A PHRASE STRUCTURE GRAMMAR OF THE ARABIC LANGUAGE 
AYMAN ELNAGGAR i 
ABSTRACT 
A lot of work has been done 
in the field of natural language 
processing (NLP) for ARABIC. Few 
researchers have tried hard to apply 
on ARABIC the methods of 
computational linguistics, as for 
example, Definite Clause Grammar 
(DCG) \[I\] and Augmented Transition 
Networks (ATN) \[2\]. Because there is 
not a modern linguistic model for 
ARABIC grammar within the frame of 
computational linguistics, the 
results achieved are not comparable 
with that achieved on ENGLISH. 
In this paper, we represent the 
phrase structures covering ARABIC. 
They are completely different from 
ENGLISH. 
i. INTRODUCTION 
Since ARABIC is a rich and 
highly inflected language \[3\], there 
are grannnatical categories in 
ARABIC which do not exist in other 
foreign languages such as in ENGLISH 
\[4\]. For example, there is an ARABIC 
category 2 called 'I/AFOOL MOTLAK' 
which has no equivalent in ENGLISH. 
Moreover, there are syntactic 
categories in ARABIC which could 
themselves be a whole sentence or 
even a psepositional phrase An 
example of such a category is the 
adverb . In addition, there are other 
differences in the order of the 
constituents within the sentence. For 
example, in ARABIC, adjectives follow 
the noun which they modify. 
In fact, ARABIC has its own 
structure, which is completely 
i 21 KASER EL-EINI St. CAIRO, EGYPT. 
2 When there is no equivalent 
translation available for any 
category in ENGLISH, the 
transliterated form will be shown 
between single quotes. 
different from ENGLISH and hence, 
requires a different treatment. 
2.ARABIC AND COMPUTATIONAL 
LINGUISTICS 
As the need increases to a 
model of A~%BIC, ARABIC linguists try 
to re-organize the granmlatic system 
in a way that may help the 
researchers in the field of 
computational linguistics. Linguists 
is traditionally concerned with 
formal, general, structural models of 
natural language. Linguists, 
therefore, have tended to build 
formal models which allow to capture, 
as much as possible, the 
regularities of language and make the 
most appropriate linguistic 
generalizations. Little or no 
attention has been paid to 
characterize the language itself, 
ignoring the mechanism that produce 
it or decipher it. 
So, the researchers spent most 
of the time in re-building the models 
given by the linguists. Their aim was 
to make these models acceptable in 
computational field (synthesis, 
generation and translation). 
2.i THE KEY 
What is new in this work is 
that we classified all the noun 
phrase modifiers into two categories. 
Category one, noun phrase post- 
modifiers, in which the ordered 
sequence of these modifiers is 
significant and it is the key to 
construct these noun phrase 
structures These modifiers are noun 
complement (NC), adjective (ADJ), 
apposition (APP) and correlation 
(COR). Category two which consists 
mainly from VP modifiers like adverb. 
The above categorization has 
given us the ability to focus on each 
category and finally the NP 
structures were obtained . 
342 1 
3. THE NOUN PHRASE STRUCTURES 
We have Eight structures 
covering all the categories of noun 
phrases. FIVE of these structures are 
giw~n here. 
The notation used in 
describing these structures is :- 
( ) = The category inside these 
brackets is optional. 
*( ) = The category inside these 
brackets could be repeated more than 
once. 
\[ \] = The category inside these 
brackets exists once at least° 
*\[ \] = The category inside these 
brackets could be repeated but it 
should exists once at least° 
Representation is in 
Chomsky's Normal Form and should be 
interpreted from right to left 
The structures given here 
consist mainly of NOUN (N) as the 
basic unit and its maximum allowable 
post-modifier8 whichcould follow it 
These post-modifiers have a 
specific sequence which is the key to 
construct these structures and 
according to the number of these 
post-modifiers, the structures vary 
The NP structures as well as some of 
the ARABIC categories they 
constitute, are given in Fig.(1) 
Example of a simple Arabic 
verbal sentence is given in Fig.(2) 
In this example, the object of the 
sentence has all the post- modifiers 
given in the structure NPI and It is 
clear that some of these modifiers 
have their own embedded structures 
shown in the embedded rectangles 
4. MODIFICATION 
'MAFOOL MOTLAK' is a special 
modifier in Arabic which has a 
different forms depending on its 
post-modifiers. We have first \[5\] 
considered its structure as: 
*(ADJ) (NC) NOUN 
But later on, we have found out that 
both NC and ADJ can not exist 
simultaneously and only one of them 
should exist alone. So, we have split 
this structure to three structures 
NP3, NP4 anti NP5 published here. 
ACKNOWLEDGMENT 
I would like to thank Dr. T. 
AMBER (~.) for her valuable discussion 
and support Also~, I would like to 
thank Dr. H~ MAHGOUB ~) for his 
comments reviewing the paper and who, 
first, implemented most of the 
structures given here° 
CONCLUS ION 
We have tried here to present 
the paper in a frame work showing 
that the study of computational 
modeling of ARABIC faces many 
problems. It is evident that:, the 
ARABIC possesses a certain symmetry 
as regards its structure, which leads 
itself easily to computation. This 
work is the basis of any further 
computational processing of ARABIC, 
leading to semantic analysis and 
eventually machine translation o 
(~) Professor of computational 
linguistics, AL-ALSON FACULTY, AIM- 
SHAMS UNIVERSITY. 
(2) Project Leader, IBM Cairo 
Scientific Center° 
2 343 
CATEGORY NP STRUCTURES 
OBJECT 
ADJECTIVE 
'MAFOOL MO TLAK' 
( CASE 1 ) 
'MAFOOL MO TLAK' 
( CASE 2 ) 
CORROLA TION 
NP1 = (COR) (APP) (ADJ) (NC) N 
NP2 = (NC) N 
NP3 = N 
"k 
NP4 = \[ADJ\] N 
NP5 = \[NC\] N 
Fig.(1) The NP structures with some of the ARABIC categories they constitute . 
ITS PROPER ENGLISH EQUIVALENT IS 
THE STUDENT MET THE NEW HEAD OF CAIRO UNIVERSITY Dr. M. SALAMA HIMSELF 
F" 
SELF I 
N 
"~ ~, .¢ .~ 
Dr. M. SALAMA 
N 
I NEW 
L_ N 
COR APP ADJ 
CAIRO 
NC 
NC 
I 
HEAD 
N 
~.~Jtta l 
THE 
STUDENT 
N 
MET 
OBJECT SUBJECT VERB 
344 
Fig.(2) Example of a simple ARABIC verbal sentence 
3 

REFERENCES 

\[i\] 'TWO ARABIC SYNTACTIC 
ANALYZERS', HISHAM EL-SHISHINY & 
AYMAN ELNAGGAR, PROCEEDINGS OF THE 
SECOND CONFERENCE ON ARABIC 
COMPUTATIONAL LINGUISTICS, KUWAIT , 
NOV., 1989 

\[2\] 'AN EXPERT SYSTEM FOR 
UNDERSTANDING ARABIC SENTENCES', 
MAHER S. AHMED, 
M.Sc. THESIS, CAIRO UNIVERSITY, 1988. 

\[3\] ' AL-NAHO AL-MOSAFFA ', 
Dr. MOHAMMED EID, CAIRO UNIVERSITY 

\[4\] 'NATURAL LANGUAGE PARSING AND 
LINGUISTIC THEORIES', U. REYLE & C° 
ROHRER, D. REIDEL PUBLISIIING COMPANY, 
1988 

\[5\] ' A FINITE STATE AUTOMATA OF THE 
ARABIC GRAMMAR', AYMAN ELNAGGAR, 
PROCEEDINGS OF THE 'IEEE 
INTERNATIONAL WORKSHOP ON TOOLS FOR 
AI', GEORGE MASON UNIVERSITY, OCT., 
1989. 
