Context-Free Grammar

CFGs are more powerful than regular

expressions. They are more powerful in the sense
that whatever can be expressed using regular
expressions can be expressed using context-free
grammars, but they can also express languages
that do not have regular expressions.
Context-Free Grammar
 Context-Free Grammar
Example:
 Context-Free Grammar
Example
 Context-Free Grammar
Construct a context free grammar for palindromes
containing 0s and 1s.
 Context-Free Grammar
Construct a context free grammar for all integers (with
sign)

G ={V, , R, S}
V = {S, G, D, I}, = {0, 1, 2, 3,., 9, +, -}
Derivation Trees
Derivation Trees
Derivation Trees
Derivation Trees
Derivation Tree
 Ambiguity
Consider the grammar for arithmetic expressions involving
addition and multiplication operators:

Consider the sentence ID+IDID.

This can be parsed in two different ways.
 Ambiguity

In the figure, the parse tree to the left gives the addition
operator precedence over multiplication. In other words, an
expression such as 3+5 9 is evaluated as (3+5) 9 with a
result of 72. Whereas, the tree to the right, does what is
considered the standard practice in programming
languages, i.e. giving precedence over +. The previous
example would be evaluated as 3+(5 9) resulting in 48.

A context free grammar G is ambiguous if there exists

some w L(G), which is ambiguous.

If G is the grammar
Show that G is ambiguous.
 Ambiguity
To prove that G is ambiguous, we need to find a w L(G),
which is ambiguous . Consider the word abababa.
Ambiguity
 Removing Ambiguity

What about a + a a ?
Removing Ambiguity
 Simplification of CFG

If G is a CFG , then we can construct a simplified

equivalent CFG G with the help of following steps:
1. Eliminate all null productions to get G1.
2. Eliminate all unit productions in G1 to get G2.
3. Construct a reduced Grammar G from G2.
 Simplification of CFG

1.Elimination of null productions

Let G1=(V, ,P,S) be the GFG having NO null

productions.

Consider the Grammar G whose productions are given

below. Construct a Grammar G1 without null productions
generating L(G)-{}
 Simplification of CFG

Step 1: Construction of the set W of all nullable variables

W1={A1 V | A1 is a production in P}={A,B}
Wi+1= Wi {K V|there exists a production K with
Wi* }
W2={A,B} {S} as S AB is a production with AB W1*
= { S,A,B}
W3= W2 = W2
Erasing from RHS
Construction of P: D b, S aS, S AB , S a, S A,
S B.
 Simplification of CFG
2.Elimination of unit productions

A unit production in a context free grammar G is a

production of the form A B where A and B are
variables(non terminals) in G.

For such a variable A the following steps have to be

applied:
Step1: Construction of the set of variables derivable
from A.
W0(A)={A}
Wi+1(A)=Wi(A) {B V| C B is in P with C Wi(A) }
 Simplification of CFG

Step2: Construction of A productions

The A-productions in G2 are either
(a) The non unit production in G1 or
(b) A whenever B is in G1 with B W(A) and
V.

Now we can define G2, where P2 is constructed using

step2 for every A V
 Simplification of CFG
Example:
Let the productions in G1 be
S AB, A a, B C|b, C D, D E and E a

W(B)={B,C,D,E}, W(C)={C,D,E}, W(D)={D,E}

The productions P2 in G2 are:

S AB, A a, B a|b, C a, D a and E a

 Simplification of CFG

3.Construction of reduced Grammars:

Many productions in P may not be useful for the purpose

of derivation.
It would be better to eliminate (1) variables that do not
derive any terminal string and (2) symbols that are not
reachable from the start symbol.

If G is a CFG, then we can find an equivalent grammar

G such that each variable in G derives some terminal
string.

Let G2=(V, ,P,S). We can define G =(V, ,P,S) as

follows.
 Simplification of CFG
a) Construction of V:
We define Wi V by recursion.
W1={A V| there exists a production A where *}
Wi+1=Wi {A V| there exists some production A with
( Wi)*}
At some point Wk=Wk+1. Then we get V=Wk.
b) Construction of P.
P={A |A, (V )*}
 Simplification of CFG
Example: Let G be
S AB, A a, B b, B C, and E c
Find G such that every variable in G derives some
terminal string.

Here we get V={S,A,B,E}.

So P= S AB, A a, B b and E c
 Simplification of CFG
If G=(V, ,P,S), we can construct an equivalent
grammar G= (V, ,P,S) such that every symbol in
V is derivable from S.

a) Construction of Wi
W1={S}.
Wi+1=Wi {X (V ) | there exists a production
A with A Wi and containing the symbol X }
At some point Wk=Wk+1.
b) Construction of V, , P.
V=V Wk , = Wk , P={A | A Wk }
 Simplification of CFG

Example: Consider G=({S,A,B,E}, {a,b,c},P,S), where

P consists of S AB, A a, B b and E c.

W3={S,A,B} {a,b}
V= {S,A,B} , ={a,b}
P= S AB, A a, B b
 Simplification of CFG

Find reduced grammars equivalent to the following

1) S AB|CA, B BC|AB, A a, C aB|b .
2) S aAa, A Sb|bCC|DaA, C abb|DD, E aC,
D aDA .

Answers:
1) S CA, A a, C b
2) S aAa, A Sb|bCC, C abb
 Normal Forms for CFG

In a context free grammar, the R.H.S of a production

can be any string of variables and terminals. When the
production in G satisfy certain restrictions, then G is said
to be in a normal form.

1. Chomsky Normal Form(CNF)

In CNF we have restrictions on the length of R.H.S and

the nature of symbols in the R.H.S
A CFG G is said to be in CNF, if every production is of
the form A a, or A BC, and S . When is in L(G)
we assume that S does not appear on the R.H.S of any
production.
 Normal Forms for CFG

Example: Consider a CFG whose productions are

S AB|, A a, B b
Is it in CNF? YES

Reduction to CNF

Consider an example. Let G be

S ABC|aC, A a, B b, C c.
Except S ABC|aC , all other productions are in the
form required for CNF. The terminal a in aC can be
replaced by a new variable D. By adding a new
production D a. So S aC becomes S DC.
 Normal Forms for CFG
S ABC is not in the required form. So it can be replaced
by two productions: S AE and EBC.
It is important to note that unit productions and null
productions are to be removed before substitutions.

Elimination of terminals on R.H.S: If there is a terminal ai

on the R.H.S of a production, then add a new variable
(non terminal) Cai ai

Restricting the number of variables on R.H.S: Consider

productions of the form A A1A2.Am . We introduce new
productions A A1C1, C1 A2C2, , Cm-2 Am-1Am
 Normal Forms for CFG
Example1: Reduce the following grammar G to CNF:
SaAD, A aB|bAB, B b, D d.

Answer: SCaC1, A CaB|CbC2, C1 AD,

C2 AB, B b, D d, Ca a, Cb b.

Example2: Reduce the following grammar G to CNF:

SaAbB, A aA|a, B bB|b.

Answer: SCaC1, C1 AC2, C2 CbB, A CaA,

B CbB, Ca a, Cb b, A a, B b.
 Normal Forms for CFG

Greibach Normal Form(GNF):

A grammar is in GNF if every production is of the form

A a , where V*. And a .
S is in G, when is in L(G) and we assume that S
does not appear on the R.H.S of any production.

Example: S aAB|, A bC, B b, C c is in GNF.

 Normal Forms for CFG
Lemma1: Let G be a CFG, and A B be a production in
P. Assume that P also has the following productions:
B 1|2|.|s
We can define P1=(P - {A B}) {A i|1 i s }
Then G1 having P1 is also a CFG equivalent to G.

Example: Consider G with the following productions:

A Bab, B aA|bB|aa|AB
G1 which is equivalent to G can be constructed with the
following productions.
A aAab| bBab|aaab|ABab,
B aA|bB|aa|AB
 Normal Forms for CFG

Lemma2: Let G contains productions of the form

A A1| A2|.| Ar|1|2|.|s (is do not start with A)
Let Z be a new variable in G1, where P1 is defined as
follows:
A 1|2|.|s
A 1Z|2Z|.|sZ.
Z 1|2|.|r
Z 1Z|2Z|.|rZ

Example: Consider G with the following productions:

A aBD|bDB|c|AB|AD
 Normal Forms for CFG

Then G1, equivalent to G contains the following

productions:
A aBD|bDB|c
A aBDZ|bDBZ|cZ
Z B|D
Z BZ|DZ
 Normal Forms for CFG

Reduction to GNF

Step1: Check whether the given grammar is in CNF. If not

in CNF, make it in CNF. Rename the variables as A1,A2,.An.
With S=A1.

Step2: Ai productions should be of the form Ai Aj such

that i < j. If there are productions of the form Ai Ai,
apply Lemma2 to get rid of such productions. Otherwise
apply Lemma1.

Step3: Modify Zi productions. Apply Lemma1 to eliminate

productions of the form Zi Ak.
 Normal Forms for CFG

Example1: Construct a grammar in GNF equivalent to the

grammar: S AA|a, A SS|b

Answer: The given grammar is in CNF. Let S be A1 and A

be A2.
A1 A2A2
A1 a
A2 A1A1
A2 b
A1 productions are in the required form. A2 b is also in
the required form.
 Normal Forms for CFG
Apply Lemma1 to A2 A1A1.

We need to apply Lemma2 to A2 A2A2A1. Let Z2 be the

new variable.

Now we can eliminate A1 A2A2 using Lemma1.

 Normal Forms for CFG

We apply Lemma1 to get

Normal Forms for CFG
 Normal Forms for CFG
Example2: Convert the following grammar into GNF.

Rename
The productions are :
Normal Forms for CFG
Normal Forms for CFG
Normal Forms for CFG