A COMPARATIVE STUDY OF THE

APPLICATION OF THETA THEORY IN

URDU AND ENGLISH LANGUAGES

THESISTHESISTHESISTHESIS

Submitted for the Degree of

Doctor of PhilosophyDoctor of PhilosophyDoctor of PhilosophyDoctor of Philosophy

InInInIn

LinguisticsLinguisticsLinguisticsLinguistics

By

SHAMIM FATMA

Under the Supervision of

Dr. Samina A.A. Surti Prof. Anjani Kumar Sinha

(Supervisor) (Co-Supervisor)

DEPARTMENT OF LINGUISTICS ALIGARH MUSLIM UNIVERSITY

ALIGARH (INDIA)

2012

ABSTRACTABSTRACTABSTRACTABSTRACT

The work is intended to study the argument structures of Urdu and

English verbs to bring out the basic difference, if any, between these two

languages in this regard. It explains why such a study is necessary. It

begins with a theoretical discussion on what has already been done within

the framework of Principles and Parameters theory and the Minimalist

Program. It discusses theta roles required by verbs in Urdu and English

and compares them to find out the ways in which they are similar to or

different from each other.

The First Chapter deals with some theoretical issues that led to the

evolution of the theta theory within the principles and parameters theory

(Chomsky 1981). It discusses the significance of theta theory for the well-

formedness and semantic interpretation of a sentence. It discusses briefly

the crucial justification for thematic relations given by Gruber (1976,

(originally 1965)) which was followed by Jackendoff’s conceptual

approach to the thematic relations (1990). Further it has discussed how

Chomsky’s concept of thematic structure is different from Jackendoff’s

conceptual structure. It argues that Chomsky’s theta theory offers a more

satisfactory solution than Jackendoff’s in so far as the thematic structures

of sentences of natural languages are concerned.

The theta criterion which is an important component of theta theory has

been discussed in details. It retains its significance even in the Minimalist

Program which questions almost every assumption of Principles and

Parameters theory. Chomsky notices that Theta roles are a fundamental

property of grammar and their proper assignment to the arguments is

necessary to satisfy the principle of Full Interpretation (FI); otherwise a

2

derivation gets crashed. The chapter presents the aims and objectives of

this work and argues that the comparative study of the thematic structures

in Urdu and English is essential to understand the problem faced by the

native speakers of Urdu learning English as a second language.

The Second Chapter contains detail of the possible theta roles of Urdu

and English verbs. It assumes the hypothesis that verbs of similar

semantic types in different languages have similar argument structures at

LF but all of them are not expressed the same way at PF. Some of them

may be implicit or theta-absorbed in a specific language. Here an attempt

has been made to find out whether the difference between Urdu and

English is due to some semantic or structural factors.

On the basis of the required number of arguments for the well-

formedness of a sentence, we have classified Urdu verbs as verbs with

one essential argument, verbs with two essential arguments, verbs with

three essential arguments and verbs which need an embedded sentential

argument. Based on their semantic affinity verbs with two arguments

have been further subdivided as inchoative transitive verbs, verbs of

creation, accomplishment, motion, physical and mental perception and

Performative verbs. The sentences with causative verbs in Urdu have

been discussed separately.

We have found that the verbs with one argument have a lot of similarity

between Urdu and English. There are some verbs which are basically

transitive in English but may be used intransitively, if only theme is to be

focused. In Urdu, on the other hand, there are separate though

morphologically related –sets of transitive and intransitive verbs (e.g. in

English open is either transitive or intransitive and in Urdu khulnaa ‘to

get open’ is intransitive and kholnaa ‘to open’ is transitive). The verb

3

with one argument indicating volitional action takes agent and the verbs

which indicate physical and mental perception take experiencer as their

argument. If an argument voluntarily does the act to draw the attention of

someone, it is an agent rather than an experiencer. Intransitive verbs

carrying medio-passive sense take theme as their essential argument in

the subject position in Urdu and have their medio-passive counterparts in

English. Such verbs are basically transitive in English but may be used

intransitively if only the theme is to be focused. Some intransitive verbs

take patient as an argument in both Urdu and English.

The verbs which need two essential arguments in Urdu may have agent

and theme, agent and patient, agent and goal or theme and experiencer as

their arguments. The verbs which take agent and theme as their

arguments are inchoative transitive verbs (e.g. paka:na: ‘to cook), verbs

of creation (e.g. ka:Rhna: ‘to embroider’), accomplishment (e.g. ji:tna:

‘to win’), motion (e.g. Dhakelna: ‘to push’), performative verbs (e.g. rad

karna: ‘to reject’) or verbs of physical (ta:kna: ‘to stare at’) and mental

perception (e.g. bhu:lna: ‘to forget’). Some other inchoative transitive

verbs (e.g. ma:rna: ‘to kill’), verbs of accomplishment (e.g. dabočna: ‘to

grab’), motion (e.g. ragedna: ‘to chase’), verbs of physical (e.g. dekhna:

‘to see’) and mental (e.g. ta:Rna: ‘to perceive’) and emotive verbs (e.g.

jhiRakna: ‘to scold’) may take agent and patient as well. The theme

argument of the verbs bearing two arguments agent and theme and the

patient argument of the verbs bearing agent and patient are obligatory in

the structures of both the languages. There are verbs of motion (e.g.

gherna: ‘to surround’) in both Urdu and English which take agent and

goal. We found the motion verbs (e.g. čaRhna: ‘to climb) of Urdu and

English appear somewhat different here in the nature of their arguments

4

i.e. in Urdu they take PoP as goal but their English counterparts take NP

as goal. Some emotive verbs (e.g. Khush karna: ‘to amuse / please’) take

theme and experiencer as their essential arguments. Some verbs of Urdu

and English take three arguments: (a) agent, theme and goal (bhejna: ‘to

send’) (b) agent, theme and location (e.g. bharna: ‘to fill’(c) agent,

theme and source (ha:sil karna:’to obtain’) (d) agent, theme and

experiencer (e.g. mãDhna: ‘to impose’(e) agent, patient and instrument

(e.g.čubho:na: ‘to pierce’) or (f) agent, patient and source (e.g. bedakhal

karna: ‘to expel’) . Something common in these verbs with three

arguments is that they have agent and theme or patient as their first two

arguments. They differ in regard to the third argument they take. Some

verbs of exchange or transaction (e.g. bečna: ‘to sell’) need θ-role of goal

both in Urdu and English. There are some verbs (bharna: ‘to fill’) in

Urdu may take the θ-role of location. The sentence may be well-formed

without the location argument in Urdu but in English, the location is

essential with agent and theme to generate grammatical sentence with

verbs such as bharna: ‘to fill’. Agent can be deleted when the sentence is

passivised in both the languages. In Urdu, a sentence may be complete

with theme only but in English both theme and location arguments are

needed with verbs such as to fill up. There are verbs of exchange (e.g.

ha:sil karna: ‘to obtain’) which require the theta role of source as their

third argument. mãDhna: ‘to impose’ type of verbs in Urdu and English

need agent, theme and experiencer. The verb čubho:na: ‘to pierce’ takes

agent, patient and instrument. Some locational verbs (e.g. haTa:na: ‘to

remove’) need agent, patient and source to complete a sentence.

The verbs which need an embedded sentential argument have been

discussed separately. Such verbs (e.g. tardi:d karna: ‘to deny’) take CP

5

or IP as one of their essential arguments which bears the theta role of

theme and the other argument may be agent or experiencer. Agent can be

dropped from Urdu sentences, as it is a pro-drop language. Verbs with

three essential arguments may also take a clause as their theme and other

two arguments are agent and experiencer. The experiencer argument of

the verb (e.g. mashwera dena: ‘to advise’) can be dropped in Urdu but its

equivalent in English should be overt for the sentence to be well-formed.

The causative verbs have been discussed separately. We have adopted the

framework of Hale and Keyser (1993) to tackle the complexities of

causatives and shown that they can be discussed the same way in which

other verbs with embedded sentences have been discussed. Superficially,

causative verbs seem to have (a) causer, agent and theme (b) causer,

agent and patient (c) causer, agent, theme and location (d) causer, agent,

theme and source or (e) causer, agent, theme and goal. The configuration

of causatives in English appears biclausal (e.g. make…to do or cause…to

do). In Urdu, causative verbs are usually formed by adding the suffix (-

w)a: to an intransitive or transitive verb. Almost all intransitive verbs

have transitive and causative counterparts (e.g. Khulna: ‘to get open’,

kholna: ‘to open’and kholwa:na: ‘to cause to open’) but all transitive

verbs do not have intransitive counterparts (e.g. paRhna: ‘to read’).

The thematic structure of causatives is complex to some extent, in the

sense that one argument is an NP which is the agent and another is an

embedded clause as its theme. This theme argument which is a clause has

its own thematic structure which may have (a) agent and theme (b) agent

and patient (c) agent, theme and location (d) agent, theme and source or

(e) agent, theme and goal. In Urdu, the agent argument can be deleted

optionally but it is not possible to do so in English. Keeping the

6

complexities of causatives in mind, we have initially adopted Grimshaw’s

approach and considered the agent of the causative sentence as ‘causer’

and the agent of the embedded sentence as ‘agent’ but we are not satisfied

with this approach, as it violates UTAH (Uniformity of Theta Assignment

Hypothesis). In the last section of the chapter, we have discussed the

drawback of Grimshaw’ hypothesis and given the priority to Hale and

Keyser’s approach (1993) over it, which enabled us to represent identical

thematic relationship in both Urdu and English.

In the Third Chapter, we have examined the reason of absence of an

argument in the PF of various sentences. If an argument is absent from

the surface but not due to its movement or because it is a pro-drop

language or it is represented by PRO then it may be due to incorporation

or theta absorption. We have discussed the concept of incorporation

proposed by Gruber (1976) and Baker (1988) and differentiated them

with Chomsky’s concept of θ-absorption (1981, 1986) Jackendoff’s

concept of argument fusion (1990). The incorporation represents the

phenomenon where two elements become one unit and are visible on the

surface (e.g. to babysit) whereas in theta absorption, an unspecified

argument is dropped from the surface if it can be inferred from the

semantics of verb (e.g. he is reading, meaning reading something

readable).

We adopted the concept of θ-absorption given by Chomsky (1981, 1986).

We have investigated the phenomenon of theta absorption in Urdu verbs

to find out which type of theta roles might be absorbed in the verb and

compared them with their English counterparts to check whether these

two languages are alike or behave differently. The most common

argument which is absorbed in the verb in both Urdu and English is

7

theme. It is also found that it is absorbed in a situation when it has an

unspecified reference or when it can be inferred from the predicate which

absorbs it. Some verbs (e.g. ga:na: ‘to sing’) can be used either

transitively or intransitively. When transitive, they absorb unspecified

theme. They indicate a profession when used in an indefinite present

tense. Some verbs (da:Rhi: bana:na: ‘to shave’) may take reflexive of an

agent as the either theme or patient. There are some verbs (e.g. pi:na: ‘to

drink’) which absorb a specified theme and are used in an indefinite

present tense. In Urdu, the verbs of transaction can omit the theme if it

can be recovered from the PF but in English it is not possible at all. Verbs

of perception in Urdu absorbs the theme if it can be recovered

contextually but it is not so in English. When these verbs are used in the

abilitative sentence, they may absorb the theme in Urdu and English.

Some other arguments which are absorbed in both the languages are

experiencer, goal, source, location and instrument. The experiencer

argument is absorbed in Urdu but not in English. Performative verbs

(iqra:r karna: ‘to testify’) may absorb experiencer in both the languages.

Urdu and English appear alike in the absorption of goal and source.

Unspecified location arguments seem to be absorbed in the verbs (e.g.

bharna: ‘to fill’) in Urdu but not necessarily so in English. However,

some verbs (e.g. bona: ‘to sow’) absorb location argument in both the

languages. The instrument argument is also found to be absorbed in the

verbs (e.g. bunna: ‘to kneat’) of both Urdu and English. The goal or

source can be absorbed in the verb of transaction but not theme. Theme is

absorbed in such verbs only in Urdu when it can be recovered

contextually. In passive sentences, the agent argument is absorbed by the

morphological marking on the verbs in both the languages but in active

sentence, it is not the case. Agent argument can be dropped in Urdu, as it

8

is a pro-drop language. The theme argument is noticed to be absorbed

more easily. The theme argument which is lowest in hierarchy, as

Grimshaw (1990) mentioned, is seem to be absorbed more easily than the

other arguments in both the languages i.e. Urdu and English.

In Chapter Four, we have focused on the relation of theta theory with

case theory. We have discussed the role of the case theory in the

derivation of a grammatical sentence. Though, the theta role assignment

is a basic property of grammar; it cannot alone generate a well-formed

sentence. All the functional elements such as case markers, tense and Agr

features should be there to arrange them structurally. The case theory

discusses the distribution of overt NPs in such a manner that they are

assigned proper case and the case filter, its basic component, says that

every lexically realized NP must be assigned proper case, for which

features of case and case assigners must be in proper configuration. We

have traced out that the interaction takes place between the case theory

and movement rather than theta theory and movement. The assignment of

theta roles is not concerned with movement because it is assigned before

movement.

During the discussion, we noticed that there is strong relation between

cases and theta roles. Hence, we have examined the types of cases taken

by a particular theta roles and checked whether an NP with a specific θ-

role (e.g. agent) is assigned a specific case (e.g. nominative) uniformly or

it is assigned a specific case in one situation and another case in a

different situation. As a result, we have found that there is no one to one

correspondence between the theta roles and the cases. In other words, one

case (e.g. nominative) may correspond to more than one θ-role (e.g.

agent, theme, patient or experiencer) and one θ-role (e.g. agent) may

9

correspond to more than one case (e.g. nominative, ergative, dative and

instrumental). The realization of agent and theme or patient is found to be

different in Urdu because of the difference in the nominative-accusative

structure in some situation and ergative-absolutive structure in some other

situation.

We started with the hypothesis that Urdu and English differ at PF in

regard to the application of theta theory but are alike at LF. We have

analyzed the argument structures of several types of Urdu verbs and

compared them with those of their English equivalents. We have reached

the conclusion that the number of arguments needed by a specific type of

verb is the same in Urdu and English but their overt realization varies

across these two languages due to the difference in the absorption of

some arguments such as, theme and experiencer. We have shown that our

hypothesis was correct. This work may be helpful in teaching English to

Urdu speakers. It may help them in using well-formed and complete

sentences in English. The section on the relation of Theta theory with

case theory may also provide useful hints about the use of specific

prepositions in English. This work may also be helpful to those working

in the area of natural language processing.

DEPARTMENT OF LINGUISTICS

ALIGARH MUSLIM UNIVERSITY

ALIGARH-202002 (U.P) INDIA

Date___________________2012

CCCCCCCCeeeeeeeerrrrrrrrttttttttiiiiiiiiffffffffiiiiiiiiccccccccaaaaaaaatttttttteeeeeeee

This is to certify that the thesis entitled “A COMPARATIVE

STUDY OF THE APPLICATION OF THETA THEORY IN URDU

AND ENGLISH LANGUAGES” submitted by Miss Shamim Fatma

for the award of the degree of Doctor of Philosophy in Linguistics has

been completed under our supervision.

It is further certified that Miss Shamim Fatma has fulfilled all the

terms and conditions laid down by the university with regard to the Ph.D.

Degree and to the best of our knowledge the thesis contains her own

research. This thesis conforms to the standard of Aligarh Muslim

University.

Dr. Samina A.A. Surti Prof. Anjani Kumar Sinha (Supervisor) (Co-Supervisor)

i

CONTENTSCONTENTSCONTENTSCONTENTS

Certificate

Acknowledgement i-iv

Transcription Symbols v-ix

Abbreviations x-xii

CHAPTERS Page No.

Chapter – I Introduction 1-27

1.0 Theoretical Background 1

1.1 Theta Theory as a Principle of Grammar 7

1.2 Assignment of Theta Roles 8

1.3 Relevance of Theta Criterion 11

1.4 Jackendoff’s Criticism of Chomsky’s Theta Criterion 14

1.5 Theta Theory in the Minimalist Program 21

1.6 Aims and Objectives of the Thesis 23

1.7 An Outline of the Thesis 25

Chapter – II: THE THEMATIC STRUCTURE OF URDU

VERBS

28-90

2.0 Essential and Non-essential Arguments in a Thematic

Structure

28

2.1 Verbs with One Argument 29

2.1.1 Verbs that need only agent as an essential argument 29

2.1.2 Verbs that need experiencer as an essential argument 32

2.1.3 Verbs that need theme as an essential argument 35

2.1.4 Verbs that need patient as an essential argument 39

2.2 Verbs with Two Essential Arguments 41

2.2.1 Verbs which take agent and theme as essential

arguments

41

ii

2.2.1.1 Inchoative Transitive verbs 41

2.2.1.2 Verbs of Creation 44

2.2.1.3 Verbs of Accomplishment 45

2.2.1.4 Verbs of Motion 47

2.2.1.5 Performative Verbs 48

2.2.1.6 Verbs of Physical and Mental Perception 50

2.2.2 Verbs which take agent and patient as essential

arguments

52

2.2.2.1 Inchoative Transitive Verbs 52

2.2.2.2 Verbs of Accomplishment 54

2.2.2.3 Emotive verbs 55

2.2.2.4 Verbs of Physical and Mental Perception 58

2.2.2.5 Verbs of Motion 58

2.2.3 Verbs which take agent and goal as their essential

arguments

59

2.2.4 Verbs which take theme and experiencer as their

essential arguments.

60

2.3 Verbs with Three Essential Arguments. 61

2.3.1 Verbs which take agent, theme and goal as their

essential arguments

61

2.3.2 Verbs which take agent, theme and location as their

essential arguments

63

2.3.3 Verbs which take agent, theme and source as their

essential argument.

65

2.3.4 Verbs which take agent, theme and experiencer as their

essential arguments

67

2.3.5 Verbs which take agent, patient and instrument as their

essential arguments

67

iii

2.3.6 Verbs which take agent, patient and source as their

essential arguments

68

2.4 Verbs which take a sentential argument 69

2.4.1 Verbs which take agent and theme as their essential

arguments

69

2.4.2 Verbs that take agent, theme and experiencer as their

essential arguments

72

2.5. The Thematic Structure of Causative Verbs 76

2.5.1. Causative verbs which take causer, agent and theme as

essential arguments

79

2.5.2. Causative verbs which take causer, agent and patient

as their essential arguments.

81

2.5.3. Causative verbs which take causer, agent, theme and

location as their essential arguments

81

2.5.4. Causative verbs which take causer, agent, theme and

source as their essential arguments.

82

2.5.5. Causative verbs which take causer, agent, theme and

goal as their essential arguments.

83

2.6. An Alternative Analysis of causatives 84

Chapter – III: THETA ABSORPTION 91-117

3.0 Arguments in a Sentence 91

3.1 Incorporation 93

3.2 Theta Absorption 95

3.2.1 Absorption of Theme 99

3.2.2 Absorption of Experiencer 108

3.2.3 Absorption of Goal 110

3.2.4 Absorption of Source 111

iv

3.2.5 Absorption of Location 112

3.2.6 Absorption of Instrument 113

3.2.7 Absorption of Agent 115

Chapter – IV: THETA THEORY AND CASE THEORY 118-141

4.0 Relation Between Theta Theory and Case Theory 118

4.1 Parallelism Between Theta Roles and Cases 129

4.1.1 Agent 130

4.1.2 Theme 132

4.1.3 Patient 133

4.1.4 Experiencer 135

4.1.5 Instrument 137

4.1.6 Source 138

4.1.7 Goal 138

4.1.8 Location 139

Chapter – V: SUMMARY AND CONCLUSION 142-152

5.0 Summary 142

5.1 Conclusion 151

Bibliography 153-162

i

AcknowledgementAcknowledgementAcknowledgementAcknowledgement

All praises and thanks are to the almighty Allah whose benign

benediction gave me the needful strength to accomplish this uphill

task.

I am greatly indebted to my supervisor Dr. Samina A. A. SurtiDr. Samina A. A. SurtiDr. Samina A. A. SurtiDr. Samina A. A. Surti

for her kindness, inspiring attitude, support and suggestions. Her

constant encouragement and help have been of great value to me.

I would like to express my sincere gratitude and utmost regards to

Prof. Anjani Kumar SinhaProf. Anjani Kumar SinhaProf. Anjani Kumar SinhaProf. Anjani Kumar Sinha, my co-supervisor for taking me under

his supervision and for his illuminating ideas, scholarly guidance

and creative supervision. I was scared of sytax but his sound

advise, good teaching and good ideas enabled me to do research in

this area. His intellectual inputs and unflinching encouragement

made it possible to accomplish this work. It has been a priviledge

to work under his supervision.

I would like to express my heartful thanks and deep gratitude to

my teachers Prof. Imtiaz HasnainProf. Imtiaz HasnainProf. Imtiaz HasnainProf. Imtiaz Hasnain, Chairman, Department of

Linguistics, Aligarh Muslim University, Prof. A. R. FatihiProf. A. R. FatihiProf. A. R. FatihiProf. A. R. Fatihi, Dr. Dr. Dr. Dr.

Masood a. BegMasood a. BegMasood a. BegMasood a. Beg, Prof. K. S. MustafaProf. K. S. MustafaProf. K. S. MustafaProf. K. S. Mustafa, Dr. Shabana HameedDr. Shabana HameedDr. Shabana HameedDr. Shabana Hameed, Dr. Dr. Dr. Dr.

ii

Aziz KhanAziz KhanAziz KhanAziz Khan, Dr. Nazreen B. LashkarDr. Nazreen B. LashkarDr. Nazreen B. LashkarDr. Nazreen B. Lashkar, whose interest in my work

was a constant source of encouragement to me.

I am indebted to Mrs. Usha Kiran SinhaMrs. Usha Kiran SinhaMrs. Usha Kiran SinhaMrs. Usha Kiran Sinha, and Dr. Dr. Dr. Dr. Anant Kumar Anant Kumar Anant Kumar Anant Kumar

SinhaSinhaSinhaSinha for their warm welcome and genuine hospitality to me

during my visit to their home. The loving and caring behaviour of

Mrs. Usha made my few hours stay memorable.

I am also grateful to the seminar incharge Mr. Mohammad Mr. Mohammad Mr. Mohammad Mr. Mohammad

Najeebul HasanNajeebul HasanNajeebul HasanNajeebul Hasan and Mr. GoyalMr. GoyalMr. GoyalMr. Goyal who supported me in different

ways and deserve special mention.

Thanks are also due to Ms. Noor BanoMs. Noor BanoMs. Noor BanoMs. Noor Bano, Mr. Mohammad Mr. Mohammad Mr. Mohammad Mr. Mohammad

Haseebur RahmanHaseebur RahmanHaseebur RahmanHaseebur Rahman and rest of the members of the staff of the

department of Linguistics.

I am extremely grateful to Dr. Ratna SanyalDr. Ratna SanyalDr. Ratna SanyalDr. Ratna Sanyal, IIIT- Allahabad,

for her support.

I wish to convey my special thanks to my friends and classmates

Afreen, Kausar, Khushbu, Naheed, Nazia Aapa, Paikar Aapa, Afreen, Kausar, Khushbu, Naheed, Nazia Aapa, Paikar Aapa, Afreen, Kausar, Khushbu, Naheed, Nazia Aapa, Paikar Aapa, Afreen, Kausar, Khushbu, Naheed, Nazia Aapa, Paikar Aapa,

Prerna Aapa, Purnima, Shagufta, , Shaista Aapa, Sheeja, Imran Prerna Aapa, Purnima, Shagufta, , Shaista Aapa, Sheeja, Imran Prerna Aapa, Purnima, Shagufta, , Shaista Aapa, Sheeja, Imran Prerna Aapa, Purnima, Shagufta, , Shaista Aapa, Sheeja, Imran

and SohelSohelSohelSohel for their affection and support.

Special thanks go to my friends NadiaNadiaNadiaNadia and SalmeenSalmeenSalmeenSalmeen for their

affection, unconditional support and hospitality.

iii

I am also thankful to the students and research scholars of the

department of linguistics for showing their interests in my work

which made me more ennthusiastic.

I am extremely grateful to Shabnam and ShaziaShabnam and ShaziaShabnam and ShaziaShabnam and Shazia, whom I consider

my best friends since high school. Their love, affection and

encouragement were source of strength for me.

I wish to convey my regards to Nazir chachoo Nazir chachoo Nazir chachoo Nazir chachoo to encourage me at

each and every step of my study.

My sincere gratitude and regards to my grandparents Late

Md. MalihUddin Ahmad Md. MalihUddin Ahmad Md. MalihUddin Ahmad Md. MalihUddin Ahmad and Late Israr Fatma Israr Fatma Israr Fatma Israr Fatma who had vital

role in raising me with their caring and gentle love.

I am forever indebted to my Father Md. SabihMd. SabihMd. SabihMd. Sabihuddin Ahmaduddin Ahmaduddin Ahmaduddin Ahmad,

beloved Mother Sayeeda KhatoonSayeeda KhatoonSayeeda KhatoonSayeeda Khatoon, my chachoo Md. Sagihuddin Md. Sagihuddin Md. Sagihuddin Md. Sagihuddin

AhmadAhmadAhmadAhmad, dear chachi Sufiya KhatoonSufiya KhatoonSufiya KhatoonSufiya Khatoon, my phuphas

Noor MohammadNoor MohammadNoor MohammadNoor Mohammad and Md. DanishMd. DanishMd. DanishMd. Danish, my loving and caring phuphis

Ibrar FatmaIbrar FatmaIbrar FatmaIbrar Fatma and Kaniz FatmaKaniz FatmaKaniz FatmaKaniz Fatma for providing me the opportunity

to make this academic pursuit. Their immeasurable sacrifices and

countless blessings are the source of my strength. I fail to find

words to express my gratitude and regards to them for their

understanding, endless patience and encouragement.

iv

Very special thanks should be recorded to my loving brothers

Tanweer, Azhar, DaudTanweer, Azhar, DaudTanweer, Azhar, DaudTanweer, Azhar, Daud, Yusuf,Yusuf,Yusuf,Yusuf, Tabish,Tabish,Tabish,Tabish, RaghibRaghibRaghibRaghib and my sisters

Saba,Saba,Saba,Saba, ShabnamShabnamShabnamShabnam, NaziaNaziaNaziaNazia and NishiNishiNishiNishi, who are very supportive and

caring siblings. They provided me enough mental and moral

strength to accomplish this task successfully.

I am taking the responsibility of all shortcomings in this thesis if

any and rendering my unconditional apology for the same.

(Shamim Fatma)(Shamim Fatma)(Shamim Fatma)(Shamim Fatma)

v

Transcription Symbols Used in This Work

Symbol

Used Description

Urdu

Character

p Voiceless, Bilabial, Stop �

ph Aspirated, Voiceless, Bilabial, Stop �

b Voiced, Bilabial, Stop �

bh Aspirated, Voiced, Bilabial, Stop �

t Voiceless, Alveolar, Stop �

th Aspirated, Voiceless, Alveolar, Stop �

d Voiced, Alveolar, Stop �

dh Aspirated, Voiced, Stop �

T Voiceless, Retroflex, Stop

vi

Th Aspirated, Voiceless, Retroflex, Stop �

D Voiced, Retroflex, Stop �

Dh Aspirated, Voiced, Retroflex, Stop �

č Voiceless, Palatal, Stop

čh Aspirated, Voiceless, Palatal, Stop �

j Voiced, Palatal, Stop �

jh Aspirated Voiced, Palatal, Stop �

k Voiceless, Velar, Stop �

kh Voiceless, Aspirated, Velar, Stop �

g Voiced, Velar, Stop �

gh Voiced, Aspirated Velar, Stop �

vii

q Voiceless, Stop, Uvular �

m Bilabial, Nasal �

n Alveolar, Nasal �

ng Velar, Nasal ��

l Alveolar, Lateral �

r Alveolar, Trill �

R Retroflex, Flap �

Rh Aspirated, Retroflex, Flap �

f Voiceless, Labia-dental, Fricative �

w Voiced, Aspirated Labio-dental, Fricative �

s Voiceless, Alveolar, Fricative � �! "

viii

sh Voiceless, Palato-alveolar, Fricative #

Kh Voiceless, Velar, Fricative $

G Voiced, Velar, Fricative %

h Voiced, Velar, Fricative , &

Y Palatal, Semi-vowel '

w Labio-dental, Semi-vowel �

z Voiced, Alveolar, Fricative �( �) �* +

i Front, Low, High, Short , -

i:

Front, High, Long , -'

e Front, Mid, Long .-

a Central, Short /-

ix

a:

Central, High 0-

u

Back, Low, High, short 1-

u: Back, High, Long �1-

o Back, Mid, Long �-

ai Diphthong ./-

au Diphthong �/-

˜ Nasalization 2

Note - Proper names have been typed the way they are usually written;

they have not been transcribed.

x

ABBREVIATIONS

The following abbreviations have been used in this work.

a – Adjunct : Argument Adjunct

A – Position : Argument Position

a – Structure : Argument structure

A’-Position : Non- Argument Position

ABL : Ablative

ABS : Absolutive

ACC : Accusative

AFF : Affect

AGR : Agreement

AGRo : Agreement- Object

AGRs : Agreement-Subject

AGRsP : Agreement Phrase

C-Command : Constituent Command

COMP : Complementizer

CP : Complimentizer Phrase

C-Selection : Constituent Selection

DAT : Dative

ECM : Exceptional Case Marking

xi

ERG : Ergative

fem : Feminine

FI : Full Interpretation

INFL : Inflection

INST

intr

:

:

Instrumental

Intransitive

LF : Logical Form

LOC : Locative

mas : Masculine

m- command : Maximal(Projection) Command

N : Noun

NOM : Nominative

NP : Noun Phrase

PF : Phonetic form

PP

PoP

:

:

Prepositional Phrases

Postpositional Phrase

PR : Present

PERF : Perfective

PROG : Progressive

sg : Singular

xii

S- Selection : Semantic Selection

SD : Structural Description

SPEC : Specifier

T : Tense

TP

tr

:

:

Tense Phrase

Transitive

V : Verb

VP : Verb Phrase

vp : Light Verb Phrase

θ- Role : Theta Role

θ-Position : Theta Position

θ’-Position : Non- Theta Position

θ-Grid : Theta Grid

φ : Zero

CChhaapptteerr--11

IIIIIIIInnnnnnnnttttttttrrrrrrrroooooooodddddddduuuuuuuuccccccccttttttttiiiiiiiioooooooonnnnnnnn

Chapter Chapter Chapter Chapter –––– I I I I

INTRODUCTION

1.0 Theoretical Background

American structural linguistics concentrated on the structures of

phonemes, morphemes, phrases and sentences and ignored semantics.

While developing generative grammar, Chomsky suggested that the

relation between syntax and semantics cannot be ignored in a grammar

that is descriptively and explanatorily adequate. The notion of “deep”

structure that he suggested in 1965, captured an important aspect of

meaning of the sentence under investigation. Despite his observation on

the nature of grammatical rules that relate syntactic structure to

semantics, Chomsky (1965) did not propose any explicit mechanism for

the analysis of meaning. Semantics was integrated with the generative

grammar at the level of “Deep Structure”. But he observed, “In general,

as syntactic description becomes deeper, what appears to be a semantic

question falls increasingly within its scope” (Chomsky 1964:36). Though

Chomsky did not directly contribute to the development of the semantic

interpretation rule (SIR), he accepted the proposal of Katz and Fodor

(1963), which determined how the structural combinations of lexical

items assigned meaning to the sentences as a whole. They claimed that

these rules help the speaker in disambiguating ambiguous sentences and

in understanding paraphrase relation between sentences. Katz and Postal

(1964) proposed the hypothesis that all information necessary for the

application of the projection rules is present in the underlying syntactic

structure. Chomsky clearly stated that transformational rules do not affect

meaning. By assuming that everything necessary for semantic

1

interpretation is present in the deep structure; they put the whole burden

of semantic interpretation on deep structure1.

Gruber (1976, (originally 1965)), developed this approach of semantic

interpretation further by claiming that the underlying structure of a

sentence generated before semantic and syntactic interpretation can be

deeper than the level of “Deep structure” in the sense that it could be

derived prior to the insertion of lexical items in the base structures. This

mechanism was termed “Pre-lexical Categorical Structure” (pp. 5-8).

Gruber developed this approach to investigate the relationship between a

verb and its argument(s). He argued that a verb needs one or more

argument(s), each of which must have a definite kind of relation with the

verb. This idea led him to the introduction of concepts such as ‘agent’,

‘theme’, ‘instrument’, ‘experiencer’, ‘accompaniment’, ‘location’, ‘goal’,

‘source’, ‘direction’, etc. The advent of this notion led to several

questions which remained unanswered in the Standard Theory (Chomsky

1965). Gruber used these relationships to account for the well-formedness

of a sentence. He claimed that sentences such as (1a) and (1b) are

synonymous not only because they have the same argument(s) and verb,

but the same relation with the verb.

1a. Akbar killed Aamir.

b. Aamir was killed by Akbar.

1 Later on, for example, Chomsky (1977:170) admitted that the term deep structure was, in some ways

“overestimated”, and that surface structure was associated “directly with semantic interpretation” (ibid,

171).

2

In (1a), the subject NP Akbar is the agent who is the performer of the

action. The object NP Aamir is the patient who is the target of the action.

In (1b), the NP Aamir is the subject; still it is the patient because the

relation between the subject Aamir and predicate was killed is the same as

in (1a). In (1b), NP Akbar is the object to the preposition by and still it

bears the theta role of agent in relation to the verb. (2) is different from

(1a,b) because the thematic relations between the predicate and its NPs

are different in it.

2. Aamir killed Akbar.

In (2), the NP Aamir is the agent as it does the action; whereas NP Akbar

is the patient as it is the entity which is affected by the action. Thus, we

notice that in (1) and (2), it is the thematic relations of NPs with the verb

which helps us in determining their meaning.

Gruber further noticed an important process which he termed as

incorporation2 (p: 9-36) in which certain elements that are overtly

expressed at the pre-lexical level are implied by the verb. For instance,

we may look at (3a) and (3b).

3a. Did the pencil pierce through the cushion?

b. No, it did not pierce it.

In (3a), through is overtly mentioned but in (3b), it is implied. Gruber

(1976:9) argued that in (3b) through is incorporated in pierce. He

emphasized that the lexical item must be the neighbour of the one into

which it is incorporated. He argued that incorporation takes place if and

only if the predicate of a sentence has the ability to reflect the

incorporated element.

2 This notion of “incorporation” is different from Baker’s (1988) use of the term.

3

4a. Akbar is eating something edible.

b. Akbar is eating.

c. Akbar is eating a mango.

d. Akbar is eating a marble.

In (4a), Akbar is eating some sort of food which is not specified. The

unspecified object NP can be incorporated in the verb, as in (4b). It is

clear that such incorporation does not change the meaning of the sentence

because the unspecified NP is recoverable from the verb. In (4c) and (4d),

we have specified objects i.e., a mango and stone respectively. If they

cannot be recovered from the verbs; they have to be overt3. This notion of

“incorporation” is different from Baker’s (1988) use of the term. As will

be discussed later on, Chomsky (1995a) uses the term theta-absorption

for a phenomenon as in (4b).

Fillmore (1968: 21-26) suggested a modification to the theory of

transformational grammar by introducing a universal system of “deep

structure cases” of “a purely syntactic nature” which was somewhat

similar to Gruber’s concept of relationship between a verb and its

arguments. He proposed cases such as agentive, instrumental, dative,

factitive, locative and objective and left the door open for other cases.

Since he tried to associate these conceptual cases with the case forms

which are found in natural language in the forms of inflections,

3 As will be discussed later on, Chomsky (1995a) uses the term theta absorption for a phenomenon as

in (4b).

4

prepositions or postpositions, he needed subjectivisation (or subject-

selection) rule and unique case assignment rule.

Jackendoff (1972: 29-36) reviewed the concept of relationship proposed

by Gruber and Fillmore and preferred the fundamental notion of thematic

relations to that of case relations on the ground that the unique case

assignment rule could not account for the possible ambiguity of (5) in

which Max can be an agent, a theme or both an agent and a theme at the

same time.

5. Max rolled down the hill.

Max could roll down the hill voluntarily and be the victim of undergoing

the motion as well. Fillmore’s case assignment rule needed a complex

solution to this problem in comparison to which Gruber’s solution was

simpler. Jackendoff (1972) accepted Gruber’s proposal of the centrality

of the theme and called the whole system thematic relations. In his view,

the concept of dual thematic relation was needed to explain the

conceptual relation between verbs of converse relation such as buy and

sell.

Jackendoff (1983) further contributed to the Gruber’s theory of thematic

relation. He agreed to Gruber’s analysis and argued that verbs may or

may not differ in their “semantic function” at the conceptual level and

this could explain the similarity between verbs of converse relation such

as buy and sell.

His concept of ‘functional structure’ represented the relation between a

predicate and its arguments at the conceptual level. Jackendoff (1990: 22)

refined his concept of thematic relations by decomposing conceptual

structure into conceptual constituents, each of which belongs to one of a

5

small set of “conceptual parts of speech” such as [Thing], [Event],

[State], [Action], [Place], [Path], [Property] and [Amount]. He argued

that each major syntactic constituent maps a conceptual constituent into

the meaning of a sentence, thus establishing a relation between syntactic

and conceptual structures. The following example illustrates this relation.

6. Akbar ran towards home.

In (6) Akbar and home correspond to Thing-constituent, the PP towards

home corresponds to a Path- constituent and the entire sentence

corresponds to an Event-constituent. He pays attention to whether the

verb shows action, event or state. Depending upon the conceptual

constituent of the verb, the number and types of arguments differ. For

instance, CHANGE is one of the conceptual constituent of action verb,

which denotes the semantic function, of taking thing as an argument from

an initial to a final stage.

7. Akbar goes to school.

In (7), Akbar and school correspond to Thing-constituent. Likewise

CAUSE, another conceptual constituent of a verb, takes ‘thing’ as an

argument in the initial and final stage.

8. Akbar forced Aamir to go away.

In (8), Akbar and Aamir are things and forced and to go away are action

and event respectively. Finally Jackendoff (1990:46-48) came to the point

that thematic roles are part of the level of conceptual structure, and not of

syntax. He claimed that the terms Theme, Agent and so on are not

primitives of a syntactic theory, rather they are relational notions defined

structurally over conceptual structure. Argument structure can be termed

as an abbreviation for the part of conceptual structure that is “visible” to

the syntax.

6

1.1 Theta Theory as a Principle of Grammar

Chomsky (1981) included the theta theory as one of the “subsystems of

principles” which is “concerned with the assignment of thematic roles (θ-

roles)4. Initially he considered Jackendoff’s work on this topic “quite

interesting” from the point of view of “descriptive semantics” but later on

he accepted it as a “unifying notion”5. The theory claims that each and

every NP receives a specific theta role. Chomsky (1981:35) used the term

“NP argument” to include names, variables, anaphors, pronouns but not

idiom chunks or elements inserted to occupy an obligatory position of

syntactic structure (e.g., expletives it or there ). He asserted that an

argument must receive a theta role from a head. The overt anaphors, R-

expressions and pronominals (including empty elements such as PRO and

pro) are all arguments because they have theta roles. The position to

which a theta role is assigned is an argument position or A-position and

this argument position is termed as a theta position. According to the

projection principle, the object position is a theta position while the

subject position [Spec, IP] may or may not be a theta position. If it is

filled by an expletive, it is not a theta position; if by an argument, it is a

theta position. [Spec, IP] is, therefore, a potential theta position. An actual

or potential theta position is an A-position. A'-position (A-bar position) is

a position where no theta role is assigned to an NP. [Spec, CP] is an A'-

position. The relation of argument(s) with its predicate is local. It is the

tightest of all grammatical relations. Only immediate sister nodes may

enter in the relation and nothing else. The verb of the main clause cannot

4. See Chomsky (1981: 50, 86).

5 . See Chomsky (1981: 50)

7

assign a theta role to the subject of an embedded clause but it can assign a

theta role to the whole clause. The direction of the theta role assignment

is dependent on the theta directionality parameter setting of a language.

As pointed out above, the theta theory is one of the core principles which

decide which element will merge to form a derivation. As it determines

the thematic relation of an NP with the verb, it plays a role in the

semantic interpretation of the sentence. While answering question on the

increasing role of the theta theory, Chomsky (1982: 85-86) observed,

“thematic role seems to be the only notion so far that has any interesting

characteristic… it is a fundamental notion and … in fact the choice of

complements to an element in the lexicon really has to do with thematic

role”. During the discussion on the operation merge Chomsky (1995a:

246-248) explains that the head of a phrase (e.g., V of VP) selects its

specifier and complement to have a “convergent derivation” (e.g.,VP).

The structure converges only if the argument has a theta role in relation to

the lexical property of the verb.

1.2 Assignment of Theta Roles

According to Chomsky (1981: 34-48) each theta role is determined by the

lexical property of its head that governs it. The lexical head takes a

phrasal category (NP, VP or PP) as a complement. The lexical head (i.e.,

V of VP) which governs its complement is called a governor.

Government is a more “local” variety of command (Chomsky 1981a,

1986a, Rizzi 1990) which applies throughout the module of grammar.

The Government theory says that: “A govern B if A c- commands B and there is no barrier

for B c-command by A.”

8

This principle has been redefined in terms of m-command which is as

follows:

“A m- commands B if A does not dominate B and B

does not dominate A and the first maximal projection

of A dominates B.”

Government and Command are two fundamental concepts which apply

throughout the module of grammar. They are essential for the assignment

of theta roles.

There are two main categories of government: (a) antecedent government

of A by an antecedent of A, and (b) head government of A by a head.

These governments are termed “Proper Government”. It is necessary to

discuss the concept of head government here because it is relevant to the

theta theory. As mentioned earlier, the locality relation between the head

and its argument is very essential for the assignment of theta roles to NPs.

As the lexical head and its complements are bound by local relationship,

a verb may theta-mark its complements only within its maximal

projection, i.e., VP. The status of [spec, IP] is different. It may or may not

be a theta position, depending on lexical choices. For examples, the verb

‘seem’ does not assign an agent theta role to its subject in (9a):

9a. Akbarx seems [ tx to have hurt himself.]

b. It seems [that Akbar has hurt himself.]

In (9a), the subject of ‘hurt’ is a trace which is the trace argument of

Akbar. The trace is the agent of hurt; as it is bound to Akbar, (i.e.,

Akbarx…tx forms a theta chain), Akbarx is assigned the same theta role as

tx. In (9b) Akbar, the subject of the embedded clause, is assigned the

agent theta role by ‘has hurt’. As the subject of the main clause is in a

9

non- theta position in (9b), it is occupied by the expletive ‘it’. Under the

VP Internal Subject Hypothesis (VPISH), the VP will have the subject of

the sentence as its specifier, which means all the theta-marked NPs or

clause will be inside the VP. We may look at (10a) which will have (10b)

as its structural representation to clarify the point.

10a. Akbar bought a book from Aamir.

b.

In (10b), within the VP-shell, spec NP ‘Akbar’ is assigned the theta role

of agent, NP ‘book’ is assigned the theta role of theme and PP ‘from

Aamir’ is assigned the theta role of source.

AGRsP

AGRs'

TP

VP

V'

V'

V

bought

NP

a book

PP

P NP

NP

Spec

Akbar

Past

NP

3,sg

Spec

NP

from Aamir

T'

10

1.3 Relevance of Theta-criterion

According to Chomsky (1981:36), theta-criterion is defined as follows:

“Each argument bears one and only one theta-role, and each theta-

role is assigned to one and only one argument.”

It means that there must be one-to-one correspondence between

arguments and their theta-roles in a sentence. We cannot have more

arguments than the number of theta-roles and more theta-roles than the

number of arguments. Furthermore, since theta-roles express particular

thematic relations, the arguments will have to be appropriate in relation to

the lexical properties of the verb. For examples, we may look at (11).

11a. Thematic structure6 of kill,

b. Akbar killed Aamir.

(11a) presents the lexical properties of the predicate kill, i.e., it must have

two essential arguments, agent and patient. In (11b), there is one-to-one

correspondence between arguments and their theta roles. As the theta-

criterion is satisfied in (11b), the sentence is well-formed. We may now

look at (12-14):

12. *Akbar killed.

13. *killed Aamir.

14. Aamir was killed.

(12) is incorrect because it lacks one argument, the patient Aamir.

Likewise, (13) is ill-formed because it has no agent. However (14) is

well-formed even though it has only one overt argument. It is because the

passive form of the predicate, was killed, suggests that some agent

brought it about. In other word, in (14) the agent is covertly indicated; it

6.What is called “thematic structure” by Chomsky (1981, 1995a) is called “semantic structure” by

Jackendoff (1990) and “argument structure” by Grimshaw (1990).

11

is understood. (15) Shows the other side of the problem: it is a sentence

with more essential arguments than needed by the verb.

15a. *Akbar killed Aamir, Anees.

In (15a), Anees does not get a theta role because killed needs only two

essential theta roles which have been already assigned to Akbar and

Aamir. Since there are three arguments in (15), it violates the first part of

the theta criterion: the requirement that every argument must have a theta

role. Thus, theta criterion filters out this sentence as ungrammatical.

However, if Aamir and Anees are combined as one NP, i.e., Aamir and

Anees, the whole structure gets the theta role of patient. For this reason,

(15b) is grammatical.

15b. Akbar killed Aamir and Anees.

It is clear from the description given above that the predicate is the

central part of a sentence and every predicate needs one or more

arguments for the completion of that sentence depending on its lexical

properties. If any argument is missing, the sentence is incomplete. For the

well-formedness of a sentence the required number of essential

arguments should be there. We can take another example to illustrate this

point. The verb put needs three arguments: agent, theme and location, as

in (16a).

16a. Put < agent, theme, location>

b. He put the book on the table.

c. *He put the book7.

7 . We put something on or at some location, which may or may not be specified by a PP. For example,

(i) has no PP but it is well-formed:

(i) He put the cup down.

Here, down, is an adverb (which means ‘a specified place below’) is used in place of a PP, it may be

treated as the residual preposition of a PP (e.g. down the stairs).

12

(16c) is ungrammatical because it does not have the locative argument.

But there is another verb ‘keep’ which in one sense is a near synonym of

‘put’; it means ‘to continue or cause to continue in a specified position,

condition or course’. In this sense, it needs only two overt arguments.

Thus, (17a) is well-formed though (17b) is also correct with ‘himself’,

where with himself is a location.

17a. He kept the book.

b. He kept the book with himself.

(17a) is complete; it conveys the sense of (17b). In other words, the verb

‘keep’ implies here, ‘in his possession’.

The assignment of theta role to an NP in a sentence is subject to a

condition. It is only the argument of the head of a VP which can occupy a

theta position in a structural configuration. Earlier, except the subject

argument, all arguments were internal to the VP. Now, because of the VP

Internal Subject Hypothesis (VPISH), even the subject originates

internally within the VP. Since expletive there and it are non-arguments,

they do not originate in the VP; they occupy only non-theta-positions and

are not theta-marked. It is the theta criterion, which puts constraints on

the movement of NP from a theta position to another theta position. If

movement is permitted, the moved NP will have two theta roles, one

assigned by lexical head and another, by virtue of being in a new theta

position, which will violate the theta criterion.

13

1.4 Jackendoff’s Criticism of Chomsky’s Theta Criterion

Chomsky’s theta criterion has the special status in theta theory as it

accounts for the theta role assignment in a sentence. However, Jackendoff

challenged the uniqueness of theta criterion. He brought forth his point by

applying it at the conceptual level. In terms of conceptual structure he

describes theta criteria as follows (1990: 59):

a. …each subcategorized NP (plus the subject) corresponds to

exactly one argument position in conceptual structure, and…

b. each open argument position in conceptual structure is

expressed by exactly one NP.

In favour of his claim, he pointed out the cases where an NP may have

more than one theta role and where multiple NPs may hold a single theta

role.

He argued this point by referring to the verb of transaction such as buy

and sell, where two actions seem to be going on at the same time. For

instance, we may consider the verb buy in (18).

18. Akbar bought a book from Aamir.

From the point of view of the transfer of the book in (18), the NP Akbar

is the agent and the NP Aamir is the source. However, if we look at the

action from the other point of view, a book is central to the action; it

involves its transfer from Aamir (source) to Akbar (goal). In other words,

the NP Akbar has two theta roles- agent and goal. If we look at (18) from

another angle, the problem is even more complex. The lexical property of

buy8 is such that it refers to not only transfer of the possession of book

from Aamir to Akbar but also the transfer of money from Akbar to Aamir.

In other words, Akbar has three theta roles (agent, goal and source) and

8 To buy means ‘to obtain something in exchange for payment of money’, etc.

14

Aamir has two theta roles (source and goal). Such counter-examples seem

to challenge the uniqueness of the theta criterion.

If we look at (19) two NPs seem to have the same theta role.

19. The bucket has water in it.

That is, the theta role of location is assigned simultaneously to the NP

‘the bucket and PP in it respectively. Though such sentences seem to be

convincing counter-examples to the uniqueness of the theta criterion, they

are not so if we look at them carefully. If we look at (18) from the point

of view of action ‘Akbar wills the transfer of book from Aamir’. Thus the

former is the agent and the latter is the source from whom the book is

obtained. Again if the focus is on the transfer of the book, and not on the

actor, the NP a book is the theme, it is transferred from Aamir (source) to

Akbar (goal). We think of two theta roles of Akbar if and only if we look

at the sentence simultaneously from two different angles, the action tier

and the thematic tier. If we look at it from either angle at a time, the

problem does not arise, as is suggested by Culicover and Wilkins (1984).

The question of assigning a third theta role to Akbar arises only if we

break the verb buy into its semantic primes and consider it as an

amalgamation of two actions, one referring to the transfer of book and the

other to the transfer of money. In (19), NP the bucket is assigned the theta

role of location. The argument in it is also assigned the same theta role

but it is not an essential argument; it is an adjunct. As it is a non-essential

component of its thematic structure, it cannot be a real counter-example

to the theta criterion.

As mentioned above, Jackendoff argues that the theta roles are not the

diacritic markers on a sentence as proposed by Chomsky; they are

15

structural relations which represent conceptual structures. He defined

thematic relations in terms of conceptual constituents, each of which

belongs to a small set of major categories such as [Thing], [Event],

[State], [Action], [Place], [Path], [Property], and [Amount]. Each of these

abstract entities can be elaborated into a function argument to assign a

particular theta role. Within the limits, each category permits a variety of

specific elaboration of the surface theta roles. The organization of these

functions, as suggested by Jackendoff (1990: 43), is given in (20):

16

(20) a. [PLACE] → [Place PLACE-FUNCTION ([THING])]

b. [PATH] →

−PLACE

THING

VIA

FROMAWAY

TOWARD

FROM

TO

Path

c. [EVENT] →

( )

])][],([][

]])[,][[

PLACETHINGSTAY

PATHTHINGGO

Event

Event

d. [STATE] →

])][],([[

])][],[([

])][],([[

PATHTHINGXTE

PATHTHINGORIENT

PLACETHINGBE

State

State

State

e. [EVENT] → )

][EVNETEVENT

THINGCAUSE

Event

17

In (20), the category PLACE is elaborated as a place function plus an

argument that belongs to the category [THING]. For example, in

syntactic constituent under the tree, the tree designates a reference object

and under expresses a place-function that maps the tree into the region

beneath it. Similarly (20b) is an extended path or trajectory, it is treated

as one of five functions that map a reference Thing or Place into a related

trajectory. Category event in (20c) can be elaborated as either of the two

event-functions GO or STAY, each of which takes two arguments. (20d)

presents three state-functions. Finally, (20e) extends an event as the

Event-function CAUSE plus two arguments.

Jackendoff (1990: 46-48) treats these functions as conceptual primitives.

Keeping in mind these primitives, he defines thematic roles in the

following manner:

(a) Actor is the first argument of CS9 and AFF.

(b) Patient is the second argument of [-AFFECT].

(c) Theme is the first argument of event and state function other than

CS.

(d) Goal is the first argument of [TO].

(e) Source is the argument of [FROM].

(f) Beneficiary is the second argument of [+ AFFECT].

9 Jackendoff (1990) uses CS as an abbreviation for ‘Conceptual Structure’. AFF (affect) is an

additional mainstream function alongside the thematic functions (p.127).

18

The process of theta assignment is shown in (21).

21a. Akbar forced Aamir to go away.

b. CS + [Akbar]1

GO [Aamir]2, [Away]

AFF [Aamir]2

AFF- ([Akbar], [Aamir]2)

c. force

V

NPj to Sk

GO ([THING2])[AWAY]

EVENT

CAUSE ([THING1])

EVENT

(21b) is an attempt to represent the conceptual structure for (21a). It

indicates that Akbar, the actor is also an instigator, exerting force on

Aamir to leave with a successful outcome. The patient or beneficiary is

19

the actor of the potential effect. The NP Akbar corresponds to the first

argument of CS+ represented by the [THING1] and is assigned the theta

role of actor; it is mapped onto the subject position of the main clause.

The NP Aamir is the second argument of [AFFECT] represented by

[THING2] and is the theta role of patient or beneficiary; and it is mapped

onto the object position of the main clause. In the embedded clause the

verb to go is indicated as [EVENT-GO], which corresponds to [EVENT-

function] that shows a motion. This function has the arguments

[AKBAR] and [AAMIR]. Akbar corresponds to the first argument of

[GO] represented by [THING] and is assigned the theta role of actor

whereas Aamir is the second argument of [GO] and is assigned the theta

role of patient or beneficiary. The NP Aamir in the structure is marked by

[AFF] indicating that he does not want to go but he had to leave. The NP

Akbar in the sentence is marked by [AFF-] indicating that he is not an

affected entity. NP Aamir is marked by [AFF] at one place indicating that

he has to leave. The verb force is conceptually interpreted as the

[EVENT-CAUSE] which corresponds to [EVENT-FUNCTION] that

expresses a motion. (21c) is the lexical entry for force. (21c) shows that

the argument [THING1] of the [EVENT-CAUSE] is co-indexed with the

argument [THING2] of the [EVENT-GO] function because it is an event

and its actor [THING1] is bound to the patient of the superordinate event.

Jackendoff’s conceptual relations led him to posit abstract relations

between the components of a predicate and its argument(s) which creates

a lot of confusion. Jackendoff faces a problem in regard to the

decomposition of a complex event as illustrated in (21). As Culicover

(1987) points out, it is difficult to justify their decomposition into two

events in a conceptual structure in a principled manner. The logic

20

regarding the mechanism of fusion and linking, adopted by Jackendoff

(1990) is not always clear and even where it is, it is too complex and

abstract to be handled smoothly. For these reasons I ignore his theory in

favour of the most widely accepted concept of theta roles developed by

Chomsky in (1981, 1993, 1995a and 2000).

1.5 Theta theory in the Minimalist Program

Chomsky (1982)10

observed that “thematic role is a fundamental notion in

semantics and may be, serves as a unifying notion”. What he means to

say is that every lexical verb carries along with it “a certain set of

thematic roles, theta roles, which have to be filled. That is a lexical entry

and from that we can determine everything in the base structure except

for what can be determined by case theory and except for some

idiosyncrasies like SVO and SVO order, which just seems to be a

parameter” (p. 86). In short, in the Principles and Parameters theory, (i.e.

GB) Chomsky considered theta theory to be a “fundamental theory”.

In 1995, Chomsky gave a minimalist perspective which questions “almost

everything” (Chomsky 2002:151)11

but he asserted that there is

“something there that is stable” (p. 152). He argues that argument

structure will remain as will the properties of scope and reconstruction”

(ibid). He talks about “the thematic properties of lexical heads” (e.g.

verbs). He observes that a verb with no θ-roles to assign to a complement

10 See his interview with Huybregts and Riemsdijk in Chomsky (1982: 86-87).

11 See his interview with Adrinna Belletti and Luigi Rizzi in Chomsky (2000:113-114, 151-152)

21

will not be able to take a complement and a verb with obligatory θ-roles

to assign will have to occur in a configuration with enough arguments to

receive these θ-roles (pp. 30-31). He also observes that, at least in part,

selectional restrictions will also be determined by thematic properties of

the verb. In other words, to receive a particular θ-role, the inherent

semantic features of an argument will have to be compatible with that θ-

role.

Chomsky (1993, 1995a :315) virtually accepts the configurational

approach to θ-theory as proposed by Hale and Keyser (1993)12

. He also

observes that an argument without a θ-role violates the principle of Full

Interpretation (FI); it causes derivation to crash (p.315). That is one of the

reasons why he asserts the significance of VP Internal Subject

Hypothesis. He argues that, if a subject is directly inserted in [Spec INFL]

configuration, and is not raised from VP, it constitutes a violation of FI

(p. 314). The shortest derivation condition entails that the violation of θ-

criterion causes the derivation to crash by failure to satisfy FI. At the

same time Chomsky asserts that θ-role is not a formal property that

permits the last resort movement as case and agreement features do.

In the Principles and Parameters theory (Chomsky 1981, 1986 and

others), assigning θ-roles was a fairly straightforward affair; it was at the

D-structures. But the Minimalist Program (Chomsky 1995a and

subsequent publications), D-structure does not exist at all and LF is

assumed to be the sole interface with semantics. Consequently θ-

12

Hale and Keyser (1993:53) assert that proper representation of “predicate-argument structure is itself

a syntax”, though later on they observe: “In an important sense there are no thematic roles. Instead,

there are just the relation determined by the categories and their projections.” (p. 68). In a sense,

Chomsky (1993) agrees to this configurational approach. However, Chomsky (1995a:389) observes

that θ-role could not be identical with structural configuration” as it raises some empirical problems.

22

relatedness is considered to be “a base property” (Chomsky 1995b). In

other words, all θ-roles are assigned within the lexical projection. It

follows from this assertion that the θ-theory does not interact with

movement. In other words, movement can never create a θ-configuration.

That is what Chomsky (1995a: 312) asserts that “…there should be no

interaction between θ-theory and theory of movement”. In other words, θ-

role cannot license movement. He also makes it clear that the domain of

thematic assignment and morphological checking are different. We

propose to discuss the thematic relations between predicates and their

arguments in English and Urdu in the light of these theoretical

developments in Chomskyan syntax

1.6 Aims and Objective of the Thesis

We begin with the hypothesis that Urdu and English are somewhat

different in regard to the number and nature of overt essential arguments

that a predicate may take even though there is no difference between

them in so far as the basic underlying thematic relation is concerned. For

example, as discussed above, in English the verb put needs three

arguments: agent, theme and location, as in (22a).

22a. Akbar put the book on the table.

b. *Akbar put the book.

(22b) is ungrammatical because it does not have the location argument.

On the other hand, the verb keep, which conveys approximately the same

sense as put in some respect, needs only two overt arguments. Thus, (23a)

is well-formed in the sense of (23b).

23a. Akbar kept the book.

b. Akbar kept the book with himself.

23

In Urdu there is only one verb rakhna: which conveys the sense of both

put and keep. Though rakhna: is semantically the equivalent of put, it is

different from put in regard to the number of overt arguments it takes; it

may have only two arguments, as in (24a,b).

24a. Asad ne kita:b rakh di:

Asad book put

b. Asad ne kita:b rakh li:

Asad book kept

(24a, b) can be expanded by adding an overt argument as in (25a, b)

respectively:

25a. Asad ne kita:b Tebal par rakh di:

Asad book table on put

‘Asad put the book on the table.’

b. Asad ne kita:b apne pa:s rakh li:

Asad book him with kept

‘Asad kept the book with him.’

(24a) has an implicit reference to a locative NP whereas (25a) has an

explicit postpositional phrase (PoP) Tebal par ‘on the table’. In (24b) li:

(the explicator verb) conveys the sense that the book was kept by the

agent whereas in (25b), the location is explicitly expressed by apne pa:s

‘with him’. (24a, b) are complete without an overt location argument.

As proficiency in English has become essential for us, it is now necessary

to know to what extent it is structurally similar to or different from our

mother tongue, i.e. Urdu. It is thus necessary for an Urdu speaker to know

the points of similarity and difference between Urdu and English in

regard to the number and nature of arguments their predicates may take

overtly. Because of the type of difference mentioned above in (24), an

24

Urdu-speaking child learning English as a second language may use more

or fewer arguments than are needed in English in a sentence.

In short, the intent of this thesis is to present a comparison between the

two languages – Urdu and English- with regard to the number and nature

of arguments needed by some frequently used verbs in these languages.

This thesis will analyze their thematic structures in order to find out

similarities and differences between them. It will discuss not only the

number of arguments but also, their nature, i.e., whether the arguments

concerned are NPs, PPs (or PoPs for Urdu), or clauses. It will pay special

attention to those verbs which may not need all essential arguments on

surface. It is our hypothesis that Urdu and English are not identical in this

respect.

A large number of studies have been done on the number and nature of

arguments in English, e.g. Grimshaw (1990), Jackendoff (1990), Stowell

(1991), Marantz (1992), Zubizarreta (1992), Williams (1992, 1994), but

to the best of my knowledge, no exhaustive work has been done on Urdu

even though some work has been done on Urdu verbs within the

framework of the traditional grammar, such as Platts (1920) and Schmidt

(1981, 1999). Even Butt (1995) and Agha (1998), who have worked on

some aspects of Urdu verbs from the transformational generative point of

view, have not discussed this issue. In short, no attempt has been made

from the point of view of thematic relation between the predicate and its

arguments in Urdu. Even though some references have been made to the

application of theta theory in Hindi, (e.g. T. Mohanan, 1990), they do not

cover the areas I propose to look into. We hope this study will

25

concentrate on various aspects of thematic relations and cover some new

ground.

1.7 An Outline of the Thesis

The second chapter of this thesis contains the details about the possible

theta roles of verbs in Urdu and English. Some transitive verbs of Urdu

and English differ in regard to the number of essential argument(s) they

take. The chapter attempts to examine whether the difference is due to

some well-defined factors mainly in terms of their structural properties or

they are idiosyncratic. For example, some arguments in English may be

realized as a PP (rather than NP) and those in Urdu as PoP, as in (26a)

and (26b) respectively.

26a. Asad bought the book [PP for me].

b. Asad ne [PoP mere liye] kita:b Khari:di:

Asad me for book bought

In some cases, an argument may be realized as NP in Urdu but PP in

English, as it is clear from (27) and its English counterpart respectively.

27. Asad [NPdehli:] gaya:

Asad Delhi went

‘ Asad went [PPto Delhi].’

In some other case Urdu may have a PoP as an argument but English has

an NP in its place, as (28) and its English counterpart illustrate the point.

28. Asad [PoP kamre mẽ] da:Khil hua:

Asad room into went

‘He entered [NP the room].’

26

I have grouped verbs of Urdu and English according to the number of the

essential arguments they take, such as:

i. Verbs that need only one essential argument (e.g. sona:, ‘to

sleep’).

ii. Verbs that need two essential arguments (e.g. ma:rna: ‘to kill.

to beat’).

iii. Verbs that need three essential arguments (e.g. bata:na:, ‘to

tell, to point out’).

Some verbs take NPs or PPs as arguments while others need a CP or IP.

For instance, we may look at (29) and (30):

29. [akbar ne] [a:m] [kha:ya:] Akbar mango ate

‘Akbar ate a mango.’

30. [ye mumkin hai [cpke [IP Asad a:j a:ye:ga:]]] It likely is that Asad today come will

‘[It is likely [cpthat [IP Asad will come today.]]]

In (29) both arguments are NPs, while in (30), the theme argument, that

Asad will come today, is a CP (i.e., a clause beginning with a

complementizer).

The third chapter discusses the role of overt- realization of theta roles for

the well-formedness of a sentence. For instance, a verb may need three

arguments for having a well-formed sentence but sometimes only two of

them may be overtly realized. When an argument is covert, it is assumed

rather than overtly expressed. It has been discussed with the help of the

concept of “theta absorption” (Chomsky 1995) as against the concept of

“incorporation” as developed by Mark Baker (1988). We will discuss

27

verbs that absorb theta roles with a view to finding out whether the

process in Urdu is different from that in English.

Chapter four shows how theta theory feeds case theory. It is important to

do so because all lexical NPs must have case and the case visibility

condition must be fulfilled for an NP with a θ-role to be legitimate. An

attempt is made to show that a θ-role may not be realized uniformly by a

single case marking and a single overt case marker may be used for NPs

with different θ-roles.

The fifth chapter includes the summary and conclusions of this research.

CChhaapptteerr--22

TTTTTTTThhhhhhhheeeeeeee TTTTTTTThhhhhhhheeeeeeeemmmmmmmmaaaaaaaattttttttiiiiiiiicccccccc SSSSSSSSttttttttrrrrrrrruuuuuuuuccccccccttttttttuuuuuuuurrrrrrrreeeeeeee ooooooooffffffff UUUUUUUUrrrrrrrrdddddddduuuuuuuu VVVVVVVVeeeeeeeerrrrrrrrbbbbbbbbssssssss

Chapter Chapter Chapter Chapter –––– IIIIIIII

THE THEMATIC STRUCTURE OF URDU VERBS

2.0 Essential and Non-essential Arguments in a Thematic Structure

A thematic structure consists of a set of thematic roles associated with a

predicate. The thematic roles are the distinguishing properties of

arguments in a constituent structure which is a key to predicting whether

or not a given sentence is structurally complete. As we treat non-essential

argument as adjunct rather than complement to the verb and they provide

only additional information, we have not discussed them in detail. This

chapter provides a detailed account of the essential theta roles of several

types of verbs in Urdu and English. We have divided Urdu verbs in

semantic subgroups on the basis of the number and nature of θ-roles their

arguments might have in their θ-grid to generate grammatical sentences.

Urdu has intransitive verbs that take one essential argument and transitive

verbs that need two essential arguments. Some transitive verbs (including

ditransitive verbs) require three arguments for having a well-formed

sentence. Some intransitive and transitive verbs of Urdu and English may

differ in regard to the number of overt essential arguments they must

have. We have tried to find out whether they differ due to some semantic

factors or mainly in terms of their structural properties. For instance, an

NP in English may be realized as a PoP in Urdu, but it must be present to

generate a grammatical sentence. However, not all essential arguments

that a verb needs for the well-formedness of a sentence are overtly

expressed. An argument may be covert, when either it is contextually

assumed, absorbed in the verb or is deleted because Urdu is a pro-drop

29

language or it may be the case that the argument has moved out and left

an empty trace behind. These points will be separately discussed later in

this work.

2.1 Verbs with One Argument

The verbs of Urdu which need only one essential argument have been

classified in sub-groups on the basis of the nature of their arguments.

Some verbs take agent; some take theme/ patient or experiencer as their

essential argument. Some linguists do not differentiate between theme

and patient but some do. We have treated them as two different theta

roles according to their properties, i.e., whether they are animate or not. If

an animate argument undergoes an action, it is a patient. If an inanimate

entity undergoes an action, it is a theme. Such Urdu verbs have been

compared with their English counterparts to find out whether they have

the same type of arguments.

2.1.1. Verbs that need only agent as an essential argument

There are intransitive verbs that take only an agent as an essential

argument. For instance, we have a verb such as dauRna: ‘to run’ which

takes an agent as its argument. We may look at (1) to illustrate this point:

1a. dauRna: ‘to run’

b. laRka: dauR raha: hai boy running is ‘The boy is running.’

In (1b), the subject NP laRka: ‘boy’ is the essential argument for the verb

dauRna: ‘to run’, it has the theta role of agent. It may be noted that the

English equivalent of (1b) is also well-formed. The verb dauRna: can

have a living-being as its agent as in (1c) or even a motor-driver

conveyance as its argument, as in (1d).

30

1c. ye ghoRa: tez dauRta: hai this horse fast runs ‘This horse runs fast.’

d. meri: isporT ka:r bahot tez dauRti: hai my sport car very fast runs ‘My sport car runs very fast.’

Some verbs that take agent as an essential argument are as follows:

2a. a:na: ‘to come’

b. mehma:n13

a:ye guest came(hon., pl.)

‘The guest came.’

3a. ja:na: ‘to go’

b. mehma:n ja: rahe haĩ guest going(hon. pl.) is

‘The guest is going.’

4a. bakna: ‘to chatter’

b. Asad hamesha: bakta: rahta: hai Asad always chatters ‘Asad always chatters.’

5a. bolna: ‘to speak’

b. wo laRka: bahot bolta: hai that boy too much speaks ‘That boy speaks too much.’

6a. bha:gna: ‘to run away’

b. čor bha:g gaya: thief ran away ‘The thief ran away.

7a. phũnka:rna: ‘to hiss’

b. sã:p phũnka:r raha: hai snake hissing is ‘The snake is hissing.’

13

Where an agent is [+honorific], it takes a plural verb in Urdu.

31

8a. čalna: ‘to walk’

b. wo laRka: ka:fi: tez čalta: hai That boy very fast walks ‘That boy walks very fast.’

9a. čahčaha:na: ‘to chirp’

b. čiRya: čahčaha: rahi: hai bird chirping is ‘The bird is chirping.’

10a. baiThna: ‘to sit’

b. laRka: baiTha: hai boy sitting is ‘The boy is sitting.’

11a. rukna: ‘to stop, to stay’

b. mehma:n ača:nak ruk gaye guest suddenly stopped/stayed(hon., pl.) ‘The guest stopped/stayed suddenly.

12a. khelna: ‘to play’

b. bačče khel rahe haĩ children playing are ‘The children are playing.’

13a. bhaũkna: ‘to bark’

b. kutta: bhaũkta: hai dog barks ‘The dog barks.’

14a. sona: ‘to sleep’

b. bačča: so gaya child slept ‘The child slept.’

15a. čilla:na: ‘to scream or cry out’

b. bačče čilla: rahe haĩ children screaming are ‘The children are screaming.’

32

16a. uThna: ‘to get up’

b. maĩ savere uThti: hũ: I early get up ‘I get up early.’

We notice that sentences (1-16) of Urdu and their English equivalents are

well-formed even when they have only one argument i.e., agent and they

all are NPs. All these verbs in Urdu have causative alternation14

. These

causatives are formed by adding the suffixes –a: or –wa:. For example:

uThna: ‘to get up’; uTha:na: ‘to cause to get up’, uThwa:na: ‘to get

someone/something picked up’.

2.1.2 Verbs that need experiencer as an essential argument

Some intransitive verbs of Urdu take experiencer, rather than agent, as

the only essential argument. The argument involuntarily experiences the

event which can be mental or physical. Culicover (2009:149) defines

‘experiencer’ as an “individual in a perceptual or cognitive state such as

seeing or knowing”.

17a. khã:sna: ‘to cough’

b. bu:Rha: a:dmi: ača:nak khã:s uTha: old man suddenly coug