Properties

of
Context-Free
languages
 Union

Context-free languages
are closed under: Union

L1 is context free
L1  L2
L2 is context free is context-free
 Example
Language Grammar

n n
L1  {a b } S1  aS1b | 

R
L2  {ww } S2  aS2 a | bS2b | 

Union
n n R S  S1 | S 2
L  {a b }  {ww }
In general:
For context-free languages L1, L2
with context-free grammars G1, G2
and start variables S1, S 2
It is assumed that variable sets of G1 and G2 are
disjoint

The grammar of the union L1  L2

has new start variable S
and additional production S  S1 | S 2
 Concatenation

Context-free languages
are closed under: Concatenation

L1 is context free
L1L2
L2 is context free is context-free
 Example
Language Grammar

n n
L1  {a b } S1  aS1b | 

R
L2  {ww } S2  aS2 a | bS2b | 

Concatenation
n n R S  S1S 2
L  {a b }{ww }
In general:

For context-free languages L1, L2

with context-free grammars G1, G2
and start variables S1, S 2

The grammar of the concatenation L1L2

has new start variable S
and additional production S  S1S 2
 Star Operation

Context-free languages
are closed under: Star-operation

L is context free *
L is context-free
 Example

Language Grammar

n n
L  {a b } S  aSb | 

Star Operation

n n
L  {a b } * S1  SS1 | 
In general:

For context-free language L

with context-free grammar G
and start variable S

The grammar of the star operation L *

has new start variable S1
and additional production S1  SS1 | 
Negative Properties
of
Context-Free
Languages
 Intersection

Context-free languages
are not closed under: intersection

L1 is context free
L1  L2
L2 is context free not necessarily
context-free
 Example
n n m n m m
L1  {a b c } L2  {a b c }
Context-free: Context-free:
S  AC S  AB
A  aAb |  A  aA | 
C  cC |  B  bBc | 
Intersection
n n n
L1  L2  {a b c } NOT context-free
 Complement

Context-free languages
are not closed under: complement

L is context free L not necessarily

context-free
 Example
n n m n m m
L1  {a b c } L2  {a b c }
Context-free: Context-free:
S  AC S  AB
A  aAb |  A  aA | 
C  cC |  B  bBc | 
Complement
n n n
L1  L2  L1  L2  {a b c }
NOT context-free
 Context-free
languages
and
Regular Languages
 The intersection of
a context-free language and
a regular language
is a context-free language

L1 context free
L1  L2
L2 regular context-free
Machine M1 Machine M2
NPDA for L1 DFA for L2
context-free regular

Construct a new NPDA machine M

that accepts L1  L2

M simulates in parallel M1 and M 2

 NPDA M1 DFA M2
a, b  c a
q1 q2 p1 p2
transition transition

NPDA M

q1, p1 a , b  c
q2 , p2
transition
NPDA M1 DFA M2

q0 p0
initial state initial state

NPDA M

q0 , p0
Initial state
Therefore: L( M1)  L( M 2 ) is context-free

(since M is NPDA)

L1  L2 is context-free
The Pumping Lemma
for
Context-Free
Languages
Take an infinite context-free language

Generates an infinite number

of different strings

Example: S  AB
A  aBb
B  Sb
Bb
 S  AB
A  aBb
B  Sb
Bb

A derivation: Variables are repeated

S  AB  aBbB  abbB 
 abbSb  abbABb  abbaBbBb 
 abbabbBb  abbabbbb
Derivation tree string abbabbbb
S

A B

a B b S b

b A B

a B b b

b
Derivation tree string abbabbbb
S

A B

a B b S b

b A B

a B b b

repeated b
B  Sb  ABb  B

 aBbBb  aBbbb S b

A B

a B b b

b
Bb
Repeated Part B

S b

A B

a B b b

B    aBbbb
Another possible derivation B

S b

A B
B    aBbbb a B b b

S b

A B

a B b b

B    aBbbb  aaBbbbbbb
 S

A B

a B b S b

b A B
B    aBbbb
a B b b

S    abbaBbbb
 S

A B

a B b S b

b A B
B    aBbbb
a B b b

S b

A B

a B b b

S    abbaBbbb    abbaaBbbbbbb
 S

A B

a B b S b

b A B

a B b b

S b

A B

a B b b Bb
b

S    abbaaBbbbbbb  abbaabbbbbbb
 S    abbaabbbbbbb

Therefore, the string

abbaabbbbbbb

is also generated by the grammar

We know: Bb
B    aBbbb
S    abbaBbbb
We also know this string is generated:

S    abbaBbbb 
 abbaabbbb
We know: Bb
B    aBbbb
S    abbaBbbb
Therefore, this string is also generated:

S    abbaBbbb 
 abbaaBbbbbbb 
 abbaabbbbbbb
We know: Bb
B    aBbbb
S    abbaBbbb
Therefore, this string is also generated:
S    abbaBbbb 
 abba(a ) B (bbb)bbb
2 2
 abba(a ) B (bbb) bbb
2 2
 abba(a ) b(bbb) bbb
We know: Bb
B    aBbbb
S    abbaBbbb
Therefore, this string is also generated:
S    abbaBbbb 

i i
 abba(a ) B (bbb) bbb
i i
 abba(a ) b(bbb) bbb
Therefore, knowing that

abbabbbb

is generated by grammar G,
we also know that
i i
abba(a ) b(bbb) bbb

is generated by G
In general:

We are given an infinite

context-free grammar G

Assume G has no unit-productions

no  -productions
Take a string w L(G )
with length bigger than

m > (Number of productions) x

(Largest right side of a production)

Consequence:
Some variable must be repeated
in the derivation of w
 u , v, x, y, z : strings of terminals
S
String w  uvxyz

u z
Last repeated variable A

v y
repeated A

x
 S
Possible
derivations: u z
 A
S  uAz
 v y
A  vAy A

A x
x
We know:
  
S  uAz A  vAy A x

This string is also generated:

 *
S  uAz  uxz

0 0
uv xy z
We know:
  
S  uAz A  vAy A x

This string is also generated:

 * *
S  uAz  uvAyz  uvxyz

1 1
The original w  uv xy z
We know:
  
S  uAz A  vAy A x

This string is also generated:

 * * *
S  uAz  uvAyz  uvvAyyz  uvvxyyz

2 2
uv xy z
We know:
  
S  uAz A  vAy A x

This string is also generated:

 * *
S  uAz  uvAyz  uvvAyyz 
* *
 uvvvAyyyz  uvvvxyyyz
3 3
uv xy z
We know:
  
S  uAz A  vAy A x
This string is also generated:
* uAz 
S * uvAyz 
* uvvAyyz *
* uvvvAyyyz 
* 
* uvvvvAy  yyyz *

* uvvvvxy  yyyz


i i
uv xy z
Therefore, any string of the form

i i
uv xy z i0

is generated by the grammar G

Therefore,

knowing that uvxyz  L(G )

i i
we also know that uv xy z  L(G )
 S

u z
A

v y
A

x
Observation: | vxy |  m
Since A is the last repeated variable
 S

u z
A

v y
A

x
Observation: | vy |  1
Since there are no unit or  productions
The Pumping Lemma:
For infinite context-free language L
there exists an integer m such that

for any string w  L, | w | m

we can write w  uvxyz
with lengths | vxy | m and | vy | 1
and it must be:
i i
uv xy z  L, for all i  0
Applications
of
The Pumping
Lemma
Non-context free languages
n n n
{a b c : n  0}

Context-free languages

n n
{a b : n  0}
Theorem: The language
n n n
L  {a b c : n  0}
is not context free

Proof: Use the Pumping Lemma

for context-free languages
 n n n
L  {a b c : n  0}

Assume for contradiction that L

is context-free

Since L is context-free and infinite

we can apply the pumping lemma
 n n n
L  {a b c : n  0}

Pumping Lemma gives a magic number m

such that:

Pick any string w L with length | w | m

m m m
We pick: wa b c
 n n n
L  {a b c : n  0}

m m m
wa b c

We can write: w  uvxyz

with lengths | vxy | m and | vy | 1

 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Pumping Lemma says:

i i
uv xy z  L for all i0
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

We examine all the possible locations

of string vxy in w
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

m
Case 1: vxy is within a

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 1: v and y consist from only a

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 1: Repeating v and y

k 1
mk m m
aaaaaa...aaaaaa bbb...bbb ccc...ccc
u 2 2 z
v xy
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

2 2
Case 1: From Pumping Lemma: uv xy z  L
k 1
mk m m
aaaaaa...aaaaaa bbb...bbb ccc...ccc
u 2 2 z
v xy
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

2 2
Case 1: From Pumping Lemma: uv xy z  L
k 1
2 2 m k m m
However: uv xy z  a b c L
Contradiction!!!
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

m
Case 2: vxy is within b

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 2: Similar analysis with case 1

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

m
Case 3: vxy is within c

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 3: Similar analysis with case 1

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

m m
Case 4: vxy overlaps a and b

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 4: Possibility 1: v contains only a

y contains only b
m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 4: Possibility 1: v contains only a

k1  k 2  1 y contains only b
m  k1 m  k2 m
aaa...aaaaaaa bbbbbbb...bbb ccc...ccc
u 2 2 z
v xy
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

2 2
Case 4: From Pumping Lemma: uv xy z  L
k1  k 2  1
m  k1 m  k2 m
aaa...aaaaaaa bbbbbbb...bbb ccc...ccc
u 2 2 z
v xy
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

2 2
Case 4: From Pumping Lemma: uv xy z  L
k1  k 2  1

2 2 m  k1 m k2 m
However: uv xy z  a b c L
Contradiction!!!
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 4: Possibility 2: v contains a and b

y contains only b
m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 4: Possibility 2: v contains a and b

k1  k 2  k  1 y contains only b
m k1 k 2 mk m
aaa...aaaaabbaabb bbbbbbb...bbb ccc...ccc
u 2
v xy 2 z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

2 2
Case 4: From Pumping Lemma: uv xy z  L
k1  k 2  k  1
m k1 k 2 mk m
aaa...aaaaabbaabb bbbbbbb...bbb ccc...ccc
u 2
v xy 2 z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

2 2
Case 4: From Pumping Lemma: uv xy z  L

However: k1  k 2  k  1
2 2 m k1 k 2 m  k m
uv xy z a b a b c L
Contradiction!!!
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 4: Possibility 3: v contains only a

y contains a and b
m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 4: Possibility 3: v contains only a

y contains a and b

Similar analysis with Possibility 2

 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

m m
Case 5: vxy overlaps b and c

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
 n n n
L  {a b c : n  0}
m m m
wa b c
w  uvxyz | vxy | m | vy | 1

Case 5: Similar analysis with case 4

m m m
aaa...aaa bbb...bbb ccc...ccc
u vxy z
There are no other cases to consider

(since | vxy | m , string vxy cannot

m m m
overlap a , b and c at the same time)
In all cases we obtained a contradiction

Therefore: The original assumption that

n n n
L  {a b c : n  0}
is context-free must be wrong

Conclusion: L is not context-free