.” We are given premises (¬r ∨ ¬f ) → (s ∧ l) , s → t , and ¬t . We
want to conclude r . We set up the proof in two columns, with reasons, as in Example 6. Note that it is valid
to replace subexpressions by other expressions logically equivalent to them.
Step
1. ¬t
2. s → t
3. ¬s
4. (¬r ∨ ¬f ) → (s ∧ l)
5. (¬(s ∧ l)) → ¬(¬r ∨ ¬f )
6. (¬s ∨ ¬l) → (r ∧ f )
7. ¬s ∨ ¬l
8. r ∧ f
9. r Reason
Hypothesis
Hypothesis
Modus tollens using (1) and (2)
Hypothesis
Contrapositive of (4)
De Morgan’s law and double negative
Addition, using (3)
Modus ponens using (6) and (7)
Simpliﬁcation using (8) 8. First we use universal instantiation to conclude from “For all x , if x is a man, then x is not an island”
the special case of interest, “If Manhattan is a man, then Manhattan is not an island.” Then we form the
contrapositive (using also double negative): “If Manhattan is an island, then Manhattan is not a man.” Finally
we use modus ponens to conclude that Manhattan is not a man. Alternatively, we could apply modus tollens. Section 1.5 Rules of Inference 23 10. a) If we use modus tollens starting from the back, then we conclude that I am not sore. Another application
of modus tollens then tells us that I did not play hockey.
b) We really can’t conclude anything speciﬁc here.
c) By universal instantiation, we conclude from the ﬁrst conditional statement by modus ponens that dragonﬂies have six legs, and we conclude by modus tollens that spiders are not insects. We could say using existential
generalization that, for example, there exists a nonsixlegged creature that eats a sixlegged creature, and
that there exists a noninsect that eats an insect.
d) We can apply universal instantiation to the conditional statement and conclude that if Homer (respectively,
Maggie) is a student, then he (she) has an Internet account. Now modus tollens tells us that Homer is not a
student. There are no conclusions to be drawn about Maggie.
e) The ﬁrst conditional statement is that if x is healthy to eat, then x does not taste good. Universal
instantiation and modus ponens therefore tell us that tofu does not taste good. The third sentence says that
if you eat x , then x tastes good. Therefore the fourth hypothesis already follows (by modus tollens) from the
ﬁrst three. No conclusions can be drawn about cheeseburgers from these statements.
f) By disjunctive syllogism, the ﬁrst two hypotheses allow us to conclude that I am hallucinating. Therefore
by modus ponens we know that I see elephants running down the road.
12. Applying Exercise 11, we want to show that the conclusion r follows from the ﬁve premises (p ∧ t) → (r ∨ s) ,
q → (u ∧ t) , u → p , ¬s , and q . From q and q → (u ∧ t) we get u ∧ t by modus ponens. From there we get
both u and t by simpliﬁcation (and the commutative law). From u and u → p we get p by modus ponens.
From p and t we get p ∧ t by conjunction. From that and (p ∧ t) → (r ∨ s) we get r ∨ s by modus ponens.
From that and ¬s we ﬁnally get r by disjunctive syllogism.
14. In each case we set up the proof in two columns, with reasons, as in Example 6.
a) Let c(x) be “ x is in this class,” let r(x) be “ x owns a red convertible,” and let t(x) be “ x has gotten
a speeding ticket.” We are given premises c(Linda) , r(Linda) , ∀x(r(x) → t(x)) , and we want to conclude
∃x(c(x) ∧ t(x)) .
Step
Reason
1. ∀x(r(x) → t(x))
Hypothesis
2. r(Linda) → t(Linda)
Universal instantiation using (1)
3. r(Linda)
Hypothesis
4. t(Linda)
Modus ponens using (2) and (3)
5. c(Linda)
Hypothesis
6. c(Linda) ∧ t(Linda)
Conjunction using (4) and (5)
Existential generalization using (6)
7. ∃x(c(x) ∧ t(x))
b) Let r(x) be “ r is one of the ﬁve roommates listed,” let d(x) be “ x has taken a course in discrete
mathematics,” and let a(x) be “ x can take a course in algorithms.” We are given premises ∀x(r(x) → d(x))
and ∀x(d(x) → a(x)) , and we want to conclude ∀x(r(x) → a(x)) . In what follows y represents an arbitrary
person.
Step
Reason
1. ∀x(r(x) → d(x))
Hypothesis
2. r(y ) → d(y )
Universal instantiation using (1)
Hypothesis
3. ∀x(d(x) → a(x))
4. d(y ) → a(y )
Universal instantiation using (3)
5. r(y ) → a(y )
Hypothetical syllogism using (2) and (4)
6. ∀x(r(x) → a(x))
Universal generalization using (5)
c) Let s(x) be “ x is a movie produced by Sayles,” let c(x) be “ x is a movie about coal miners,” and let 24 Chapter 1 The Foundations: Logic and Proofs w(x) be “movie x is wonderful.” We are given premises ∀x(s(x) → w(x)) and ∃x(s(x) ∧ c(x)) , and we want
to conclude ∃x(c(x) ∧ w(x)) . In our proof, y represents an unspeciﬁed particular movie.
Step
Reason
1. ∃x(s(x) ∧ c(x))
Hypothesis
2. s(y ) ∧ c(y )
Existential instantiation using (1)
3. s(y )
Simpliﬁcation using (2)
4. ∀x(s(x) → w(x))
Hypothesis
5. s(y ) → w(y )
Universal instantiation using (4)
6. w(y )
Modus ponens using (3) and (5)
7. c(y )
Simpliﬁcation using (2)
8. w(y ) ∧ c(y )
Conjunction using (6) and (7)
9. ∃x(c(x) ∧ w(x))
Existential generalization using (8)
d) Let c(x) be “ x is in this class,” let f (x) be “ x has been to France,” and let l(x) be “ x has visited the
Louvre.” We are given premises ∃x(c(x) ∧ f (x)) , ∀x(f (x) → l(x)) , and we want to conclude ∃x(c(x) ∧ l(x)) .
In our proof, y represents an unspeciﬁed particular person.
Step
Reason
Hypothesis
1. ∃x(c(x) ∧ f (x))
2. c(y ) ∧ f (y )
Existential instantiation using (1)
3. f (y )
Simpliﬁcation using (2)
4. c(y )
Simpliﬁcation using (2)
5. ∀x(f (x) → l(x))
Hypothesis
6. f (y ) → l(y )
Universal instantiation using (5)
7. l(y )
Modus ponens using (3) and (6)
8. c(y ) ∧ l(y )
Conjunction using (4) and (7)
9. ∃x(c(x) ∧ l(x))
Existential generalization using (8)
16. a) This is correct, using universal instantiation and modus tollens.
b) This is not correct. After applying universal instantiation, it contains the fallacy of denying the hypothesis.
c) After applying universal instantiation, it contains the fallacy of aﬃrming the conclusion.
d) This is correct, using universal instantiation and modus ponens.
18. We know that some s exists that makes S (s, Max) true, but we cannot conclude that Max is one such s .
Therefore this ﬁrst step is invalid.
20. a) This is invalid. It is the fallacy of aﬃrming the conclusion. Letting a = −2 provides a counterexample.
b) This is valid; it is modus ponens.
22. We will give an argument establishing the conclusion. We want to show that all hummingbirds are small. Let
Tweety be an arbitrary hummingbird. We must show that Tweety is small. The ﬁrst premise implies that
if Tweety is a hummingbird, then Tweety is richly colored. Therefore by (universal) modus ponens we can
conclude that Tweety is richly colored. The third premise implies that if Tweety does not live on honey, then
Tweety is not richly colored. Therefore by (universal) modus tollens we can now conclude that Tweety does
live on honey. Finally, the second premise implies that if Tweety is a large bird, then Tweety does not live
on honey. Therefore again by (universal) modus tollens we can now conclude that Tweety is not a large bird,
i.e., that Tweety is small, as desired. Notice that we invoke universal generalization as the last step.
24. Steps 3 and 5 are incorrect; simpliﬁcation applies to conjunctions, not disjunctions. Section 1.5 Rules of Inference 25 26. We want to show that the conditional statement P (a) → R(a) is true for all a in the domain; the desired
conclusion then follows by universal generalization. Thus we want to show that if P (a) is true for a particular a , then R(a) is also true. For such an a , by universal modus ponens from the ﬁrst premise we have Q(a) ,
and then by universal modus ponens from the second premise we have R(a) , as desired.
28. We want to show that the conditional statement ¬R(a) → P (a) is true for all a in the domain; the desired
conclusion then follows by universal generalization. Thus we want to show that if ¬R(a) is true for a particular a , then P (a) is also true. For such an a , universal modus tollens applied to the second premise gives us
¬(¬P (a) ∧ Q(a)) . By rules from propositional logic, this gives us P (a) ∨ ¬Q(a) . By universal generalization
from the ﬁrst premise, we have P (a) ∨ Q(a) . Now by resolution we can conclude P (a) ∨ P (a) , which is logically
equivalent to P (a) , as desired.
30. Let a be “Allen is a good boy”; let h be “Hillary is a good girl”; let d be “David is happy.” Then our
assumptions are ¬a ∨ h and a ∨ d . Using resolution gives us h ∨ d , as desired.
32. We apply resolution to give the tautology (p ∨ F) ∧ (¬p ∨ F) → (F ∨ F) . The lefthand side is equivalent to
p ∧ ¬p , since p ∨ F is equivalent to p , and ¬p ∨ F is equivalent to ¬p . The righthand side is equivalent to F .
Since the conditional statement is true, and the conclusion is false, it follows that the hypothesis, p ∧ ¬p , is
false, as desired.
34. Let us use the following letters to stand for the relevant propositions: d for “logic is diﬃcult”; s for “many
students like logic”; and e for “mathematics is easy.” Then the assumptions are d ∨ ¬s and e → ¬d . Note
that the ﬁrst of these is equivalent to s → d , since both forms are false if and only if s is true and d is false.
In addition, let us note that the second assumption is equivalent to its contrapositive, d → ¬e . And ﬁnally,
by combining these two conditional statements, we see that s → ¬e also follows from our assumptions.
a) Here we are asked whether we can conclude that s → ¬e . As we noted above, the answer is yes, this
conclusion is valid.
b) The question concerns ¬e → ¬s . This is equivalent to its contrapositive, s → e . That doesn’t seem to
follow from our assumptions, so let’s ﬁnd a case in which the assumptions hold but this conditional statement
does not. This conditional statement fails in the case in which s is true and e is false. If we take d to be true
as well, then both of our assumptions are true. Therefore this conclusion is not valid.
c) The issue is ¬e ∨ d , which is equivalent to the conditional statement e → d . This does not follow from our
assumptions. If we take d to be false, e to be true, and s to be false, then this proposition is false but our
assumptions are true.
d) The issue is ¬d ∨ ¬e , which is equivalent to the conditional statement d → ¬e . We noted above that this
validly follows from our assumptions.
e) This sentence says ¬s → (¬e ∨ ¬d) . The only case in which this is false is when s is false and both e and
d are true. But in this case, our assumption e → ¬d is also violated. Therefore, in all cases in which the
assumptions hold, this statement holds as well, so it is a valid conclusion. 26 SECTION 1.6 Chapter 1 The Foundations: Logic and Proofs Introduction to Proofs 2. We must show that whenever we have two even integers, their sum is even. Suppose that a and b are
two even integers. Then there exist integers s and t such that a = 2s and b = 2t . Adding, we obtain
a + b = 2s + 2t = 2(s + t) . Since this represents a + b as 2 times the integer s + t , we conclude that a + b is
even, as desired.
4. We must show that whenever we have an even integer, its negative is even. Suppose that a is an even integer.
Then there exists an integer s such that a = 2s . Its additive inverse is −2s , which by rules of arithmetic and
algebra (see Appendix 1) equals 2(−s) . Since this is 2 times the integer −s , it is even, as desired.
6. An odd number is one of the form 2n + 1 , where n is an integer. We are given two odd numbers, say 2a + 1
and 2b + 1 . Their product is (2a + 1)(2b + 1) = 4ab + 2a + 2b + 1 = 2(2ab + a + b) + 1 . This last expression
shows that the product is odd, since it is of the form 2n + 1 , with n = 2ab + a + b .
8. Let n = m2 . If m = 0 , then n + 2 = 2 , which is not a perfect square, so we can assume that m ≥ 1 . The
smallest perfect square greater than n is (m + 1)2 , and we have (m + 1)2 = m2 + 2m + 1 = n + 2m + 1 >
n + 2 · 1 + 1 > n + 2 . Therefore n + 2 cannot be a perfect square.
10. A rational number is a number that can be written in the form x/y where x and y are integers and y = 0.
Suppose that we have two rational numbers, say a/b and c/d . Then their product is, by the usual rules for
multiplication of fractions, (ac)/(bd) . Note that both the numerator and the denominator are integers, and
that bd = 0 since b and d were both nonzero. Therefore the product is, by deﬁnition, a rational number.
12. This is true. Suppose that a/b is a nonzero rational number and that x is an irrational number. We must
prove that the product xa/b is also irrational. We give a proof by contradiction. Suppose that xa/b were
rational. Since a/b = 0 , we know that a = 0, so b/a is also a rational number. Let us multiply this rational
number b/a by the assumed rational number xa/b . By Exercise 26, the product is rational. But the product is
(b/a)(xa/b) = x , which is irrational by hypothesis. This is a contradiction, so in fact xa/b must be irrational,
as desired.
14. If x is rational and not zero, then by deﬁnition we can write x = p/q , where p and q are nonzero integers.
Since 1/x is then q/p and p = 0 , we can conclude that 1/x is rational.
16. We give a proof by contraposition. If it is not true than m is even or n is even, then m and n are both odd.
By Exercise 6, this tells us that mn is odd, and our proof is complete.
18. a) We must prove the contrapositive: If n is odd, then 3n + 2 is odd. Assume that n is odd. Then we can
write n = 2k + 1 for some integer k . Then 3n + 2 = 3(2k + 1) + 2 = 6k + 5 = 2(3k + 2) + 1 . Thus 3n + 2 is
two times some integer plus 1, so it is odd.
b) Suppose that 3n + 2 is even and that n is odd. Since 3n + 2 is even, so is 3n . If we add subtract an odd
number from an even number, we get an odd number, so 3n − n = 2n is odd. But this is obviously not true.
Therefore our supposition was wrong, and the proof by contradiction is complete.
20. We need to prove the proposition “If 1 is a positive integer, then 12 ≥ 1 .” The conclusion is the true statement
1 ≥ 1 . Therefore the conditional statement is true. This is an example of a trivial proof, since we merely
showed that the conclusion was true. Section 1.6 Introduction to Proofs 27 22. We give a proof by contradiction. Suppose that we don’t get a pair of blue socks or a pair of black socks.
Then we drew at most one of each color. This accounts for only two socks. But we are drawing three socks.
Therefore our supposition that we did not get a pair of blue socks or a pair of black socks is incorrect, and
our proof is complete.
24. We give a proof by contradiction. If there were at most two days falling in the same month, then we could
have at most 2 · 12 = 24 days, since there are 12 months. Since we have chosen 25 days, at least three of
them must fall in the same month.
26. We need to prove two things, since this is an “if and only if” statement. First let us prove directly that
if n is even then 7n + 4 is even. Since n is even, it can be written as 2k for some integer k . Then
7n + 4 = 14k + 4 = 2(7k + 2) . This is 2 times an integer, so it is even, as desired. Next we give a proof by
contraposition that if 7n + 4 is even then n is even. So suppose that n is not even, i.e., that n is odd. Then
n can be written as 2k + 1 for some integer k . Thus 7n + 4 = 14k + 11 = 2(7k + 5) + 1 . This is 1 more than
2 times an integer, so it is odd. That completes the proof by contraposition.
28. There are two things to prove. For the “if” part, there are two cases. If m = n , then of course m2 = n2 ;
if m = −n , then m2 = (−n)2 = (−1)2 n2 = n2 . For the “only if” part, we suppose that m2 = n2 . Putting
everything on the left and factoring, we have (m + n)(m − n) = 0 . Now the only way that a product of two
numbers can be zero is if one of them is zero. Therefore we conclude that either m + n = 0 (in which case
m = −n ), or else m − n = 0 (in which case m = n ), and our proof is complete.
30. We write these in symbols: a < b , (a + b)/2 > a , and (a + b)/2 < b . The latter two are equivalent to
a + b > 2a and a + b < 2b , respectively, and these are in turn equivalent to b > a and a < b , respectively. It
is now clear that all three statements are equivalent.
32. We give direct proofs that (i ) implies (ii ), that (ii ) implies (iii ), and that (iii ) implies (i ). That will suﬃce.
For the ﬁrst, suppose that x = p/q where p and q are integers with q = 0 . Then x/2 = p/(2q ) , and this is
rational, since p and 2q are integers with 2q = 0 . For the second, suppose that x/2 = p/q where p and q
are integers with q = 0 . Then x = (2p)/q , so 3x − 1 = (6p)/q − 1 = (6p − q )/q and this is rational, since
6p − q and q are integers with q = 0 . For the last, suppose that 3x − 1 = p/q where p and q are integers
with q = 0 . Then x = (p/q + 1)/3 = (p + q )/(3q ) , and this is rational, since p + q and 3q are integers with
3q = 0.
√
34. No. This line of reasoning shows that if
2x2 − 1 = x , then we must have x = 1 or x = −1 . These are
therefore the only possible solutions, but we have no guarantee that they are solutions, since not all of our
steps were reversible (in particular, squaring both sides). Therefore we must substitute these values back into
the original equation to determine whether they do indeed satisfy it.
36. The only conditional statements not shown directly are p1 ↔ p2 , p2 ↔ p4 , and p3 ↔ p4 . But these each
follow with one or more intermediate steps: p1 ↔ p2 , since p1 ↔ p3 and p3 ↔ p2 ; p2 ↔ p4 , since p2 ↔ p1
(just established) and p1 ↔ p4 ; and p3 ↔ p4 , since p3 ↔ p1 and p1 ↔ p4 .
38. We must ﬁnd a number that cannot be written as the sum of the squares of three integers. We claim that 7
is such a number (in fact, it is the smallest such number). The only squares that can be used to contribute
to the sum are 0 , 1 , and 4 . We cannot use two 4’s , because their sum exceeds 7 . Therefore we can use at
most one 4 , which means that we must get 3 using just 0’s and 1’s . Clearly three 1’s are required for this,
bringing the total number of squares used to four. Thus 7 cannot be written as the sum of three squares. 28 Chapter 1 The Foundations: Logic and Proofs 40. Suppose that we look at the ten groups of integers in three consecutive locations around the circle (ﬁrstsecondthird, secondthirdfourth, . . . , eighthninthtenth, ninthtenthﬁrst, and tenthﬁrstsecond). Since
each number from 1 to 10 gets used three times in these groups, the sum of the sums of the ten groups must
equal three times the sum of the numbers from 1 to 10 , namely 3 · 55 = 165 . Therefore the average sum is
165/10 = 16.5 . By Exercise 39, at least one of the sums must be greater than or equal to 16.5 , and since the
sums are whole numbers, this means that at least one of the sums must be greater than or equal to 17 .
42. We show that each of these is equivalent to the statement (v ) n is odd, say n = 2k +1 . Example 1 showed that
(v ) implies (i ), and Example 8 showed that (i ) implies (v ). For (v ) → (ii ) we see that 1 − n = 1 − (2k + 1) =
2(−k ) is even. Conversely, if n were even, say n = 2m , then we would have 1 − n = 1 − 2m = 2(−m) + 1 , so
1 − n would be odd, and this completes the proof by contraposition that (ii ) → (v ). For (v ) → (iii ), we see
that n3 = (2k +1)3 = 8k 3 +12k 2 +6k +1 = 2(4k 3 +6k 2 +3k )+1 is odd. Conversely, if n were even, say n = 2m ,
then we would have n3 = 2(4m3 ) , so n3 would be even, and this completes the proof by contraposition that
(iii ) → (v ). Finally, for (v ) → (iv ), we see that n2 + 1 = (2k + 1)2 + 1 = 4k 2 + 4k + 2 = 2(2k 2 + 2k + 1) is
even. Conversely, if n were even, say n = 2m , then we would have n2 + 1 = 2(2m2 ) + 1, so n2 + 1 would be
odd, and this completes the proof by contraposition that (iv ) → (v ). SECTION 1.7 Proof Methods and Strategy 2. The cubes that might go into the sum are 1 , 8 , 27 , 64 , 125 , 216 , 343 , 512 , and 729 . We must show that
no two of these sum to a number on this list. If we try the 45 combinations ( 1 + 1 , 1 + 8 , . . . , 1 + 729 , 8 + 8 ,
8 + 27 , . . . 8 + 729 , . . . , 729 + 729 ), we see that none of them works. Having exhausted the possibilities, we
conclude that no cube less than 1000 is the sum of two cubes.
4. There are three main cases, depending on which of the three numbers is smallest. If a is smallest (or tied for
smallest), then clearly a ≤ min(b, c) , and so the lefthand side equals a . On the other hand, for the righthand
side we have min(a, c) = a as well. In the second case, b is smallest (or tied for smallest). The same reasoning
shows us that the righthand side equals b ; and the lefthand side is min(a, b) = b as well. In the ﬁnal case,
in which c is smallest (or tied for smallest), the lefthand side is min(a, c) = c , whereas the righthand side is
clearly also c . Since one of the three has to be smallest we have taken care of all the cases.
6. The number 1 has this property, since the only positive integer not exceeding 1 is 1 itself, and therefore the
sum is 1 . This is a constructive proof.
8. The only perfect squares that diﬀer by 1 are 0 and 1 . Therefore these two consecutive integers cannot both
be perfect squares. This is a nonconstructive proof—we do not know which of them meets the requirement.
(In fact, a computer algebra system will tell us that neither of them is a perfect square.)
10. Of these three numbers, at least two must have the same sign (both positive or both negative), since there are
only two signs. (It is conceivable that some of them are zero, but we view zero as positive for the purposes of
this problem.) The product of two with the same sign is nonnegative. This was a nonconstructive proof, since
we have not identiﬁed which product is nonnegative. (In fact, a computer algebra system will tell us that all
three are positive, so all three products are positive.)
12. An assertion like this one is implicitly universally quantiﬁed—it means that for all rational numbers a and b ,
ab is rational. To disprove such a statement it suﬃces to provide one counterexample. Take a = 2 and
√
√
b = 1/2 . Then ab = 21/2 = 2 , and we know from Example 10 in Section 1.6 that 2 is not rational. Section 1.7 Proof Methods and Strategy 29 14. We know from algebra that the following equations are equivalent: ax + b = c , ax = c − b . x = (c − b)/a .
This shows, constructively, what the unique solution of the given equation is.
16. Given r , let a be the closest integer to r less than r , and let b be the closest integer to r greater than r . In
the notation to be introduced in Section 2.3, a = r and b = r . In fact, b = a + 1 . Clearly the distance
between r and any integer other than a or b is greater than 1 so cannot be less than 1/2 . Furthermore, since
r is irrational, it cannot be exactly halfway between a and b , so exactly one of r − a < 1/2 and b − r < 1/2
holds.
18. Given x , let n be the greatest integer less than or equal to x , and let = x − n . In the notation to be
introduced in Section 2.3, n = x . Clearly 0 ≤ < 1 , and is unique for this n . Any other choice of n
would cause the required to be less than 0 or greater than or equal to 1 , so n is unique as well.
20. We follow the hint. The square of every real number is nonnegative, so (x − 1/x)2 ≥ 0 . Multiplying this out
and simplifying, we obtain x2 − 2 + 1/x2 ≥ 0 , so x2 + 1/x2 ≥ 2 , as desired.
22. If a = 5 and b = 8 , then the quadratic mean is (52 + 82 )/2 ≈ 6.67 , and the arithmetic mean is (5 + 8)/2 =
6.5. If a = 10 and b = 100 , then the quadratic mean is
(102 + 1002 )/2 ≈ 71.06 , and the arithmetic
mean is (10 + 100)/2 = 55 . We conjecture that the quadratic mean of a and b is always greater than
their arithmetic mean if a and b are distinct positive real numbers (clearly if a = b then both means are
this common value). So we want to verify the inequality
(a2 + b2 )/2 > (a + b)/2 . Squaring both sides
(this is legal because everything in sight is positive) and multiplying by 4 gives us the equivalent inequality
2a2 + 2b2 > a2 + 2ab + b2 , which is in turn equivalent to (a − b)2 > 0 after putting everything on the lefthand
side and factoring. This is clearly always true, and our proof is complete.
24. If we were to end up with nine 0’s , then in the step before this we must have had either nine 0’s or nine
1’s , since each adjacent pair of bits must have been equal and therefore all the bits must have been the same.
Thus if we are to start with something other than nine 0’s and yet end up with nine 0’s , we must have had
nine 1’s at some point. But in the step before that each adjacent pair of bits must have been diﬀerent; in
other words, they must have alternated 0 , 1 , 0 , 1 , and so on. This is impossible with an odd number of bits.
This contradiction shows that we can never get nine 0’s .
26. Clearly only the last two digits of n contribute to the last two digits of n2 . So we can compute 02 , 12 , 22 ,
32 , . . . , 992 , and record the last two digits, omitting repetitions. We obtain 00 , 01 , 04 , 09 , 16 , 25 , 36 , 49 ,
64 , 81 , 21 , 44 , 69 , 96 , 56 , 89 , 24 , 61 , 41 , 84 , 29 , 76 . From that point on, the list repeats in reverse order
(as we take the squares from 252 to 492 , and then it all repeats again as we take the squares from 502 to
992 ). The reason for these last two statements are that (50 − n)2 = 2500 − 100n + n2 , so (50 − n)2 and n2
have the same two ﬁnal digits, and (50 + n)2 = 2500 + 100n + n2 , so (50 + n)2 and n2 have the same two
ﬁnal digits. Thus our list (which contains 22 numbers) is complete.
28. If y  ≥ 2 , then 2x2 + 5y 2 ≥ 2x2 + 20 ≥ 20 , so the only possible values of y to try are 0 and ±1 . In the
former case we would be looking for solutions to 2x2 = 14 and in the latter case to 2x2 = 9 . Clearly there
are no integer solutions to these equations, so there are no solutions to the original equation.
30. Following the hint, we let x = m2 − n2 , y = 2mn , and z = m2 + n2 . Then x2 + y 2 = (m2 − n2 )2 + (2mn)2 =
m4 − 2m2 n2 + n4 + 4m2 n2 = m4 + 2m2 n2 + n4 = (m2 + n2 )2 = z 2 . Thus we have found inﬁnitely many
solutions, since m and n can be arbitrarily large. 30 Chapter 1 The Foundations: Logic and Proofs √
√
32. One proof that 3 2 is irrational is similar to the proof that 2 is irrational, given in Example 10 in Section 1.6.
√
It is a proof by contradiction. Suppose that 21/3 (or 3 2 , which is the same thing) is the rational number
p/q , where p and q are positive integers with no common factors (the fraction is in lowest terms). Cubing,
we see that 2 = p3 /q 3 , or, equivalently, p3 = 2q 3 . Thus p3 is even. Since the product of odd numbers is
odd, this means that p is even, so we can write p = 2s . Substituting into the equation p3 = 2q 3 , we obtain
8s3 = 2q 3 , which simpliﬁes to 4s3 = q 3 .
Now we play the same game with q . Since q 3 is even, q must be even. We have now concluded that p
and q are both even, that is, that 2 is a common divisor of p and q . This contradicts the choice of p/q to be
√
in lowest terms. Therefore our original assumption—that 3 2 is rational—is in error, so we have proved that
√
3
2 is irrational.
34. The average of two diﬀerent numbers is certainly always between the two numbers. Furthermore, the average
a of rational number x and irrational number y must be irrational, because the equation a = (x + y )/2 leads
to y = 2a − x , which would be rational if a were rational.
36. The solution is not unique, but here is one way to measure out four gallons. Fill the 5gallon jug from the
8gallon jug, leaving the contents (3, 5, 0) , where we are using the ordered triple to record the amount of water
in the 8gallon jug, the 5gallon jug, and the 3gallon jug, respectively. Next ﬁll the 3gallon jug from the
5gallon jug, leaving (3, 2, 3) . Pour the contents of the 3gallon jug back into the 8gallon jug, leaving (6, 2, 0) .
Empty the 5gallon jug’s contents into the 3gallon jug, leaving (6, 0, 2) , and then ﬁll the 5gallon jug from
the 8gallon jug, producing (1, 5, 2) . Finally, top oﬀ the 3gallon jug from the 5gallon jug, and we’ll have
(1, 4, 3) , with four gallons in the 5gallon jug.
38. a) 16 → 8 → 4 → 2 → 1
b) 11 → 34 → 17 → 52 → 26 → 13 → 40 → 20 → 10 → 5 → 16 → 8 → 4 → 2 → 1
c) 35 → 106 → 53 → 160 → 80 → 40 → 20 → 10 → 5 → 16 → 8 → 4 → 2 → 1
d) 113 → 340 → 170 → 85 → 256 → 128 → 64 → 32 → 16 → 8 → 4 → 2 → 1
40. This is easily done, by laying the dominoes horizontally, three in the ﬁrst and last rows and four in each of
the other six rows.
42. Without loss of generality, we number the squares from 1 to 25, starting in the top row and proceeding left to
right in each row; and we assume that squares 5 (upper right corner), 21 (lower left corner), and 25 (lower right
corner) are the missing ones. We argue that there is no way to cover the remaining squares with dominoes.
By symmetry we can assume that there is a domino placed in 12 (using the obvious notation). If square
3 is covered by 38, then the following dominoes are forced in turn: 49, 1015, 1920, 2324, 1722, and 1318,
and now no domino can cover square 14. Therefore we must use 34 along with 12. If we use all of 1722,
1823, and 1924, then we are again quickly forced into a sequence of placements that lead to a contradiction.
Therefore without loss of generality, we can assume that we use 2223, which then forces 1924, 1520, 910,
1314, 78, 611, and 1217, and we are stuck once again. This completes the proof by contradiction that no
placement is possible.
44. The barriers shown in the diagram split the board into one continuous closed path of 64 squares, each adjacent
to the next (for example, start at the upper left corner, go all the way to the right, then all the way down,
then all the way to the left, and then weave your way back up to the starting point). Because each square in
the path is adjacent to its neighbors, the colors alternate. Therefore, if we remove one black square and one
white square, this closed path decomposes into two paths, each of which starts in one color and ends in the
other color (and therefore has even length). Clearly each such path can be covered by dominoes by starting
at one end. This completes the proof. Supplementary Exercises 31 46. If we study Figure 7, we see that by rotating or reﬂecting the board, we can make any square we wish
nonwhite, with the exception of the squares with coordinates (3, 3) , (3, 6) , (6, 3) , and (6, 6) . Therefore the
same argument as was used in Example 22 shows that we cannot tile the board using straight triominoes if
any one of those other 60 squares is removed. The following drawing (rotated as necessary) shows that we can
tile the board using straight triominoes if one of those four squares is removed. 48. We will use a coloring of the 10 × 10 board with four colors as the basis for a proof by contradiction showing
that no such tiling exists. Assume that 25 straight tetrominoes can cover the board. Some will be placed
horizontally and some vertically. Because there is an odd number of tiles, the number placed horizontally and
the number placed vertically cannot both be odd, so assume without loss of generality that an even number
of tiles are placed horizontally. Color the squares in order using the colors red, blue, green, yellow in that
order repeatedly, starting in the upper left corner and proceeding row by row, from left to right in each row.
Then it is clear that every horizontally placed tile covers one square of each color and each vertically placed
tile covers either zero or two squares of each color. It follows that in this tiling an even number of squares of
each color are covered. But this contradicts the fact that there are 25 squares of each color. Therefore no
such coloring exists. SUPPLEMENTARY EXERCISES FOR CHAPTER 1
2. The truth table is as follows.
p
T
T
T
T
F
F
F
F q r TT
TF
FT
FF
TT
TF
FT
FF p∨q p ∧ ¬r (p ∨ q ) → (p ∧ ¬r) T
T
T
T
T
T
F
F F
T
F
T
F
F
F
F F
T
F
T
F
F
T
T 4. a) The converse is “If I drive to work today, then it will rain.” The contrapositive is “If I do not drive to work
today, then it will not rain.” The inverse is “If it does not rain today, then I will not drive to work.”
b) The converse is “If x ≥ 0 then x = x .” The contrapositive is “If x < 0 then x = x .” The inverse is “If
x = x , then x < 0 .”
c) The converse is “If n2 is greater than 9 , then n is greater than 3 .” The contrapositive is “If n2 is not
greater than 9 , then n is not greater than 3 .” The inverse is “If n is not greater than 3 , then n2 is not
greater than 9 .” 32 Chapter 1 The Foundations: Logic and Proofs 6. The inverse of p → q is ¬p → ¬q . Therefore the inverse of the inverse is ¬¬p → ¬¬q , which is equivalent to
p → q (the original proposition). The converse of p → q is q → p . Therefore the inverse of the converse is
¬q → ¬p , which is the contrapositive of the original proposition. The inverse of the contrapositive is q → p ,
which is the same as the converse of the original statement.
8. Let t be “Sergei takes the job oﬀer”; let b be “Sergei gets a signing bonus”; and let h be “Sergei will receive a
higher salary.” The given statements are t → b , t → h , b → ¬h , and t . By modus ponens we can conclude b
and h from the ﬁrst two conditional statements, and therefore we can conclude ¬h from the third conditional
statement. We now have the contradiction h ∧ ¬h , so these statements are inconsistent.
10. Since both knights and knaves claim that they are knights (the former truthfully and the latter deceivingly),
we know that A is a knave. But since A’s statement must be false, and the ﬁrst part of the conjunction is
true, the second part must be false, so we know that B must be a knave as well. If C were a knight, then
B ’s statement would be true, and knaves must lie, so C must also be a knave. Thus all three are knaves.
12. If S is a proposition, then it is either true or false. If S is false, then the statement “If S is true, then unicorns
live” is vacuously true; but this statement is S , so we would have a contradiction. Therefore S is true, so the
statement “If S is true, then unicorns live” is true and has a true hypothesis. Hence it has a true conclusion
(modus ponens), and so unicorns live. But we know that unicorns do not live. It follows that S cannot be a
proposition.
14. a) The answer is ∃xP (x) if we do not read any signiﬁcance into the use of the plural, and ∃x∃y (P (x) ∧ P (y ) ∧
x = y ) if we do.
b) ¬∀xP (x) , or, equivalently, ∃x¬P (x)
c) ∀yQ(y )
d) ∀xP (x) (the class has nothing to do with it) e) ∃y ¬Q(y ) 16. The given statement tells us that there are exactly two elements in the domain. Therefore the statement will
be true as long as we choose the domain to be anything with size 2 , such as the United States presidents
named Bush.
18. We want to say that for every y , there do not exist four diﬀerent people each of whom is the grandmother of y .
Thus we have ∀x¬∃a∃b∃c∃d(a = b ∧ a = c ∧ a = d ∧ b = c ∧ b = d ∧ c = d ∧ G(a, y ) ∧ G(b, y ) ∧ G(c, y ) ∧ G(d, y )) .
20. a) Since there is no real number whose square is −1 , it is true that there exist exactly 0 values of x such
that x2 = −1 .
b) This is true, because 0 is the one and only value of x such that x = 0.
√
√
c) This is true, because 2 and − 2 are the only values of x such that x2 = 2.
d) This is false, because there are more than three values of x such that x = x , namely all positive real
numbers.
22. Let us assume the hypothesis. This means that there is some x0 such that P (x0 , y ) holds for all y . Then
it is certainly true that for all y there exists an x such that P (x, y ) is true, since in each case we can take
x = x0 . Note that the converse is not always a tautology, since the x in ∀y ∃xP (x, y ) can depend on y .
24. No. Here is an example. Let P (x, y ) be x > y , where we are talking about integers. Then for every y there
does exist an x such that x > y ; we could take x = y + 1 , for example. However, there does not exist an x
such that for every y , x > y ; in other words, there is no superlarge integer (if for no other reason than that
no integer can be larger than itself). Supplementary Exercises 33 26. a) It will snow today, but I will not go skiing tomorrow.
b) Some person in this class does not understand mathematical induction.
c) All students in this class like discrete mathematics.
d) There is some mathematics class in which all the students stay awake during lectures.
28. Let W (r) means that room r is painted white. Let I (r, b) mean that room r is in building b . Let L(b, u)
mean that building b is on the campus of United States university u . Then the statement is that there is
some university u and some building on the campus of u such that every room in b is painted white. In
symbols this is ∃u∃b(L(b, u) ∧ ∀r(I (r, b) → W (r))) .
30. To say that there are exactly two elements that make the statement true is to say that two elements exist that
make the statement true, and that every element that makes the statement true is one of these two elements.
More compactly, we can phrase the last part by saying that an element makes the statement true if and only
if it is one of these two elements. In symbols this is ∃x∃y (x = y ∧ ∀z (P (z ) ↔ (z = x ∨ z = y ))) . In English
we might express the rule as follows. The hypotheses are that P (x) and P (y ) are both true, that x = y , and
that every z that satisﬁes P (z ) must be either x or y . The conclusion is that there are exactly two elements
that make P true.
32. We give a proof by contraposition. If x is rational, then x = p/q for some integers p and q with q = 0.
Then x3 = p3 /q 3 , and we have expressed x3 as the quotient of two integers, the second of which is not zero.
This by deﬁnition means that x3 is rational, and that completes the proof of the contrapositive of the original
statement.
34. Let m be the square root of n , rounded down if it is not a whole number. (In the notation to be introduced in
√
Section 2.3, we are letting m =
n .) We can see that this is the unique solution in a couple of ways. First,
clearly the diﬀerent choices of m correspond to a partition of N , namely into {0} , {1, 2, 3} , {4, 5, 6, 7, 8} ,
{9, 10, 11, 12, 13, 14, 15} , . . . . So every n is in exactly one of these sets. Alternatively, take the square root
√
√
of the given inequalities to give m ≤ n < m + 1 . That m is then the ﬂoor of n (and that m is unique)
follows from statement (1a) of Table 1 in Section 2.3.
36. A constructive proof seems indicated. We can look for examples by hand or with a computer program. The
smallest ones to be found are 50 = 52 + 52 = 12 + 72 and 65 = 42 + 72 = 12 + 82 .
38. We claim that the number 7 is not the sum of at most two squares and a cube. The ﬁrst two positive squares
are 1 and 4 , and the ﬁrst positive cube is 1 , and these are the only numbers that could be used in forming
the sum. Clearly no sum of three or fewer of these is 7 . This counterexample disproves the statement.
√
√
√
40. We give a proof by contradiction. If 2 + 3 were rational, then so would be its square, which is 5 + 2 6 .
√
Subtracting 5 and dividing by 2 then shows that 6 is rational, but this contradicts the theorem we are told
to assume. ...
View
Full Document
 Spring '10
 JEFFHAROUTUNIAN
 Math, Logic, Proposition, Logical connective

Click to edit the document details