DiscreteMathComputer 06 - Propositional Logic

Kanak Kshetri

1 Chapter Outline
• Difficulties with informal logic

• Propositional logic (one type of formal logic)

• Three mathematical systems for reasoning about propositions:

– Truth Tables: easy to understand but unwieldy for large expressions; “semantic
technique”
– Natural Deduction: “syntactic technique”, uses inference rules
– Boolean Algebra: axiomatic approach

2 Need for Formalism

• Natural language has ambiguities and subtleties

– e.g. “and” can sometimes suggest an implication as in “The sun is shining and I
am happy”

• Proposition: variable whose value must be either True or False

• Logical Operators: and, or, not, implies . Operate on propositions

• Statements that can’t be represented by propositional logic (due to lack of context):

– e.g. “That cloud looks like a motor bike”

– context required: which cloud? who thinks that it looks like a motor bike?

3 Basic Logical Operators

3.1 Logical And (∧)
• aka Conjunction

1
3.1.1 Truth Table
A B A∧B
False False False
False True True
True False False
True True True

3.2 Inclusive Or (∨)

• aka Disjunction

3.2.1 Truth Table

A B A∨B
False False False
False True True
True False True
True True True

3.3 Exclusive Or (⊕)

• Either one or the other, but not both

3.3.1 Truth Table

A B A⊕B
False False False
False True True
True False True
True True False

3.4 Logical Not (¬)

3.4.1 Truth Table
A ¬A
True False
False True

3.5 Logical Implication (⇒)

• Use to represent conditional phrases

• e.g. “If it is sunny today, then there will be a picnic.”

• Implication does not say anything about cause-and-effect

2
3.5.1 Truth Table
A B A⇒B
False False True
False True True
True False False
True True True

3.5.2 Understanding the truth table

• Why is an implication true even if A is False?

• One way to think about is that an implication is a statement that if something happens
then something else must also happen

– So, if the first event happens, but the second doesn’t, the implication must have
been faulty
– But the first event doesn’t happen at all, then we’re indifferent about the implica-
tion

• Rephrasing the above point, “If A then B” is like saying that “A” is sufficient for B to
occur

– So, if we see an A but not a B, clearly A isn’t sufficient for B

– But if we see a B on its own, maybe something other than A caused it.
– e.g. Matchsticks are sufficient for fire. But just because you use a lighter to make
a fire doesn’t mean matchsticks can’t cause fire. However, if you have a matchstick
and still can’t produce fire, it clearly isn’t sufficient.

• Book is a little handwavy and just says “this is the way it is”

3.6 Logical Equivalence (⇔)

• Use to make a claim that two propositions always have the same value, or are equivalent

• aka “If and only if”, “necessary and sufficient”

3.6.1 Truth Table

A B A⇔B
False False True
False True False
True False False
True True True

3.6.2 "If and only If"

• A ⇔ B can be written as (A ⇒ B) ∧ (B ⇒ A)

3
4 The Language of Propositional Logic
4.1 Well-Formedness
• Have a grammar for propositional logic

• If a formula doesn’t follow this grammar, it isn’t well formed

• e.g. Well-Formed Formula: A ⇒ (B ∧ (¬A))

• e.g. not Well-Formed Formula: ∧AB¬C

• Well-formedness is only about the syntax and says nothing about the semantics
of the formula

– e.g. “if it compiles”, then it is well-formed. Just because something compiles doesn’t
mean it’ll pass all the tests

4.2 Syntax of Well-Formed Formulas

• Constants F alse and T rue are WFF.

• Any propositional variable is WFF.

• Negation: If a is WFF, then ¬a is WFF

• Conjunction: If a and b are WFF, then a ∧ b is WFF

• Disjunction: If a and b are WFF, then a ∨ b is WFF

• Implication: If a and b are WFF, then a ⇒ b is WFF

• Equivalence: If a and b are WFF, then a ⇔ b is WFF

4.2.1 Example: P ⇒ (Q ∧ R) is a WFF

• P, Q, R are WFF because they’re propositional variables

• So, Q ∧ R is a WFF (conjunction rule)

• So, P ⇒ (Q ∧ R) is a WFF (implication rule)

4.3 Order of Precedence

• Negation (¬) is the most tightly binding operator

• Conjunction (∧)

• Disjunction (∨)

• Implication (⇒)

• Equivalence (⇔) is the least tightly binding operator

4
4.4 Object Language and Meta Language
• WFFs of propositional language will be called object language

– They refer to sentences in propositional logic

• Language of equations, substitutions, and justifications will be called the meta lan-
guage

– The surround the object language

– They’re used to talk about propositions

4.5 Computing with Boolean Expressions

• ¬x is written not x

• a ∧ b is written a && b

• a ∨ b is written a || b

• a ⇒ b is written a ==> b

• a ⇔ b is written a <=> b

4.6 (Exercise 1) : Lying Bus Driver

4.6.1 Scenario
• Want to go to the airport, are at the bus stop

• Either this bus or the next is going to the airport, but not both

• Can only ask one “yes-or-no” question

• Some drivers will answer “yes” when the correct answer is “no”

– You don’t know if this bus driver is one of those people

• Ask one “yes or no” question that can help you decide whether to take this bus or the
next

4.6.2 Solution 1
• I first gave names to the situations:

– A: this bus goes to the airport

– So, ¬A means the next one goes to the airport
– L: this driver is a liar
– So, ¬L means this driver answers truthfully

• I want to formulate my question so that if the answer is “True”, then A.

5
 • I then made a table of the possible situations:

L A Desired Answer
True True True
True False False
False True True
False False False

• So, clearly just asking A gives me the desired answer regardless of what L is

4.7 (Exercise 2) : Evaluating propositional formulas

• F alse ∧ T rue

– F alse

• T rue ∨ (¬T rue)

– T rue

• ¬(F alse ∨ T rue)

– ¬T rue
– F alse

• (¬(F alse ∧ T rue)) ∨ F alse

– (¬F alse) ∨ F alse

– T rue ∨ F alse
– T rue

• (¬T rue) ⇒ T rue

– F alse ⇒ T rue
– T rue

• T rue ∨ F alse ⇒ T rue

– T rue ⇒ T rue
– T rue

• T rue ⇒ (T rue ∧ F alse)

– T rue ⇒ F alse
– F alse

• F alse ⇒ F alse

– T rue

6
 • ¬F alse ⇔ T rue

– T rue ⇔ T rue
– T rue

• T rue ⇔ (F alse ⇒ F alse)

– T rue ⇔ T rue
– T rue

• F alse ⇔ (T rue ∧ (F alse ⇒ T rue))

– F alse ⇔ (T rue ∧ T rue)

– F alse ⇔ T rue
– F alse

• ¬(T rue ∨ F alse) ⇔ F alse ∧ T rue

– ¬T rue ⇔ F alse
– F alse ⇔ F alse
– T rue

4.7.1 Verification using the Book’s Stdm Library

∗Stdm> False / \ True

False
∗Stdm> True \ / ( not True )
True
∗Stdm> not ( False \ / True )
False
∗Stdm> ( not ( False / \ True ) ) \ / False
True
∗Stdm> ( not True ) ==> True
True
∗Stdm> True \ / False ==> True
True
∗Stdm> True ==> ( True / \ False )
False
∗Stdm> False ==> False
True
∗Stdm> ( not False ) <=> True
True
∗Stdm> True <=> ( False ==> False )
True
∗Stdm> False <=> ( True / \ ( False ==> True ) )
False

7
∗Stdm> ( not ( True \ / False ) ) <=> False / \ True
True
∗Stdm>

5 Truth Tables: Semantic Reasoning

• Table
• Column headers are all the variables (and expressions)
• Add a row for each possible setting of values for variables
– 3 variables means 23 = 8 rows
– because 2 possible settings for each of the 3 variables

5.1 Example: Calculating using a Truth Table

• Suppose we want to know when the proposition ((A ⇒ B) ∧ ¬B) ⇒ ¬A is true
• Make a Truth Table
A B A⇒B ¬B (A ⇒ B) ∧ ¬B ¬A ((A ⇒ B) ∧ ¬B) ⇒ ¬A
F F T T T T T
F T T F F T T
T F F T F F T
T T T F F F T

5.2 Definition: Tautology

• Proposition that is always True, regardless of its variables

5.3 Definition: Contradiction

• Proposition that is always False, regardless of its variables

5.4 Definition: Entailment ()

• P1 , P2 , · · · , Pn  Q means that if all the propositions P1 , P2 , · · · , Pn are True, then Q is
also True

5.4.1 Discussion
•  makes a statement about actual meanings of propositions
• Model of a logic system: Set of truth values along with the method for calculating
meaning of any well-formed formula
• Different values of n
– n = 1 gives P  Q, which has the same semantics as ⇒
– n = 0 gives  Q, which means that Q is always True, or is a tautology

8
5.5 Limitations of Truth Tables
• Size is exponential in number of variables
• Other techniques (natural deduction, Boolean algebra) give much better insight into
why a proposition is true

5.6 Exercise 3: Determine which formulas are tautologies using truth ta-
bles
5.6.1 (T rue ∧ P ) ∨ Q
P Q T rue ∧ P (T rue ∧ P ) ∨ Q
T T T T
T F T T
F T F T
F F F F
• not a tautology
• without truth tables: (T rue ∧ P ) ∨ Q becomes P ∨ Q, which is False when both P and Q
are False

5.6.2 (P ∨ Q) ⇒ (P ∧ Q)
P Q P ∨Q P ∧Q (P ∨ Q) ⇒ (P ∧ Q)
T T T T T
T F T F F
F T T F F
F F F F T
• not a tautology

5.6.3 (P ∧ Q) ⇒ (P ∨ Q)
P Q P ∧Q P ∨Q (P ∨ Q) ⇒ (P ∧ Q)
T T T T T
T F F T T
F T F T T
F F F F T
• Tautology

5.6.4 (P ∨ Q) ⇒ (Q ∨ P )
P Q P ∨Q Q∨P (P ∨ Q) ⇒ (P ∧ Q)
T T T T T
T F T T T
F T T T T
F F F F T

9
 • Tautology

• Without Truth Table: by commutativity of ∨

5.6.5 ((P ∨ Q) ∧ (P ∨ R)) ⇔ (P ∧ (Q ∨ R))

P Q R P ∨Q P ∨R LHS Q∨R RHS LHS ⇔ RHS
F F F F F F F F T
F F T F T F T F T
F T F T F F T F T
F T T T T T T F F
T F F T T T F F F
T F T T T T T T T
T T F T T T T T T
T T T T T T T T T

• Not a tautology

5.7 Exercise 4: Determine Well-formedness, and build a truth table

5.7.1 (T rue ∧ P ) ∨ Q
• Truth table from Ex 3.1

• See that it is satisfiable, but not a tautology

5.7.2 ((P ∨ Q) ∧ (P ∨ R)) ⇔ (P ∧ (Q ∨ R))

• Truth table from Ex 3.5

• Satisfiable, but not a tautology

5.7.3 ((P ∧ ¬Q) ∨ (Q ∧ ¬P )) ⇒ ¬(P ⇔ Q)

P Q A=P ∧ ¬Q B=Q ∧ ¬P A∨B C=(P ⇔ Q) ¬C (A ∨ B) ⇒ C
T T F F F T F T
T F T F T F T T
F T F T T F T T
F F F F F T F T

• Tautology

5.7.4 (P ⇒ Q) ∧ (P ⇒ ¬Q)
P Q LHS=P ⇒ Q RHS=P ⇒ ¬Q LHS ∧ RHS
T T T F F
T F F T F
F T T T T
F F T T T

10
 • Satisfiable, but not a tautology

5.7.5 (P ⇒ Q) ∧ (¬P ⇒ Q)
P Q LHS=P ⇒ Q RHS=¬P ⇒ Q LHS ∧ RHS
T T T T T
T F F T F
F T T T T
F F T F F

• Satisfiable, but not a tautology

5.7.6 (P ⇒ Q) ⇔ (¬Q ⇒ ¬P )
P Q LHS=P ⇒ Q RHS=¬Q ⇒ ¬P LHS ⇔ RHS
T T T T T
T F F F T
F T T T T
F F T T T

• Tautology

• (This is just the contrapositive)

6 Natural Deduction: Inference Reasoning

6.1 Introduction
• Natural deduction allows us to reason about logical propositions without substituting
values for variables

• Logical Inference: reasoning formally about a set of statements. Requires:

– Object Language: to talk about the statements we’re reasoning about. e.g. Propo-
sitional expressions
– Inference Rules: Rules to infer new facts from information given
– Meta Language: Defines the “form of argument” so that we can determine
whether arguments are actually valid

6.1.1 Definition: Sequent (`)

• P1 , P2 , · · · , Pn ` Q is called a sequent

• Means: if all propositions P1 , · · · , Pn are known, then proposition Q can be inferred

formally using the rules of natural deduction

• e.g. Difference between P  Q ⇒ P and P ` Q ⇒ P

11
 – The  version doesn’t say how we know that Q ⇒ P when we assume P . It just
says that it is True.
– The ` version says that there is a proof that shows Q ⇒ P when we assume P

6.1.2 Meta Variables

• Meta variables can stand for any WFF proposition

• e.g. we can have a be P ∧ Q

• Convention: meta variables will be written in lower case, propositional variables in

upper case

6.1.3 Writing an Inference Rule

• Assumptions are written above a horizontal line

• Conclusion written below the line.

• (For latex simplicity, I’ll just use the sequent notation to talk about inference rules)

• Inference rule only works in one direction

6.2 Definitions of True, Not and Equivalence

6.2.1 Primitives
• Only primitive objects are constant False, and ∨, ∧, ⇒

• All other objects are built using these

• Below we give definitions for some primitives

– can be verified using truth tables

– But we’ll only manipulate them using rules of logical deduction

6.2.2 Definition: True

• T rue = F alse ⇒ F alse

• (Verify using truth table, but this particular definition is “trivial” because that’s how
we defined implication)

6.2.3 Definition: Negation

• ¬a = a ⇒ F alse

• Proof (via truth tables):

– Suppose a is T rue, then right side becomes T rue → F alse, which is F alse

12
 – Suppose a is F alse, then right side becomes F alse → F alse, which we defined to
be T rue
– So in both cases the definition is consistent

• Consistency (via logical inference only):

– If a is T rue, then ¬T rue is T rue ⇒ F alse which is (F alse ⇒ F alse) ⇒ F alse

– We have a rule called Reduction Ad Absurdum which allows us to infer F alse from
(F alse ⇒ F alse) ⇒ F alse
– If a is F alse, then ¬F alse is $False ⇒ False$< which we defined to be T rue

6.2.4 Definition: Equivalence

• a ⇔ b = (a ⇒ b) ∧ (b ⇒ a)

• (Note that’s the definition of if and only if)

• Proof by Truth Table:

a b a⇒b b⇒a LHS RHS

T T T T T T
T F F T F F
F T T F F F
F F T T T T

• Hence, this formula is also correct

6.3 IMPORTANT: Note on Proofs of the Inference Rules

• I’ve “proven” each of the inference rules by making references to the Truth Table

• But it looks like that might be circular atleast based on the discussion in the section
on “Inferring the Operator Truth Tables”

• Assume that you have a system with FALSE, and three operators OR, AND, IMPLIES

• Pretend that it is just a coincidence that they match the names of Boolean operators

6.4 And Introduction

6.4.1 Definition: a, b ` a ∧ b
• Proof:

– We know that a and b are True

– By definition of ∧, a ∧ b is True

13
6.4.2 Usage Example 1
• Statement: (R ⇒ S), ¬P ` (R ⇒ S) ∧ ¬P

• Proof:

– Use and introduction

6.4.3 Usage Example 2

• Statement: P, Q, R ` (P ∧ Q) ∧ R

• Proof:

– If P and Q are True, P ∧ Q is True by and introduction

– Since, R is True, another usage of and introduction means (P ∧ Q) ∧ R is True as
well

• Note: This can be written in a tree diagram type form (Page 130)

6.4.4 Exercise 10:

• Statement: Prove P, Q, R ` P ∧ (Q ∧ R)

• Proof:

– If Q and R are True, Q ∧ R is true by and introduction

– If, P is true, another application of and introduction means P ∧ (Q ∧ R) is True as
well

• Note, we could have proven this simply by noting the associativity of ∧

6.4.5 Exercise 11
• Shape of x = A ∧ (B ∧ (C ∧ D))

– Right-heavy tree
– If there are 2n variables in x, then the height of the proof tree will also be 2n
– (At each level, we can reason about one of the variables in x)

• Shape of y = (A ∧ B) ∧ (C ∧ D))

– Balanced tree: height is logarithmic in number of variables

– (At each level, we can reason about half of the variables in y)

14
6.5 And Elimination
6.5.1 Definition: (a ∧ b) ` a, (a ∧ b) ` b
• Proof:

– We know that a ∧ b is True

– But a ∧ b is True exactly when both a and b are True
– So, if the conjunction is True, they must individually be True as well

6.5.2 Usage Example

• Statement: P, Q ∧ R ` P ∧ Q

• Proof:

– P is True by assumption
– Q ∧ R means Q is True by and elimination
– Since both P and Q are true, P ∧ Q is True (by and introduction)

6.5.3 Usage Example

• Statement: (P ∧ Q) ∧ R ` R ∧ Q

• Proof:

– By and elimination, both P ∧ Q and R are True

– Another and elimination on P ∧ Q yields True for P and Q
– So, both Q and R are True, by and introduction, Q ∧ R is True

6.5.4 Usage Example

• Statement: P ∧ Q ` Q ∧ P

• Proof:

– By and elimination, both P and Q are True

– By and introduction, Q ∧ P is True

• This proves the commutativity of And (which we could have also done using Truth
Tables)

15
6.5.5 Usage Example
• Statement: a ∧ (b ∧ c) ` (a ∧ b) ∧ c

• Proof:

– By and elimination on a ∧ (b ∧ c), both a and b ∧ c are True

– By and elimination on b ∧ c, both b and c are True
– By and introduction, a ∧ b is True
– By and introduction, (a ∧ b) ∧ c is True

• We’ve proved the associativity of and

6.5.6 Exercise 12
• Statement: Prove: (P ∧ Q) ∧ R ` P ∧ (Q ∧ R)

• Proof:

– We proved that “and” is associative for well formed propositions

– This is a well formed proposition

6.6 Imply Elimination

6.6.1 Definition: a, a ⇒ b ` b
• Also known as Modus Ponens

• Proof:

– Know a is True
– Know a ⇒ b is True
∗ If b is False, we get T rue ⇒ F alse which would be False
∗ But if b is True, we get T rue ⇒ T rue, which is True
– So, b is True

6.6.2 Usage Example

• Statement: Q ∧ P, P ∧ Q ⇒ R ` R

• Proof:

– Q ∧ P is True which means that P ∧ Q is true by the commutativity of and (proved

earlier)
– By Modus Ponens on P ∧ Q ⇒ R, we get R is True

16
6.6.3 Usage Example
• Statement: a, a ⇒ b, b ⇒ c ` c

• Proof:

– By Modus Ponens on a, a ⇒ b, we see that b is True

– By Modus Ponens on b, b ⇒ c, we see that c is True

6.6.4 Exercise 13
• Statement: P, P ⇒ Q, (P ∧ Q) ⇒ (R ∧ S) ` S

• Proof:

– By Modus Ponens on P, P ⇒ Q, we see that Q is True

– By And Introduction, P ∧ Q is True
– By Modus Ponens on P ∧ Q, P ∧ Q ⇒ R ∧ S, we see that R ∧ S is True
– By And Elimination on R ∧ S, we see that both R and S are True

6.6.5 Exercise 14
• Statement: P ⇒ Q, R ⇒ S, P ∧ R ` S ∧ R

• Proof:

– By And Elimination on P ∧ R, we see that P and R are True

– By Modes Ponens on R, R ⇒ S, we see that S is True
– By And Introduction on R and S, we see that S ∧ R is True

6.7 Imply Introduction

6.7.1 Definition: (a ` b) ` (a ⇒ b)
• “In order to imply the logical implication a ⇒ b, you must have a proof of b using a as
an assumption”

6.7.2 Example Usage

• Statement: ` (P ∧ Q) ⇒ Q

• Since the ` doesn’t have anything to the left, it is saying that this is a tautology

• What proof do we have for this?

– Consider the sequent P ∧ Q ` Q, which is just and elimination

– So, we have a proof for P ∧ Q ` Q, which means we can write ` P ∧ Q ⇒ Q by
Implies introduction

17
 • Assumptions Made

– And Elimination assumed P ∧ Q is True.

– We can’t infer Q without this assumption.
– So, this appears in the sequent to the left of the ` in P ∧ Q ` Q
– But the entire sequent P ∧ Q ` Q doesn’t have any further assumptions
∗ Whenever P ∧ Q is true, Q is true.
– So, this sequent can be put in as an assumption without adding any other assump-
tions
– This is why when we do implies introduction, we get ` (P ∧ Q) ⇒ Q

6.7.3 Discharged assumptions

• Assumptions that are made temporarily to establish a sequent, and which is then
thrown away

• E.g. to establish the sequent P ∧ Q ` Q, we made an assumption P ∧ Q

• Once we got P ∧ Q ` Q, we don’t need to assumption for any other purpose

• Indicate discharged assumptions by boxes

6.7.4 Implication Chain Rule

• Statement: a ⇒ b, b ⇒ c ` a ⇒ c

• Proof:

– Since we are trying to prove a ⇒ c, we must use Imply Introduction somewhere

– This means that we need to show a ` c
– Assume a is True, then by Modus Ponens on a ⇒ b, b is True
– Since b is True, by Modus Ponens on b ⇒ c, c is True
– So, we have a sequent a ` c, and we can discharge the assumption that a is True
– By Imply Introduction on a ` c, we get ac

6.7.5 Modus Tollens

• Statement: a ⇒ b, ¬b ` ¬a

• Proof:

– ¬a is an abbreviation for a ⇒ F alse

– This means that we need to use Imply Introduction
– Expand ¬a and ¬b to get: a ⇒ b, b ⇒ F alse ` a ⇒ F alse
– This is just implication chain rule

18
6.7.6 Exercise 15
• Statement: P ` Q ⇒ P ∧ Q

• Proof:

– We are trying to prove an implication Q ⇒ P ∧Q, means we have to Imply Introduction

– To introduce imply, we need to prove the sequent Q ` P ∧ Q
– We know that P is true from what is given
– Assume that Q is true.
– By and introduction, we have P, Q ` P ∧ Q
– This sequent does not have any further assumptions, so we can discharge the as-
sumption that B is True
– So, we have the sequent Q ` P ∧ Q, which means we have Q ⇒ P ∧ Q by
introduce imply

• (Maybe incorrect)

6.7.7 Exercise 16
• Statement: ` P ∧ Q ⇒ Q ∧ P

• Proof:

– We need to use introduce imply on the sequent P ∧ Q ` Q ∧ P

– But that’s just commutativity of ∧, which we proved earlier, the sequent holds
without any further assumptions
– So, we can introduce imply on the sequent to get the desired answer

6.8 Or Introduction
6.8.1 Definition: a ` a ∨ b and b ` a ∨ b
• Proof

– By the definition of `, if one of its arguments is true, then the whole expression
becomes true

6.8.2 Example Usage

• Statement: P ∧ Q ` P ∨ Q

• Proof

– By and elimination on P ∧ Q, we have P is True, Q is True

– By or introduction, we have P ∨ Q is True

19
6.8.3 Exercise 17
• Statement: P ⇒ F alse ∨ P

• Proof

– To prove this implication, we need to prove the sequent P ` F alse ∨ P

– But that sequent is True by or introduction
– So, by introduce implication, we have P ⇒ F alse ∨ P

6.8.4 Exercise 18
• Statement: P, Q ` (P ∧ Q) ∨ (Q ∨ R)

• Proof:

– By and introduction, we have P ∧ Q is True

– By or introduction, we have Q ∨ R is True
– By or introduction, (P ∧ Q) ∨ (Q ∨ R) is True
– (Actually, knowing that P ∧ Q is True is sufficient because we can or introduce that
to get True)

6.9 Or Elimination
6.9.1 Definition (((a ∨ b), (a ` c), (b ` c)) ` c)
• This is Proof by Case Analysis

• Proof:

– Assumption is that a ∨ b is True

– By the definition of ∨ (from Truth Table), atleast one of a and b must be True
– Suppose a is True
∗ We have a ` c, i.e. a proof that c is True when a holds
– Otherwise, b is True
∗ We have b ` c, i.e. a proof that c is True when a holds
– So, c is True whenever the assumptions hold

• (Did i just prove Proof by Case Analysis by DOING proof of case analysis? Something
is weird here. . . )

20
6.9.2 Example Usage
• Statement: (P ∧ Q) ∨ (P ∧ R) ` P

• Proof:

– Suppose P ∧ Q is True, then by and elimination, P is True

– Suppose instead that P ∧ R is True, then by and elimination, P is True
– So, we have ((P ∧ Q) ∨ (P ∧ R)), (P ∧ Q) ` P, (P ∧ R) ` P
– Use or elimination to get that P is True

6.9.3 Example Usage

• Statement: (a ∧ b) ∨ (a ∧ c) ` (b ∨ c)

• Proof:

– Assume a ∧ b by and elimination, we get a ∧ b ` b

– Discharge the assumption a ∧ b because we have a sequent
– By or introduction, we have b ` b ∨ c
– Assume a ∧ c by and elimination, we get a ∧ c ` c
– Discharge the assumption a ∧ c
– By or introduction, we have c ` b ∨ c
– We can no do or elimination on (a ∧ b) ∨ (a ∧ c) ` (b ∨ c), ((a ∧ b) ` b) ` b ∨ c, ((a ∧ c) `
c) ` b ∨ c
– This gives us b ∨ c

6.10 Identity
6.10.1 Definition a ` a
• It is necessary because we can’t prove P ` P without it.

6.10.2 Theorem: ` T rue

• By definition, T rue = F alse ⇒ F alse

• So, we need to introduce imply to prove this

• To introduce imply, we need to prove the sequent F alse ` F alse

• This is true by identity

• So, we have F alse ⇒ F alse or equivalently, T rue

21
6.11 Contradiction
6.11.1 Definition: F alse ` a
• Anything can be inferred if the assumption is False

• (Remember the Truth table for P ⇒ Q gave True when P was False?)

• Expresses the fact that False is untrue purely using logical inference

– Can’t just say False is not True, because we defined True that way
– Also doesn’t rely on us making “intuitionistic” statements like “False is Wrong”

6.11.2 Example Usage

• Statement: P, ¬P ` Q

• Proof:

– ¬P is shorthand for P ⇒ F alse

– So, from P, P ⇒ F alse by modus ponens, we have F alse
– F alse ` Q follows from contradiction

6.11.3 Example Usage

• Statement; a ∨ b, ¬a ` b

• Proof:

– ¬a is shorthand for a ⇒ F alse

– Performing or elimination on a ∨ b,
∗ Assume a is True, then we have a, a ⇒ F alse, which by Modus Ponens gives
us F alse
∗ So, by contradiction, b is True
∗ (Discharge the assumption that a is True)
∗ Assume b is True, the by identity, b is True
∗ (Discharge the assumption that b is True)
∗ Combining, we have (a ∨ b, ¬a), (((a ∨ b) ` a), a ⇒ F alse) ` b, (((a ∨ b) ` b) ` b),
we can now perform or elimination

6.12 Reduction Ad Absurdum

6.12.1 Definition: (¬a ` F alse) ` a
• If you can infer False from an assumption ¬a, then you must conclude that a is true

• This is the Proof by Contradiction strategy

22
6.12.2 Example Usage
• Statement: ¬¬a ` a

• Proof:

– Strategy: Assume ¬a and show that it leads us to infer False (proof by contradic-
tion)
– Assuming ¬a, we have ¬a, ¬¬a
– Expand out the abbreviations to get: a ⇒ F alse, (a ⇒ F alse) ⇒ F alse
– Since we have a ⇒ F alse and (a ⇒ F alse) ⇒ F alse, by Modus Ponens, we have
F alse
– So, by reductio ad absurdum, a Must be True if ¬¬a is True

6.13 Inferring the Operator Truth Tables

• Two fundamental questions:

– Are inference rules powerful enough?

– Will they ever allow us to make mistakes?

• profound and difficult questions, modern research has shown some astonishing results

• Easy version of first question: Are inference rules powerful enough to infer the truth
tables of logical operators?

6.13.1 Example: Deriving the Truth Table value of T rue ∧ F alse

• Want to show that it is logically equivalent to False, and not logically equivalent to True

• To show equivalence with False, we need to show: F alse ⇒ (T rue ∧ F alse), T rue ∧
F alse ⇒ F alse

– F alse ` (T rue ∧ F alse) follows from contradiction

– By introduce implication, we have F alse ⇒ (T rue ∧ F alse)
– T rue ∧ F alse ` F alse
∗ Expand out T rue the expression is now F alse ∧ (F alse ⇒ F alse)
∗ By and elimination, we get F alse ∧ (F alse ⇒ F alse) ` F alse, which by identity
gives us F alse ` F alse
∗ By introduce implication, we have T rue ∧ F alse ⇒ F alse

23
6.13.2 Exercise: Calculating value of T rue ∧ T rue
• Proof that T rue ∧ T rue ⇒ T rue

– By and elimination, we have T rue ∧ T rue ` T rue

– By identity, we have T rue ` T rue
– By introduce implication, we have T rue ∧ T rue ⇒ T rue

• Proof that T rue ⇒ (T rue ∧ T rue)

– Need to prove the sequent T rue ` T rue ∧ T rue

– By and introduction, we have T rue ` T rue ∧ T rue
– By introduce implication, we have T rue ⇒ T rue ∧ T rue

6.13.3 Exercise: Calculating the value of T rue ∨ F alse

• Proof that T rue ∨ F alse ⇒ T rue

– We have T rue ` T rue from identity

– We have F alse ` T rue from contradiction
– So, we have (T rue ∨ F alse), (T rue ` T rue), (F alse ` T rue)
– By or elimination, we have (T rue ∨ F alse) ` T rue
– By introduce implication, we have T rue ∨ F alse ⇒ T rue

• Proof that T rue ⇒ (T rue ∨ F alse)

– From or introduction, we have T rue ` (T rue ∨ F alse)

– By introduce implication, we have T rue ⇒ (T rue ∨ F alse)

6.13.4 Extra Credit: What happens when we try to say T rue ∨ F alse is F alse?
• Need to show equivalence

• Proof that F alse ⇒ (T rue ∨ F alse)

– F alse ` (T rue ∨ F alse) by contradiction

– By introduce implication, we have F alse ⇒ (T rue ∨ F alse)

• Proof that (T rue ∨ F alse) ⇒ F alse

– Need to do an or elimination
– F alse ` F alse comes from identity
– T rue ` F alse needs to be shown
– Expand out T rue: F alse ⇒ F alse ` F alse needs to be shown
– Since F alse is not T rue, we can’t use Modus Ponens
– We have no applicable inference rule to use

24
6.13.5 Exercise
• When we computed value of T rue ∧ F alse, we used and elimination to say that (F alse ⇒
F alse) ∧ F alse ` F alse

• But we could have said (F alse ⇒ F alse) ` F alse ` (F alse ⇒ F alse) which is True

• What would have happened then?

– We would have had to show that T rue ` F alse

– That is, F alse ⇒ F alse ` F alse
– Same situation as in extra credit: Have no valid inference law to use; so can’t
deduce.

7 Proof Checking by Computer

7.1 Proof Checkers and Theorem Provers
• Proof Checker: reads in a theorem and a proof. Determines if the proof is valid

• Theorem Prover: reads in a theorem and attempts to generate a proof

7.2 Example of Proof Checking

• See 06-proofs.hs

8 Boolean Algebra: Equational Reasoning

8.1 Introduction
8.1.1 Approaches to Propositional Logic
• Truth Tables: semantic approach

• Natural Deduction: syntactic approach

• Boolean Algebra: axiomatic approach

8.1.2 Equational Reasoning

• Show that two values are the same by building chains of equalities

– e.g. show that a = b and b = c, so a = c

• can substitute equals for equals

– if x = y, can substitute in x wherever y is used

25
8.2 Laws of Boolean Algebra
8.2.1 Operations with Constants
• a ∧ F alse = F alse

• a ∧ T rue = T rue

• Identity for or: a ∨ T rue = T rue

• Identity for and: a ∨ F alse = a

• Exercise 25: Simplification

– (P ∧ F alse) ∨ (Q ∧ T rue)
– F alse ∨ (Q ∧ T rue)
– Q ∧ T rue
– Q

• Exercise 26

– Prove that: (P ∧ F alse) ∧ T rue) = F alse

– Direct Proof:
∗ (P ∧ F alse) ∧ T rue)
∗ F alse ∧ T rue
∗ F alse

8.2.2 Basic Properties of ∧ and ∨

• Disjunctive Implication: a ⇒ a ∨ b (Or Introduction)

• Conjunctive Implication: a ∧ b ⇒ a (And Elimination)

• Idempotence: a ∧ a = a

• Idempotence: a ∧ a = a

• Commutative: a ∧ b = b ∧ a

• Commutative: a ∨ b = b ∨ a

• Associativity: (a ∧ b) ∧ c = a ∧ (b ∧ c)

• Associativity: (a ∨ b) ∨ c = a ∧ (b ∨ c)

• Exercise 27

26
 – Prove: (P ∧ ((Q ∨ R) ∨ Q)) ∧ S = S ∧ ((R ∨ Q) ∧ P )
– Direct Proof starting from LHS
– (P ∧ ((Q ∨ R) ∨ Q)) ∧ S
– (P ∧ ((R ∨ Q) ∨ Q)) ∧ S (by commutativity of or)
– (P ∧ (R ∨ (Q ∨ Q)) ∧ S (by associativity of or)
– (P ∧ (R ∨ Q)) ∧ S (by idempotence of or)
– ((R ∨ Q) ∧ P ) ∧ S (by commutivatiy of and)
– S ∧ ((R ∨ Q) ∧ P ) (by commutivatiy of and)

• Exercise 28

– Prove: P ∧ (Q ∧ (R ∧ S)) = ((P ∧ Q) ∧ R) ∧ S

– Direct proof starting from LHS
– P ∧ (Q ∧ (R ∧ S))
– ((P ∧ Q) ∧ (R ∧ S)) (by associativity of and)
– ((P ∧ Q) ∧ R) ∧ S (by associativity of and)

8.2.3 Distributive and De Morgan’s Laws

• And distributes over Or: a ∧ (b ∨ c) = (a ∧ b) ∨ (a ∧ c)

• Or distributes over And: a ∨ (b ∧ c) = (a ∨ b) ∧ (a ∨ c)

• De Morgan: ¬(a ∧ b) = ¬a ∨ ¬b

• De Morgan: ¬(a ∨ b) = ¬a ∧ ¬b

• Exercise 29

– Give an intuitive explanation of second De Morgan’s laws

– If none of (a, b) are true, it means that a is not true and b is not true

8.2.4 Laws on Negation

• ¬T rue = F alse

• ¬F alse = T rue

• a ∧ ¬a = F alse

• a ∨ ¬a = T rue

• ¬(¬a) = a

27
8.2.5 Laws on Implication
• Modus Ponens: a ∧ (a ⇒ b) ⇒ b

• Modus Tollens: ¬b ∧ (a ⇒ b) ⇒ ¬a

• Disjunctive Syllogism: (a ∨ b) ∧ ¬ ⇒ b

• Implication Chain: (a ⇒ b) ∧ (b ⇒ c) ⇒ (a ⇒ c)

• Implication Combination: (a ⇒ b) ∧ (c ⇒ d) ⇒ ((a ∧ c) ⇒ (b ∧ d))

• Currying: (a ∧ b) ⇒ c = a ⇒ b ⇒ c

• Implication: a ⇒ b = ¬a ∧ b

• Contrapositive: a ⇒ b = ¬b ⇒ ¬a

• Absurdity: (a ⇒ b) ∧ (a ⇒ ¬b) = ¬a

• Understanding the currying law

– Says there are two equivalent ways of establishing that a and b both hold
– One of them is to say a ∧ b
– The other is to have an implication on a and an implication on b
– If either of a or b is false,
∗ a ∧ b is false, so (a ∧ b) ⇒ c is vacuous
∗ Either a ∧ b or b ∧ c is vacuous (if a is False, then the first one is vacuous, if b
is false, the second one)

• Understanding the Implication Law

– allows us to prove a ⇒ b
– Note that we don’t have an “Introduce Implication” rule like in natural deduction
– So we use this rule

8.2.6 Logical Equivalence

• a ⇔ b is the same as (a ⇒ b) ∧ (b ⇒ a)

• Subtle difference between ⇔ and =

– = is an operator in the meta language that says that two sides of an equation have
the same value
– ⇔ is in the object language itself and it gives a new proposition

• Exercise 30

28
 – Prove: (A ∨ B) ∧ BB
– Strategy: We’ll reduce the left side of the equivalence to be the right side
– (A ∨ B) ∧ B
– (A ∨ B) ∧ (B ∨ F alse) (by Identity of OR)
– (B ∨ A) ∧ (B ∨ F alse) (commutativity of AND)
– B ∨ (A ∧ F alse) (distributivity of OR)
– B ∨ F alse
– B

• Exercise 31

– Prove: ((¬A ∧ B) ∨ (A ∧ ¬B)) ⇔ (A ∨ B) ∧ (¬(A ∧ B))

– Direct proof starting from “left side”
– (¬A ∧ B) ∨ (A ∧ ¬B)
– (¬A ∨ (A ∧ ¬B)) ∧ (B ∨ (A ∧ ¬B)) (by distributing the OR over the AND)
– ((¬A ∨ A) ∧ (¬A ∨ ¬B)) ∧ ((B ∨ A) ∧ (B ∨ ¬B)) (by distributing the OR over the AND)
– (T rue ∧ (¬A ∨ ¬B)) ∧ ((B ∨ A) ∧ T rue)
– (¬A ∨ ¬B) ∧ (B ∨ A)
– ¬(A ∧ B) ∧ (B ∨ A) (by Demorgan’s laws)

• Exercise 32

– Prove: ¬(A ∧ B) ≡ ¬A ∨ ¬B
– Direct proof starting from “right side”
– ¬A ∨ ¬B
– ¬(¬(¬A ∨ ¬B)) (Double negation is just identity)
– ¬(A ∧ B) (by De Morgan’s other law)

• Exercise 33

– Prove: (A ∨ B) ∧ (¬A ∨ C) ∧ (B ∨ C) ⇔ (A ∨ B) ∧ (¬A ∨ C)

– Proof is by cases on A
– Case: A is True
∗ LHS: (T rue ∨ B) ∧ ((¬T rue) ∨ C) ∧ (B ∨ C)
∗ i.e. T rue ∧ C ∧ (B ∨ C)
∗ i.e. C ∧ (B ∨ C)
∗ i.e. C (proven in Review Exercise 45)
∗ RHS: (T rue ∨ B) ∧ ((¬T rue) ∧ C)

29
 ∗ i.e. T rue ∧ C
∗ i.e. C
∗ So the equality holds in this case
– Case: A is False
∗ LHS: (F alse ∨ B) ∧ ((¬F alse) ∨ C) ∧ (B ∨ C)
∗ i.e. B ∧ T rue ∧ (B ∨ C)
∗ i.e. B ∧ (B ∨ C)
∗ i.e. B
∗ RHS: (F alse ∨ B) ∧ ((¬F alse) ∨ C)
∗ i.e. B ∧ T rue
∗ i.e. B
∗ So, the equality holds in this case
– Since the equality holds in both cases, the statement is proven.

9 Logic in Computer Science

• Talks about applications of theorem provers

10 Meta Logic
• Making statements about the logical system itself

• Fundamental Meta-Logic operators:

– P1 , P2 , · · · , Pn ` Q means there is a proof that shows Q from the assumptions

P1 , · · · , Pn
– P1 , P2 , · · · , Pn  Q means Q must be tree if all assumptions are true, but says
nothing about the proof

10.1 Definition: Consistency of a Logical System

• A formal system is consistent if, for all well-formed formulas a and b, if a ` b then a  b

• That is, each proposition that is provable is actually True

10.2 Definition: Completeness of a Logical System

• A formal system is complete if, for all well-formed formulas a and b, if a  b then a ` b

• That is, every proposition that is true can be proved using the system

10.3 Theorem: Propositional Logic is consistent and complete

• No proof given

30
10.4 Godel’s Theorem: A logical system powerful enough to express ordi-
nary arithmetic must be either inconsistent or incomplete
• It is impossible to capture all of mathematics in a safe logical system

11 Further Reading
• Forever Undecided: A Puzzle Guide to Godel by Raymond Smullyan

– “Don’t miss it”

• Godel, Escher, Bach: An Eternal Braid by Douglas Hofstadter

– “Another unmissable classic”

• How to Prove it: A Structured approach by Daniel Velleman

– Good source for hints on carrying out proofs with lots of example

• Computer-Aided Reasoning: An approach: by Matt Kaufmann, Panagiotis Mano-

lios, and J Strother Moore

– Describes the ACL2 theorem prover

• Introduction to HOL: by M Gordon and T Melham

– Describes the HOL theorem prover

• Logic for Mathematics and Computer Science by Burris

– “more advanced treatement of mathematical logic”

• A Mathematical Introduction to Logic by Enderton

– “standard presentation of logic from a mathematical perspective”

• Type Theorey and Functional Programming by Thompson

– “detailed development of relationship between inference rules and rules used to

define type systems”

• Logic and Declarative Language by Downward

• Proofs and Types by Girard, Lafont and Taylor

– relationship between inference proofs and type systems at a research level

31