Deterministic Finite State Automaton of Arabic Verb System: A Morphological Study

Mohammad Mahyoob
International Journal of Computational Linguistics (IJCL), Volume (9) : Issue (1) : 2018 13
Deterministic Finite State Automaton of Arabic Verb System:
A Morphological Study
Mohammad Mahyoob eflu2010@gmail.com
Department of Languages & Translation
Faculty of Science & Arts, Alula
Taibah University,
Madina, Saudi Arabia
Abstract
Finite State Morphology serves as an important tool for investigators of natural language processing.
Morphological Analysis forms an essential preprocessing step in natural language processing. This
paper discusses the morphological analysis and processing of verb forms in Arabic. It focuses on the
inflected verb forms and discusses the perfective, imperfective and imperatives. The deterministic
finite state morphological parser for the verb forms can deal with Morphological and orthographic
features of Arabic and the morphological processes which are involved in Arabic verb formation and
conjugation. We use this model to generate and add all the necessary information (prefix, suffix,
stem, etc.) to each morpheme of the words; so we need subtags for each morpheme. Using Finite
State tool to build the computational lexicon that are usually structured with a list of the stems and
affixes of the language together with a representation that tells us how words can be structured
together and how the network of all forms can be represented.
Keywords: Computational Morphology, Finite-State, Arabic Verb Forms, Morphological Analysis.
1. INTRODUCTION
Semitic words can be viewed as a simple mechanism consisting of two lists: a relatively short list of
templates, no more than a few hundred, for forming nouns, verbs, etc. in all their inflected forms; and
a much longer list of several thousand roots [1].
Arabic language belongs to Semitic group of languages. Other languages belonging to this group
are: Amharic, Aramaic, Hebrew, Tigrinya and Maltese [2]. The Arabic language grammarians
organized words into three main divisions. These divisions also have sub-divisions that contain
every word in Arabic language. Seen in this perspective a verb form in Arabic consists of a root and
a vocalic melody in addition to the agreement affixes.
As pointed in the above lines the Semitic morphology is different in many ways. A unique
characteristic of this morphology is the non-concatenative merging of roots and patterns to form
words or word stems [3], [4].
In view of the facts given above the Arabic morphological analysis needs to add all the necessary
information (prefix, infix, suffix, etc.) to each root or stem of the words. Further, we need technical
applications that analyze Arabic words and deal with internal structure of a given word [5],[6],[ 7].
2. RELATED WORKS
There are many morphological analyzers have been developed and built for Arabic morphology.
Many techniques have been offered by the authors. The discussion here will be for the most
important and in the evaluation part the results have been shown and compared with our work.

Mohammad Mahyoob
2.1 Buckwalter Arabic Morphological Analyzer
This analyzer is considered as one of the most cited work in the literature. Many tool developers
make use of the data of this analyzer for developing other computational applications [8]. The
problem of this system is the long output analysis.
2.2 Xerox Arabic Morphological Analysis and Generation
It is one of the good morphological analyzers for Arabic language it is a root based using
FiniteStateTechnology [9]; [10]. One of the good advantages of this analyzer is the ability to cover
most of the lexical features. However, it is a rule- based and English gloss is provided for each
lexeme. The generation of words those are not available in the language is one of the disadvantages
of this system too [11].
2.3 ElixirFM: An Arabic Morphological Analyzer by Otkar Smrz
Otakar Smrz developed an online Arabic Morphological Analyzer for Modern Standard Arabic [12].
The author made use of Buckwalter lexicon [13]. The advantage of this system is that the output of
the analyzed word is processed in four different layers (Resolve, Inflect, Derive and Lookup). But the
system is limited to coverage because of certain analysis.
3. VERB PARADIGMS
According to Wright [14], a great majority of the Arabic verbs is trilateral. That is to say; it contains
three radical letters, though quadrilateral verbs are by no means rare. In English the infinitive form
of the verb is ‘to + verb’ in the bare form of the verb.
But in Arabic trilateral verbs can be derived according to these scales َ‫ل‬َ‫ع‬َ‫ف‬, /faʕal َ‫ل‬ُ‫ع‬َ‫ف‬, / fuʕal/ َ‫ل‬ِ‫ع‬َ‫ف‬
/faʕil/ ‘to work’ . These forms constitute the first stage of verb derivation. All the derived forms can
function as stems. Therefore as observed by Wright [15] , the 3
rd
per. sing.masc.perf. being the
bare form of the verb is commonly used as paradigm. The translational equivalent of this form is
the bare form of English verb which is in present tense.
The Arabic grammarians considered the verb ‘fʕ l’ ‫فعل‬ (to work) as basic to develop a paradigm.
The first radical of this trilateral verb is called by them as fa, the second is the ʕ ain, and the third is
the lam. If we are utilizing these three base letters fa, ʕ ain and la we will get the noun ‘action or
verb’. The same thing holds true in other situation too. Thus for example, if we have three letters
viz ‘ra, sa and ma’ we will get the noun ‘drawing’ and so on. These base letters need to be woven
into a pattern from the morphological system. One of these patterns is ‘fʔʕ l’ which is used for the
active participle.
3.1 Verb Conjunction
In order to show the verb conjugation we take a set of base letters and place them on the pattern
‘fʕ l’ ‫فعل‬) ). These three base letters are ( ‫ر‬,‫س‬,‫م‬ ) r, s, m. ‘rsm’, ‘rasama’(draw). We will study the
variations of the simple past and see how the morphological inflection of the verb can change
according to the verb-subject agreement.
We noticed that the root or radical form is ‘r s m’. These three consonants letters are repeated in all
the forms of the verb. We also find that the first and the third letters or radicals of this simple
trilateral verb in the active tense is always vowelled with ‘a’ fathah. The second letter may be
vowelled by ‘a’ or it can be ‘u’ Damah or ‘i’ Kassra. The change occurs only in short vowels.
However the radical letters do not change. If we compare the verb (rsm) with the radical (f, ʕ and
l), the letter that corresponds to ‘f ’ is ‘r’, to ‘ʕ ’ corresponds ‘s’ and to ‘l’ corresponds ‘m’ as shown
in figure 1 below :

Mohammad Mahyoob
f ʕ l
r s m
FIGURE 1: Radical Form of The Verb
In the imperfective active form radical form ‘f ʕ l’ (faʕ al) do not change but the vocalic melody
changes. Thus we have forms ‘jafʕ al’, ‘nafʕ al’, ‘ʔafʕ al’ and ‘tafʕ al’ etc. We notice that the
conjugation of the imperfective form is different from the base form of perfective form. There is no
prefix in the perfective, only suffix is added while in the imperfective we have both the prefix and
the suffix in addition to the vocalic change.
3.2 Affixation In Arabic Verbs
As we have seen above in Standard Arabic, the verb occurs in two morphological forms: perfective
and imperfective. The main difference between the two is in the realization of their agreement
features. In the perfective all agreement morphology is expressed by suffixes while in the
imperfective, agreement features are realized by both suffixes and prefixes. The prefixes carry
person features, except the first person plural, where number is also realized on the prefix; the
suffixes mainly carry number features. Gender feature is also realized on the person prefix, except
in the second person singular feminine, where it is realized by a suffix [16], [17].
3.2.1 Perfective Form and Affixation
The following table shows the perfective form which is realized by suffix. The verb form consists of
a root and vocalic melody in addition to the agreement suffix as shown in table 1 below:
Person Number Gender Affix Verb forms
1 Singular F/M -tu daras-tu ‘(I)
studied’
2 Singular M -ta daras-ta
2 Singular F -ti daras-ti
3 Singular M -a daras-a
3 Singular F -at daras-at
2 Dual M/F -tummaa daras-tuma
3 Dual M -aa daras-aa
3 Dual F -ataa daras-ataa
1 Plural M/F -naa daras-naa
2 Plural M -tum daras-tum
2 Plural F -tunna daras-tunna
3 Plural M -uu daras-uu
3 Plural F -na daras-na
TABLE 1: Arabic Perfective Form and The Affixation.
It is claimed that there are two ways of morphological realization of the past tense:
a. “The agreement morphology suffixed to the verb realizes both tense and agreement”.
b. “The vocalic melody realizes the past tense; the suffix is just a realization of the agreement
morphology” [18].
Benmamoun [16], went against these two claims and he approved that the phonological realization
of past tense morpheme is abstract like English simple present tense morpheme.
3.2.2 The Imperfective Form and Affixation
The imperfective in Standard Arabic occurs in different morphological forms, usually referred to as
moods distinguished by their endings [19], [20].

Mohammad Mahyoob
The indicative form is represented by the sound ‘u’ if the verb ends with a consonant and ni/na if
the verb ends with a long vowel.
The subjunctive form is expressed by ‘a’ if the verb ends with a consonant, but if the verb ends with
a long vowel there is zero suffix. Jussive form is represented by zero morphemes. In table 2, the
bare imperfective forms are shown without mood endings.
Person Number Gender Affix Verb forms
1 singular F/M ʔa- ʔa-ktub
I write/ am writing.
2 singular M ta- ta-ktub
you write/ are writing.
2 singular F ta----ii ta- ktub-ii
3 singular M ya- ya- ktub
he writes/ is writing.
3 singular F ta- ta- ktub
she writes/ is writing
2 Dual M/F ta----aa ta- ktub-aa
3 Dual M/F ya----aa ya- ktub-aa
they write/ are writing..
1 Plural M/F na- na-ktub
2 Plural M ta----uu ta-ktub-uu
you study/ are studying
2 Plural F ta----na ta-ktub-na
3 Plural M ya----uu ya-ktub-uu
they write/ are writing.
3 Plural F ta---na ta-ktub-na
they write/ are writing.
TABLE 2: Imperfective Form and Its Affixation.
3.3 Imperative
The imperative (the order or command) is formed from the imperfective form in Arabic, but there are
some features for this form as stated below:
If the first radical letter is a consonant, the glottal stop is inserted at the beginning and to avoid
cluster (there is no cluster in Standard Arabic word-initially) a vowel is also inserted. The insertion of
this vowel is according to the vowel which follows the second radical letter of the root:
If the vowel is ‘u’ the glottal stop is rendered to „u’.
If the vowel is ‘a’ or ‘I’ the glottal stop is rendered to ‘I’ for example:
ta-ktubu ‘you write/ are writing’ changes to ʔuktub
taftaH „you open/ are opening’ changes to ʔiftaH
taDrib ‘ you beat/ are beating’ changes to ʔiDrib
3.4 Tense and Aspect In Standard Arabic
There is no specific indication for the tense and aspect in Arabic verb forms (perfective and
imperfective) [21].
Arabic does not grammaticalize the perfective /imperfective distinction, nor does it have any
particular progressive morphology [22]. As Fassi Fehri pointed out that, we will check these two
examples:

Mohammad Mahyoob
a- katab-a wrote-3.s.m
b- ya-ktub-u write-3.s.m.Indic
In (a), we noticed that the verb carries lexical meaning of the verb, the past tense and the active
voice. In (b), the verb form indicates the imperfective; the suffix indicates the indicative mood and the
agreement. In these two examples we noticed that the tense morpheme in both cases is abstract.
Fehri shows the relation between the agreement and affixation. Two kinds of contrasts contribute to
the identification of temporal morphemes; on the one hand, we have the internal vocalic pattern, on
the other, the position of the agreement morpheme. With the past forms, the agreement (with the
subject) is exclusively by suffixes. With non-past forms, the agreement is both by prefixes and
suffixes.
The past tense morpheme in Arabic is not realized by the overt affixes of the perfective form that
seems to carry agreement only. The vocalic melody of the verb does not carry the past tense as
well. It seems to be an abstract morpheme located in tense which can be hosted by negation or by
the verb [23].
4. DETERMINISTIC FINITE STATE AUTOMATON (FSA)
Deterministic Finite State Automaton is a finite state machine that accepts/rejects finite strings of
symbols and only produces a unique computation of the automaton for each input string.
'Deterministic' refers to the uniqueness of the computation. The behaviour of the deterministic finite
state automaton during the recognition is fully determined by the state it is in and the symbol it is
looking at. For example, the figure 2 illustrates a deterministic finite automaton using a state
diagram. There are three states: S0, S1 and S2 which are called nodes. The automaton takes a
finite sequence of 0s and 1s as input. For each state, there is a transition arrow leading to a next
state for both 0 and 1. A DFA jumps deterministically from a state to another by following the
transition arrow. For example, if the automaton is currently in state S0 and current input symbol is 1
then it deterministically jumps to state S1. A DFA has a start state (denoted graphically by an arrow
coming in from nowhere) where computations begin, and a set of accept states (denoted graphically
by a double circle) which helps define when a computation is successful1.
FIGURE 2: An Example of A Deterministic Finite State Automaton.
A deterministic finite automaton is a 5-tuple, (Q, ∑, , q0, F), consisting of
 (Q) a finite set of states
 (∑) a finite set of input symbols called the alphabet
 δ a transition function (δ : Q × Σ → Q)
 q0 a start state (q0 ∈ Q)
 F a set of accept states (F ⊆ Q)
The machine starts in the start state q0 or s0, the machine will transit from state to state with the data
according to the transition function δ. Finally, the machine accepts data if the last input of this
data causes the machine to halt in one of the accepting states. Otherwise, it is said that the
automaton rejects the string [24], [25].
4.1 Morphological Parser
To Build a Morphological Parser, we need at least the following:

Mohammad Mahyoob
 Lexicon (the list of stem and affixes together with basic information).
This basic information is about the word stem. Lexicon is a repository for words.
 Morphotactics refers to the model of morpheme ordering. This model explains which classes
of morphemes are there inside the word. In other words which morphemes precede and
which follow.
There are many ways to model morphotactics . Finite State Automaton is one of these models which
is discussed in this paper.
 Graphotactics (spelling rules).these rules include the deletion, the addition or transformation
processes.
4.1.1 Developing Finite State Lexicon
A lexicon is a repository for words. The simplest lexicon would consist of an explicit list of every
word of the language; by every word we mean every word, including abbreviations and proper
nouns. It is impossible to list all the words in the language, computational lexicons are usually
structured with a list of the stems and affixes of the language together with a representation of the
mophotactics that tells us how they can fit together. There are many ways to model morphotactics;
one of the most common is the finite state automaton, [26],[27],[28],[29], [30].
The following table is a representation of the inflection system and affixation summary in Arabic
Verb Forms.
TABLE 3: The Affixation Summary In Arabic Verb Forms.
4.2 Experiment
In the following table we will see an example of building lexicon for some verbs in perfective forms.
In this lexicon we have a list of three verbs: katab (write), rasam (draw) and ishtra (buy) and a list of
all possible affixation for the perfective forms which can be shown in subject- verb agreement for
gender, person and number as shown in table 4.
Aspect Tense Root Agreement
IMPERFECTIVE
Future+
present
Sa-
0
ta- Ktb
rsm
nsr
-ii/aa/uu/na/0
0
-uu/aa/0
0
ʔa-
ya-
na-
Imperative ʔu-
ʔi-
Ktb -ii/aa/na/uu/0
PERFECTIVE
Zero morpheme Katab
rasam
nasar
nasar
nassar
kasar
kassar
-
t+u/a/i/ummaa/u
m/unna
a+t/a/taa
-na+a
Uu

Mohammad Mahyoob
Multichar_Symbols +SG +DL +PL +N +V +1P +2P +3P +PERF +NOM +ACC
+MA +FE [ +SG stands for singular, +DL stands for dual +PL stands for plural,
+N stands for Noun, +V stands for Verb, +1P stands for 1st person
singular|dual|plural, +2P stands for 2nd person singular|dual|plural, +3P
stands for 3rd person singular|dual|plural, +PERF stands for perfective, +NOM
stands for nominative case, +ACC stands for accusative case, +MA stands for
masculine, +FE stands for feminine]
LEXICON Root Verbs;
LEXICON Verbs
katab Vend;
rasam Vend;
ishtara Vend;
LEXICON Vend
+V:0 #;
+V+1P+SG+PERF:tu #;
+V+2P+SG+MA+PERF:ta #;
+V+2P+SG+FE+PERF:ti #;
+V+3P+SG+MA+PERF:a #;
+V+3P+SG+FE+PERF:at #;
+V+2P+DL+MA+PERF:tuma #;
+V+2P+DL+FE+PERF:tuma #;
+V+3P+DL+MA+PERF:aa #;
+V+3P+DL+FE+PERF:ata #;
+V+1P+PL+PERF:na #;
+V+2P+PL+MA+PERF:tum #;
+V+2P+PL+FE+PERF:tunna #;
+V+3P+PL+MA+PERF:u #;
+V+3P+pL+FE+PERF:nna #;
TABLE 4: An example of Lexicon for some verbs in perfective forms
The first column contains the bare form of the verb and it’s tense. The second column contains the
stem of each word and its entire morphological features. These features give additional information
about each word stem. The feature +V indicates that the word is verb; +SG indicates that the word is
singular;+DL means that the word is dual; +PL means that the word is in the plural form; +MAS
means that the word is masculine; +FEM means the word is feminine in gender.
According to table4 we believe that the task or the main goal of morphological analysis is to list all
possible analysis of the words.
The following figure shows the model of processing verb morphology in Arabic language.
FIGURE 3: Model of Our Study.
Output
Morpholo
gical
analyzer
Text
Input

Mohammad Mahyoob
4.3 Results and Analysis
The finite state automaton for Arabic verb forms given in Table 3 can be represented
diagrammatically as follows:
FIGURE 4: A Deterministic Finite State Automation for the verb forms in Arabic
In figure 4: we have seen the following:
The number of the states are: 5
The number of the transitions (arcs) are: 14
q0 is the initial state
q3 and q4 are the final states
: {sa (future), ʔu(Imperative-1), ʔ I ( Imperative-2), ta ( Prsnt-1), ʔa ( Prsnt-2), ya (Prsnt-3), ta/ ʔ
a/ya/na(fut-psn-agr), ksr (Past-1), kssr (past-2), ktb (root-1), rsm (root-2), -ii/aa/uu/na (Agr-1), -
t+i/u/a/ummaa/uu/unna/, a+t/a/taa, na+a,uu (agr-2)}
4.4 The Symbols
The following are the symbols which are used in the Finite State Automation (Figure 2):
1- Past-1: {ksr} this symbol refers to the past tense form which can be the first bare form of the
verb conjugation in Standard Arabic, no prefix precedes past tense*.
2- Past-2: {kssr} this symbol refers to the past tense form which can be the second bare form of
the verb conjugation in Standard Arabic.
3- Future: {sa} is applicable for future tense marker. It has to be followed by the prefix of present
markers.
4- Imperative-1: the imperative marker which can be added to the root directly, it has two
morphemes:
The morpheme {ʔu} is applicable for the imperative when the second radical consonant is
followed by ‘u’ as shown in this example:
ta-ktubu ‘you write/ are writing’ changes to ʔuktub
5- Imperative -2 {ʔ i} is applicable when the second radical letter of the root is either ‘i’ or ‘a’ as in
the following examples:
taftaH ‘you open/ are opening’ changes to ʔ iftaH
taDrib ‘ you beat/ are beating’ changes to ʔ iDrib
6- Fut-psn-Agr (future-person-agreement) { ta/ ʔ a/ya/na} the future tense morpheme has to be
followed by the agreement of persons which are the same as in present tense morphemes.
q0 q1
q3future
Root-1
Root-2
Agr-1
Agr-2
q2
Past-1/Stem
Prsnt-3
Prsnt-2
Prsnt-4
Imperat-1
Imperat-2
Fut-Psn-Agr
Prsnt-1
Past-2/Stem
q3
Start q3
q4

Mohammad Mahyoob
7- Prsnt-1: {ta} is applicable for present tense form with the following person, gender and number:
second singular feminine, second dual masculine/feminine, second plural masculine/feminine,
third singular female, and third plural feminine.
8- Prsnt-2: {ʔa} is applicable for first singular feminine/masculine.
9- Prsnt-3: {ya} is applicable for third singular masculine, third dual masculine / feminine and third
plural masculine.
10- Prsnt-4: {na} is applicable for first plural masculine/feminine.
11- Root-1: {k t b}
12- Root-2: {r s m}
13- Agr-1: {-ii/aa/uu/na} is applicable for present, future and imperative.
14- Agr-2: {-t+u/i/ummaa/um/unna, -a+t/a/taa, -na+a, -uu} is applicable for perfective (past tense).
4.5 Transition Function Matrix
Transition function matrix between the states indicates how the transition moved from one state to
another carrying some data.
In the following table we will show the number of states and how the transition function matrix moves
from one state to another according to figure 4 above.
From To Output
0 1 sa -(future)
0 2 ta- (Prsnt-1)
0 2 ʔa- (Prsnt-2)
0 2 ta- (prsnt-3)
0 2 na- (prsnt-4)
0 2 ʔ u- (imperative-1)
0 2 ʔ i -(imperative-2)
0 3 ksr (past-1)
0 3 kssr (past-2)
1 2 Fut-Pre-Agr
2 3 Roots
3 4 -ii/aa/uu/na (Agreement-1)
3 4 -t+u/i/ummaa/um/unna, -a+t/a/taa, -na+a, -uu
(Agreement-2)
TABLE 5: The Transition Table.
3.7 Example
In Figure (4) we have shown diagrammatically how the finite state machine works. The transition of
the three verbs; katab (wrote), rasam (drew) and saafar (travelled) is illustrated in the following figure
and the output is shown in table 6 below.
Sigma= {l, a, b, h, I, k, m, n, p, r, s, t, u, + +DL +FE +MA +PERF +PL +SG +V +1P +2P +3Pa}

Mohammad Mahyoob
FIGURE 5: States Transition Diagram for katab, rasam, and saafar.
The above diagram shows the automatic network representation and how the Finite State machine
is working to produce the inflected forms of the verbs.
Table 6 below shows the output of our morphological analysis using FSA, by testing our developed
program the outcomes are possibly all the forms of the verbs according the tense either perfective or
imperfective.
Input Morphological Parsed Output
r a s a m (drew) perfective
forms
r a s a m +V
r a s a m 0
r a s a m +V +1P +SG +PERF
r a s a m t u 0 0
r a s a m +V +2P +SG +MA +PERF
r a s a m t a 0 0 0
r a s a m +V +2P +SG +FE +PERF
r a s a m t i 0 0 0
r a s a m +V +2P +DL +MA +PERF
r a s a m t u m a 0
r a s a m +V +2P +DL +FE +PERF
r a s a m t u m a 0
r a s a m +V +2P +PL +MA +PERF
r a s a m t u m 0 0
r a s a m +V +2P +PL +FE +PERF
r a s a m t u n n a
r a s a m +V +3P +SG +MA +PERF
r a s a m a 0 0 0 0
r a s a m +V +3P +SG +FE +PERF
Q0
Q3
Q2
Q9
Q6
Q5
Q4
Q7
Q10
Q11
Q12
Q13
rasam
saafar
<+3P+SG+M>
<+V:aa>
<+2P+PL+M>
<+3P+PL+F>
<+2P+DL+M>
Q2
Q1
Q2
Q8
<+2P+SG+F>
Q14

Mohammad Mahyoob
TABLE 6: Output of The Perfective Forms of The Verb drew ‘rasam’.
The goal is to process an input form, from those in the first column and produce output forms, like
those in the second column as shown in table 6.
4. EVALUATION
The following table brings the obtained results of the tested data in our system.
Data Correct output Generated forms F-Score
1500 96.00 50.00 65.75
750 97.00 50.00 74.08
TABLE 7: Obtained Results.
The evaluation of our work is compared with other previous morphological analyzers; Tim
Buckwalter Morphological Analyzer, Tri-literal Root algorithm, Khoja Stemmer, Xerox Morphological
Analyzer and ElixirFM. We performed the experimental analysis to show that the developed program
outcomes are all possible forms of the verbs according the tense; either perfective or imperfective. The
results generated by proposed methodology are sufficient and concrete with high accuracy of 96.00%.
Our system brings the best results compared to previous systems. The advantages and
disadvantages of some previous systems are discussed in the related work above. The following
table shows the accuracy percentage of our system and other morphological analyzers:
Morphological
Analyzers
Buckwalter
morph.
Analyzer
Tri-literal
Root
algorithm
Khoja
stemmer
Xerox
Morpho.
Analyzer
ElixirFMS Our
system
Accuracy 33.91% 65.00% 71.25% 88.91% 89.58% 96.00%
TABLE 8: The Evaluation Process Results.
5. CONCLUSION AND FUTURE WORK
The present study, however, made several noteworthy contribution to the field of Arabic
computational morphology by presenting the language analysis in a new and easy methods, by
using FSA tool, and testing the obtained output.
This paper discussed how to build Finite-state machines based on the linguistic principles for the
verb system of Arabic language.
r a s a m a t 0 0 0
r a s a m +V +3P +DL +FE +PERF
r a s a m a t a 0 0
r a s a m +V +3P +DL +MA +PERF
r a s a m a a 0 0 0
r a s a m +V +1P +PL +PERF
r a s a m n a 0 0
r a s a m +V +3P + p L +FE +PERF
r a s a m n n a 0 0 0 0
r a s a m +V +3P +PL +MA +PERF
r a s a m u 0 0 0 0

Mohammad Mahyoob
It also describes the morphological analysis and processing of verb forms in Arabic using finite state
machine It focuses on the inflected verb forms. It shows the methods of analyzing Arabic verbs with
the morphological and orthographic features of Arabic and the morphological processes which are
involved in Arabic verb formation and conjugation. The Morphological analyzer adds all the linguistic
information to each morpheme of a word. We use the Finite State tool to build the computational
lexicon that is usually structured with a list of the stems and affixes of the language together with a
representation that tells us how they can be combined together.
This paper is a representation of Arabic Verb Forms by developing a morphological analyzer for
these forms. Future plan is to cover all Arabic categories and forms .i.e. developing a
morphological analyzer for Nouns forms, Pronouns forms and other Particles. This work will be
extended to develop a POS Tagger and Parser for Arabic Language categories.
6. REFERENCES
[1] Shimron, J. Language Processing and Aquisition in Language of Semitic, Root-based,
Morphology. Amsterdam: John Benjammins Publishing Company, 2002.
[2] Haywood, J.A. and Nahmad, H.M. A New Arabic Grammar of the written Language. London:
Lund Humphries, 1965.
[3] Farghaly, A., and K. Shaalan. "Arabic Natural Language Processing: Challenges and
Solutions", ACM Transactions on Asian Language Information Processing, 2009.
[4] Greenberg,J.1950.'The patterning of root morphemes in Semitic'.Word 6:162-81.
[5] Forsberg M. and Ranta A. Functional Morphology ICFP'04, Proceedings of the Ninth ACM
SIGPLAN International Conference of Functional Programming, September 19-21, Snowbird,
Utah, 2004.
[6] Atwell E., Al-Sulaiti L., Al-Osaimi S., Abu Shawar B.“A Review of Arabic Corpus Analysis
Tools”, JEP-TALN 04, Arabic Language Processing, Fès, 19-22 April, 2004.
[7] Bender, M. L. Amharic Verb Morphology. East Lansing, Michigan: Michigan State University,
1978.
[8] Bat-El, O. “Semitic verb structure within a universal perspective”. Amesterdam: Language
Aquisition and Language Disorder:28, 2002.
[9] Beesley K.R. “Arabic Finite-State Morphological Analysis and Generation”, Proceedings the
16th conference on Computational linguistics, Vol 1. Copenhagen, Denmark: Association for
Computational Linguistics, 1996, pp 89-94.
[10] Aronoff, M. Morphology by Itself: Stem and Inflectional Classes. Cambridge: The MIT Press,
1994.
[11] Beesley KR. Finite-State Non-Concatenative Morphotactics, SIGPHON-2000, Proceedings of
the Fifth Workshop of the ACL Special Interest Group in Computational Phonology,
Luxembourg, August 6, 2000, p. 1-12.
[12] Darwish K. Building a Shallow Morphological Analyzer in One Day, Proceedings of the
workshop on Computational Approaches to Semitic Languages in the 40th Annual Meeting of
the Association for Computational Linguistics (ACL-02). Philadelphia, PA, USA, 2002.
[13] Buckwalter T. Buckwalter Arabic Morphological Analyzer Version 1.0. Linguistic Data
Consortium, University of Pennsylvania, LDC Catalog No.: LDC2002L49, 2002.

Mohammad Mahyoob
[14] W. Wright, L. A Grammar of the Arabic Language. New Delhi: Munshiram Manoharlal
Publishers Pvt. Ltd, 2004.
[15] Gridach, M., & Chenfour, N. “Developing a new system for Arabic morphological analysis and
generation”. In Proceedings of the 2nd Workshop on South Southeast Asian Natural
Language Processing (WSSANLP) 2011, (pp. 52-57).
[16] Bnmamoun, E. The Feature Structure of Functional Categories: A Comparative Study of
Arabic Dailects. Oxford: Oxford University Press, 2000.
[17] Bnmamoun, Elabbas. “The role of the imperfective template in Arabic morphology”. Language
Acquisition and Language Disorders, 2003, 28: 99-114
[18] McCarthy, J. “A Prosodic theory of nonconcatenative Morphology” Linguistic Inquiry, 1981,12
(3): 373-418.
[19] Benmamoun, E. The Feature Structure of Functional Categories: A Comparative Study of
Arabic Dailects. Oxford: Oxford University Press, 2000.
[20] Hassan, A. Al-NaHw Al-Wafii, . vol. 4. Cairo: Daar Al-Maarif, 1973.
[21] Fehri, F. Issues in the Sructure of Arabic Clauses and Words. Dordrecht: Kluwer, 1993.
[22] Fehri, F. Issues in the Sructure of Arabic Clauses and Words. Dordrecht: Kluwer, 1993.
[23] Beesley, K. R. and L. Karttunen, Finite State Morphology. Stanford, Calif., Csli, 2003.
[24] Jurafsky, D. and J.H.Martin. Speech and Language Processing, Prentice-Hall, New, 2000.
Jersy.
[25] McCarthy, J. 1979. Formal Problems in Semitic Phonology and Morphology . Ph.D.
dissertation. Camberidge:MIT.
[26] Nizar Y. Habash. 2010. Introduction to Arabic Natural Language Processing. Morgan &
Claypool
[27] Roy Bar-Haim, K. S. 2006. Part-of-Speech Tagging of Modern Hebrew Text. Cambridge:
Cambridge University Press.
[28] Sawalha, M., & Atwell, E. (2008). Comparative evaluation of Arabic language morphological
Analysers and stemmers. Coling 2008: Companion volume: Posters, 107-110.
[29] Sproat, B. R. 2007. Computational Approaches to Morphology and Syntax. Oxford: Oxford
University Press.
[30] Wikipedia contributors, 'Deterministic finite automaton', Wikipedia, The Free Encyclopedia, ,
07:24 UTC: https://en.wikipedia.org/w/index.php?title=Deterministic_finite_automaton&oldid=
836856582> 17 April 2018 [4 May 2018].

Deterministic Finite State Automaton of Arabic Verb System: A Morphological Study

More Related Content

What's hot (20)

Similar to Deterministic Finite State Automaton of Arabic Verb System: A Morphological Study (20)

Recently uploaded (20)

Deterministic Finite State Automaton of Arabic Verb System: A Morphological Study