Unit-I Proposional Logic & First Order Logic

October 25, 2015

Unit-I Proposional Logic & First Order Logic

Satisfiability, Validity and Consequence

Definition
Let A F.
A is satisfiable iff vI (A) = T for some interpretation I. A
satisfying interpretation is a model for A.
A is valid, denoted |= A, iff vI = T for all interpretations I. A
valid propositional formula is also called a tautology.
A is unsatisfiable iff it is not satisfiable, that is, if vI (A) = T
for all interpretations I.
A is falsifiable, denoted 2A, iff it is not valid, that is, if vI (A)
= F for some interpretation v.

Unit-I Proposional Logic & First Order Logic

Satisfiability and Validity of Formulas

Figure: Satisfiability and validity of formulas

Unit-I Proposional Logic & First Order Logic

Satisfiability and Validity of Formulas

Theorem
Let A F . A is valid if and only if k A is unsatisfiable. A is
satisfiable if and only if k A is falsifiable

Unit-I Proposional Logic & First Order Logic

Proof.
Let I be an arbitrary interpretation. vI (A) = T if and only if vI
(kA) = F by the definition of the truth value of a negation.
Since I was arbitrary, A is true in all interpretations if and only if k
A is false in all interpretations, that is, iff kA is unsatisfiable.
If A is satisfiable then for some interpretation I, vI (A) = T.
By definition of the truth value of a negation, vI (kA) = F so that
k A is falsifiable.
Conversely, if vI (kA) = F then vI (A) = T.

Unit-I Proposional Logic & First Order Logic

Decision Procedures in Propositional Logic

Definition (Decision Procedure)

Let U F be a set of formulas. An algorithm is a decision
procedure for U if given an arbitrary formula A F, it terminates
and returns the answer yes if A U and the answer no if A 6 U .
If U is the set of satisfiable formulas, a decision procedure for U is
called a decision procedure for satisfiability, and similarly for
validity.

Unit-I Proposional Logic & First Order Logic

Note
1

To decide if A is valid, apply the decision procedure for

satisfiability to k A

If it reports that k A is satisfiable, then A is not valid; if it

reports that k A is not satisfiable, then A is valid. Such an
decision procedure is called a refutation procedure, because
we prove the validity of a formula by refuting its negation

Refutation procedures can be efficient algorithms for deciding

validity, because instead of checking that the formula is always
true, we need only search for a falsifying counterexample

The existence of a decision procedure for satisfiability in

propositional logic is trivial, because we can build a truth
table for any formula

Unit-I Proposional Logic & First Order Logic

Example for Satisfiability

p
T
Truth table for the formula pq T
F
F

Unit-I Proposional Logic & First Order Logic

q
T
F
T
F

pq
T
F
T
T

Example for Validity and Unsatisfiable

Validity of (p q)(kq k p)
Proof Refer BB
Prove that (p q) k p k q is unsatisfiable.
Proof Refer BB

Unit-I Proposional Logic & First Order Logic

Satisfiability of Set of Formulas

Definition
A set of formulas U={A1 ,...}s (simultaneously) satisfiable iff there
exists an interpretation I such that vI (Ai ) = T for all i. The
satisfying interpretation is a model of U . U is unsatisfiable iff for
every interpretation I , there exists an i such that vI (Ai ) = F
Example:
The set U1 = {p, kp q, q r} is simultaneously satisfiable by
the interpretation which assigns T to each atom, while the set
U2 = {p, kp q, k p} is unsatisfiable. Each formula in U2 is
satisfiable by itself, but the set is not simultaneously satisfiable

Unit-I Proposional Logic & First Order Logic

Logical Consequence
Definition (Logical Consequence)
Let U be a set of formulas and A a formula. A is a logical
consequence of U , denoted U  A, iff every model of U is a model
of A
Example:
Let A = (pr) (kq kr). Then A is a logical consequence
of {p, k q}, denoted {p, kq} A, since A is true in all
interpretations I such that I(p) = T and I(q) = F . However,
A is not valid, since it is not true in the interpretation I0 where
I0 (p) = F , I0 (q)=T , I0 (r) = T

Unit-I Proposional Logic & First Order Logic

Theorem

Theorem
U  A if and only if 

Ai A.

Example:
{p, k q}  (p r) (k q k r), so by Theorem,  (p kq)
(p r) (k q k r)

Unit-I Proposional Logic & First Order Logic

Semantic Tableaux

The method of semantic tableaux is an efficient decision

procedure for satisfiability (and by duality validity) in
propositional logic
The principle behind semantic tableaux is very simple: search
for a model (satisfying interpretation) by decomposing the
formula into sets of atoms and negations of atoms
It is easy to check if there is an interpretation for each set: a
set of atoms and negations of atoms is satisfiable iff the set
does not contain an atom p and its negation kp
The formula is satisfiable iff one of these sets is satisfiable

Unit-I Proposional Logic & First Order Logic

Decomposing Formulas into Sets of Literals

Definition (Literals)
A literal is an atom or the negation of an atom. An atom is a
positive literal and the negation of an atom is a negative literal.
For any atom p, {p,kp} is a complementary pair of literals
For any formula A, {A,kA} is a complementary pair of
formulas. A is the complement of kA and kA is the
complement of A

Unit-I Proposional Logic & First Order Logic

Problems

Analyze satisfiability of the formula:A = p (kq kp).

Solution Refer BB

Prove that B = (pq) (kp kq)is unsatisfiable

Theorem
A set of literals is satisfiable if and only if it does not contain a
complementary pair of literals - Proof: Refer BB

Unit-I Proposional Logic & First Order Logic

SAT Solvers
Definition (SAT Solver)
A computer program that searches for a model for a propositional
formula is called a SAT Solver
Properties of Clausal Form
Definition
Let S, S be sets of clauses. S S denotes that S is satisfiable if
and only if S is satisfiable
It is important to understand that S S (S is satisfiable if
and only if S is satisfiable) does not imply that S S (S is
logically equivalent to S)
Unit-I Proposional Logic & First Order Logic

Pure Literals and Renaming

Definition (Pure Literals)

Let S be a set of clauses. A Pure literal in S is a literal l that
appears in atleast one clause of S, but its complement lc does not
appear in any clause of S
Definition (Renaming)
Let S be a set of clauses and U a set of atomic propositions RU (S)
the renaming of S by U is obtained from S by replacing each literal
l on an atomic proposition in U by lc

Unit-I Proposional Logic & First Order Logic

Example

If S={pqr, pq, qr ,r}, find Rp,q (S)

Unit-I Proposional Logic & First Order Logic

Example

If S={pqr, pq, qr ,r}, find Rp,q (S)

Solution: Rp,q (S)={
p q,p
q ,qr ,r},

Unit-I Proposional Logic & First Order Logic

Davis Putnam Algorithm

Input : A formula A in Clausal Form
Output: Report A is Satisfiable or Unsatisfiable
Perform the following rules repeatedly but the third rule is used
only if the first two do not apply.
1 Unit Literal rule: If there is a unit clause (l), delete all clauses
containing l and delete all occurrences of lc from other clauses
2

Pure Literal rule:If there is a pure literal l, delete all clauses

containg l

Eliminate a variable by resolution:Choose an atom p and

perform all possible resolutions on clauses that clash on p and
p. Add these resolvents to the set of clauses and then delete
all clauses containing p or p
Algorithm 1: Davis Putnam Algorithm

Unit-I Proposional Logic & First Order Logic

Termination condition for the Algorithm

If empty clause  is produced, report the formula is

unsatisfiable

If no more rules are applicable, report that formula is

satisfiable

Unit-I Proposional Logic & First Order Logic

Example
Consider the set of clauses {p, pq,
q r,r st}. Apply DP
algorithm.

Unit-I Proposional Logic & First Order Logic

Example
Consider the set of clauses {p, pq,
q r,r st}. Apply DP
algorithm.
Step(i) According to step(i) of D.P algorithm, p is a unit
literal. Hence it must be deleted in all clauses containing p
and delete all occurrences of pc from other clauses. the
given set becomes {q,
q r,r st}

Unit-I Proposional Logic & First Order Logic

Example
Consider the set of clauses {p, pq,
q r,r st}. Apply DP
algorithm.
Step(i) According to step(i) of D.P algorithm, p is a unit
literal. Hence it must be deleted in all clauses containing p
and delete all occurrences of pc from other clauses. the
given set becomes {q,
q r,r st}
Step(ii) Like the previous step, remove q and qc . the set
becomes {r,r st}

Unit-I Proposional Logic & First Order Logic

Example
Consider the set of clauses {p, pq,
q r,r st}. Apply DP
algorithm.
Step(i) According to step(i) of D.P algorithm, p is a unit
literal. Hence it must be deleted in all clauses containing p
and delete all occurrences of pc from other clauses. the
given set becomes {q,
q r,r st}
Step(ii) Like the previous step, remove q and qc . the set
becomes {r,r st}
Step(iii) Like the previous step, remove r and its complement.
the set becomes {st}

Unit-I Proposional Logic & First Order Logic

Example
Consider the set of clauses {p, pq,
q r,r st}. Apply DP
algorithm.
Step(i) According to step(i) of D.P algorithm, p is a unit
literal. Hence it must be deleted in all clauses containing p
and delete all occurrences of pc from other clauses. the
given set becomes {q,
q r,r st}
Step(ii) Like the previous step, remove q and qc . the set
becomes {r,r st}
Step(iii) Like the previous step, remove r and its complement.
the set becomes {st}
No more rules are applicable for the set {st}, the set of
clauses is satisfiable
Unit-I Proposional Logic & First Order Logic

Hilbert System (H)

Gentzen System vs Hilbert System

Gentzen
Hilbert
One Axiom
Several Axioms
Many Inference Rules Only one rule of inference
Notations used
A,B,C denote arbitrary formulas in propositional logic
`AA means AA can be proved

Unit-I Proposional Logic & First Order Logic

Hilbert System(H)
Definition (Axioms of H & Rules of Inference)
1

`(A(BA))

`(A(BC))((AB)(AC))

`(B A)(AB)

Rule of Inference: (Modus Ponens)(MP)

Theorem
`A A
Proof.
Refer BB
Unit-I Proposional Logic & First Order Logic

`A,`AB
`B

Assignment Due Date:27.10.15

Prove `(AB)(B A)

Prove `(A A) A in H

Unit-I Proposional Logic & First Order Logic

First Order Logic: Deductive Systems

Definition (First Order Logic)
An extension of propositional logic that includes predicates
interpreted as relations on domains
Notations used
P,A,V denotes the countable sets of predicate symbols,
constant symbols and variables
pn P n-ary predicate
n=1 or 2 unary or binary
Universal Quantifier
Existential Quantifier

Unit-I Proposional Logic & First Order Logic

Example
x

p(x,y)

p(y,x)

Unit-I Proposional Logic & First Order Logic

First Order Logic:Deductive System

Gentzen System:
Deductive System
One Axiom
4 rules of Inference

Two Special rules: and

xA(x)
xA(x)

(a)
A(a)
A(a)

U{,(a)}
U{}
U{(a)}
U{}

Unit-I Proposional Logic & First Order Logic

xA(x)
xA(x)

(a)
A(a)
A(a)

Hilbert System
Definition (Axioms of H for the First Order Logic)
1

Apart from the three axioms of H for Propositional Logic we

have:
1
2

` xA(x)A(a)
` x(AB(x))(A xB(x))

Rules of Inference(Modus Ponens and Generalization)

`AB,`A
`B
`A(a)
`xA(x)

C-Rule:
(i) U ` xA(x) (Existential Quantifier)
(i+1) U ` A(a) (C rule)

Deduction Rule:
U{A}`B
U`AB
Unit-I Proposional Logic & First Order Logic

Theorems
Theorem
`A(a) xA(x)
Proof.
Refer BB
Theorem
` x A(x) x A(x)
Proof.
Refer BB

Unit-I Proposional Logic & First Order Logic

Skolems Algorithm

An algorithm to transform a formula A into a formula Ain

clausal form x(p(x) q(x)) (xp(x) xq(x))

Unit-I Proposional Logic & First Order Logic

Skolems Algorithm I
Input : A closed formula A of first-order logic
Output: A formula Ain clausal form such that AA
1 Rename bound variables so that no variable appears in two
quantifiers
x(p(x) q(x)) (yp(y) zq(z))
2

Eliminate all binary Boolean operators other than and

x( p(x) q(x)) yp(y) zq(z)

Push negation operators inward, collapsing double negation,

until they apply to atomic formulas only. Use the equivalences:
xA(x) x A(x), xA(x) x A(x)

Unit-I Proposional Logic & First Order Logic

Skolems Algorithm Contd..

4

The given formula is transformed to:

x(p(x) q(x)) y p(y) zq(z)

Extract quantifiers from the matrix. Choose an outermost

quantifier, that is, a quantifier in the matrix that is not within
the scope of another quantifier still in the matrix. Extract the
quantifier using the following equivalences, where Q is a
quantifier and op is either or:
A op QxB(x) Qx(A op B(x)), QxA(x) op B Qx(A(x) op
B)
Repeat until all quantifiers appear in the prefix and the matrix
is quantifier-free. The equivalences are applicable because
since no variable appears in two quantifiers.

Unit-I Proposional Logic & First Order Logic

Skolems Algorithm Contd..

6

In the example, no quantifier appears within the scope of

another, so we can extract them in any order, for example, x,
y, z: xyz((p(x) q(x)) p(y) q(z))

Use the distributive laws to transform the matrix into CNF.

The formula is now in PCNF
xyz((p(x) p(y) q(z)) ( q(x) p(y) q(z)))

For every existential quantifier x in A, let y1 , . . . , yn be

the universally quantified variables preceding x and let f be a
new n-ary function symbol. Delete x and replace every
occurrence of x by f(y1 , . . . , yn ). If there are no universal
quantifiers preceding x, replace x by a new constant (0-ary
function). These new function symbols are Skolem functions
and the process of replacing existential quantifiers by
functions is Skolemization

Unit-I Proposional Logic & First Order Logic

Skolem Algorithm Contd..

For Example, z((p(a) p(b) q(z)) ( q(a) p(b)

q(z))), where a and b are the Skolem functions (constants)
corresponding to the existentially quantified variables x and y,
respectively

10

The formula can be written in clausal form by dropping the

(universal) quantifiers and writing the matrix as sets of
clauses:
{{p(a), p(b), q(z)}, { q(a), p(b), q(z)}}
Algorithm 2: Skolems Algorithm

Unit-I Proposional Logic & First Order Logic

Skolems Algorithm Example

Step
Original formula
Rename bound variables
Eliminate Boolean operators
Push negation inwards
Extract quantifiers
Distribute matrix
Replace existential quantifiers
Write in clausal form

Unit-I Proposional Logic & First Order Logic

Transformation
xyp(x,y) yxp(x,y)
xyp(x,y) wzp(z,w)
xyp(x,y) wzp(z,w)
xy p(x,y) wzp(z, w)
xywz( p(x,y) p(z,w))
(no change)
xw( p(x,f(x)) p(g(x,w), w))
{{ p(x,f(x)), p(g(x,w), w)}}

Skolems Theorem

Theorem
Let A be a closed formula. Then there exists a formula A in
clausal form such that A A

Unit-I Proposional Logic & First Order Logic