faculty of philosophy & computer science &

faculty of mathematics mathematics

and natural sciences

SYLLABUS INTRODUCTION TO LOGIC

Rineke Verbrugge,
Joris IJsselmuiden &
Barteld Kooi

September 2016
Contents

Part A
1 Introduction 2
2 Atomic sentences 2
3 Propositional logic: conjunction, disjunction and negation 3
4 Propositional logic: material conditional and biconditional 4
5 First-order logic: introduction to quantifiers 5
6 First-order logic: sentences with multiple quantifiers 7

Part B
7 Introduction 8
8 Set theory 8
9 Propositional logic: normal forms 12
10 First-order logic: syntax and semantics 12
11 First-order logic: normal forms 13

selected exercises: Part A 18

Solutions to selected exercises: Part B 19

Additional proof exercises with solutions 20

2

Part A

1 Introduction
This syllabus complements the book “Language, Proof and Logic” by Bar-
wise and Etchemendy. It provides some alternative notations, additional
exercises and background information. These are mainly drawn from the
following existing sources:

– Syllabus Parvulae Logicales by Phoebe Kradavis

– Syllabus Elementaire Verzamelingenleer by Riegholt Hilbrands

Special thanks goes out to Vı́ctor Sánchez for providing some translation ex-
ercises, Erik Krabbe for his remarks about an earlier version of this syllabus,
Jan Misker for providing solutions to some of the exercises in this syllabus,
Ayla Kangur and Sybren Jansen for helping in translating the syllabus.
LPL Introduction pp. 1 – 11 The introduction of “Language, Proof and Logic” describes what logic en-
tails, why we study logic, and provides instructions on how to use the book
properly.

2 Atomic sentences
Alternative pp. 23 LPL The book LPL (Language, Proof and Logic) is exceptional in the way that
definition of atomic sentences of propositional logic are defined. In most books the con-
atomic sentences vention is to write atomic sentences of propositional logic with a single up-
percase letter. Instead of writing Taller(claire, max) those books simply
write it as H provided with the translation key:

H = Claire is taller than Max.

A similar notation with lowercase instead of uppercase letters often occurs

as well. You will encounter several exercises in this syllabus in which you
will learn to work with these more conventional atomic sentences.
Other books provide a clearer distinction between propositional and first-
order logic than LPL does. In propositional logic, arguments are studied
 3

in which the validity depends on the meaning of connectives like not, and,
or, if ... then and if and only if. Everything else is disregarded. Other
books are therefore not interested in the construction of atomic sentences:
abbreviating them with a single uppercase letter is sufficient.
In first-order logic, arguments of which the validity depends on the mean-
ing of quantifiers, including all and some, is studied as well. In this case the
structure of the atoms is of great importance. In LPL, logic is presented as
a single system in which atomic propositions are decomposed using pred-
icate symbols and constants. For the final exam it is important that you
can distinguish between both logics. You should know that in propositional
logic, atomic sentences are represented by a single uppercase letter and that
in first-order logic, these sentences are further decomposed using predicate
symbols and constants.

Formal proofs using Fitch

To create formal proofs you can make use of Fitch, software which is already
installed on the university computers. There you will also find practice
exercises and an extensive user manual.
EXERCISE 2.1 Given the premises SameColumn(a,b) and a = c, use Fitch to prove that
SameColumn(c,b) is the case. Use the same premises to prove that c = c.

3 Propositional logic: conjunction,

disjunction and negation
3.1 Satisfiability and logical truth
Spurious rows pp. 94 – 103 LPL On page 94 of LPL, Barwise and Etchemendy make the distinction between
a tautology and a logical truth. A row in a truth table is called a spurious
row if that row cannot possibly be true because of the meaning of the atomic
sentences in the sentence. Suppose that the atomic sentence A represents
d = d. Assignments which make A false cannot really exist. All rows
containing such assignments are spurious rows.
If you want to check whether a sentence is logically true, you first cross
out all spurious rows and check whether the formula is true in all remaining
rows. If you just want to know whether a formula is satisfiable, you first
cross out all the spurious rows and check whether the sentence is true in at
least one remaining row.
EXERCISE 3.1 Suppose A, B and C are atomic sentences.

a. Build a truth table for the sentence ¬(A ∧ B) ↔ (¬C ∨ ¬B).

b. Is the sentence ¬(A ∧ B) ↔ (¬C ∨ ¬B) a tautology? Explain your

answer.

For c, d and e suppose that the atomic sentences A and C are logically equiv-
alent, for example A stands for Larger(e,d) and C stands for Smaller(d,e).

c. Specify which rows of the truth table constructed in part a are spurious
rows.

d. Is the sentence ¬(A ∧ B) ↔ (¬C ∨ ¬B) logically true? Explain your

answer.

e. Is the sentence ¬(A ∧ B) ↔ (¬C ∨ ¬B) satisfiable? If so, provide a truth

assignment which makes the sentence true. If not, explain why.
4

3.2 Formal proofs

Previously proven During lectures and tutorials, we occasionally refer to previously proven
theorems theorems. It is important to understand these occurrences. During the
final exam, however, it is not allowed to use them. With Fitch you can
refer to these previously proven theorems by using the Taut Con rule. For
propositional logic these are among others the following:

a. P ∨ ¬P without premises
b. ¬(P ∧ Q) from premise ¬P ∨ ¬Q
c. ¬P ∨ ¬Q from premise ¬(P ∧ Q)
d. ¬(P ∨ Q) from premise ¬P ∧ ¬Q
e. ¬P ∧ ¬Q from premise ¬(P ∨ Q)
f. Q from premises P ∨ Q and ¬P
g. ¬¬P from premise P

EXERCISE 3.2 Prove by natural deduction, if necessary use Fitch:

a. ¬(A ∧ ¬A)
b. B from premise A ∧ ¬A

EXERCISE 3.3 Prove by natural deduction, if necessary use Fitch:

a. A ∨ B ∨ C from premise A ∧ C
b. B from premise B ∨ (A ∧ B)
c. B from premise B ∧ (A ∨ B)
d. B ∧ (A ∨ B) from premise B
e. C from premise (A ∧ C) ∨ (B ∧ C)

4 Propositional logic: material conditional

and biconditional
4.1 Formal proofs with conditionals

EXERCISE 4.1 Prove by natural deduction, if necessary use Fitch:

a. A → A
b. (A ∧ ¬A) → B
c. (A → ¬A) → ¬A
d. ((A → B) ∧ (A → ¬B)) → ¬A
EXERCISE 4.2 Prove by natural deduction, if necessary use Fitch:
a. A → (B ∧ C) from premises A → B and A → C
b. A → ¬¬A, without premises
c. A → C from premises (A ∧ B) → C and B
 5

EXERCISE 4.3 Prove by natural deduction, if necessary use Fitch:

a. A → (B → C) from premise B → (A → C)
b. A → D from premises A → (B ∨ C), B → D and C → D

c. ¬C from premises ¬(A ∧ B), C → A and C → B

d. (A ∨ B) ↔ A from premise ¬B
e. B from premises A → B and ¬A → B
f. B from premises A, (A ∧ C) → B and ((A → B) → B) → (A → C)

g. (A → B) ∨ (A ∧ ¬B), without premises

5 First-order logic: introduction to

quantifiers
5.1 Translate
EXERCISE 5.1 Translate the following sentences into first-order logic. Provide as much
structure as possible and give a translation key for each domain of discourse.
First check which atomic sentences you will need per domain of discourse.
Domain of discourse 1: All animals.
a. Some primates are not old world monkeys.

b. No primate is an ape and not intelligent.

c. All flying mammals are bats.
d. If an animal is an ape, it is not an old world monkey.

e. Bats are flying mammals, but they are not intelligent.

Domain of discourse 2: All companies and licenses.

f. Not every company can finance themselves.

g. Companies that do not work with us are our competitors.
h. Companies that cooperate with us make profit. Some companies that
do not cooperate with us also make profit.

i. There is a license that Microsoft has not granted to Sun.

j. Microsoft has granted no license to Sun.

Domain of discourse 3: All computer software.

k. Everything is open source or closed source.
l. Everything is open source or everything is closed source.

m. Some open source and closed source computer programs are affordable.
n. No open source computer program is unaffordable.
6

EXERCISE 5.2 Translate the following sentences into first-order logic. Provide as much
structure as possible and give a translation key for each sentence. The
domain of discourse is the set of integers: . . . , −2, −1, 0, 1, 2, . . .
a. 2 is an even number.

b. 2 is greater than 3.
c. If 5 is greater than 4, then 5 is also greater than 3.
d. Every number greater than 4 is also greater than 3.

e. There exists an even number greater than 3.

f. No number is greater than itself.

5.2 Formal proofs with quantifiers

EXERCISE 5.3 Prove by natural deduction, if necessary use Fitch:

a. ∀x(P (x) → R(x)) from the premises ∀x(P (x) → Q(x)) and
∀x(Q(x) → R(x))
b. ∀xP (x) → ∀xQ(x) from the premise ∀x(P (x) → Q(x))
c. ∀xP (x) ∧ ∀xQ(x) from the premise ∀x(P (x) ∧ Q(x))
d. ∀x(P (x) ∧ Q(x)) from the premise ∀xP (x) ∧ ∀xQ(x)

e. ∃x(P (x) ∨ Q(x)) from the premise ∃xP (x) ∨ ∃xQ(x)

f. ∃xP (x) ∨ ∃xQ(x) from the premise ∃x(P (x) ∨ Q(x))
g. ∃xP (x) ∧ ∃xQ(x) from the premise ∃x(P (x) ∧ Q(x))
(Take care: the reverse is not valid.)

h. ¬∃x(P (x) ∧ R(x)) from the premises ∀x(P (x) → Q(x)) and
¬∃x(Q(x) ∧ R(x))
EXERCISE 5.4 Prove by natural deduction, if necessary use Fitch:
a. c = i from the premises ∀x(O(b, x) → x = i) and O(b, c).

This is a possible formalization of the riddle “Brothers and sisters I

have none, but that man’s father is my father’s son. Who am I?”,
with the following translation key:
b = my father
c = that man’s father
i = me
O(x, y) = x is a parent of y

b. ∃xH(a, x) from hypothesis ∀xH(x, a)

 7

6 First-order logic: sentences with

multiple quantifiers
EXERCISE 6.1 Translate the following sentences into first-order logic. Provide as much
structure as possible and give a translation key for each domain of discourse.
First check which atomic sentences you will need per domain of discourse.
Domain of discourse 1: All primates.
a. If a primate is not an ape, it is an old world monkey.
b. Kanzi acquires something from Washoe, but not from Matata.
c. Every ape loves another ape.

d. Kanzi does not steal food from another primate.

e. No ape is smarter than Washoe, except Kanzi.

Domain of discourse 2: All humans.

f. Every student knows a student that admires him/her.
g. Every student knows a student that admires himself/herself.

h. Some people are kind and some are unkind.

i. Nobody admires everyone.
j. No one admires someone who admires everyone who admires no one.

k. Someone whom no one sees is invisible.

l. Someone who sees no one is blind.
m. Every lecturer is satisfied if even one student is satisfied.
n. John loves nobody except himself.
8

Part B

7 Introduction
Part B of this syllabus explores some topics which are of special interest for
students Artificial Intelligence, Computing Science and Mathematics. For
example, normal forms are essential for logic programming, and the extra
attention to models of first-order logic is useful for courses such as Advanced
Logic, Artificial Intelligence, Multiagent Systems, Program Correctness and
Discrete Structures.

8 Set theory
Informal pp. 37 – 38 LPL You have encountered many different sets in mathematics classes in high
introduction to school and university, for example N, the set of natural numbers 0, 1, 2, . . .,
sets and Z, the set of integers . . . , −2, −1, 0, 1, 2, . . .. Strangely enough one can-
not precisely define the concept of “set”. However, since the end of the 19th
century Cantor, Zermelo, Fraenkel and other mathematicians have made
axiomatic descriptions of set theory. More on this topic can be found in
Chapter 15.9 of LPL (not part of the course material). We provide a short
but more intuitive introduction to the most common concepts encountered
in set theory, including a number of examples and exercises.

An example of a set is {1, 2, 3}. This set contains the members (or elements)
1, 2 and 3. You write 1 ∈ {1, 2, 3} to denote that 1 is a member of the set
{1, 2, 3} (and similarly for the other members), but 4 6∈ {1, 2, 3}. Sets can
have all sorts of members; for example {Aristoteles, P lato} is the set that
contains exactly two philosophers, namely Aristoteles and Plato, as its mem-
bers. Sets can also contain other sets as their members. For example, the
set {1, {2, 3}} contains two members: 1 ∈ {1, {2, 3}} and {2, 3} ∈ {1, {2, 3}}.
 9

8.1 Set operations and statements

In the descriptions below, A and B are both sets.
– A and B are equal, written as A = B, if A and B contain exactly the
same members. This is formally written as: for all x it holds that
x ∈ A if and only if x ∈ B.
Examples: {1, 2} = {2, 1} and {{1, 2}, 0, {2, 1}} = {{1, 2}, 0}, but it
is not the case that {{1, 2}, 1} = {1, 2}.
– A is a subset of B, written as A ⊆ B, if every member of A is also a
member of B. This is formally written as: for all x it holds that if
x ∈ A then also x ∈ B.
Examples: {1} ⊆ {1, 2}, and {1} ⊆ {1}, but not 1 ⊆ {1, 2}.
– A is a proper subset of B, written as A ⊂ B, if every member of A is
also a member of B, but B contains at least one other member. This
is formally written as: A ⊆ B and there exists an x such that x ∈ B
and x 6∈ A.
– The empty set ∅ is the set containing no element. An odd but true
property of the empty set is that it is a subset of all sets!
Example: ∅ ⊆ {1, 2}. Don’t you believe that for every x it holds that
if x ∈ ∅ then also x ∈ {1, 2}? Then name an object that is a member
of ∅ but not a member of {1, 2}!
– If E(x) is a particular property of objects x, then the abstraction
{x | E(x)} is the set of all objects having the property E. Often
you use an earlier given set A as a basis for an abstraction. Then
{x ∈ A | E(x)} is the set containing all members of A that share the
property E.
Example: {x | x ∈ N and x is even} is the set containing the even
natural numbers; this is equivalent to {x ∈ N | x is even}.
– The intersection A ∩ B is {x | x ∈ A and x ∈ B}.
Example: {1, 2} ∩ {1, 3} = {1}.
– The union A ∪ B is {x | x ∈ A or x ∈ B}.
Example: {1, 2} ∪ {1, 3} = {1, 2, 3}. The “or” in this definition is the
inclusive or. If an object is a member of both A and B, then it is also
definitely a member of A ∪ B.
– The difference A \ B is equal to {x | x ∈ A and x 6∈ B}.
Example: {1, 2} \ {1, 3} = {2}.
– The complement Ac is {x | x 6∈ A}. In general, you consider the
complement not with respect to everything (numbers, sets, cookies,
philosophers, etc.), but with respect to some given superset U , called
the universe. In that case Ac = U \ A.
Example: Let U be the set of natural numbers N. Then {n | n ∈ N
and n is even }c = {n | n ∈ N and n is odd}.
10

EXERCISE 8.1 Which of the following statements are true, which are not? Explain your
answers.
a. 1 ∈ {1}
b. 1 ⊆ {1}

c. {{1}} ∈ {1, {1}}

d. {{1}} ⊆ {1, {1}}
e. ∅ ∈ {∅}

f. ∅ ⊆ {∅}

For the next three exercises (8.2, 8.3 and 8.4), we make use of the following
sets:

A = {a, b, c, 2, 3, 4}

B = {a, b}
C = {c, 2}
D = {b, c}

E = {a, b, {c}}
F = ∅
G = {{a, b}, {c, 2}}

We assume that a, b and c are not natural numbers and that they differ
from one another (so a 6= b, a 6= c and b 6= c).
EXERCISE 8.2 Which of the following statements are true, which are not? Explain your
answers.
a. c ∈ A

b. c ∈ F
c. c ∈ E
d. {c} ⊆ E
e. b ∈ G

f. B ⊆ G

EXERCISE 8.3 Write down the following sets (listing all members):
a. B ∪ C
b. D ∪ F
c. A ∩ B

d. A \ B
e. B \ A
f. G \ B
 11

EXERCISE 8.4 As universe we take U = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, a, b, c, d, e, f }. Write

down the following sets (listing all members):

a. Ac

b. (A ∩ C)c

c. ((A \ D) \ C)

d. (A ∩ B) ∪ C

e. ((C ∪ D) ∩ B)c

f. ((A ∪ D) ∪ B) ∩ F

g. ((A \ D) \ C)c

8.2 Ordered pairs, relations and functions

The members of a set have no order: {1, 2} = {2, 1}. Sometimes we do
want to represent order in set theory. For this we use ordered pairs or more
generally ordered n-tuples. We write them as h1, 2i, or h1, 1, 2, 3, 5, 8, 13i. In
this situation, order does matter and repetitions as well: h1, 2i =6 h2, 1i and
h1, 1, 2, 3, 5, 8, 13i =
6 h1, 2, 3, 5, 8, 13i.
We can now use sets of ordered pairs to model predicates of any arity.
A (binary) relation is a set of ordered pairs. Consider the following set
R = {hx, yi | x ∈ N, y ∈ N and x = y + 5}. A pair hx, yi is a member of this
set if x is five greater than y. The set R is thus the relation “five greater
than”. If hx, yi ∈ R we write it as xRy.
Some relations have a special property: for every x, y, z: if xRy and xRz,
then y = z. In other words, for every x there is a unique y such that xRy.
Such a relation is called a function, which is often displayed in lowercase.
Consider the set f = {hx, yi | x ∈ N, y ∈ N and x2 = y}. This relation is
a function, namely the square function. Instead of writing hx, yi ∈ f , we
write f (x) = y.

EXERCISE 8.5 Make a graph of the following relations; represent ordered pairs by an arrow
between their members.

a. R1 = {h1, 1i}

b. R2 = {h1, 2i}

c. R3 = {h1, 1i, h2, 2i}

d. R4 = {h1, 2i, h2, 2i}

e. R5 = {h1, 2i, h2, 1i}

f. R6 = {h1, 1i, h2, 2i, h1, 2i, h2, 1i}

g. R7 = ∅

h. R8 = {h1, 1i, h2, 3i}

EXERCISE 8.6 Which relations from Exercise 8.5 are functions?

12

9 Propositional logic: normal forms

LPL 4.6 pp. 122 – 125 Put the following sentences in disjunctive normal form:
EXERCISE 9.1 a. ¬(P ↔ (Q ∨ R))
b. P ↔ (Q ↔ R)

c. (¬P ∨ Q ∨ ¬R) ∧ (P ∨ ¬Q)

EXERCISE 9.2 Put the following sentences in conjunctive normal form:
a. ¬(P ∨ Q) ∨ ¬(¬R ∧ S)
b. (P ∧ Q ∧ R) ∨ ¬S

c. ¬((P → Q) ∧ (R → S))

10 First-order logic: syntax and semantics

10.1 Syntax
LPL 1.5 pp. 31 – 34 LPL mentions many examples of terms, but does not give an inductive
definition. The definition is provided below:
LPL 9.1 pp. 230 – 231
– constants are terms

– variables are terms

– If t1 , . . . , tn are terms and f is an n-ary function symbol, then f (t1 , . . . , tn )
is also a term
– nothing else is a term

This shows that you can reapply function symbols over and over again
(termed “stacking”). The term father(mother(max)), for example, refers to
the maternal grandfather of Max. But also (x + 5) ∗ (y − 2) is a stacking of
function symbols. This can be easily observed if you use the prefix notation
instead of the infix notation: ∗(+(x, 5), −(y, 2))
Note: predicate symbols cannot be stacked. Big(Dodec(a)) is total non-
sense. If you want to express in first-order logic that a is a big dodecahedron
you will need a conjunction: Big(a) ∧ Dodec(a).

10.2 Semantics
LPL 18.1, 18.2 pp. 511 – 522 This exercise is about the three musketeers. Suppose that we have a struc-
ture M with the domain of discourse D = {Aramis, Athos, Porthos}.
EXERCISE 10.1
We know the following facts about the predicates and constants:
– M(Tall ) = {Porthos},

– M(Rich) = {Aramis},
– M(Fast) = {Aramis, Porthos},
– M(Old ) = {Athos},
– (a)M = Aramis
 13

Let the variable assignment h be defined as

– h(x1 ) = Aramis
– h(x2 ) = Athos

– h(x3 ) = Porthos
Which of the following statements are true, which are not? Explain your
answers for c, f , j and o.
a. M |= Fast(a) [h]

b. M |= Fast(x2 ) [h]
c. M |= Fast(x3 ) [h[x3 /Aramis]]
Explain your answer
d. M |= Fast(x3 ) [h[x3 /Porthos]]

e. M |= Rich(x2 ) [h]
f. M |= Fast(x1 ) ↔ ¬Old (x1 ) [h[x1 /Aramis]]
Explain your answer
g. M |= Fast(x1 ) ↔ ¬Old (x1 ) [h[x1 /Athos]]

h. M |= Fast(x1 ) ↔ ¬Old (x1 ) [h[x1 /Porthos]]

i. M |= ∀x1 (Fast(x1 ) ↔ ¬Old (x1 )) [h]
j. M |= ∀x1 (Fast(x1 ) → Rich(x1 )) [h]
Explain your answer

k. M |= ∀x1 (Rich(x1 ) → Fast(x1 )) [h]

l. M |= ∃x1 ∀x2 (¬Tall (x2 ) ↔ x1 = x2 ) [h]
m. M |= ∀x1 (Fast(x1 ) → (Rich(x1 ) ↔ Tall (x1 ))) [h]

n. M |= ∀x1 (Tall (x1 ) → (Fast(x1 ) ↔ Rich(x1 ))) [h]

o. M |= ∃x2 ∀x1 (x1 6= x2 → Fast(x1 )) [h]
Explain your answer
EXERCISE 10.2 Which of the statements of Exercise 10.1 change truth value if we add
D’Artagnan to the domain of discourse D? It is known that D’Artagnan is
not tall. He is rich and old, but not fast.

11 First-order logic: normal forms

11.1 Prenex normal form
LPL 11.7 pp. 320 – 323 A formula is in prenex normal form if either of the following holds:
Definition prenex LPL p. 320 – it contains no quantifiers at all, or
normal form
– it is of the form Q1 v1 . . . Qn vn P , where Qi is either ∀ or ∃, each vi is
a variable, and P is a quantifier-free formula.
14

Formulas that are in prenex normal form are, for example:

– P (c), and

– ∀x∃y(P (x, y) → R(x, y)).

Not in prenex normal form are:

– ¬∃xP (x) (due to the ¬ at the beginning of the formula), and

– ∃xP (x) → P (c) (this formula does not conform to the form quantifier
+ quantifier-free formula, because ∃x only applies to P (x)).

Equivalences for LPL p. 285 In addition to the equivalences provided on page 285, there are some other
prenex normal equivalences that you can use to convert formulas to prenex normal form.
form They are all listed below. See also LPL 10.3 and 10.4 (pp. 277–285).
To create an alphabetical variant.
Note: y is free in P (y) exactly where x is free in P (x).

a. ∀xP (x) ⇔ ∀yP (y)

b. ∃xP (x) ⇔ ∃yP (y)

To push negation signs past a quantifier (LPL p. 281):

c. ¬∀xP (x) ⇔ ∃x¬P (x)

d. ¬∃xP (x) ⇔ ∀x¬P (x)

To push quantifiers to the front in formulas containing ∧, ∨:

e. ∀xP ∧ ∀xQ ⇔ ∀x(P ∧ Q)

f. ∃xP ∨ ∃xQ ⇔ ∃x(P ∨ Q)

g. ∀xP ∨ Q ⇔ ∀x(P ∨ Q) if x is not free in Q.

h. ∃xP ∧ Q ⇔ ∃x(P ∧ Q) if x is not free in Q.

i. ∀xP ∧ Q ⇔ ∀x(P ∧ Q) if x is not free in Q.

j. ∃xP ∨ Q ⇔ ∃x(P ∨ Q) if x is not free in Q.

k. Q ⇔ ∀xQ if x is not free in Q.

l. Q ⇔ ∃xQ if x is not free in Q.

The equivalences i and j are convenient because they provide more effi-
ciency. Besides, i follows from e and k, and j follows from f and l.
To push quantifiers to the front in formulas containing →:

m. ∀xP → Q ⇔ ∃x(P → Q) if x is not free in Q.

n. ∃xP → Q ⇔ ∀x(P → Q) if x is not free in Q.

o. P → ∀xQ ⇔ ∀x(P → Q) if x is not free in P .

p. P → ∃xQ ⇔ ∃x(P → Q) if x is not free in P .

 15

Here follows an example how to put a formula in prenex normal form using
the above-mentioned equivalences. Suppose we start with the following
formula:
¬∃xP (x) ∨ ∀xQ(x).

We first push the negation sign past the quantifier, by rule d. This leads to
∀x¬P (x) ∨ ∀xQ(x).
We canno push all the quantifiers to the front by using rules e–l at this mo-
ment: we first have to make sure that both disjuncts have different quanti-
fiers as main connective. Using rule a we can create an alphabetical variant
∀x¬P (x) ∨ ∀yQ(y).
By rule g we get
∀x(¬P (x) ∨ ∀yQ(y)),

and applying rule g again gives us

∀x∀y(¬P (x) ∨ Q(y)).
This formula is in prenex normal form and is equivalent to the formula we
started with.
Here is a more difficult example with less explanation (the rules that we
applied can be found directly after the ⇔ sign).

∀xA(x) → [∃yB(y) → ∀xB(x)] ⇔ (a)

∀xA(x) → [∃yB(y) → ∀zB(z)] ⇔ (n)

∀xA(x) → ∀y[B(y) → ∀zB(z)] ⇔ (o)

∀xA(x) → ∀y∀z[B(y) → B(z)] ⇔ (o)

∀y(∀xA(x) → ∀z[B(y) → B(z))] ⇔ (o)

∀y∀z(∀xA(x) → [B(y) → B(z)]) ⇔ (m)

∀y∀z∃x(A(x) → [B(y) → B(z)])

Note that we are using square brackets here. Square brackets are interpreted
in the same way as normal parentheses, but in some situations provide more
clarity.
EXERCISE 11.1 Put the following formulas into prenex normal form:

a. ∀xA(x) ∧ B(c)
b. B ∨ ¬∃xA(x)
c. ∀xA(x) → B(x)
d. ∀xA(x, y) → ∃xA(x, y)

e. B → ∀x(∃y(A(y) → C(x, y)) → D(x))

f. ∀x(∀y(R(x, y) → A(y)) → A(x)) → ∀xA(x)
16

11.2 Skolem normal form

LPL 18.5 pp. 530 – 532 A formula is in Skolem normal form if it is in prenex normal form and only
contains universal quantifiers.
Definition Skolem LPL p. 530
LPL provides only the simplest case of the Skolem normal form: A
normal form
Skolem normal form of the formula ∀x∃yP (x, y) is given by the (non-
equivalent) formula ∀xP (x, f (x)).
Here are two other cases:
– The formula ∀x∀y∃zP (x, y, z) has as its Skolem normal form the for-
mula
∀x∀yP (x, y, g(x, y)); note that g is a binary function! With more uni-
versal quantifiers you will need corresponding n-ary Skolem functions.

– The formula ∃zP (z) has as its Skolem normal form the formula P (d);
in this case the constant d can be viewed as a nullary Skolem function.
EXERCISE 11.2 Skolemize the following formulas. If necessary, put them in prenex normal
form first.

a. ∀x∃y∀zP (x, y, z)
b. ∀x∃y∀z∃uQ(x, y, z, u)
c. ∀x∀y(A(x, y) → ∃z(B(z) ∧ ∃uC(z, y, u)))

11.3 Horn sentences

LPL 17.3 (partly) pp. 495 - 500 Section 17.3 of LPL up to the point where “correctness of algorithm” is
shown in the margin, is part of the course material.
Now that you know prenex and Skolem normal forms, we can extend the
definition of a Horn sentence. A Horn sentence:
– is in prenex normal form,

– is in Skolem normal form (so only with universal quantifiers), and

– its quantifier-free part is in Horn form, as defined in LPL: it is in con-
junctive normal form with the additional property: every disjunction
of literals in the sentence contains at most one positive literal.
Examples of Horn sentences are:

– ∀x(¬R(x) ∨ T (x))
– ∀xR(x)
– ∀x((¬A(x) ∨ B(f (x))) ∧ (¬B(f (x)) ∨ ¬C(x) ∨ A(x))).

The following are not Horn sentences:

– ∀x(A(x)∨B(x)∨C(x)) (this formula has too many positive disjuncts.)
– ∀x∃yR(x, y) (this formula contains an existential quantifier.)
Universal quantifiers are often left out of Horn sentences, like in the formula
(¬A(x) ∨ B(f (x))) ∧ (¬B(f (x)) ∨ ¬C(x) ∨ A(x))); in these cases you must
think of all the free variables as being universally quantified.
 17

Here follows an example to put a formula in Horn form:

¬∀x∀yR(x, y) ∨ ∀zA(z) ⇔ (c)

∃x¬∀yR(x, y) ∨ ∀zA(z) ⇔ (c)

∃x∃y¬R(x, y) ∨ ∀zA(z) ⇔ (o)

∃x∃y∀z(¬R(x, y) ∨ A(z)).

This formula is now in prenex normal form. The next step is to Skolemize
this formula to create a (non-equivalent) Skolem normal form:

∀z(¬R(c, d) ∨ A(z)).
This sentence is already in Horn form, because the quantifier-free part is
a conjunction of a single disjunction, which indeed contains at most one
positive disjunct, namely A(z).
EXERCISE 11.3 Put the following formulas in Horn form; if necessary use Skolem normal
forms:

a. ∀x∀y(R(x, y) → ∀z(R(y, z) → R(x, z)))

b. ∀x∀y(R(x, y) → ¬R(y, x))
c. ∀x(A(x) → B(x)) → (∀yA(y) → ∀zB(z))
18

selected exercises: Part A

LPL ex. 1.9 p. 30

1. Owned(claire, f olly, 2:00)

2. Gave(claire, pris, max, 2:05)

3. Student(max)

4. F ed(claire, carl, 2:00)

5. Owned(max, f olly, 3:05)

6. 2:00 < 2:05

LPL ex. 3.23 p. 88

1. Student(max) ∧ ¬P et(max)

2. F ed(claire, f olly, 2:00) ∧ Gave(claire, f olly, max, 2:00)

The second 2:00 is open to discussion.

3. Owned(max, f olly, 2:05) ∨ Owned(claire, f olly, 2:05)

4. ¬F ed(max, f olly, 2:00) ∧ ¬F ed(max, f olly, 2:05) ∧

¬F ed(claire, f olly, 2:00) ∧ ¬F ed(claire, f olly, 2:05).
Alternatively:
¬(F ed(max, f olly, 2:00) ∨ F ed(max, f olly, 2:05)) ∧
¬(F ed(claire, f olly, 2:00) ∨ F ed(claire, f olly, 2:05))

5. 1:55 < 2:00 ∧ 2:00 < 2:05

Warning: it is not possible to translate this sentence as
1:55 < 2:00 < 2:05.

6. Gave(max, f olly, claire, 2:00) ∧ ¬Hungry(f olly, 2:00) ∧

Hungry(f olly, 3:00)

LPL ex. 7.18 p. 189

1. Gave(claire, f olly, max, 2:03) →

(Owned(claire, f olly, 2:00) ∧ Owned(max, f olly, 2:05))

2. F ed(max, f olly, 2:00) ∧

(Gave(max, f olly, claire, 2:00) → ¬Hungry(f olly, 2:05))

3. (¬F ed(max, f olly, 2:00) ∧ ¬F ed(claire, f olly, 2:00)) →

Hungry(f olly, 2:00)

4. Angry(max, claire, 2:05) →

(F ed(claire, f olly, 2:00) ∨ F ed(claire, silly, 2:00))

5. Student(max) ↔ ¬Student(claire)

Syllabus ex. 5.2 p. 6

a. Translation key:
Even(x) = x is an even number
two = the number 2

Translation: Even(two)
 19

b. Translation key:
Greater(x, y) = x is greater than y
two = the number 2
three = the number 3

Translation: Greater(two, three)

c. Translation key:
four = the number 4
three = the number 3
five = the number 5
Greater(x, y) = x is greater than y

Translation: Greater(five, four) → Greater(five, three)

d. Translation key:
four = the number 4
three = the number 3
Greater(x, y) = x is greater than y

Translation: ∀x(Greater(x, four) → Greater(x, three))

e. Translation key:
three = the number 3
Even(x) = x is an even number
Greater(x, y) = x is greater than y

Translation: ∃x(Even(x) ∧ Greater(x, three))

f. Translation key:
Greater(x, y) = x is greater than y

Translation: ∀x¬Greater(x, x), or: ¬∃x Greater(x, x)

Solutions to selected exercises: Part B

Syllabus ex. 8.1 p. 10

1. True: 1 is a member of the set {1}.

2. False: 1 is not a set, so definitely not a subset of {1}.

3. False: the set {1, {1}} contains only the members 1 and {1}, it does not
contain the member {{1}}.
4. True: all members of the set {{1}} (which is only one, namely {1}) occur
in the set {1, {1}}. The set {{1}} is thus a subset of the set {1, {1}}.
5. True: the empty set is a member of the set {∅}.
6. True: the empty set is a subset of all sets, because all members of the
empty set (none) occur in all sets.
20

Syllabus ex. 8.3 p. 10

a. {a, b, c, 2}
b. {b, c}
c. {a, b}
d. {c, 2, 3, 4}
e. ∅
f. {{a, b}, {c, 2}}

Additional proof exercises with solutions

ADD. EXERCISE 1 Prove by natural deduction:
a. P ∨ Q from premise P ∧ Q
b. a = c from premise (a = b) ∧ (b = c)
c. C ∨ B from premise (A ∧ B) ∨ C
d. (A ∧ B) ∨ (A ∧ C) from premises A and B ∨ C

ADD. EXERCISE 2 Prove by natural deduction:

a. A → (B → A) without premises
b. (A → (B → C)) ↔ ((A ∧ B) → C) without premises
c. C ∧ D from premises A ∨ (B ∧ C), ¬E, (A ∨ B) → (D ∨ E) and ¬A

ADD. EXERCISE 3 Prove by natural deduction:

(P → Q) ↔ (¬P ∨ Q)

ADD. EXERCISE 4 Prove by natural deduction:

∃xDodec(x)
from premises:
∀y(Cube(y) ∨ Dodec(y)),

∀x(Cube(x) → Large(x)),

∃x¬Large(x)

ADD. EXERCISE 5 Prove by natural deduction:

a. ∀xP (x) from premise ∀yP (y)
b. ∃xP (x) from premise ∃yP (y)

ADD. EXERCISE 6 Prove by natural deduction:

a. ¬∀xP (x) from premise ∃x¬P (x)
b. ¬∃xP (x) from premise ∀x¬P (x)
c. ∀x¬P (x) from premise ¬∃P (x)
 21

ADD. EXERCISE 7 Prove by natural deduction:

a. ∀x∃yLikes(x, y) from premise ∀x∀yLikes(x, y)

b. ∃xCube(x) from premises ∀x(Small(x) → Cube(x)) and

∃x¬Cube(x) → ∃xSmall(x)

c. ∃xLikes(x, carl) from premises Likes(carl, max) and

∀x(∃y(Likes(y, x) ∨ Likes(x, y)) → Likes(x, x))

d. ∀x∃yLikes(x, y) from premises ∀x∀y(Likes(x, y) → Likes(y, x)) and

∃x∀yLikes(x, y)

ADD. EXERCISE 8 Prove by natural deduction:

a. P ∧ (Q ∨ R) from premise (P ∧ Q) ∨ (P ∧ R)

b. (P ∧ Q) ∨ (P ∧ R) from premise P ∧ (Q ∨ R)

c. P ∨ (Q ∧ R) from premise (P ∨ Q) ∧ (P ∨ R)

d. (P ∨ Q) ∧ (P ∨ R) from premise P ∨ (Q ∧ R)

Solutions to the additional proof exercises

Add. exercise 1

a. 1. P ∧ Q
2. P ∧ Elim: 1
3. P ∨ Q ∨ Intro: 2

b. 1. (a = b) ∧ (b = c)
2. a = b ∧ Elim: 1
3. b = c ∧ Elim: 1
4. a = c = Elim: 2,3

c. 1. (A ∧ B) ∨ C
2. A ∧ B
3. B ∧ Elim: 2
4. C ∨ B ∨ Intro: 3
5. C
6. C ∨ B ∨ Intro: 5
7. C ∨ B ∨ Elim: 1, 2–4, 5–6

d. 1. A
2. B ∨ C
3. B
4. A ∧ B ∧ Intro: 1, 3
5. (A ∧ B) ∨ (A ∧ C) ∨ Intro: 4
6. C
7. A ∧ C ∧ Intro: 1, 6
8. (A ∧ B) ∨ (A ∧ C) ∨ Intro: 7
9. (A ∧ B) ∨ (A ∧ C) ∨ Elim: 2, 3–5, 6–8
22

Add. exercise 2

a.

1. A
2. B
3. A Reit: 1
4. B → A → Intro: 2–3
5. A → (B → A) → Intro: 1–4

b.

1. A → (B → C)
2. A ∧ B
3. A ∧ Elim: 2
4. B → C → Elim: 1, 3
5. B ∧ Elim: 2
6. C → Elim: 4, 5
7. (A ∧ B) → C → Intro: 2–6
8. (A ∧ B) → C
9. A
10. B
11. A ∧ B ∧ Intro: 9, 10
12. C → Elim: 8, 11
13. B → C → Intro: 10–12
14. A → (B → C) → Intro: 9–13
15. (A → (B → C)) ↔ ((A ∧ B) → C) ↔ Intro: 1–7, 8–14

c. 1. A ∨ (B ∧ C)
2. ¬E
3. (A ∨ B) → (D ∨ E)
4. ¬A
5. A
6. ⊥ ⊥ Intro: 4,5
7. C ∧ D ⊥ Elim: 6
8. B ∧ C
9. B ∧ Elim: 8
10. C ∧ Elim: 8
11. A ∨ B ∨ Intro: 9
12. D ∨ E → Elim: 3, 11
13. D
14. D Reit: 13
15. E
16. ⊥ ⊥ Intro: 2, 15
17. D ⊥ Elim: 16
18. D ∨ Elim: 12, 13–14, 15–17
19. C ∧ D ∧ Intro: 10, 18
20. C ∧ D ∨ Elim: 1, 5–7, 8–19
 23

Add. exercise 3

1. P → Q
2. ¬(¬P ∨ Q)
3. P
4. Q → Elim: 1, 3
5. ¬P ∨ Q ∨ Intro: 4
6. ⊥ ⊥ Intro: 2, 5
7. ¬P ¬ Intro: 3–6
8. ¬P ∨ Q ∨ Intro: 7
9. ⊥ ⊥ Intro: 2, 8
10. ¬¬(¬P ∨ Q) ¬ Intro: 2–9
11. ¬P ∨ Q ¬ Elim: 10
12. ¬P ∨ Q
13. P
14. ¬P
15. ⊥ ⊥ Intro: 13,14
16. Q ⊥ Elim: 15
17. Q
18. Q Reit: 17
19. Q ∨ Elim: 12, 14–16, 17–18
20. P → Q → Intro: 13–19
21. (P → Q) ↔ (¬P ∨ Q) ↔ Intro: 1–11, 12–20

Add. exercise 4 1. ∀y[Cube(y) ∨ Dodec(y)]

2. ∀x[Cube(x) → Large(x)]
3. ∃x¬Large(x)
4. a ¬Large(a)
5. Cube(a) ∨ Dodec(a) ∀ Elim: 1
6. Cube(a) → Large(a) ∀ Elim: 2
7. Cube(a)
8. Large(a) → Elim: 6, 7
9. ⊥ ⊥ Intro: 4, 8
10. Dodec(a) ⊥ Elim: 9
11. Dodec(a)
12. Dodec(a) Reit: 11
13. Dodec(a) ∨ Elim: 5, 7–10, 11–12
14. ∃xDodec(x) ∃ Intro: 13
15. ∃xDodec(x) ∃ Elim: 3, 4–14
Add. exercise. 5

a. 1. ∀yP(y)
2. a
3. P(a) ∀ Elim: 1
4. ∀xP(x) ∀ Intro: 2–3
24

b. 1. ∃yP(y)
2. a P(a)
3. ∃xP(x) ∃ Intro: 2
4. ∃xP(x) ∃ Elim: 1, 2–3

Add. exercise 6

a. 1. ∃x¬P(x)
2. a ¬P(a)
3. ∀xP(x)
4. P(a) ∀ Elim: 3
5. ⊥ ⊥ Intro: 2, 4
6. ¬∀xP(x) ¬ Intro: 3–5
7. ¬∀xP(x) ∃ Elim: 1, 2–6

b. 1. ∀x¬P(x)
2. ∃xP(x)
3. a P(a)
4. ¬P(a) ∀ Elim: 1
5. ⊥ ⊥ Intro: 3, 4
6. ⊥ ∃ Elim: 2, 3–5
7. ¬∃xP(x) ¬ Intro: 2–6

c. 1. ¬∃xP(x)
2. a
3. P(a)
4. ∃xP(x) ∃ Intro: 3
5. ⊥ ⊥ Intro: 1, 4
6. ¬P(a) ¬ Intro: 3–5
7. ∀x¬P(x) ∀ Intro: 2–6

Add. exercise 7

a. 1. ∀x∀yLikes(x, y)
2. a
3. ∀yLikes(a, y) ∀ Elim: 1
4. Likes(a, b) ∀ Elim: 3
5. ∃yLikes(a, y) ∃ Intro: 4
6. ∀x∃yLikes(x, y) ∀ Intro: 2–5
 25

b. 1. ∀x(Small(x) → Cube(x))
2. ∃x¬Cube(x) → ∃xSmall(x)
3. ¬∃xCube(x)
4. ¬Cube(a)
5. ∃x¬Cube(x) ∃ Intro: 4
6. ∃xSmall(x) → Elim: 2, 5
7. b Small(b)
8. Small(b) → Cube(b) ∀ Elim: 1
9. Cube(b) → Elim: 8
10. ∃xCube(x) ∃ Intro: 9
11. ⊥ ⊥ Intro: 3, 10
12. ⊥ ∃ Elim: 6, 7–11
13. ¬¬Cube(a) ¬ Intro: 4–12
14. Cube(a) ¬ Elim: 13
15. ∃xCube(x) ∃ Intro: 14
16. ⊥ ⊥ Intro: 3, 15
17. ¬¬∃xCube(x) ¬ Intro: 3–16
18. ∃xCube(x) ¬ Elim: 17

c. 1. Likes(carl, max)
2. ∀x[∃y(Likes(y, x) ∨ Likes(x, y)) → Likes(x, x)]
3. Likes(max, carl) ∨ Likes(carl, max) ∨ Intro: 1
4. ∃y(Likes(y, carl) ∨ Likes(carl, y)) ∃ Intro: 3
5. ∃y(Likes(y, carl) ∨ Likes(carl, y)) → Likes(carl, carl) ∀ Elim: 2
6. Likes(carl, carl) → Elim: 4, 5
7. ∃xLikes(x, carl) ∃ Intro: 6

d. 1. ∀x∀y[Likes(x, y) → Likes(y, x)]

2. ∃x∀yLikes(x, y)
3. a

4. b ∀yLikes(b, y)
5. Likes(b, a) ∀ Elim: 4
6. ∀yLikes(b, y) → Likes(y, b) ∀ Elim: 1
7. Likes(b, a) → Likes(a, b) ∀ Elim: 6
8. Likes(a, b) → Elim: 5, 7
9. ∃yLikes(a, y) ∃ Intro: 8
10. ∃yLikes(a, y) ∃ Elim: 2, 4–9
11. ∀x∃yLikes(x, y) ∀ Intro: 3–10
26

Add. exercise 8

a. 1. (P ∧ Q) ∨ (P ∧ R)
2. P ∧ Q
3. P ∧ Elim: 2
4. Q ∧ Elim: 2
5. Q∨R ∨ Intro: 4
6. P ∧ (Q ∨ R) ∧ Intro: 3, 5
7. P ∧ R
8. P ∧ Elim: 7
9. R ∧ Elim: 7
10. Q ∨ R ∨ Intro: 9
11. P ∧ (Q ∨ R) ∧ Intro: 8, 10
12. P ∧ (Q ∨ R) ∨ Elim: 1, 2–6, 7–11

b. 1. P ∧ (Q ∨ R)
2. P ∧ Elim: 1
3. Q ∨ R ∧ Elim: 1
4. Q
5. P ∧ Q ∧ Intro: 2, 4
6. (P ∧ Q) ∨ (P ∧ R) ∨ Intro: 5
7. R
8. P ∧ R ∧ Intro: 2, 7
9. (P ∧ Q) ∨ (P ∧ R) ∨ Intro: 8
10. (P ∧ Q) ∨ (P ∧ R) ∨ Elim: 3, 4–6, 7–9

c. 1. (P ∨ Q) ∧ (P ∨ R)
2. P ∨ Q ∧ Elim: 1
3. P ∨ R ∧ Elim: 1
4. P
5. P ∨ (Q ∧ R) ∨ Intro: 4
6. Q
7. P
8. P ∨ (Q ∧ R) ∨ Intro: 7
9. R
10. Q ∧ R ∧ Intro: 6, 9
11. P ∨ (Q ∧ R) ∨ Intro: 10
12. P ∨ (Q ∧ R) ∨ Elim: 3, 7–8, 9–11
13. P ∨ (Q ∧ R) ∨ Elim: 2, 4–5, 6–12
 27

d. 1. P ∨ (Q ∧ R)
2. P
3. P ∨ Q ∨ Intro: 2
4. P ∨ R ∨ Intro: 2
5. (P ∨ Q) ∧ (P ∨ R) ∧ Intro: 3, 4
6. Q ∧ R
7. Q ∧ Elim: 6
8. R ∧ Elim: 6
9. P ∨ Q ∨ Intro: 7
10. P ∨ R ∨ Intro: 8
11. (P ∨ Q) ∧ (P ∨ R) ∧ Intro: 9, 10
12. (P ∨ Q) ∧ (P ∨ R) ∨ Elim: 1, 2–5, 6–11