Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

13 Sep, 2004

4.1.0.1

created.

Tafseer Ahmed

27 Oct, 2004

4.1.0.2

Analysis is modified after

meeting with Dr. Miriam Butt.

Tafseer Ahmed

4 Nov, 2004

4.1.0.3

Rules updated

Sara Hussain

17th Jan, 2005

4.1.0.4

Sara Hussain

7th Feb, 2005

4.1.0.5

Aasim Ali

05-AUG-2005

5.1.0.1

Added Kar phrase (KarP) and

sentence level adjunct
(S_Adjunct) in each rule
Added gender agreement
check for infinitivals in
Subject and object position
(KPmain)
Documenting the analyses
and modifications (on the
basis of these analyses)
made by Nayyara

Rule ID: UGR103

Rule Syntax: Following is the constituent description of the rule.

Sdec -> KPmain;

Sdec -> KPmain;
Sdec -> KPmain;
Sdec -> KPmain;

@SAP
@SAP
@SAP
@SAP

@OBLIQUE
@OBLIQUE
@OBLIQUE
@OBLIQUE

(KPmain; )
(KPmain; )
KPmain;
(KPmain; )

VPperf ;.
VPperf;.
VPperf;.
VPperf;.

Sdec -> KPmain;

Sdec -> KPmain;

@SAP
@SAP

@OBLIQUE (KPmain; )
@OBLIQUE (KPmain; )

VPnonperf;.
VPnonperf;.

Sdec -> KPmain;

Sdec -> KPmain;

@SAP
@SAP

@OBLIQUE (KPmain; )
@OBLIQUE (KPmain; )

VPraha;.
VPraha;.

Sdec -> KPmain;

Sdec -> KPmain;
Sdec -> KPmain;

@SAP
@SAP
@SAP

@OBLIQUE KPmain;
@OBLIQUE KPmain;
@OBLIQUE KPmain;

VPperf;.
VPnonperf;.
VPraha;.

Sdec -> KPmain;

Sdec -> KPmain;
Sdec -> KPmain;

@SAP
@SAP
@SAP

@OBLIQUE KPmain;
@OBLIQUE (KPmain; )
@OBLIQUE (KPmain; )

VPmodal;.
VPmodal;.
VPmodal;.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Functional Description:

Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;
@SAP
@OBLIQUE
(KPmain: ! CASE =c NOM, ! SEM_TYPE =c{UNANIM_CONC,ANIMAL,ABSTRACT},^ OBJ =!; )
VPperf : ^ =!, ^ SUBJ CASE =c ! _SUBJ_CASE, ^ TNS_ASP MODAL =c NONE,
^ TNS_ASP NEED =c NEG, ^ TNS_ASP PERF =c POS, ^ TNS_ASP PROG =c NEG,
^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND, ^ SUBJ NUM =c ! NUM,
^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;
@SAP
@OBLIQUE
(KPmain: ! CASE =c ACC,! SEM_TYPE =c HUMAN, ^ OBJ =!; )
VPperf : ^ =!, ^ SUBJ CASE =c ! _SUBJ_CASE, ^ TNS_ASP MODAL =c NONE,
^ TNS_ASP NEED =c NEG,^ TNS_ASP PERF =c POS, ^ TNS_ASP PROG =c NEG,
^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND, ^ SUBJ NUM =c ! NUM,
^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c ERG;
@SAP
@OBLIQUE
KPmain: ! CASE =c NOM, ! SEM_TYPE =c {UNANIM_CONC,ANIMAL,ABSTRACT}, ^ OBJ =!;
VPperf : ^ =!, ^ SUBJ CASE =c ! _SUBJ_CASE, ^ TNS_ASP MODAL =c NONE,
^ TNS_ASP NEED =c NEG, ^ TNS_ASP PERF =c POS, ^ TNS_ASP PROG =c NEG,
^ OBJ PERS =c ! PERS, ^ OBJ GEND =c ! GEND, ^ OBJ NUM =c ! NUM,
^ OBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c ERG;
@SAP
@OBLIQUE
(KPmain: ! CASE =c ACC,! SEM_TYPE =c HUMAN, ^ OBJ =!;)
VPperf : ^ =!, ^ SUBJ CASE =c ! _SUBJ_CASE, ^ TNS_ASP MODAL =c NONE,
^ TNS_ASP NEED =c NEG, ^ TNS_ASP PERF =c POS, ^ TNS_ASP PROG =c NEG,
! NUM =c SG, ! GEND =c M, ! PERS =c 3, ! RESPECT =c NORESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;
@SAP
@OBLIQUE
(KPmain: ! CASE =c ACC,! SEM_TYPE =c HUMAN, ^ OBJ =!; )
VPnonperf : ^ =!,! _SUBJ_CASE =c {NOM,ERG}, ^ TNS_ASP MODAL =c NONE,
^ TNS_ASP PERF =c NEG,^ TNS_ASP NEED =c NEG, ^ TNS_ASP PROG =c NEG,
^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND, ^ SUBJ NUM =c ! NUM,
^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;

@SAP
@OBLIQUE
(KPmain: ! CASE =c NOM, ! SEM_TYPE =c {UNANIM_CONC,ANIMAL,ABSTRACT}, ^ OBJ =!; )
VPnonperf : ^ =!,! _SUBJ_CASE =c {NOM,ERG}, ^ TNS_ASP MODAL =c NONE,
^ TNS_ASP PERF =c NEG,^ TNS_ASP NEED =c NEG, ^ TNS_ASP PROG =c NEG,
^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND, ^ SUBJ NUM =c ! NUM,
^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;
@SAP
@OBLIQUE
(KPmain: ! CASE =c ACC,! SEM_TYPE =c HUMAN, ^ OBJ =!; )
VPraha : ^ =!,! _SUBJ_CASE =c {NOM,ERG},^ TNS_ASP MODAL =c NONE,^ TNS_ASP NEED =c NEG,
^ TNS_ASP PROG =c POS, ^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND,
^ SUBJ NUM =c ! NUM, ^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;
@SAP
@OBLIQUE
(KPmain: ! CASE =c NOM,! SEM_TYPE =c {UNANIM_CONC,ANIMAL,ABSTRACT}, ^ OBJ =!; )
VPraha : ^ =!,! _SUBJ_CASE =c {NOM,ERG},^ TNS_ASP MODAL =c NONE,^ TNS_ASP NEED =c NEG,
^ TNS_ASP PROG =c POS, ^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND,
^ SUBJ NUM =c ! NUM, ^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c DAT;
@SAP
@OBLIQUE
KPmain: ! CASE =c NOM, ^ OBJ =!;
VPperf : ^ =!, ! _SUBJ_CASE =c DAT, ^ TNS_ASP MODAL =c NONE,^ TNS_ASP NEED =c NEG,
^ TNS_ASP PERF =c POS, ^ TNS_ASP PROG =c NEG, ^ OBJ PERS =c ! PERS,
^ OBJ GEND =c ! GEND, ^ OBJ NUM =c ! NUM, ^ OBJ RESPECT =c ! RESPECT,
!VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c DAT;
@SAP
@OBLIQUE
KPmain: ! CASE =c NOM, ^ OBJ =!;
VPnonperf : ^ =!,! _SUBJ_CASE =c DAT, ^ TNS_ASP MODAL =c NONE, ^ TNS_ASP PERF =c NEG,
^ TNS_ASP NEED =c NEG, ^ TNS_ASP PROG =c NEG, ^ OBJ PERS =c ! PERS,
^ OBJ GEND =c ! GEND, ^ OBJ NUM =c ! NUM, ^ OBJ RESPECT =c ! RESPECT,
! VOICE =c ACTIVE;.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Sdec -> KPmain:^ SUBJ =!,! CASE =c DAT;

@SAP
@OBLIQUE
KPmain: ! CASE =c NOM, ^ OBJ =!;
VPraha : ^ =!,! _SUBJ_CASE =c DAT, ^ TNS_ASP MODAL =c NONE,^ TNS_ASP NEED =c NEG,
^ TNS_ASP PROG =c POS, ^ OBJ PERS =c ! PERS, ^ OBJ GEND =c ! GEND,
^ OBJ NUM =c ! NUM, ^ OBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c DAT;
@SAP
@OBLIQUE
KPmain: ! CASE =c NOM, ^ OBJ =!;
VPmodal : ^ =!,! _SUBJ_CASE =c DAT, ^ TNS_ASP MODAL =c ! TNS_ASP MODAL,
^ TNS_ASP NEED =c NEG, ^ OBJ PERS =c ! PERS, ^ OBJ GEND =c ! GEND,
^ OBJ NUM =c ! NUM, ^ OBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;
@SAP
@OBLIQUE
(KPmain: ! CASE =c ACC,! SEM_TYPE =c HUMAN, ^ OBJ =!; )
VPmodal : ^ =!,^ TNS_ASP MODAL =c ! TNS_ASP MODAL,^ TNS_ASP NEED =c NEG,
^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND, ^ SUBJ NUM =c ! NUM,
^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.
Sdec -> KPmain:^ SUBJ =!,! CASE =c NOM;
@SAP
@OBLIQUE
(KPmain: ! CASE =c NOM,! SEM_TYPE =c {UNANIM_CONC,ANIMAL,ABSTRACT}, ^ OBJ =!; )
VPmodal : ^ =!,^ TNS_ASP MODAL =c ! TNS_ASP MODAL,^ TNS_ASP NEED =c NEG,
^ SUBJ PERS =c ! PERS, ^ SUBJ GEND =c ! GEND, ^ SUBJ NUM =c ! NUM,
^ SUBJ RESPECT =c ! RESPECT, !VOICE =c ACTIVE;.

Frequency: Description: This rule shows the sentence level production of Urdu Grammar.
c-structure: Sentence consists of Case Phrases (KP) and a Verb Phrase (VP). Sentence level adjunct phrase
(@SAP) and oblique (@OBLIQUE) are optional and handled through macro.
f-structure: The Case Phrases can be Subject or Predlink according to their Case. The adjunct phrase and Xadjunct
phrase act as adjuncts of a sentence.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Example:

Rule Status: Active

Reference: [1] Butt and King, The Status of case
[2] Tara Mohanan, Argument Structure in Hindi Page 55.
[3] Rajesh Bhatt, Survey, Dative Subject, Passivization.
http://web.mit.edu/rbhatt/www/24.956/
[4] Mirium Butt, Discussion at EGD_ULP meetings.

Related Rules:
Related POS:
Replaces: - UGR001
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis 1:
Urdu is free-order language. It means that phrases can change their place in a sentence. We have analyzed only
the most commonly used phrase order of Urdu.
On the basis of Verbal Phrase structure and CASE of Subject, Urdu Sentences can be divided into following types:

Usual Sentences: It is the most commonly used type of Urdu Sentences. It consist of default behavior of all its
constituent i.e. Subject is in Nominative case, and any verb can be part of the Complex Predicate. Example
sentences are:

Perfective Sentences: Urdu like other South Asian language has special syntax for Perfective Sentences. The
simplest model for this type of sentence is that Transitive Verbs requires Ergative Subject and Intransitive verbs like
Nominative Subject. For Example,
Ergative Subject:

Nominative Subject:

But there are few examples in which the above rule fails. For Example,
-

Hence the Rule is that: if Subject of a Verb has control on action, then it will have Ergative Case, Otherwise it Verb
will have Nominative case. We denote this property by PerfControl feature.

Predicate Sentences: The sentences having fail-e-naqis ( copular verb) has a subject and its modifier(called
PREDLINK in LFG). There are two types of PREDLINK sentences.
()
()
(3)

is SUBJ and

In sentence 1,

is PREDLINK. In sentences 2 and 3, Grammatical Function identification is

disputed. Tara Mohanan says that is SUBJ in sentence 3 and

sentence 2,

is case marker. But the problem is that in

becomes the SUBJ. Both sentences 2 and 3 are same except the difference in phrase order. This

is why, we inferred that 2 is commonly order of sentence 3, and we will analyze both the sentences in similar way.
i.e.

is SUBJ and

is PREDLINK. Another argument in favor of this analysis is that there are many

similar sentences that has postposition and nominal postposition phrase at start of the sentence. For example,

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

If we consider as Case Marker then we have to classify about all the postpositions and nominal postpositions as
Case Marker.
Following sentence is an example of sentence having PredLink. It has Genitive Subject (gari) and a Predlink
(paihay).

Passive Sentences: Passive constructions in Urdu involves Auxiliary verb Ja. The Subject of passive sentence
(usually Logical Object) is in Nominative form. For Example:
(4)An alternate analysis is that saib is Object in above sentence, and a NULL subject is present in the sentence. This
analysis says that larka is Subject and saib is Object in following sentence.
(5)

The problem with first sentence is the multiple usage of word ja. It is used as Passive Auxiliary as in (4), but it is
also used as Ability Auxiliary and Habitual Auxiliary as in following examples.

Ability
Habitual

Sentence (5) is similar to Ability sentence. Hence we will analyze it as different type of sentence. Sentence (6) is a
passive sentence, but ke zarye is used as an ADJUNCT. Historically ke zarye examples are not present in
traditional text. It is introduced after interaction with English in which by Somebody is present in passive sentence.

Ability Sentences: The syntax of Ability sentences is similar to Passive Sentences, but these sentences require
Dative Subject. The sentences shows ability/disability of Subject to perform the action.
Dative Subject Sentences: The verbs of these sentences require a Dative Subject. Dative subject is required
due to different types of Verbal Phrases. Modal verbs like chahiye and par need dative subject as:

Similarly verb ho has a special usage which needs dative subject. For example:

The verb ho named as hay_dat is different from the verb ho in following examples.

Other verbs like dikh and lag also needs Dative Subject.
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

In Summary, Subjects can be Nominative, Dative, Genitive and Ergative. Object can be Nominative or Accusative.
The transitive Verb can be used in a sentence without its object. For example:

Analysis 2:
This was observed that resolving the type of sentence on the basis of syntax in the early stages is more efficient
and effective for the application; therefore, some checks have been introduced at sentence level (which were being
performed at some deeper level in the parse tree) to prune the huge production-set to the most suitable subset, as
early as possible.
Following are the major types of sentences on the basis of Verb Phrase:
Perfective (VPperf), Non-Perfecetive (VPnonperf), Progressive (VPraha), Modal (VPmodal), Infinitival (VPinf);

VPperf: There are subtle differences among multiple productions of VPperf:

CASE
Subject
Nominative

Object
Accusative

Nominative

Nominative

Ergative

Nominative

Ergative

Accusative

Dative

Nominative

Example

/
/

Everything else is more or less same in all productions of VPpref.

VPnonperf: There are subtle differences among multiple productions of VPnonperf:

CASE

Example

Subject
Nominative

Object
Accusative

Nominative

Nominative

Dative

Nominative

Everything else is same in all productions of VPnonpref.

EGD Urdu Localization Project - Grammar Rule Document

/
/

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

VPraha: There are subtle differences among multiple productions of VPraha:

CASE
Subject
Nominative

Object
Accusative

Nominative

Nominative

Dative

Nominative

Example

Everything else is same in all productions of VPraha.

VPmodal: There are subtle differences among multiple productions of VPmodal:

CASE
Subject
Nominative

Object
Accusative

Nominative

Nominative

Dative

Nominative

Example

/
/

Everything else is same in all productions of VPmodal.

Infinitival Agreement
Gender agreement of infinitival verb with its embedded object is dependant on its position in the sentence. When
ever the subject or the object is an infinitival occurring in nominative case this agreement varies. When an infinitival
act as an object and its subject is in ergative, dative or instrumental case then this agreement is positive. For other
cases it is negative. Reasons behind these agreements are discussed in detail in the document UGR104.

Result: We decided on analysis 2.

Future Work:
VPchahiyay

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

13 Sep, 2004

4.1.0.1

created.

Tafseer Ahmed

30 Oct, 2004

4.1.1.1

Rule modified for VPinf.

th

Sara Hussain

7 Feb, 2005

4.1.1.2

Rule modified

Aasim Ali

08-AUG-2005

5.1.0.1

Documentation of changes
made in UGR file by Tafseer

Rule ID: UGR104

Rule Syntax: Following is the constituent description of the rule.
VPinf -> (KP) pro v

Rule Functional Description: Following are the functional specifications of the rule.
VPinf -> ([KP:^OBJ=!, ! SEM_TYPE =c {UNANIM_CONC,ANIMAL,ABSTRACT}, !CASE =c NOM;
|
KP:^OBJ=!, ! SEM_TYPE =c HUMAN, !CASE =c ACC;])
pro:^SUBJ = !;
v:^=!, !_MORPH_FORM =c INFINITIVE, !_INFL_AGREEMENT = c NEGATIVE, ^GEND =c M, ^NUM=SG; .

Frequency: Description: This rule models infinitival clause. Infinitival clauses are clauses ending with a verb in infinitival form.
c-structure: Infinitival clause consists of optional Case phrase (KP), compulsory pronoun and infinitival verb.
f-structure: This rule handles only non-agreeing infinitival form as shown in the column Dialect 2 of the table given
below in the Analysis section.

Example:

]
]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: 'Long Distance Agreement in Hindi-Urdu' by Rajesh Bhatt
Related Rules:
Related POS:
Replaces: - UGR001
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis 1:
There are three aspects of infinitival clauses that have been analyzed separately: (1) The position of infinitival
clause in a sentence, (2) The occurrence of infinitival verb in Nominative and Oblique case, and (3) the gender
agreement between the infinitival verb and its embedded object in an infinitival clause.
(1) It has been decided that infinitivals are verbs which have nominal behavior. That is they occur as either subjects
or objects in a sentence depending on their position in a sentence. For this reason rule for infinitival clause is called
from the main noun rule (i.e. UGR106 NPnoun rule). Some occurrences of infinitivals in different nominal positions
are given below.

(as a specifer in genitive phrase)

(genitive phrase)

(nominal post position phrase)

(before case marker)

(in wala phrase)

However infinitival verbs like verbs have sub-categorization frame but they lack a subject argument. They can
further have additional description in form of adjuncts (as shown by KarP and S_adjunct in the above rule).
(2) Infinitival verb can occur in Nominative or Oblique case. Nominal case rules govern change in case of infinitival
verbs. Thus an infinitival verb occurring before a case marker or a post position is in oblique form. Similarly other
such rules are also followed by infinitival verbs. Different forms of the infinitival verb

are given below. This

information about cases can be embedded in the lexicon.

Masculine Singular Nominative:
Masculine Singular Oblique:

Masculine Plural Nominative/Oblique:

Feminine Singular/Plural Nominative/Oblique:

(3) In this case we are determining the presence or absence of gender agreement between the infinitival verb and
its embedded object in an infinitival clause. Speakers show variation in this regard. That is some speakers prefer
gender agreement in cases where other speakers prefer no gender agreement. While both the sentences below
seem correct to a group of speakers.

Gender agreement between feminine object (

) and Infinitival verb (

Absence of gender agreement between feminine object (

).

) and Infinitival verb (

Variation and similarity in dialect in different sentence structure can be seen in the following table.

EGD Urdu Localization Project - Grammar Rule Document

).

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Ser.
No.
1

Dialect 1

Gender
Agreement
no agreement

6
7
8

Agreement

no agreement

no agreement

Agreement
no agreement

no agreement

no agreement

no agreement

agreement

no agreement

Agreement

Gender Agreement

no agreement

Dialect 2

no agreement
agreement
no agreement

no agreement

no agreement

10

no agreement

no agreement

11

no agreement

no agreement

12

no agreement

no agreement

It has been noticed that most speakers prefer the dialect 1. The analysis given below tries to model this dialect.
We agree with Marium Butt's analysis of infinitives which introduces two independent instances of local agreement.
The first instance involves agreement between object of the embedded infinitival clause and the infinitival verb and
the second instance involves agreement between the infinitival clause and the matrix (outer main) verb [1]. The
second agreement is the same as between an object and the main verb in which the main verb agrees with the
object only when the subject is blocked (i.e. subject is not in nominative form). So for infinitivals we only consider
the agreement between object of the embedded infinitival clause and the infinitival verb.
As a rule the infinitival verb does not agree with its object when the infinitival clause is followed by case markers,
post-positions, nominal post position or the word 'wala'. Also whenever the infinitival clause occurs as a subject this
agreement does not hold (sentence 8-12). Thus the presence or absence of this agreement is questionable only
when the infinitival clause occurs as an object in nominative case.
From the above sentences it can be deduced that the agreement holds when the sentence having infinitival object
has a blocking subject. Thus when an infinitival act as an object and its subject is in ergative, dative or instrumental
case then this agreement is positive. For other cases it is negative.
Sentence number 11 and 12 in the above table indicates absence of gender agreement plus it shows absence of
gender in the infinitival phrase. The absence of gender is indicated by the fact that the pred-link (

)does

not agrees with the subject (sentence 12). In general it can be said that whenever gender agreement is absent
gender is missing. To model this lexicon can have separate entries for infinitivals having positive gender agreement

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

(like Masculine

and Feminine

) and for the ones that have negative gender agreement. Entries with

negative agreement will have no gender feature in them (like

without any gender feature).

Result: We decided on analysis 1 & 2 above.

Future Work: To generate all accent variation.
.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

23 Sep, 2004

4.1.0.1

Created

Sara Hussain

16 Dec, 2004

4.1.0.2

Sara Hussain

7th Feb, 2005

4.1.0.3

Sara Hussain

22nd Jan, 2005

4.1.0.4

Nayyara Karamat

07-Jul-2005

5.1.0.1

Tafseer Ahmed

07-Jul-2005

5.1.0.2

Aasim Ali

13-Jul-2005

5.1.0.3

Aasim Ali

16-Jul-05

5.1.0.4

rule modified for compatibility

with grammar file (changed
name of AdjP to AdjPmain)
Modified to add infinitival rule
(VPinf)
Modified to add checks for
Deverbal
Current NPnoun checked
and rectified
Current NPnoun checked
and rectified
WP (wala phrase) is shifted in
the start (right after GP)
AdjPmain is changed to AdjP*
Changed GP to GPmain

Rule ID: UGR106

Rule Syntax: Following is the constituent description of the rule.
NPnoun -> (GPmain) (WP) (DeverbalVP ) (dem) ([QuantP | NumberP] ) [AdjP]* (n) [n | VPinf].
Rule Functional Description: Following are the functional specifications of the rule.
NPnoun -> (GPmani:^ SPEC GEN=!, ^NUM = !NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT,
^FORM = ! FORM ;)
(WP: ^ SPEC WAL = ! , ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^
FORM = ! FORM ;)
(DeverbalVP: ^ SPEC DVER =!; )
(dem: ^ SPEC DEM = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ FORM = ! FORM;)
([QuantP: ^ SPEC QUANT = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = !
RESPECT , ^ FORM = ! FORM;
| NumberP: ^ SPEC NUMBER = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT
, ^ FORM = ! FORM;])
(AdjP: ^ ADJUNCT ADJ = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = !
RESPECT , ^ FORM = ! FORM ; )
[n: ! $ ^ ADJUNCT MOD_N, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^ FORM
= ! FORM;]#0#3
[ n: ^=!, ^ PERS = 3, !_INFL_AGREEMENT = NA; | VPinf:^=!, ^ PERS = 3, ^_INFL_AGREEMENT =
{NEGATIVE,POSITIVE}, ^ SEM_TYPE = ABSTRACT;].

Frequency: 1
Description: The rule shows the Noun Phrase having noun with its modifiers and specifiers.
c-structure: Noun Phrase can have Genitive Phrase, Deverbal Phrase, Determiner, Number, Demonstrative,
Quantifier Phrase, Number Phrase, Adjective Phrase nodes with compulsory Noun or infinitival node.
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

f-structure: Features present in any of the daughter nodes are copied to NP.

Example:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under-process

Reference: [1] Javaid, Ismat

1985, 2nd Ed
[2] Platts, John T
A Grammar of the Hinsustani or Urdu Language
1909, 5th Ed.

Related Rules: UGR008, UGR010, UGR011, UGR013, UGR016

Replaces: - UGR008, UGR010, UGR011, UGR013, UGR016, UGR104
Reason: - NPcommon and NPproper are merged. Previous rules have hierarchal structure, which are
merged into a flat structure rule.
Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis 1: The previous analysis discussed in UGR008, 010, 011, 013, 016 that has different rules for Common
and Proper Noun.
Analysis 2: Common and Proper noun are different in semantic properties, but their usage in sentence is similar.
The Noun Phrase consists of a noun that is head word. It has modifiers and specifiers that describe it. Genitive
Phrase, Deverbal Phrase, Demonstrative, Quantifier, Number and Wala phrase are used to specify the noun.
Adjectives and Noun acting as adjective are used to describe properties of the head noun. The order of these
modifiers/specifiers in a noun phrase can change but it has a more frequently used order.
Infinitivals have been modeled as noun since they occur in nominal positions and can take same modifiers (GP,
adjectives, deverbals etc.) as nouns do. For this reason infinitival phrase has been placed in this rule.
WP (Wala Phrase) can handle translation of English phrase like Top 3 books as

. This also works for

position (i.e, right after GP) in the Noun.

Long list of adjective (without any separator), as in

, and thus with every noun before

in this

, caused the

use of AdjP* in place of AdjPmain. Experiencing the problem that * allows only 0 to 3 occurences changed the * to
become #0#7 instead. Though in this example is not a pair of simple adjectives (rather it holds the sense
of numerous as well) yet they are left as it is for the sake of reducing ambiguity / time-complexity for Generator.

Result: We decided on analysis 2.

Future Work:
Examples as given below are not included in current analysis because they have not been encountered yet while
translating from English, and they are complicated enough that may cause a marginal delay in response time of the
Generator:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

25 Oct, 2004

4.1.0.1

Created

Tafseer Ahmed

4th Nov, 2004

4.1.0.2

F-Description are added to

cover various types of
VerbalP.

Rule ID: UGR109

Rule Syntax: Following is the constituent description of the rule.
VerbalP -> ComplexP (TenseAuxP).
VerbalP -> ComplexP asp_aux (TenseAuxP)
VerbalP -> ComplexP TenseAuxP
VerbalP -> ComplexP

Rule Functional Description: Following are the functional specifications of the rule.
VerbalP -> ComplexP:^ = !, ~[! _MORPH_FORM =c PERFECTIVE], ^ VTYPE = Usual,! _SUBJ_CASE =c NOM, ^
_MORPH FORM = ! _MORPH_FORM; (TenseAuxP:^ TNS_ASP = ! TNS_ASP,^ _MORPH FORM = ! _MORPH
ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND;).
VerbalP -> ComplexP:^ = !, ~[! _MORPH_FORM =c PERFECTIVE], ^ VTYPE = ERG,! _SUBJ_CASE =c ERG, ^
_MORPH FORM = ! _MORPH_FORM; (TenseAuxP:^ TNS_ASP = ! TNS_ASP, ^ _MORPH FORM = ! _MORPH
ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND;).
VerbalP -> ComplexP:^ = !, ^ VTYPE = Usual,! _SUBJ_CASE =c NOM; asp_aux:^ TNS_ASP = ! TNS_ASP,^
_MORPH_FORM =c ! _ALLOWED_FORM, ^ _MORPH FORM = ! _MORPH_FORM, ^ NUM = ! NUM, ^ RESPECT
= ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND; (TenseAuxP:^ TNS_ASP = ! TNS_ASP, ^ _MORPH FORM =
! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = !
GEND;).
VerbalP -> ComplexP:^ = !, ^ VTYPE = ERG,! PERF_CONTROL = c POS; asp_aux:^ TNS_ASP = ! TNS_ASP,^
_MORPH_FORM =c ! _ALLOWED_FORM, ^ _MORPH FORM = ! _MORPH_FORM, ^ NUM = ! NUM, ^ RESPECT
= ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND; (TenseAuxP:^ TNS_ASP = ! TNS_ASP, ^ _MORPH FORM =
! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = !
GEND;).
VerbalP -> ComplexP:^ = !, ~[! _MORPH_FORM =c PERFECTIVE], ^ VType = DAT,! _SUBJ_CASE =c DAT, ^
_MORPH FORM = ! _MORPH_FORM; (TenseAuxP:^ TNS_ASP = ! TNS_ASP,^ _MORPH FORM = ! _MORPH
ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND;).
VerbalP -> ComplexP:^ = !, ^ VType = DAT,! _SUBJ_CASE =c DAT; asp_aux:^ TNS_ASP = ! TNS_ASP,^
_MORPH_FORM =c ! _ALLOWED_FORM, ^ _MORPH FORM = ! _MORPH_FORM, ^ NUM = ! NUM, ^ RESPECT
= ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND; (TenseAuxP:^ TNS_ASP = ! TNS_ASP, ^ _MORPH FORM =
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = !

GEND;).
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VTYPE = Usual,! _SUBJ_CASE =c NOM, ^
_MORPH FORM = ! _MORPH_FORM; TenseAuxP:^ TNS_ASP TENSE = PAST, ^ TNS_ASP FAR_TYPE = !
FAR_TYPE, ^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^
PERS = ! PERS, ^ GEND = ! GEND;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VTYPE = Usual,! _SUBJ_CASE =c NOM, ^
_MORPH FORM = ! _MORPH_FORM; TenseAuxP:! TNS_ASP TENSE =c FUTURE, ^ TNS_ASP = ! TNS_ASP,^
_MORPH FORM = ! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS,
^ GEND = ! GEND;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VTYPE = Usual,! _SUBJ_CASE =c NOM, ^
_MORPH FORM = ! _MORPH_FORM,^ TNS_ASP TENSE = PAST, ^ TNS_ASP FAR_TYPE = NEUTRAL;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VTYPE = ERG,! _SUBJ_CASE =c ERG, ^
_MORPH FORM = ! _MORPH_FORM; TenseAuxP:^ TNS_ASP TENSE = PAST, ^ TNS_ASP FAR_TYPE = !
FAR_TYPE, ^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^
PERS = ! PERS, ^ GEND = ! GEND;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VTYPE = ERG,! _SUBJ_CASE =c ERG, ^
_MORPH FORM = ! _MORPH_FORM; TenseAuxP:! TNS_ASP TENSE =c FUTURE, ^ TNS_ASP = ! TNS_ASP, ^
_MORPH FORM = ! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS,
^ GEND = ! GEND;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VTYPE = ERG,! _SUBJ_CASE =c ERG, ^
_MORPH FORM = ! _MORPH_FORM,^ TNS_ASP TENSE = PAST, ^ TNS_ASP FAR_TYPE = NEUTRAL;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VType = DAT,! _SUBJ_CASE =c DAT, ^
_MORPH FORM = ! _MORPH_FORM; TenseAuxP:^ TNS_ASP TENSE = PAST, ^ TNS_ASP FAR_TYPE = !
FAR_TYPE,^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^
PERS = ! PERS, ^ GEND = ! GEND;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VType = DAT,! _SUBJ_CASE =c DAT, ^
_MORPH FORM = ! _MORPH_FORM; TenseAuxP:! TNS_ASP TENSE =c FUTURE, ^ TNS_ASP = ! TNS_ASP, ^
_MORPH FORM = ! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS,
^ GEND = ! GEND;.
VerbalP -> ComplexP:^ = !, ! _MORPH_FORM =c PERFECTIVE, ^ VType = DAT,! _SUBJ_CASE =c DAT, ^
_MORPH FORM = ! _MORPH_FORM,^ TNS_ASP TENSE = PAST, ^ TNS_ASP FAR_TYPE = NEUTRAL;.

Frequency: 1
Description: This rule shows main rule for verb phrase.
c-structure: Verb phrase consists of a complex predicate phrase followed by optional aspectual auxiliaries and
tense auxiliary phrase.
f-structure: Complex predicate phrase, auxiliary and tense auxiliary have agreement on the basis of number,
gender, person and morphological form of verb.
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Examples:
[

Rule Status: Under Process

Reference:
[1] UGR103
[2] Miriam Butt, Discussion at EGD_ULP meetings

Related Rules: UGR116, UGR024

Related POS: UPOS104
Replaces: - UGR020
Reason: - UGR020 was a general rule for all types of VP.
Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: The Usual Verb Phrase consists of Complex Predicate, Optional Auxiliary Phrase and Tense Auxiliary
Phrase. In syntax of commonly used Urdu, only one aspectual auxiliary is used after Complex Predicate.
Morphological form of each node depends on form of the next node. For example,

and

allow

PERFECTIVE or HABITUAL form. allows SUBJUNCTIVE form before it. The Complex Predicate provides the
Sub-Categorization frame for the phrase. The case of subject is provided by last main or light verb of the phrase.

Result: We decided on above analysis.

Future Work: Verbal Phrase should deal all possible combinations of Aspectual auxiliaries so the sentences like

should be parsed successfully.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

25 Oct, 2004

4.1.0.1

Created

Tafseer Ahmed

4th Nov, 2004

4.1.0.1

Sara Hussain

20th Dec, 2004

4.1.0.2

F-Description is added to the

rule.
Renamed FORM feature of
asp_aux to _AUX_FORM

Rule ID: UGR110

Rule Syntax: Following is the constituent description of the rule.
VerbalPpassive -> ComplexP asp_aux (asp_aux) (TenseAuxP)

Rule Functional Description: Following are the functional specifications of the rule.
VerbalPpassive -> ComplexP:^ = !; asp_aux: [!_AUX_FORM =c 'passive_ja'],^ TNS_ASP = ! TNS_ASP,^
_MORPH_FORM =c ! _ALLOWED_FORM, ^ _MORPH FORM = ! _MORPH_FORM,^ NUM = ! NUM, ^ RESPECT
= ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND; (TenseAuxP:^ TNS_ASP = ! TNS_ASP, ^ _MORPH FORM =
! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = !
GEND;).
VerbalPpassive -> ComplexP:^ = !; asp_aux: [!_AUX_FORM =c 'passive_ja'],^ TNS_ASP = ! TNS_ASP,^
_MORPH_FORM =c ! _ALLOWED_FORM, ^ _MORPH JA_FORM = ! _MORPH_FORM,^ NUM = ! NUM, ^
RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND; asp_aux:^ TNS_ASP = ! TNS_ASP, ^ _MORPH
JA_FORM =c ! _ALLOWED_FORM, ^ _MORPH FORM = ! _MORPH_FORM, ^ NUM = ! NUM, ^ RESPECT = !
RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND; (TenseAuxP:^ TNS_ASP = ! TNS_ASP,^ _MORPH FORM = !
_MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND;).

Frequency: 1
Description: This rule shows main rule for verb phrase.
c-structure: Verb phrase consists of a complex predicate phrase followed by optional aspectual auxiliaries and
tense auxiliary phrase.
f-structure: Complex predicate phrase, auxiliary and tense auxiliary have agreement on the basis of number,
gender, person and morphological form of verb.

Examples:
[
[

]
]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: UGR103, UGR109
Related Rules: UGR116, UGR024
Related POS: UPOS013
Replaces: - UGR020
Reason: - UGR020 was a general rule for all types of VP.
Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: Passive Sentence has a special syntax that is different from syntax of Usual Verb Phrase. It has
Aspectual Auxiliary passive_ja after Complex Predicate. The form of passive_ja is determined by tense and
aspect of the sentence. The Agreements in this kind of Verbal Phrase is same as discussed in UGR109.

Result: We decided on above analysis.

Future Work: Sentences having more than two aspectual auxiliaries will be modeled.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

25 Oct, 2004

4.1.0.1

Created

Tafseer Ahmed

4th Nov, 2004

4.1.0.2

F-Description is added to the

rule.

Rule ID: UGR112

Rule Syntax: Following is the constituent description of the rule.
VerbalPchahiye -> v (TenseAuxP)

Rule Functional Description: Following are the functional specifications of the rule.
VerbalPchahiye -> v: ^=!, !FORM =c 'chahiye'; (TenseAuxP:~[! TENSE =c PRES],^ TNS_ASP = ! TNS_ASP, ^
_MORPH FORM = ! _MORPH ALLOWED_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS,
^ GEND = ! GEND;).

Frequency: 1
Description: This rule shows main rule for verb phrase.
c-structure: Verb phrase consists of verb chaihye followed by optional tense auxiliary phrase.
f-structure: verb and tense auxiliary have agreement on the basis of number, gender, person and morphological
form of verb.

Examples:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: UGR103, UGR109
Related Rules: UGR024
Related POS: UPOS013, UPOS014
Replaces: - UGR020
Reason: - UGR020 was a general rule for all types of VP.
Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: Chahiye is a special modal of Urdu. It do not allow aspectual auxiliaries after it. The verb chahiye in this
context has only one morphological form i.e. chahiye used in all contexts. The behavior of this modal is different
from verb chah, so it is considered as a separate lexical entry with seperate predicate. The Agreements in this
kind of Verbal Phrase is same as discussed in UGR109.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

25 Oct, 2004

4.1.0.1

Created

4.1.0.2

rule modified for compatibility

with grammar file (square
brackets added) Modified
tns_aux to tense_aux and
FORM of tense_aux to
_AUX_FORM

Sara Hussain

16 Dec, 2004

Rule ID: UGR114

Rule Syntax: Following is the constituent description of the rule.
VerbalPhay -> [ v | [v tense_aux]]

Rule Functional Description: Following are the functional specifications of the rule.
VerbalPhay -> [ v: ^ = !, ! _VERB_FORM =c hay_dat; | [v: ! _VERB_FORM =c hay_dat; tense_aux: !
_AUX_FORM =c ga;]].

Frequency: 1
Description: This rule shows main rule for verb phrase.
c-structure: Verb phrase consists of a verb.
f-structure:.

Examples:
[

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: UGR103, UGR109
Related Rules:
Related POS: UPOS013, UPOS016
Replaces: - UGR020
Reason: - UGR020 was a general rule for all types of VP.
Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: DATIVE hay is a special verb that need dative subject. It is different from Copula hay that needs
nominative Subject and Predlink. The forms of hay_dat is hay(all forms), tha(all forms) and ho(followed by
tense auxiliary ga).

Result: We decided on above analysis.

Future Work

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

25 Oct, 2004

4.1.0.1

Created

Sara Hussain

20th Dec,2004

4.1.0.2

Modified tns_aux to
tense_aux and FORM of
tense_aux to _AUX_FORM

Rule ID: UGR115

Rule Syntax: Following is the constituent description of the rule.
VerbalPpred -> [ v | v tense_aux]

Rule Functional Description: Following are the functional specifications of the rule.
VerbalPpred -> [ v: ^ = !, ! FORM =c hay_pred; | [v: ! FORM =c hay_pred; tense_aux: ! _AUX_FORM =c ga;] ]

Frequency: 1
Description: This rule shows main rule for verb phrase.
c-structure: Verb phrase consists of a verb node and an optional tense auxiliary.
f-structure:

Examples:
[

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules:
Related POS: UPOS013, UPOS016
Replaces: - UGR026
Reason: - New Rules are introduced for all Verb Phrases.
Replaced by: -

Analysis: Following is the in-depth analysis of the rule.

Analysis: Predicative or Copular hay is a special verb that needs nominative subject. It is different from Dative
hay that needs dative Subject. The forms of hay_pred are hay(all forms), tha(all forms) and ho(followed by
tense auxiliary ga)

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

27 Oct, 2004

4.1.0.1

Created

Tafseer Ahmed

4th Nov, 2004

4.1.0.2

Sara Hussain

23 Feb, 2005

4.1.0.3

Sara Hussain

3 March, 2005

4.1.0.4

Aasim Ali

08-AUG-2005

5.1.0.1

F-Description is added to the

rule.
Added check so that
deverbals do not pass
through this rule
Added rule for verbal
adjective/adverb
Documentation of changes
made in UGR file by Nayyara

Rule ID: UGR0116

Rule Syntax: Following is the constituent description of the rule.
ComplexP -> v
ComplexP -> v_n ltv
Rule Functional Description: Following are the functional specifications of the rule.
ComplexP -> v: ^ =!;.
ComplexP -> v_n: ^ = !; ltv: ^ = !;.

Frequency: 1
Description: This rule shows rule for main verb phrase.
c-structure: main verb phrase consists of a simple verb or light verb with some adjective or noun.
f-structure: light verb is head of phrase and noun or adjective coming before it becomes DPRED.

Examples:

[ ]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
[1] UPOS014(Light Verb Document )
[2] Miriam Butt, Discussion at EGD_ULP Meetings

Related Rules:
Related POS: UPOS013, UPOS015, UPOS017
Replaces: - UGR023
Reason: - Structure of Verbal Phrase is made flatter. Previously auxiliaries were part of ComplexVP.
Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is of the rule.

Analysis 1:
In Urdu, Complex Predicate can be a simple main verb or it can be a Noun-Verb or Adj-Verb Complex. In second
case, Noun/Adjective followed by a light verb like hona, karna, aana etc. There are two issues regarding
Complex Predicate.
First issue is its definition. It is discussed in UPOS014 that aspectual auxiliaries are not part of Complex Predicate.
For example, In

parta is not a light verb.

The Complex Predicate involves karna and hona etc. as light verbs. But all instances of these verbs are not light
verbs. For example, karna in following sentence is not a light verb.
-

In this sentence, kaam is object and kiya is the main verb. Because of this reason, Form of the verb is depending
on the Object. If we use a feminine object, the form of verb will be changed. For example:

On the other hand, yaad karna is a complex predicate because it requires Subject and Object and it agreement of
karma light verb depends on the Object and not on yaad. For example:

If we introduce an accusative object, even then the light verb karna will not agree with feminine yaad but it will
have default masculine agreement.

It is important to note that yaad does not always involve in a complex predicate. For Example,

is not a complex predicate. Because form of verb aana is depending on the Object meri yaad.
The second issue is of Subcategorization Frame and Agreement requirements. In case of a single verb (for
example hay_dat as discussed in UGR11x, the verb tells that it needs Subject and Object (in Subcategorization
Frame). Subject will be dative and the object will be Nominative in all tenses. For example:

hamid is Subject and bukhaar is Object.

But in case of Complex Predicate, these responsibilities are divided between verbal_noun and the light verb. The
verbal noun has the SubCategorization information. Hence yaad tells that the complex predicate will require
Subject and Object. The light verb determines Subject and Object Agreement and verbal_noun has no contribution
in it. For example,

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Since deverbals are also verbs, an extra check has been added so that deverbals cannot pass through this rule.
For all deverbals feature DEVERBAL is present having the value of either NOUN or ADJECTIVE.
Analysis 2:
There was another solution to these N-V complex predicates. It was that to list all such combinations as separate
words. But it is not a proper solution because light verbs are very productive in making new complex verbs.

Result: We decided on analysis 1.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

24 Sep, 2004

4.1.0.1

Created

Aasim Ali

08-Jul-2005

Status Marked Deprecated

Rule ID: UGR118

Rule Syntax: Following is the constituent description of the rule.
NPnoun -> (GP) (DeverbalVP ) (dem) ([QuantP | NumberP] ) WP.
NPnoun -> (GP) DeverbalVP .

Rule Functional Description: Following are the functional specifications of the rule.
NPnoun -> (GP:^ SPEC GEN=!, ^NUM = !NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT, ^FORM =
! FORM ;)
(DeverbalVP: ^ SPEC DVER =!; )
(dem: ^ SPEC DEM = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^
FORM = ! FORM;)
([QuantP: ^ SPEC QUANT = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = !
RESPECT , ^ FORM = ! FORM;
| NumberP: ^ SPEC NUM = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = !
RESPECT , ^ FORM = ! FORM;])
WP: ^ = ! , ^ PERS = 3, ^ NFORM =c POS;
NPnoun -> (GP:^ SPEC GEN=!, ^NUM = !NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT, ^FORM =
! FORM ;)
DeverbalVP: ^ SPEC DVER =!, ^ PERS = 3, ^ NFORM =c POS;

Frequency: 1
Description: This rule shows pre-adjectival phrase.
c-structure: Noun Phrase can have Genitive Phrase, Deverbal Phrase, Demonstrative, Wala Phrase as nodes.
f-structure: Features present in any of the daughter nodes are copied to NP.

Examples:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process ----- It has been Deprecated, please see UGR106
Reference:
Related Rules:
Related POS: UPOS011
Replaces: Reason: Replaced by: - UGR106

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis 1: Usually Noun acts as head of Noun Phrase. In absence of noun demonstratives, Deverbals, quantifiers,
adjective etc. can act as noun (and head of noun phrase). For example,

In above examples, adjective, demonstrative and number act as noun.

Analysis 2: An alternate analysis of above phenomenon is that , ,are nouns. A point favoring the
analysis is that these words have noun morphology and do not show morphology of adjective or demonstrative etc.
For example, Plural Oblique of adjective is , but it is in above example that follows nouns rule. Hence,
it can be said that adjectives, demonstratives, numbers, quantifiers are noun also.
Deverbal and wala phrase, act as noun in the absence of noun head word.

Result: We decided on Analysis 2.

Future Work:
Genitive Phrase used as noun is not dealt here because its evidence is not available. The sentences like

seems to be unacceptable. But it requires search of this phenomenon in written text and spoken language.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Sara Hussain

16 Dec, 2004

4.1.0.2

rule modified for compatibility

with grammar/Lexicon file
(removed square brackets
and removed NUM,
RESPECT, GEND and
FORM checks from intens)

Aasim Ali

07-Jul-2005

4.1.0.3

Updated for multiple adverbs

Aasim Ali

15-Jul-05

5.1.0.4

Analysis about
added

Rule ID: UGR119

Rule Syntax: Following is the constituent description of the rule.
N_AdvP -> PP.
N_AdvP -> A_advP.
N_AdvP -> PP A_advP.
Rule Functional Description: Following are the functional specifications of the rule.
N_AdvP -> PP: ^ = !, !PFORM =c 'se', ^ NUM = {SG,PL}, ^ RESPECT =
{NORESPECT,FAMILIAR,USUAL,EXTRA}, ^ GEND = {M,F}, ^ FORM = {NOM,OBL};.
N_AdvP -> A_AdvP: ^ ADV=!, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ GEND = ! GEND,
^ FORM = ! FORM;.
N_AdvP -> PP: ^ = !, !PFORM =c 'se', ^ NUM = {SG,PL}, ^ RESPECT =
{NORESPECT,FAMILIAR,USUAL,EXTRA}, ^ GEND = {M,F}, ^ FORM = {NOM,OBL};
A_AdvP: ^ ADV=!, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ GEND = ! GEND,
^ FORM = ! FORM;.

Frequency:
Description: This rule shows Adverbial effects on Adjective used in Noun Phrase.
c-structure: Pre-Nominal Adverbial Phrase consists of a adverb or postpositional phrase having se. The adverb is
followed by optional intensifier.
f-structure: Pre-Nominal Adverbial Phrase usually becomes ADJUNT of an adjective.

Examples:
[

] [

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

[ ] [

[ ] [[

][

[ ] [[

[[[ ]

]]
]]

]]

Rule Status: Under Process

Reference: Discussion with Mr Tafseer and Ms Nayyara
Related Rules: UGR128, UGR146, UGR120
Related POS:
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis1: Adverbial Phrases have many functions in Urdu grammar. These can modify sentences, verbs or
adjectives. Pre-Nominal Adverbial Phrase is commonly used before adjective in a noun phrase. The adverb
commonly modifies the quality of adjective e.g.
postposition. Here se is used for comparison like
intensify the effect of adeverb.

. The postposition phrase can have only se as

. The intensifier e.g. hi is used to

Two adverbs may come together in a situation when 1st adverb (sub-adverb) tells the degree/intensity of the 2nd one
(main-adverb), like

. It is also observed that comparison (PP with pform=se), if exists in the

N_AdvP, occurs only in the start of the phrase. All elements of this rule are kept at parallel level because no single
element is dependent on the other, as:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis2:
In case of two adverbs, 1st one is qualifier of 2nd one. Thus the f-structure of phrase

be as discussed in the rule describing A_AdvP (UGR146).

In phrase like

, POS for

Same observation is noted for

(

) work as noun.

is analyzed to be noun because this position is nominal, as:

should

. In these examples the starting words

Result: We decided on above analysis1 in all respects except where Analysis2 overrides (more than one adverb).
Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Aasim Ali

25-Jul-05

5.1.0.1

Added a sub-rule AdjP1

Rule ID: UGR120

Rule Syntax: Following is the constituent description of the rule.
AdjP -> [ AdjP1 ] #1#4 AdjP1
AdjP -> AdjP1
AdjP1 -> N_AdvP adj
AdjP1 -> adj

Rule Functional Description: Following are the functional specifications of the rule.
AdjP1 -> N_AdvP:^ SPEC =!, ^NUM=!NUM, ^GEND = !GEND, ^RESPECT = !RESPECT , ^FORM = !FORM;
adj:^=!, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^ FORM = ! FORM,
^NDERIVED = !NDERIVED; .
AdjP1 -> adj:^=!, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^ FORM = ! FORM,
^NDERIVED = !NDERIVED; .
AdjP -> AdjP1: ^=!, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^ FORM = ! FORM; .
AdjP -> [AdjP1: !$^, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^ FORM = ! FORM; ] #1#4
AdjP1: !$^, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^ FORM = ! FORM; .

Frequency:
Description: This rule shows Adjectival Phrase used in Noun Phrase.
c-structure: Adjectival Phrase consists of an adjective that is followed by an optional Pre-Nominal Adverbial Phrase.
f-structure: Adjectival Phrase usually becomes member of ADJUNT ADJ set of the noun.

Examples:
[

[[

[]

]]

]]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules: UGR119
Related POS: UPOS110
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: Adjectival Phrase consists of an adjective and the adverbial phrase modifying it.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Sara Hussain

6th Jan, 2005

4.1.0.2

Sara Hussain

7th Feb, 2005

4.1.0.3

Aasim Ali

16-Jul-05

5.1.0.1

Modified to add adjunct

(adverb) cases
Modified to add gender
agreement check for
infinitivals in NPnoun
Removd adv, which was
added to accommodate
( their POS has
already been changed to
noun)

Rule ID: UGR121

Rule Syntax: Following is the constituent description of the rule.
GP -> [NPnoun | NPpronoun ] cm
GP -> gen_pro
Rule Functional Description: Following are the functional specifications of the rule.
GP -> [NPnoun: ^ GENOBJ =!, ! FORM =c OBL, !_INFL_AGREEMENT =c NEGATIVE ; | NPpronoun: ^ GENOBJ
=!, ! FORM =c OBL, ! CASE =c GEN;]
cm: ^=!, ! CASE = GEN;.
GP -> gen_pro: ^ =!;.

Frequency:
Description: This rule shows Genitive Phrase that is special type of Case Phrase.
c-structure: Genitive Phrase consists of an noun phrase/pronoun followed by genitive case marker or a genitive
pronoun in accusative/dative form.
f-structure: Genitive Phrase has case marker as head and the noun phrase is present as its object.

Examples:

[ ]
[
[

]
]
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules: UGR1017, UPOS106
Related POS: UPOS120, UPOS001
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis 1: Genitive Phrase is a special type of case phrase that can occur in a noun phrase. It acts as a specifier,
and specifies a known. For example, in

] , it is specified that book(kitaab) belongs to girl. Genitive

Pronoun is special type of pronoun that has Genitive case and do not need a case marker. We make a separate
phrase for Genitives, because unlike other case phrases, these modify noun and become its adjunct. Other modify
verbs and act as subject, object, obliq and adjunct etc. of it.
Adverbs which indicate place (direction) or time can also act as a specifier. For example in
specified that we are taking about the fields (khet) that exist here (in this place).
In the phrase

] , it is

], an infinitival clause precedes the genitive case marker ka. The gender agreement

is absent between the infinitival verb and its embedded object whenever the infinitival clause is followed by a case
marker. For this reason a check (~[!_INFL_AGREEMENT =c POSITIVE]) has been added to allow only negative
agreement to pass through this rule.
Analysis 2: Genitive Phrase is a case phrase, so a KP with genitive case as constraint should be used for it.

Result: We decided on analysis 1.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Sara Hussain

7th Feb, 2005

4.1.0.2

Aasim Ali

13-Jul-2005

5.1.0.1

Modified to add gender

agreement check for
infinitivals in NPnoun
Using NPmain instead of
NPnoun
Introducing AdjP and ordinal

Rule ID: UGR122

Rule Syntax: Following is the constituent description of the rule.
WP -> NPmain wala
WP -> AdjP wala
WP -> ordinal wala
Rule Functional Description: Following are the functional specifications of the rule.
WP -> NPmain: ^ WALOBJ = !, ! FORM =c OBL, !_INFL_AGREEMENT =c NEGATIVE;
wala: ^=!;.
WP -> AdjP: ^ WALOBJ = !, ^GEND=!GEND, ^NUM=!NUM, ^FORM=!FORM, ^RESPECT=!RESPECT;
wala: ^=!;.
WP -> ordinal: ^ WALOBJ = !, ^GEND=!GEND, ^NUM=!NUM, ^FORM=!FORM, ^RESPECT=!RESPECT;
wala: ^=!;.

Frequency:
Description: This rule shows Wala Phrase that is an adjunct of noun.
c-structure: Wala Phrase consists of a noun phrase followed by wala.
f-structure: Wala Phrase has wala as head and the noun phrase is present as its object.

Examples:
[]
[ ]
[

[ [

[ [

]]

]]

]][ [
[[

]]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: [1] UPOS119 (Wala POS Document)
[2] Javaid, Ismat

1985, 2nd Ed

Related Rules: UGR106

Related POS: UPOS119
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis : Wala Phrase is a specifier used to modify the noun. The phrase is unique as it has wala word as its
head. wala is a word that is also a POS too. See POS document of wala( UPOS119). The structure of wala
phrase is similar to Genetive phrase and its semantics are also some-what similar.
In the phrase [

], an infinitival clause precedes 'wala'. In such cases the gender agreement is absent

between the infinitival verb and its embedded object. Additional check (~[!_INFL_AGREEMENT =c POSITIVE]) has
thus been added to allow only negative agreement to pass through this rule.
Agreement of gender, number, form, and respect is observed only in case of Adjective-Wala phrase, like

so disagreement in this case is incorrect as

*.

In case of ordinal-Wala phrase, the agreement is optional (dpending upon the context), as in .
However, for restricting the over-generation, this agreement is also defined as required.
As far as NPmain-Wala phrase is concerned, there is no need/concept of agreement between the NPmain and the
following wala, as in:

Result: We decided on analysis.

Future Work:
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Rule ID: UGR123

Rule Syntax: Following is the constituent description of the rule.
PPmain -> PP
PPmain -> PP [[comma PP]]* coord_conj
Rule Functional Description: Following are the functional specifications of the rule.
PPmain -> PP: ^=!;.
PPmain -> PP: ! $ ^; [[comma:; PP: ! $ ^;]]* coord_conj:^=!; PP:! $ ^;.

Frequency:
Description: This rule represents the co-ordination of Postpositional Phrase.
c-structure: It consists of a single postpositional phrase or a list of comma separated postpositional phrases.
f-structure: A set of postpositional phrase having co-ordination type at root..

Examples:

Rule Status: Under Process

Reference:
Related Rules: UGR106
Related POS: UPOS121, UPOS122
Replaces: Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis : Postpositional Phrases can be conjuncted like other phrases. The co-ordination consists of a commaseperated PP list having a co-ordinating conjunction before last PP.

Result: We decided on analysis.

Future Work: Conjunction between disimiliar phrases e.g. Case Phrase and Postpositional Phrase, or Case
Phrase and Nominal Postpositional Phrase are to be modeled. For example,
[
[

[ ]
[ ]

] KP and NomPP

]PP and NomPP

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Sara Hussain

16 Dec, 2004

4.1.0.2

Aasim Ali

29-Jun-2005

5.1.0.1

Aasim Ali

18-Jul-2005

5.1.0.2

Aasim Ali

19-Jul-2005

5.1.0.3

Changed ~AUR to YA

Aasim Ali

28-Jul-05

5.1.0.4

Removed redundant checks

on Form, Case, Num, Gend
etc.

comments added for

compatibility with grammar
file and better understanding
Removed redundant rule for
coordinate conjuntion of 2nd &
3rd Person
Removed constraint for
PERSon

Rule ID: UGR124

Rule Syntax: Following is the constituent description of the rule.
KPmain -> KP
KPmain -> KP [[comma KP]]* coord_conj KP

Rule Functional Description: Following are the functional specifications of the rule.
KPmain -> KP:^=!; .
//AUR Conjuncted
KPmain -> KP: ! $ ^; [[comma:; KP: ! $ ^;]]* coord_conj:^=!, !CONJ_FORM =c AUR; KP: ! $ ^; .
//YA Conjuncted
KPmain -> KP: ! $ ^; [[comma:; KP: ! $ ^;]]* coord_conj:^=!, !CONJ_FORM =c YA; KP: ! $ ^; .

Frequency:
Description: This rule represents the co-ordination of Case Phrase.
c-structure: It consists of a case phrase or a list of comma separated case phrases.
f-structure: A set of case phrases having co-ordination type and agreement features at the root.

Examples:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules: UGR002
Related POS: UPOS121, UPOS122
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis : Case Phrase can be coordinated by using coordination conjunction. So a Case Phrase(KP) in sentence
can be a single case phrase or a list of comma separated case phrase having a coordinating conjunction before
last case phrase.
There are two issues in the coordination of KP. There is a favorable order of KPs depending upon their types. For

example,

is more acceptable than

. The natural order is that 1st person will be

followed by 2nd Person and that will be followed by third person.

The other issue is the features of coordinated KP. A case phrase(KP) has person, number, gender, form, case and
respect features. The verbal phrase agrees with all these features(except case). So the features for coordinated KP
as a whole should be known.
If coordination is made by using AUR, then it shows a collection which results in plural number. For example,

Similarly, we can say that respect feature will be usual because usual repect is equivalent to plural in Urdu syntax.
The Gender of AUR coordinated KP is masculine, because coordination of masculine KP and feminine KP results
in masculine KP(that is default gender of Urdu.) For example,

Co-ordination of KP has a condition that KPs of same form and case can be coordinated. For example, following
are not valid

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rules for these features are given in following table.
Conjunction Type
Number
Gender
Person
Form
Respect
Case

Aur
Plural
Masculine
1st then 2nd then 3rd
All KPs should have
same.
Usual
All KPs should have
same.

Others( e.g. ya)

Of last KP
Of last KP
Of last KP
All KPs should have
same.
Of last KP
All KPs should have
same.

The rules for coordination like above are not present in grammar books.The rules are extracted from a small survey
from native speakers. The rules for AUR coordination seems to be well established among these people. But for
others, the rules are confusing and more than one rules can exist. For example,

In both of the above examples, verb is agreeing with the last KP. But there are examples that are confusing fro the
speakers. For example, one can predict both aaya and aayen in following sentence.

The sentence is also confusing, because it seems that Urdu prefers singular phrases before plural ones. So a more
natural version of above sentence in Urdu will be

The sentence also show agreement of verb with last KP.

Constraint for person has been removed keeping in view to prefer the order of the original sentence in English.
All checks related to Form, Case, Num, Gend etc have been removed because they will be coming from English
and will fill the Urdu / Mapping f-structure accordingly.

Result: We decided on analysis.

Future Work: The Gender rule for AUR conjunction is an over-simplification, that says that gender of aur
conjuncted phrase is always Masculine. The rule is true for masc-masc and fem-masc conjunction, but it is not true
for fem-fem conjunction. New rules should be written to successfully parse sentences like

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Sara Hussain

16 Dec, 2004

4.1.0.2

Aasim Ali

29-Jun-2005

5.1.0.1

Aasim Ali

18-Jul-2005

5.1.0.2

Aasim Ali

19-Jul-2005

5.1.0.3

comments added for

compatibility with grammar
file and better understanding
Removed redundant rule for
coordinate conjuntion of 2nd &
3rd Person
Removed constraint for
PERSon
Changed ~AUR to YA

Rule ID: UGR125

Rule Syntax: Following is the constituent description of the rule.
NPmain -> NP
NPmain -> NP [[comma NP]]* coord_conj NP

Rule Functional Description: Following are the functional specifications of the rule.
NPmain -> NP: ^=!; .
//AUR Conjuncted
NPmain -> NP: ! $ ^; [comma:; NP: ! $ ^, ^ FORM = ! FORM;]* coord_conj:^=!, !CONJ_FORM =c AUR; NP: ! $ ^; .
//YA Conjuncted
NPmain -> NP: ! $ ^; [comma:; NP: ! $ ^, ^ FORM = ! FORM;]* coord_conj:^=!, !CONJ_FORM =c YA; NP: ! $ ^; .

Frequency:
Description: This rule represents the co-ordination of Noun Phrase.
c-structure: It consists of a noun phrase or a list of comma separated noun phrases.
f-structure: A set of noun phrases having co-ordination type and agreement features at the root.

Examples:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: UGR137(KP Coordination Document)
Related Rules: UGR138
Related POS: UPOS121, UPOS122
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis : Noun Phrase can be coordinated by using coordination conjunction. So a Noun Phrase (NP) can be a
single noun phrase or a list of comma separated noun phrases having a coordinating conjunction before last case
phrases.
The features of coordinated noun phrase are determined by the grammar rules. The rules are similar to the rules of
KP discussed in UGR137. The only except is that NP do not deal CASE feature. The feature rules are following:
Conjunction Type
Number
Gender
Person
Form
Respect

Aur
Plural
Masculine
1st then 2nd then 3rd
All NPs should have
same.
Usual

Others( e.g. ya)

Of last NP
Of last NP
Of last NP
All NPs should have
same.
Of last NP

Constraint for person has been removed keeping in view to prefer the order of the original sentence in English.

Result: We decided on analysis.

Future Work: (1) The Gender rule for aur conjunction is an over-simplification, that says that gender of aur
conjuncted phrase is always Masculine. The rule is true for masc-masc and fem-masc conjunction, but it is not true
for fem-fem conjunction. New rules should be written to successfully parse sentences like

(2) We assume that NP do not have CASE feature, the assumption is not true for pronouns having
accusative/dative case. For example,
phrase.

. It is required that these pronouns should be constrained in this

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

Aasim Ali

08-Jul-2005

5.1.0.1

Aasim Ali

15-Jul-05

5.1.0.2

Verified for correctness and

fixed by few minor changes
Modified to allow ONLY YA
( )conjunction.

Rule ID: UGR126

Rule Syntax: Following is the constituent description of the rule.
AdjPmain -> AdjP [[comma AdjP]]* coord_conj AdjP .
AdjPmain -> AdjP

Rule Functional Description: Following are the functional specifications of the rule.
AdjPmain -> AdjP: ! $ ^, ^NUM = !NUM, ^GEND = !GEND, ^FORM = !FORM, ^RESPECT = !RESPECT,
^NDERIVED = !NDERIVED;
[[comma:; AdjP: ! $ ^, ^NUM = !NUM, ^GEND = !GEND, ^FORM = !FORM, ^RESPECT = !RESPECT, ^NDERIVED
= !NDERIVED;]]*
coord_conj: ^=!, !CONJ_FORM =c YA; AdjP: ! $ ^, ^NUM = !NUM, ^GEND = !GEND, ^FORM = !FORM,
^RESPECT = !RESPECT, ^NDERIVED = !NDERIVED; .
AdjPmain -> AdjP:^ =!, ^NUM = !NUM, ^GEND = !GEND, ^FORM = !FORM, ^RESPECT = !RESPECT,
^NDERIVED = !NDERIVED; .

Frequency:
Description: This rule represents the co-ordination of Adjective Phrase.
c-structure: It consists of a adjective phrase or a list of comma separated adjective phrases.
f-structure: A set of adjective phrases having co-ordination type and agreement features at the root.

Examples:
[

]
[

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules: UGR120
Related POS: UPOS121, UPOS122
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis : Adjective Phrase can be coordinated by using coordination conjunction. So a Adjective Phrase (AdjP)
can be a single adjective phrase or a list of comma separated adjective phrases having a coordinating conjunction
before last adjective phrases. As features of adjectives are used for agreement in noun phrase and copular
sentences, the conjuncted phrase have features whose values are same as the values of the constituent. Hence
adjectives having similar features can be co-ordinated. For example, following phrase is not valid.

It has been observed that AND conjunction ( )hardly occurs when combining Adjectives (using replication of
AdjP would suffice in such cases). For example:

So keeping the majority cases in view, AdjPmain has been modified to accommodate ONLY YA conjunction ().

Result: We decided on analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Tafseer Ahmed

5 Oct, 2004

4.1.0.1

Created

4.1.0.2

Modified

th

Sara Hussain

18 Feb, 2005

Rule ID: UGR127

Rule Syntax: Following is the constituent description of the rule.
DeverbalVP -> (KPmain) v v
DeverbalVP -> (KPmain) v_n ltv v

Rule Functional Description: Following are the functional specifications of the rule.
DeverbalVP -> (KPmain: ^OBJ=!;) v:^=!, !_MORPH_FORM =c {PERFECTIVE, HABITUAL}, !DEVERBAL =c
{NOUN, ADJECTIVE}; v: ^HO=!, !_VERB_FORM =c 'ho', !DEVERBAL =c {NOUN, ADJECTIVE}, ^NUM = !NUM,
^GEND = !GEND, ^FORM = !FORM;.
DeverbalVP -> (KPmain: ^OBJ=!;) v_n:^=!; ltv: ^=!, !_MORPH_FORM =c {PERFECTIVE, HABITUAL}, !DEVERBAL
=c {NOUN, ADJECTIVE}; v: ^HO=!, !_VERB_FORM =c 'ho', !DEVERBAL =c {NOUN, ADJECTIVE}, ^NUM = !NUM,
^GEND = !GEND, ^FORM = !FORM;.

Frequency:
Description: Deverbal Verb Phrase can be used as adjective in a noun phrase, a noun phrase or as an adjunct.
c-structure: Deverbal Verb Phrase consists of an optional case phrase and two verbs.
f-structure: Deverbal Verb Phrase checks that the first verb is in either in habitual or perfective form and the second
verb is a morphological form of verb 'ho' .

Examples:
[

]
[

]
]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: 54
Related Rules: UGR007, UGR109
Related POS:
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis : Deverbial Phrase can act as an adjective, a noun or as an adjunct phrase in a sentence. The phrase is
of two type: Habitual and Perfective. Perfective phrase is used to indicate the completion of an action. e.g.

. Habitual phrase is used to indicate the continuity of an action e.g.

. This verb is optionally

followed by the verb 'ho'. Since there is no clear semantic difference introduced by adding the verb 'ho' we can
chose to generate (optional) verb 'ho' with every deverbial Phrase. This choice is based on the observation that the
verb 'ho' completes meaning of this phrase when it occurs as a noun. For this reason verbal form of 'ho' is always
generated in a deverbial phrase.
Deverbial Phrase has another difference with normal verbal phrases. The verbs in deverbal phrase have
nominative and oblique forms. So a deverbial phrase can be in oblique form in a case phrase or postpositional
phrase. For Example:

]]

]]

The table below shows different forms of a deverbal phrase with verb in habitual form, when it occurs as modifiers
of noun phrases.
Sample phrases
singular
nominative

Other adjective
/ noun

ho

Habitual

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

plural
nominative

singular
oblique

plural
oblique

Deverbial Phrase can also act as adverb. For example,

The following sentences show that case of Deverbial Phrase is not affected by number and gender. For example,

The table below shows different forms of a deverbal phrase with verb in habitual form, when it occurs as a noun.
ho
Habitual
Sample phrases
singular

nominative

plural
nominative

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

sing
oblique

plural
oblique

Use default
(masculine) case i.e.

Using default gender

Result: We decided on analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Sara Hussain

27th Dec, 2004

4.1.0.0

Sara Hussain

7th Jan, 2005

4.1.0.1

Sara Hussain

7th Feb, 2005

4.1.0.2

Aasim Ali

07-Jul-2005

5.1.0.3

Aasim Ali

15-Jul-05

5.1.0.4

rule modified for compatibility

with grammar file (changed
KP to NP). This approach
also constraints over
generation of cm with p cases
Modified rule to add adjunct
(adverb) cases
Modified to add gender
agreement check for
infinitivals in NP
Adjunct (adverb) rule
eliminated
NP changed to NPmain

Rule ID: UGR128

Rule Syntax: Following is the constituent description of the rule.
PP NPmain p
Rule Functional Description: Following are the functional specifications of the rule.
PP -> NPmain: ^OBJ=!, ! FORM =c OBL, !_INFL_AGREEMENT =c NEGATIVE; p: ^=!;.

Frequency: 1
Description: This rule shows the Postpositional Phrase of Urdu Grammar. This phrase can act as Oblique or
Adjunct.
c-structure: This phrase consists of a Case Phrase and Postposition. See POS Definition Document for details
about postposition.
f-structure: All features present in postposition are moved to Postposition Phrase. KP becomes object of the
phrase.

Example:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Active

Reference:
[1] John T. Platts, A Grammar of the Hindustani or Urdu Language
Related Rules: UGR002, UGR129, UPOS112 (adv)
Replaces: - UGR004
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
Postpositional phrase can consist of (1) case phrase and a postposition, or (2) an adverb and a postposition. Case
Phrase can have Nominative (NOM) or Genitive (GEN) case as specified by postpositions lexicon entry. Case
Phrase has two possible forms, Nominative (NOM) and Oblique (OBL), e.g
Postpositional phrase, case phrase can be only Oblique form.
In the phrase

respectively. In

, an infinitival clause precedes the post position. In such cases the gender agreement is

absent between the infinitival verb and its embedded object. Additional check (~[!_INFL_AGREEMENT =c
POSITIVE]) has thus been added to allow only negative agreement to pass through this rule.
Adverbs which indicate place (direction) or time can also occur before postpositions. For example in

is

the postposition while is the place adverb. [The POS of these adverbs has been modified to be noun so the
rule taking care of adverbs in place of NP is no more required.]

NP has been modified to be NPmain to accommodate

of generating it incorrectly as

single PP clause correctly, instead

*.

Result: We decided on above analysis.

Future Work: Some of postposition also act as preposition, e.g

. This is not catered in current rule.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Sara Hussain

16 Dec, 04

4.1.0.0

modified (added a semi

colon)

Rule ID: UGR129

Rule Syntax: Following is the constituent description of the rule.
NomPP (KP) nomp
Rule Functional Description: Following are the functional specifications of the rule.
NomPP (KP: (^OBJ)=!;) nomp: ^=!;.

Frequency: 1
Description: This rule shows the Nominal Prepositional Phrase. This phrase can act as Oblique or Adjunct of
sentence.
c-structure: This phrase consists of a Case Phrase and an Nominal Preposition. See POS Definition Document for
details about nominal preposition (UPOS03).
f-structure: All features present in preposition are moved to the phrase. KP becomes object of the phrase.

Example:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Active

Reference:
[1] John T. Platts, A Grammar of the Hindustani or Urdu Language
Related Rules: UGR002, UGR128
Related POS: UPOS103
Replaces: - UGR005
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
Nominal Postpositional phrase consists of an optional case phrase and a postposition. These postpositions can
occur without any object. While taking Case Phrase as object, Case Phrase can have Genitive (GEN), or Ablative
(ABL) case, i.e. case denoted by

, as specified by postpositions lexicon entry. It can have more than one

possible case values with different semantics. Examples below show this variation. Case Phrase has two possible
forms, Nominative (NOM) and Oblique (OBL), e.g.
case phrase can be only Oblique form.

respectively. In Nominal Postpositional phrase,

An alternative analysis for words likes ander is to give them two separate POS for with object and without

object cases. This way we can call them adverbs when they occur separate and preposition when occur with
object. This analysis was rejected in discussion because it does not seem natural to give two POS to a word when
it is playing same role in sentence in both cases. e.g. who bahir geya , woh gher ke bahir geya

Result: We decided on above analysis.

Future Work: Some of postpositions also act as preposition, e.g.

. This is not catered in current rule.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Rule ID: UGR130

Rule Syntax: Following is the constituent description of the rule.
KPP NomPP cm
Rule Functional Description: Following are the functional specifications of the rule.
KPP -> NomPP: ^=!, ! ALLOWED_CASE =c ! CASE; cm: ^=!;.

Frequency: Description: This rule shows the prepositional phrase marked with case.
c-structure: This phrase consists of an Independent Prepositional Phrase and Case Marker.
f-structure: All features present in Independent Prepositional Phrase and Case Marker are moved to this phrase.

Example:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Active

Reference:
[1] John T. Platts, A Grammar of the Hindustani or Urdu Language
Related Rules: UGR002, UGR129
Related POS: UPOS101, UPOS103
Replaces: - UGR006
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
Nominal Prepositional Phrase NomPP (UGR005) can take case marker after them. For example
NomPP, and there can be a case marker after it, i.e

is a

. Possible cases which can occur after any

Nominal Postpositional Phrase are determined by head nominal postposition. Case Marker is headless and it will
provide case feature to phrase. This information is mentioned in corresponding lexicon entry of nominal
postposition in form of an attribute ALLOWED_CASE, which can take set of possible case values.

Result: We decided on above analysis.

Future Work: Feature ALLOWED_CASE may be replaced with some common semantic features of case
markers after analysis.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Sara Hussain

16 Dec, 2004

4.1.0.0

rule modified for compatibility

with grammar file (renamed
the L.H.S non-terminal)

Rule ID: UGR131

Rule Syntax: Following is the constituent description of the rule.
NPpronoun pro
Rule Functional Description: Following are the functional specifications of the rule.
NPpronoun -> pro:^ = !; .

Frequency: 1
Description: This rule shows pronoun acting as noun phrase.
Examples:

[ ]

Rule Status: Active

Reference:
Related Rules: UGR007
Related POS: UPOS105
Replaces: - UGR017
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
Pronouns can act as noun phrase. Unlike nouns modifiers cannot occur before pronouns.

Result: We decided on the above analysis.

Future Work: Some specifiers like seb, log etc are used with second person plural, these will be handled later.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Sara Hussain

16 Dec, 2004

4.1.0.0

rule modified for compatibility

with grammar file (checks
added)

Rule ID: UGR132

Rule Syntax: Following is the constituent description of the rule.
NumberP (ordinal) (cardinal).
Rule Functional Description: Following are the functional specifications of the rule.
NumberP -> ordinal:^ SPEC = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT,^FORM=!FORM;
cardinal:^ = !, ^ GEND = {M, F}, ^ FORM = {NOM, OBL}, ^ RESPECT ={NORESPECT,FAMILIAR,USUAL,EXTRA};.
NumberP -> ordinal:^ = !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^ RESPECT = ! RESPECT , ^ FORM =
! FORM;.
NumberP -> cardinal:^ = !, ^ GEND = {M, F}, ^ FORM = {NOM, OBL}, ^ RESPECT = {NORESPECT,
FAMILIAR, USUAL, EXTRA};.

Frequency: 1
Description: This rule shows number phrase.
Examples:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Active

Reference:
Related Rules:
Related POS: UPOS108, UPOS109
Replaces: - UGR018
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
Number phrase shows number on any noun phrase. Number information can be of two types. Cardinal numbers
such as

and ordinal numbers such as

becomes specifier of the other, as in

. When these two types come together ordinal number

is more attached with than

so it is specifier of .

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Sara Hussain

16 Dec, 2004

4.1.0.2

rule modified for compatibility

with grammar file (checks
added

Rule ID: UGR133

Rule Syntax: Following is the constituent description of the rule.
QuantP quant
Rule Functional Description: Following are the functional specifications of the rule.
QuantP -> quant: ^=! ;.

Frequency: 1
Description: This rule shows quantifier phrase.
Examples:

Rule Status: Active

Reference:
Related Rules: UGR016
Related POS: UPOS107
Replaces: - UGR019
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
This rule shows quantifier phrase. It consists of a single quantifier. It specifies quantity of noun occurring after it.
Detail about quantifiers will be in UPOS107.

Result: We decided on above analysis.

Future Work: Quantifier phrase can have more than one quantifiers which will be dealt later. Examples are

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Rule ID: UGR134

Rule Syntax: Following is the constituent description of the rule.
S [Sint | Sdec | Simp | Sconj | Scomp].
Rule Functional Description: Following are the functional specifications of the rule.
S -> [Sint:^ = !; | Sdec: ^ = !; | Simp: ^ = !; | Sconj: ^ = !;| Scomp: ^ = !;].

Frequency: 1
Description: This is starting rule of the grammar.
Examples:

Rule Status: Active

Reference:
Related Rules: UGR01, UGR028, UGR029, UGR033, UGR036
Related POS:
Replaces: - UGR030
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis:
This is initial rule of grammar. S is starting symbol of grammar. Sentences can be of declarative, imperative or
interrogative type. They can also be joined by coordinating conjunction or subordinating conjunction. This rule is
merely added to make a singly starting point for grammar.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Rule ID: UGR135

Rule Syntax: Following is the constituent description of the rule.
Sconj [Sdec comma]* Sdec coord_conj Sdec.
Sconj [Sint comma]* Sint coord_conj Sint.
Sconj [Simp comma]* Simp coord_conj Simp.
Rule Functional Description: Following are the functional specifications of the rule.
Sconj -> [Sdec:! $ ^; comma:;]* Sdec:! $ ^; coord_conj: ^ = !; Sdec:! $ ^;.
Sconj -> [Sint:! $ ^; comma:;]* Sint:! $ ^; coord_conj: ^ = !; Sint:! $ ^;.
Sconj -> [Simp:! $ ^; comma:;]* Simp:! $ ^; coord_conj: ^ = !; Simp:! $ ^;.

Frequency: 1
Description: This rule describes sentence level coordinate conjunction.
Examples:

Rule Status: Active

Reference:
Related Rules: UGR001, UGR028, UGR029
Related POS:
Replaces: - UGR032
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
Sentence level coordinate conjunction is handled in this rule. Two coordinate conjunctions, and , are
considered. Three kinds of sentences are joined with each other. When two sentences are being joined, they are
separated by a conjunction. If there are more than two sentences, they are separated by comma , followed by
conjunction and the last sentence. A sentence can only be joined with sentences of same sentence type.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Rule ID: UGR136

Rule Syntax: Following is the constituent description of the rule.
Scomp Sdec sub_conjP.
Scomp Simp sub_conjP.
sub_conjP sub_conj Sdec.
Rule Functional Description: Following are the functional specifications of the rule.
Scomp -> Sdec: ^=!; Sub_conjP: ^=!;.
Scomp -> Simp: ^=!; Sub_conjP: ^=!;.
Sub_conjP -> Sub_conj: ^=!; Sdec: ^ SUBCOMP = !;.

Frequency: 1
Description: This rule describes sentence level subordinate conjunction.
Examples:

Rule Status: Active

Reference:
Related Rules: UGR025, UGR028
Related POS:
Replaces: Reason: Replaced by: -UGR034

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis:
Sentence level subordinate conjunction is handled in this rule. In subordinate conjunction first sentence must be
declarative or imperative and the second one must be declarative. Second sentence becomes complement
(COMP) to the first one.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

6th Nov, 2004

4.1.0.3

Created

Sara Hussain

21st Feb, 2004

4.1.0.4

Aasim Ali

08-Jul-2005

5.1.0.1

Tafseer Ahmed

07-Jul-2005

5.1.0.2

Aasim Ali

08-Aug-2005

5.1.0.3

Modified for Infinitival

agreement check and
Deverbal entry
Relative Clause added, and
Deverbal-only production
removed
Current NPnoun checked
and rectified
Documented the NOUN
feature added by Nayyara

Rule ID: UGR137

Rule Syntax: Following is the constituent description of the rule.
KP -> NP cm (Rel_Cl)
KP -> NPnoun (Rel_Cl)
KP -> NPpronoun (Rel_Cl) (DeverbalVP)
Rule Functional Description: Following are the functional specifications of the rule.
KP -> NP: ^=!, !FORM =c OBL; cm: ^= !; (Rel_Cl:^ADJUNCT REL_CL=!;).
KP -> NPnoun: ^=!, !FORM =c NOM, ^CASE = NOM, ^ NOUN = POS; ( Rel_Cl: ^ ADJUNCT REL_CL= !; ).
KP -> NPpronoun:^=!,!FORM=c NOM,!CASE=c {NOM,ACC,DAT,GEN},!_INFL_AGREEMENT=c NA,^NOUN=NEG;
( Rel_Cl: ^ ADJUNCT REL_CL= !; )
(DeverbalVP: ^SPEC DEM= !, ^ NUM = ! NUM, ^ GEND = ! GEND, ^FORM = ! FORM;).

Frequency: 1
Description: This rule shows the Case Phrase of Urdu Grammar. This Case Phrase can act as Subject, Object or
Object2.
c-structure: Case phrase consists of a Noun phrase and an optional case marker. In case the noun phrase is a
pronoun it can be followed by an optional adverbial phrase.
f-structure: All features present in Noun Phrase and case marker are moved to Case Phrase. Since these features
are unified, if there is any inconsistent feature in NP and cm, production will fail.

Example:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

([

) ]

[[

]]
]

Rule Status: Under Process

Reference:
[1] Butt and King, The Status of case
[2] 42-41
[3] Bhatt, Rajesh
Topics in the Syntax of the Modern Indo-Aryan Languages
April 4, 2003, Correlative Clauses
[4] Bhatt, Rajesh
Topics in the Syntax of the Modern Indo-Aryan Languages
April 11, 2003, Other Correlative Clauses
Related Rules: UGR138, UGR131, UGR106, UGR144
Related POS: UPOS101
Replaces: -UGR002
Reason: Replaced by: -

Analysis: Following is the in-depth analysis of the rule.

Analysis:
Case phrase can consist of only noun phrase or noun phrase with a case marker. It consists of only noun phrase
when its case is nominative, as in Urdu there is no Nominative case marker. Other possibility is that noun phrase
itself contains case. This situation is with pronouns like mujhe

, tumhen

etc. For definition of case marker

and possible cases in Urdu, see POS document for case markers (ref: UPOS01).
Correct form of noun with case marker is determined on the basis of their lexical features. Nouns have two forms,
Nominative and one altered form which can take any case marker. Some pronouns take case markers to make KP,
i.e men ney

is ergative case of first person singular. Some pronouns contain case themselves and form KP
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

without adding case marker, i.e mujhe

, tumhen

are accusative and dative. Adding case marker with such

is wrong. To check this error FORM feature is used.

pronouns is syntax error, i.e mujhe ko

Deverbals can occur in case phrase as optional specifiers (adjective) for pronouns. In this case number, gender
and form of deverbal and pronoun should be same. Deverbals can also occur in a sentence as a noun in
nominative case.
Whenever the Relative Clause appears adjacent to its related Noun Phrase, it carries post-nominal form, in Urdu.
This rule follows the fixed order of the sentence as already decided and implemented in the rule UGR103 (Sdec).
Thus the Relative Clause is only allowed to occur at its syntactically fixed position in the sentence.

[
]

] - Not Handled

- Handled

Result: We decided on above analysis.

Future Work:
NPconj will be added rules to deal NP co-ordination.
In addition to the hybrid relatives following tasks are also to be handled in future:
- Relatives in place other post-NP
- Sentence-level relative

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Sara Hussain

23 Feb,05

4.1.0.1

Modified to remove infinitival

phase and add deverbal
phrase

Aasim Ali

29-Jun-05

5.1.0.1

Added Relative Clause

Tafseer Ahmed

07-Jul-2005

5.1.0.2

Aasim Ali

08-Jul-05

5.1.0.3

Current NPnoun checked

and rectified
Transferred Relative Clause
to KP (UGR137), and
Transferred DeverbalVP to
NPnoun

Rule ID: UGR138

Rule Syntax: Following is the constituent description of the rule.
NP -> NPnoun | NPpronoun
Rule Functional Description: Following are the functional specifications of the rule.
NP -> [NPnoun: ^=!; | NPpronoun: ^= !, !_INFL_AGREEMENT =c NA;].

Frequency: 1
Description: This rule shows Noun phrase.
c-structure: Noun Phrase can consist of Noun Phrase(covering Common Noun, Proper Noun, Verb acting as noun
(gerund) or Adjective acting as Noun), Pronoun Phrase or deverbal phrase.
f-structure: Features present in any of the daughter nodes are copied to NP.

Example:
(common noun)
(common noun)
(proper noun)
(pronoun)
(Deverbal Phrase)

[
[

]
]

][
[

Above mentioned examples are simplest forms of corresponding phrases. Complex phrases will be discussed in
their respective rules.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under-process

Reference:
Related Rules: UGR131, UGR106
Replaces: -UGR007
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis 1:
Noun phrase can consist of common noun phrase, in which head is a common noun and there are modifier
attached to it. In the same way it can be a Proper Noun phrase or pronoun. Sometimes verb acts like a noun
phrase, it is called gerund. Verb acting as noun is in masder
position, e.g jahil

, accha

form. Sometime adjectives occur at noun phrase

Analysis 2:
Common and Proper Noun do not have any syntactical difference. Therefore, there is no need of two separate
rules for common and proper noun. Noun Phrase can be divided as phrases having Noun, phrases having pronoun
and deverbal phrase.

Result: We decided on analysis 2.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Omar Javed

06 Nov, 04

0.1

Created

Rule ID: UGR139

Rule Syntax: Following is the constituent description of the rule.
Sint pro Sdec
Rule Functional Description: Following are the functional specifications of the rule.
Sint -> pro: !PRONFORM =c 'KIYA'; Sdec: ^ = !, ^ STMT_TYPE = INT; .

Frequency: 1
Description: This rule is for yes/no interrogative questions.
Examples:

Rule Status: Under Process

Reference: [1] John T. Platts, A Grammar of the Hindustani or Urdu Language
Related Rules:
Related POS:
Replaces: - UGR029
Reason: Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the analysis of the rule.

Analysis:
Yes/no question can be formed using word kiya before a declarative sentence [1].

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name
Tafseer Ahmed
Aasim Ali

Change Date
18th May, 04
08-AUG-2005

Version
0.1
5.1.0.1

Description of Changes
Created, replaced UGR028
Documentation of changes made in
UGR file by Nayyara

Rule ID: UGR140

Rule Syntax: Following is the constituent description of the rule.
Simp pro (KPmain) v
Rule Functional Description: Following are the functional specifications of the rule.
Simp -> pro:^ SUBJ=!, !PRONTYPE =c NULL;
(KPmain:^OBJ=!, !CASE =c NOM;)
v:^=!, !_MORPH_FORM =c PERCATIVE; .

Frequency: 1
Description: This rule shows imperative sentences.
Examples:

Rule Status: Active

Reference:
Related Rules: UGR002, UGR004, UGR005, UGR006, UGR012, UGR020, UGR034
Related POS:
Replaces: - UGR028
Reason: - previous release document is deprecated.
Replaced by: -

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Analysis: Following is the in-depth analysis of the rule.

Analysis:
This rule shows imperative sentences. In imperative sentence there is no explicit subject, therefore, it is kept NULL.
There can be four ways of imperative sentences with different level of respect which is shown through morphology
of verb being used, as shown in the Future Work section below.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Sara Hussain

21st Dec, 2004

4.1.0.0

Created

Aasim Ali

08-AUG-2005

5.1.0.1

Documentation of changes
made in UGR file by Nayyara

Rule ID: UGR141

Rule Syntax: Following is the constituent description of the rule.
TenseAuxP -> tense_aux

Rule Functional Description: Following are the functional specifications of the rule.
TenseAuxP -> tense_aux: ^ = !, [! _MORPH_FORM =c SUBJUNCTIVE && ! _AUX_FORM =c 'ho'] ,
^ TNS_ASP PROBABLE =c NEG, ^ _MORPH ALLOWED_FORM =c ! _ALLOWED_FORM,
^ TENSE =c ! TENSE;.

Frequency: 1
Description: This rule shows main rule for tense auxiliary phrase.
c-structure: tense auxiliary phrase consists of simple tense auxiliaries or verb ho followed by 'ga'/ 'gi' etc..
f-structure:

Examples:
-[

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules:
Related POS: UPOS116
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: There are three undisputed tense auxiliaries: hay, tha and ga with there inflectional forms. They show
present, past and future tense respectively. ho ga also acts as tense auxiliary. Combination of two tense
auxiliaries also act as tense auxiliary phrase, where first one is ho in bare form and second one is ga. This
combination shows agreement on the basis of their number and person. Gender agreement is only reflected on ga.

Result: We decided on above analysis.

Future Work:
[

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Sara Hussain

17th Jan 2005

4.1.0.0

Created

Aasim Ali

08-AUG-2005

5.1.0.1

Documentation of changes
made in UGR file by Nayyara

Rule ID: UGR142

Rule Syntax: Following is the constituent description of the rule.
KarP -> (KP) pro v v

Rule Functional Description: Following are the functional specifications of the rule.
KarP -> ([KP:^OBJ=!, ! SEM_TYPE =c {UNANIM_CONC,ANIMAL,ABSTRACT}, !CASE =c NOM;
|
KP:^OBJ=!, ! SEM_TYPE =c HUMAN, !CASE =c ACC;])
pro:^SUBJ = !;
v:^=!, !_MORPH_FORM =c BARE, ^ GEND =c M, ^ NUM=SG;
v:^KAR=!, !_VERB_FORM =c 'kar', ! _MORPH_FORM =c BARE; .

Frequency: 1
Description: Complex sentences are those sentences which contain more than one clause (more than one verb):
a main clause and complement clause(s).
c-structure: Infinitival clause consists of optional Case phrase (KP), compulsory pronoun and infinitival verb in its
BARE form. The clause ends with the verb kar in its BARE form.
f-structure:

Examples:

[
]

[ ]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules:
Related POS: UPOS103
Replaces: Reason: - Sentences containing the verb kar as an adjunct was not analyzed before.
Replaced by: -

Analysis: Following is the in-depth analysis of the rule.

Analysis: The Kar clause can acts as a Xadjunct and/or a conjunction (meaning after) in a sentence. Syntactically
there is no different between these two constructions. This rule models the Xadjunct form of the Kar clause.
The Kar clause consists of a main verb in BARE form followed by the verb Kar. Objects may occur for some verbs.
Also additional description can be added through further adjuncts.
In complex predicates, if the verbal phrase ends with the light verb kar, then the clausal kar is modified to kae.
Consider the following example:

Result: We decided on above analysis.

Future Work:
(1) To handle multiple Kar Clauses in a sentence
(2)

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Sara Hussain

18th Jan 2005

4.1.0.0

Created

Sara Hussain

22 Feb 2005

4.1.0.1

Added deverbal entry

Rule ID: UGR143

Rule Syntax: Following is the constituent description of the rule.
S_Adjunct -> (NPmain) (adv)* (PPmain)* //temporal
(adv)* (PPmain)*
//place
(adv)* (PPmain)*
//manner
(NomPPmain)*
//manner / purpose
(DeverbalVP)
//manner

Rule Functional Description: Following are the functional specifications of the rule.
S_Adjunct -> (NPmain: ! $ ^TIME, !SEM_TYPE =c TEMPORAL;)
//temporal
([adv: ! $ ^ TIME,!SEM_TYPE =c TEMPORAL;]*)
([PPmain: ! $ ^ TIME, !OBJ SEM_TYPE =c TEMPORAL;]*)
([adv: ! $ ^ PLACE, !SEM_TYPE =c PLACE;]*)
//place
([PPmain: ! $ ^PLACE, !OBJ SEM_TYPE =c PLACE;]*)
([adv: ! $ ^, ~[!SEM_TYPE =c TEMPORAL], ~[!SEM_TYPE =c PLACE];]*)
//manner and others
([PPmain: ! $ ^, ~[!OBJ SEM_TYPE =c PLACE], ~[!OBJ SEM_TYPE =c TEMPORAL];]*)
([NomPPmain: ! $ ^;]*)
//manner / purpose
(DeverbalVP: ! $ ^MANNER, !DEVERBAL =c NOUN, !FORM =c OBL, !NUM =c SG, !GEND =c M;)
.

Frequency: 1
Description: Additional (optional) information about time or place (or manner or purpose, etc.) of an activity is
said to serve as an adjunct. This rule shows main order of verbal adjuncts in a sentence.
c-structure: Adjuncts consist of temporal nouns, adverbs, postpositions phrases and deverbals. Since adjuncts are
additional information all nodes are optional.
f-structure: Temporal adjuncts are followed by place, manner and then other adjuncts.

Examples:

phrase)

] ( order: temporal noun, place adverb and then place postposition phrase)

] ( order: temporal adverb, manner adverb and nominal postposition

]

(order: multiple place postposition phrases)

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

phrase)

[
[
]

( ] order: temporal noun, temporal postposition phrase and then place postposition
]

(order: temporal noun then place adverb)

(occurrence of deverbal as adjunct)

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
Related Rules:
Related POS: UGR103 (S_dec), UPOS129 (NomPP), UGR125 (NPmain), UGR123 (PPmain),
UPOS112 (adv)

Replaces: Reason: - Order of sentence level adjuncts was not analyzed before.
Replaced by: -

Analysis: Following is the in-depth analysis of the rule.

Analysis: Urdu is free-order language. For this reason various sentence level adjunct phrases can change their
place in a sentence. Thus both sentences given below are correct.

]
]

This rule models commonly used adjunct order in Urdu. The following lexical and phrase categories can occur as
sentence level adjuncts in Urdu: Few nouns (like temporal nouns), adverbs, prepositional phrases and nominal post
position phrases. Their description is given below. Relative order of these adjuncts is discussed next.
Noun Adjuncts
Nouns having temporal properties can act as adjuncts. Some examples of temporal nouns are

. These

words could be modeled in two ways. (1) Either they could be assigned two lexical entries: one in which they are
nouns and the other in which they are adverb. And it is the adverb entry which can act as an adjunct. (2) Or such
words could be assigned a single noun entry in lexicon having temporal property. And rules (such as the one given
above) govern when a noun can occur as an adjunct.
We chose the second option, since there are many temporal nouns. Two entries for each one of them can easily be
avoided by rule formation.
Adverbial Adjuncts
Meaning of some adverbs can equivalently be expressed by certain prepositional phrases. Consider the table
below.
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Adverbs Equivalent prepositional phrases

During generation there exits a choice between using such adverbs (formed through morphology) and using
equivalent prepositional phrases. Since the prepositional phrases form simpler Urdu sentences therefore it was
decided that they will be preferred.
Postpositional Adjuncts
Postpositional phrases (including nominal post position phrases) also act as adjuncts in a sentence. This can be
seen in the following examples:

]
[

Deverbal phrase as Adjuncts

Deverbal phrase can also be used to describe manner of a verb. Such deverbals always occur in oblique form.
Their number and gender is not affected by context. This can be seen in the following examples:

Duplication of words and Adjuncts

At times meaning generated by duplicated words is equivalent to the occurrence of the same word with a
preposition and / or quantifier. This can be seen in the first two examples given below. Other examples show
behavior of duplicated words as adjuncts.

Duplicated words

Sample Sentence

Equivalent meaning phrases

Duplication shows state

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Duplication shows state

Duplication intensifies meaning

Duplication intensifies meaning
Duplication completes sentence

During generation there exits a choice between using such duplicated words and using equivalent phrases. Since
these phrases form simpler Urdu sentences therefore it was decided that they will be preferred.
Order of Adjuncts:
After analyzing couple of dozen sentences the following order of adjuncts was observed. First temporal adjuncts
occur. They are followed by place (spatial) adjuncts. Adjuncts depicting manner occur next. Adjuncts indicating the
purpose (usually nominal post positions phrases) are last to occur in sentence level adjuncts.

Result: We decided on above analysis.

Future Work:
(1) Some temporal nouns can take another noun / adverb after it to form an adjunct. For example:

Other such words are:

These words need to be analyzed.

(2) Position of intensifier phrase, which may act as adjunct, need to be determined. Relative position of NomPP and
few adverbs like () may need to be reanalyzed.
(3) In actual practice, it is highly unusual to have a number of adjunct phrases. Because the placement of adjuncts
is so flexible, when many adjunct phrases occur in a sentence, some of them would probably move to the
beginning of the sentence. This phenomenon can to be analyzed. Also in case of such movements long distance
dependency may occur.

Reference No:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name
Aasim Ali

Change Date
27-Jun-05

Version
0.1

Description of Changes
Created

Rule ID: UGR144

Rule Syntax: Following is the constituent description of the rule.
Rel_Cl -> Sdec

Rule Functional Description: Following is the functional description of the rule.

Rel_Cl -> Sdec: ^ =! ; .

Frequency: Description: This rule shows the relative clause that modifies NP.
c-structure: Relative Clause is a clause having a complete sentence modifying NP. The sentence can be transitive
or intransitive. Relative clause is introduced by relative pronoun.
f-structure: Initially it has been kept same as Sdec. Thus forcing that SUBJ must come in the start of the clause.

Examples:

[
[

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Active

Reference: [1] Bhatt, Rajesh
Topics in the Syntax of the Modern Indo-Aryan Languages
April 4, 2003, Correlative Clauses
[2] Bhatt, Rajesh
Topics in the Syntax of the Modern Indo-Aryan Languages
April 11, 2003, Other Correlative Clauses
[3] Pam Peters, The Cambridge English Guide to Usage, p 468
[4] Miriam Butt, Tracy Holloway King, A Grammar Writers Cookbook

Related Rules: UGR138

Related POS: UPOS105
Replaces: Reason: Replaced by: Reason: -

Analysis: Following is the in-depth analysis of the rule.

Analysis1: There are two types of relative clauses in Urdu: (1) Personal Relative Pronoun, (2) Adverbial Relative.
Personal Relative Pronoun clause serves either to define, or to describe and evaluate the noun to which it is
attached. Examples 1 to 3 given above illustrate Personal Relative Pronoun, and others illustrate Adverbial
Relative. Relative Clause in Urdu may occur at any position in the sentence even away from its related Noun
Phrase, but this grammar only handles post-NP relative clause.
Relative clauses are often introduced by the relative pronouns such as

etc. However, in some informal

speech, relative pronoun may be omitted:

] [

Some relative clauses are linked to the main clause by adverbial relatives, such as

. For example:

The adverbial relatives act as relaters of the second clause to a noun of time, place or reason in the main clause. In
less formal styles, the relative adverbs can be replaced by

]
]

, as in

Handled in JO form

Handled

Result: The above analysis was finalized except:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

The relative pronoun

is word translation of English relative which, but in Urdu these relative

pronouns can be equally represented by with one relative clause

so later option has been chosen for simplicity.

Future Work:
Some relatives, in Urdu, are of hybrid nature, which are not dealt with, because they may occur at Sentence level
that will be handled in future. They may be used as personal relative pronoun and as adverbial relative, depending

on the context. These are

In addition to the hybrid relatives following task is also to be handled in future:
- Multiple head relatives

Reference No:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name
Aasim Ali
Aasim Ali

Change Date
15-Jul-2005
18-Jul-2005

Version
5.1.0.1
5.1.0.2

Description of Changes
Created
Removed constraint for PERSon

Rule ID: UGR145

Rule Syntax: Following is the constituent description of the rule.
GPmain -> GP
GPmain -> GP [[comma GP]]* coord_conj GP

Rule Functional Description: Following is the functional description of the rule.

GPmain -> GP: ^=!; .
//AUR Conjuncted
GPmain -> GP: ! $ ^, ^PERS=!PERS, ^ FORM = ! FORM, ^GEND=!GEND;
[[comma:; GP: ! $ ^, ^ FORM = ! FORM, ^GEND=!GEND;]]*
coord_conj:^=!, !CONJ_FORM =c AUR, ^ NUM = PL, ^RESPECT = USUAL;
GP: ! $ ^, ^ FORM = !FORM, ^GEND=!GEND; .
//YA Conjuncted
GPmain -> GP: ! $ ^, ^ FORM = ! FORM, ^GEND=!GEND;
[[comma:; GP: ! $ ^, ^ FORM = ! FORM, ^GEND=!GEND;]]*
coord_conj:^=!, !CONJ_FORM =c YA;
GP: ! $ ^, ^PERS=!PERS, ^ GEND = !GEND, ^NUM = !NUM, ^ RESPECT = !RESPECT, ^ FORM =
!FORM; .

Frequency: Description: This rule represents the co-ordination of Genitive Phrase.

c-structure: It consists of a genitive phrase or a list of comma separated genitive phrases.
f-structure: A set of genitive phrases having co-ordination type and agreement features at the root.

Examples:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Active

Reference:
Related Rules: UGR106
Related POS: UPOS121, UPOS122
Replaces: Reason: Replaced by: Reason: -

Analysis: Following is the analysis of the rule.

Analysis : Genitive Phrase can be coordinated by using coordination conjunction. So a Genitive Phrase (GP) can
be a single genitive phrase or a list of comma separated genitive phrases having a coordinating conjunction before
last case phrases. Gender of all the participating genitive phrases must be same like:

The features of coordinated genitive phrase are determined by the grammar rules. The rules are similar to the rules
of KP discussed in UGR137. The only except is that GP do not deal CASE feature. The feature rules are following:
Conjunction Type
Aur
Others( e.g. ya)
Plural
Of last GP
Number
Masculine
Of last GP
Gender
1st then 2nd then 3rd Of last GP
Person
All GPs should have All GPs should have
Form
same.
same.
Usual
Of last GP
Respect
Constraint for person has been removed keeping in view to prefer the order of the original sentence in English.

Result: We decided on analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Aasim Ali

08-Jul-2005

5.1.0.1

Created

Rule ID: UGR146

Rule Syntax: Following is the constituent description of the rule.
A_AdvP -> (adv) adv (intens)
A_AdvP -> adv intens adv (intens)

Rule Functional Description: Following are the functional specifications of the rule.
A_AdvP ->

(adv: ^ MOD_ADV = !, !ADVTYPE =c ADVERB;)

adv: ^ =!, !ADVTYPE =c ADJECTIVE, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ GEND = !
GEND, ^ FORM = ! FORM;
(intens: ^ INSFORM=!INSFORM, !INSTYPE =c ADVERB ;).

A_AdvP ->

adv: ^ MOD_ADV = !, !ADVTYPE =c ADVERB;

intens: ^ MOD_ADV INSFORM=! INSFORM, !INSTYPE =c ADVERB ;
adv: ^ =!, !ADVTYPE =c ADJECTIVE, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ GEND = !
GEND, ^ FORM = ! FORM;
(intens: ^ INSFORM =! INSFORM, ! INSTYPE =c ADVERB ;).

Frequency:
Description: This rule shows Adverbial Intensifier and Specifier used in Adverbial Phrase.
c-structure: Pre-Nominal Adverbial Phrase consists of a adverb or postpositional phrase having se. The adverb is
followed by optional intensifier.
f-structure: Pre-Nominal Adverbial Phrase usually becomes ADJUNT of an adjective.

Examples:
[

[[ ]

] []

][

] [[ ]

][

[[ ]

] [[ ]

][

]
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference: Discussion with Mr Tafseer and Ms Nayyara
Related Rules: UGR119
Related POS: UPOS112, UPOS118
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: Two adverbs may come together in a situation when 1st adverb (sub-adverb) tells the degree/intensity of

the 2nd one (main-adverb), like

. However example of more than two adverbs used in formal

writing has not been encountered yet. Therefore maximum of two adverbs are allowed in this rule.

* is invalid in Urdu. Therefore 1st one is qualifier of 2nd one. So, if two

It was also analyzed that

adverbs are present in the phrase then 1st one will become the SPEC feature of the 2nd one. In addition, PP should
be recorded as SPEC of the Adjective since it compares the adjective of the head noun. Thus the f-structure of

phrase

should be as under:

be recorded as ADJUNCT ADJ of

be recorded as SPEC ADV of

be recorded as MOD_ADV of

The Intensifier (intens) simply sends up its lexicon feature INSFORM to its related head (ADJUNCT or SPEC).

Result: We decided on above Analysis.

Future Work:
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Aasim Ali

08-AUG-2005

5.1.0.1

Created to document the

changes made in UGR file by
Nayyara

Rule ID: UGR147

Rule Syntax: Following is the constituent description of the rule.
VPperf -> ComplexP.
VPperf -> ComplexP TenseAuxP.

Rule Functional Description: Following are the functional specifications of the rule.
VPperf -> ComplexP:^ = !, ^ _MORPH_FORM =c PERFECTIVE,^ TNS_ASP NEARNESS =c INDEFINITE, ^ NUM
=c ! NUM_END;.
VPperf -> ComplexP:^ = !, ^ _MORPH_FORM =c PERFECTIVE, ^ NUM =c ! NUM_MID;
[[
TenseAuxP:^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ TNS_ASP
NEARNESS =c POS, ! _MORPH_FORM =c SUBJUNCTIVE, ! _AUX_FORM =c 'ho', ^ TNS_ASP TENSE
=c PAST, ^ NUM =c ! NUM, ^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;
|
TenseAuxP:^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ TNS_ASP
NEARNESS =c NEG, ! _MORPH_FORM =c PERFECTIVE, ! _AUX_FORM =c 'tha', ^ TNS_ASP =c !
TNS_ASP, ^ NUM =c ! NUM, ^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;]
|
[TenseAuxP: ^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ TNS_ASP
NEARNESS =c NA, ! TNS_ASP PROBABLE =c POS, !, ^ TNS_ASP =c ! TNS_ASP, ^ NUM =c ! NUM, ^
RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;
]].

Frequency: 1
Description: This rule shows main rule for verb phrase in PERFECTIVE form.
c-structure: Verb phrase consists of a complex predicate phrase followed by optional tense auxiliaries phrase.
f-structure: Complex predicate phrase and tense auxiliary have agreement on the basis of number, gender, person
and morphological form of verb.

Examples:

[TNS_ASP NEARNESS =c INDEFINITE]

[TNS_ASP NEARNESS =c POS]

[TNS_ASP NEARNESS =c NEG]

[TNS_ASP NEARNESS =c NA]

[TNS_ASP NEARNESS =c INDEFINITE]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
[1] UGR103
[2] Miriam Butt, Discussion at EGD_ULP meetings

Related Rules: UGR103 UGR116, UGR141

Related POS:
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: The Usual Verb Phrase consists of Complex Predicate and Optional Tense Auxiliary Phrase. In syntax of
commonly used Urdu, only one aspectual auxiliary is used after Complex Predicate. Morphological form of each
node depends on form of the next node. For example,

and

allow PERFECTIVE or HABITUAL form.

allows SUBJUNCTIVE form before it. The Complex Predicate provides the Sub-Categorization frame for the
phrase. The case of subject is provided by last main or light verb of the phrase.
NUMber of the VPnonperf is chosen from two features, namely, NUM_MID and NUM_END. NUM_MID means the
NUMber of the middle element in the composition, whereas NUM_END means the NUMber of the last element.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Aasim Ali

08-AUG-2005

5.1.0.1

Created to document the

changes made in UGR file by
Nayyara

Rule ID: UGR148

Rule Syntax: Following is the constituent description of the rule.
VPnonperf -> ComplexP TenseAuxP.

Rule Functional Description: Following are the functional specifications of the rule.
VPnonperf -> ComplexP:^ = !, ^ _MORPH_FORM =c {HABITUAL,SUBJUNCTIVE}, ^ NUM = ! NUM_MID;
TenseAuxP: ^ _MORPH_FORM = ! _MORPH ALLOWED_FORM, ^ TNS_ASP TENSE = ! TENSE,
^ TNS_ASP PROBABLE = ! TNS_ASP PROBABLE, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT,
^ PERS = ! PERS, ^ GEND = ! GEND;.

Frequency: 1
Description: This rule shows main rule for verb phrase in NON-PERFECTIVE forms.
c-structure: Verb phrase consists of a complex predicate phrase followed by tense auxiliaries phrase.
f-structure: Complex predicate phrase and tense auxiliary have agreement on the basis of number, gender, person
and morphological form of verb.

Examples:
[TNS_ASP TENSE = FUTURE]

[TNS_ASP TENSE = PRES]

[TNS_ASP TENSE = PRES]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
[1] UGR103
[2] Miriam Butt, Discussion at EGD_ULP meetings

Related Rules: UGR103 UGR116, UGR141

Related POS:
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: The Usual Verb Phrase consists of Complex Predicate and Tense Auxiliary Phrase. In syntax of
commonly used Urdu, only one aspectual auxiliary is used after Complex Predicate. Morphological form of each
node depends on form of the next node. For example, , and allow PERFECTIVE or HABITUAL form.
allows SUBJUNCTIVE form before it. The Complex Predicate provides the Sub-Categorization frame for the
phrase. The case of subject is provided by last main or light verb of the phrase.
NUMber of the VPnonperf is chosen from two features, namely, NUM_MID and NUM_END. NUM_MID means the
NUMber of the middle element in the composition, whereas NUM_END means the NUMber of the last element.

Result: We decided on above analysis.

Future Work:

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Aasim Ali

08-AUG-2005

5.1.0.1

Created to document the

changes made in UGR file by
Nayyara

Rule ID: UGR149

Rule Syntax: Following is the constituent description of the rule.
VPraha -> ComplexP asp_aux TenseAuxP.

Rule Functional Description: Following are the functional specifications of the rule.
VPraha -> ComplexP:^ = !, ^ _MORPH_FORM =c BARE;
asp_aux: ^ _MORPH_FORM = ! _ALLOWED_FORM, ! _AUX_FORM =c 'raha',
^ _MORPH FORM = ! _MORPH_FORM, ^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS,
^ GEND = ! GEND;
TenseAuxP: ^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ TNS_ASP TENSE = ! TENSE,
^ NUM = ! NUM, ^ RESPECT = ! RESPECT, ^ PERS = ! PERS, ^ GEND = ! GEND;.

Frequency: 1
Description: This rule shows main rule for verb phrase in NON-PERFECTIVE forms.
c-structure: Verb phrase consists of a complex predicate phrase followed by aspect auxiliary and tense auxiliary
phrase.
f-structure: Aspect auxiliary and tense auxiliary have agreement on the basis of number, gender, person and
morphological form.

Examples:

[TNS_ASP TENSE = PRES]

[TNS_ASP TENSE = FUTURE]
[TNS_ASP TENSE = PAST]

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Rule Status: Under Process

Reference:
[1] UGR103
[2] Miriam Butt, Discussion at EGD_ULP meetings

Related Rules: UGR103 UGR116, UGR141

Related POS: UPOS114
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: The Usual Verb Phrase consists of Complex Predicate and Tense Auxiliary Phrase. In syntax of Urdu
progressive sentences the main verb occurs in BARE form, and is followed by aspectual auxiliary raha. The
Complex Predicate provides the Sub-Categorization frame for the phrase. The case of subject is provided by last
main or light verb of the phrase.

Result: We decided on above analysis.

Future Work: Verbal Phrase should deal all possible combinations of Aspectual auxiliaries so the sentences like

should be parsed successfully.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Aasim Ali

08-AUG-2005

5.1.0.1

Created to document the

changes made in UGR file by
Nayyara

Rule ID: UGR150

Rule Syntax: Following is the constituent description of the rule.
VPmodal -> ComplexP asp_aux TenseAuxP.
VPmodal -> ComplexP asp_aux.

Rule Functional Description: Following are the functional specifications of the rule.
VPmodal -> ComplexP:^ = !, ^ _MORPH_FORM =c BARE;
asp_aux: ^ _MORPH_FORM = ! _ALLOWED_FORM, ^ TNS_ASP MODAL =c ! TNS_ASP MODAL,
^ TNS_ASP PERF =c NEG,^ TNS_ASP NEED =c NEG, ^ TNS_ASP PROG =c NEG,
^ _MORPH FORM =c {HABITUAL,SUBJUNCTIVE}, ^ _MORPH FORM =c ! _MORPH_FORM,
^ NUM =c ! NUM_MID, ^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;
TenseAuxP: ^ _MORPH FORM = ! _MORPH ALLOWED_FORM, ^ TNS_ASP TENSE =c ! TENSE,
^ NUM =c ! NUM, ^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;.
VPmodal -> ComplexP:^ = !, ^ _MORPH_FORM =c BARE;
asp_aux: ^ _MORPH_FORM = ! _ALLOWED_FORM, ^ TNS_ASP MODAL =c ! TNS_ASP MODAL,
^ TNS_ASP PERF =c POS, ^ TNS_ASP PROG =c NEG, ^ _MORPH FORM =c PERFECTIVE,
^ _MORPH FORM =c ! _MORPH_FORM, ^ TNS_ASP NEARNESS =c INDEFINITE,
^ NUM =c ! NUM_END, ^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;.
VPmodal -> ComplexP:^ = !, ^ _MORPH_FORM =c BARE;
asp_aux: ^ _MORPH_FORM = ! _ALLOWED_FORM, ^ TNS_ASP MODAL =c ! TNS_ASP MODAL,
^ TNS_ASP PERF =c POS, ^ TNS_ASP PROG =c NEG, ^ _MORPH FORM =c PERFECTIVE,
^ _MORPH FORM =c ! _MORPH_FORM, ^ NUM =c ! NUM_MID, ^ RESPECT =c ! RESPECT,
^ PERS =c ! PERS, ^ GEND =c ! GEND;
[[
TenseAuxP:^ _MORPH FORM = ! _MORPH ALLOWED_FORM,
^ TNS_ASP NEARNESS =c POS, ! _MORPH_FORM =c SUBJUNCTIVE, ! _AUX_FORM =c 'ho',
^ TNS_ASP TENSE =c PAST, ^ NUM =c ! NUM, ^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS,
^ GEND =c ! GEND;
|
TenseAuxP: ^ _MORPH FORM = ! _MORPH ALLOWED_FORM,
^ TNS_ASP NEARNESS =c NEG, ! _AUX_FORM =c 'tha',^ TNS_ASP =c ! TNS_ASP, ^ NUM =c ! NUM,
^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;]
|
[TenseAuxP:^ _MORPH FORM = ! _MORPH ALLOWED_FORM,
^ TNS_ASP NEARNESS =c NA, ! TNS_ASP PROBABLE =c POS, !, ^ TNS_ASP =c ! TNS_ASP,
^ NUM =c ! NUM, ^ RESPECT =c ! RESPECT, ^ PERS =c ! PERS, ^ GEND =c ! GEND;
]].

Frequency: 1
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Description: This rule shows main rule for verb phrase in NON-PERFECTIVE forms.
c-structure: Verb phrase consists of a complex predicate phrase followed by aspect auxiliary and tense auxiliary
phrase.
f-structure: Aspect auxiliary and tense auxiliary have agreement on the basis of number, gender, person and
morphological form.

Examples:

Rule Status: Under Process

Reference:
[1] UGR103
[2] Miriam Butt, Discussion at EGD_ULP meetings

Related Rules: UGR103 UGR116, UGR141

Related POS: UPOS114
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: The Usual Verb Phrase consists of Complex Predicate and Tense Auxiliary Phrase. In syntax of Urdu
progressive sentences the main verb occurs in BARE form only and is followed by aspect auxiliary sak. The
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Complex Predicate provides the Sub-Categorization frame for the phrase. The case of subject is provided by last
main or light verb of the phrase.
The aspectual auxiliaries other than the one shown above (

) in examples, are

etc.

NUMber of the VPmodal is chosen from two features, namely, NUM_MID and NUM_END. NUM_MID means the
NUMber of the middle element in the composition, whereas NUM_END means the NUMber of the last element.

Result: We decided on above analysis.

Future Work: Verbal Phrase should deal all possible combinations of Aspectual auxiliaries so the sentences like

should be parsed successfully.

EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Revision History:
Name

Change Date

Version

Description of Changes

Aasim Ali

08-AUG-2005

5.1.0.1

Created to document the

changes made in UGR file by
Nayyara

Rule ID: UGR151

Rule Syntax: Following is the constituent description of the rule.
VPhay -> (neg_adv) v .
VPhay -> (neg_adv) v tense_aux .
VPhay -> VPperf .
VPhay -> VPraha .

Rule Functional Description: Following are the functional specifications of the rule.
VPhay -> (neg_adv:^ADJUNCT NEG=!, !NEGFORM =c nahin;)
v:^=!, !_MORPH_FORM =c {PERFECTIVE,SUBJUNCTIVE}, ^ TNS_ASP MODAL =c NONE,
^ PROBABLE=c NEG,^TNS_ASP PERF=c NEG,^TNS_ASP NEARNESS=c NA,^TNS_ASP NEED=c NEG,
^ TNS_ASP PROG =c NEG, ^TNS_ASP TENSE =c !TENSE, ^TNS_ASP TENSE =c {PAST,PRES}; .
VPhay -> (neg_adv:^ADJUNCT NEG=!, !NEGFORM =c nahin;)
v:^=!, , !_MORPH_FORM =c PRE_GA, ^ TNS_ASP MODAL =c NONE, ^ PROBABLE =c NEG,
^ TNS_ASP PERF =c NEG, ^ TNS_ASP NEARNESS =c NA, ^ TNS_ASP NEED =c NEG,
^ TNS_ASP PROG =c NEG, ^TNS_ASP TENSE =c !TENSE, ^TNS_ASP TENSE =c FUTURE;
tense_aux: ^NUM=!NUM,^GEND=!GEND,^PERS=!PERS,^RESPECT=c !RESPECT,!_AUX_FORM=c 'ga';
.
//woh khush hoa, hoa hay, hoa tha
VPhay -> VPperf:^=!, ^ TNS_ASP MODAL =c NONE, ^ PROBABLE =c NEG, ^ TNS_ASP PERF =c POS,
^ TNS_ASP NEED =c NEG, ^ TNS_ASP PROG =c NEG; .
//woh khush ho raha hay, tha
VPhay -> VPraha: ^=!, ^ TNS_ASP MODAL =c NONE, ^ PROBABLE =c NEG, ^ TNS_ASP PERF =c NEG,
^ TNS_ASP NEARNESS =c NA, ^ TNS_ASP NEED =c NEG, ^ TNS_ASP PROG =c POS; .

Frequency: 1
Description: This rule shows main rule for verb phrase in NON-PERFECTIVE forms.
c-structure: Verb phrase consists of a complex predicate phrase followed by aspect auxiliary and tense auxiliary
phrase.
f-structure: Aspect auxiliary and tense auxiliary have agreement on the basis of number, gender, person and
morphological form.
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Examples:

Rule Status: Under Process

Reference:
[1] UGR103
[2] Miriam Butt, Discussion at EGD_ULP meetings

Related Rules: UGR103, UGR147, UGR149

Related POS: UPOS113, UPOS116, neg_adv
Replaces: Reason: Replaced by: -

Analysis: Following is the analysis of the rule.

Analysis: The Usual Verb Phrase consists of Complex Predicate and Tense Auxiliary Phrase. In syntax of Urdu
progressive sentences the main verb occurs in BARE form only and is followed by aspect auxiliary sak. The
Complex Predicate provides the Sub-Categorization frame for the phrase. The case of subject is provided by last
main or light verb of the phrase.
Predicative or Copular hay is a special verb that needs nominative subject and Predlink. The forms of hay_dat is
hay(all forms), tha(all forms) and ho(followed by tense auxiliary ga).
EGD Urdu Localization Project - Grammar Rule Document

Center for Research in Urdu Language Processing

National University of Computer and Emerging Sciences, Lahore Pakistan

Result: We decided on above analysis.

Future Work: Verbal Phrase should deal all possible combinations of Aspectual auxiliaries so the sentences like

should be parsed successfully.

EGD Urdu Localization Project - Grammar Rule Document