 

 

LIBRARY
Mlchigan State
University

——

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES Mum on Of baton data due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU Is An Atﬂnnativo Action/Equal Oppoctunity Institution
cMMmS-nt

 

 

——'

 

ANAPHORIC RELATIONS IN ARABIC
AND ENGLISH
SPOKEN NARRATIVES

By

Abdallah H. Al-Kahtany

A DISSERTATION

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of English

1992

, '90“

XI

"5

ora‘

ABSTRACT
ANAPHORIC RELATIONS IN ARABIC
AND ENGLISH
SPOKEN NARRATIVES
By

Abdallah H. Al-Kahtany

This study deals with anaphoric relations in spoken Arabic and English narratives.
The analysis was based on 20 spoken narratives involving 40 Arabic and English native
speakers. I have developed a combined multidimensional system that included, within its
domain, a combination of major approaches previously applied to study anaphoric
relations. In this approach, I combined Clancy's (1980) method of studying cognitive
constraints, Givon's (1983) technique of studying referential distance and referential
choices, Fox's (1987) patterning of anaphora and their distribution and approaches to
studying interaction between discourse units and anaphoric relations, and Payne's (1988)
methodology of measuring the influence of contextual factors.

This study aims to provide answers to the following research questions;

1. How are anaphoric patterns in Arabic and English oral narratives distributed
across the narration situations, and how did these anaphoric patterns interact with episodes,
as discourse boundaries?

2. What impact do topicworthiness and discourse boundaries in narratives have on
the selection of anaphoric devices in Arabic and English narratives?

3. How is referential gap of sequential and return-pop anaphoric relations affected
by contextual and discourse factors? What implications do such crosslinguistic ﬁndin gs
have for the theory of anaphora in discourse?

Our study of anaphoric relations in Arabic and English spoken narratives support

the assumptions made by Payne (1988) and Blass (1990) that contextual factors exert

strong effect on the relations among parts of discourse, regardless of the variation in
linguistic and cultural backgrounds of speakers. This effect was illustrated through the
way referents with different topicworthiness values have interacted with anaphaphoric
relations in terms of referential gap and referential choices. My ﬁndings revealed that it is
not only discourse boundaries that determine the selection of marked members of the
morpho-syntactic anaphoric devices, but rather both discourse boundaries (represented in
narratives by episodes) and topicworthiness of referents involved in the narration process
as a contextual factor. This interaction between topicworthiness of referents and episode
boundaries play signiﬁcant roles in determining the kind of anaphoric patterns involved, the

average referential gap exerted and the type of referential coding devices being selected.

 

 

IN 1m: NAME or ALLAH
nu: mm THE COMPASSIONATE

I dedicate this work for the sake of Allah

ACKNOWLEDGMENTS

Although only one author's name appears on the cover on this dissertation, it is
only because of the help of Allah and that of many pc0ple who have contributed directly or
indirectly for this work exists.

I extend my thanks to my advisor and major professor Dr. Nancy Ainsworth-
Vaghn whose support, encouragement and crucial input were so crucial to bring this work
to reality, that no words of gratitude could describe.

My respected committee members, Dr. David Dwyer, Dr. Susan Gass and Dr.
Marilyn \Vllson deserve my appreciation for their support and continuous feedback that
they were generous to give, regardless of their numerous duties.

I also thank Dr. Dennis Preston for his enlightening suggestions on the final draft
of this dissertation and his acceptance to serve as an outside reader. _

Dr. Abdulrahman Assiri, a friend and a brother who chose to spend the last days of
his stay in East Lansing at the computer lab assisting me to enter difficult data with
enthusiasm and dedication. May Allah reward him.

The thousands of miles of water and land did not stand in the way of my beloved
parents to extend their hands to the skies praying for my success that they have nourished
since the ﬁrst day I came to this world. May Allah expand their lives and reward them with
paradise.

For those who accompanied me through the years of my studies in MSU second by
second and shared with me happiness and sorrow, I say THANK YOU. Aysha was not
only a wife but a friend and a refuge also. My beloved children Areej, Osamah, Fatimah,
and Anas never ran short of their smiles and kisses for daddy even when he says “ Leave

me alone, I am very busy ..”

TABLE OF CONTENTS

Page
LIST OF TABLES ............................................................................ vii
LIST OF FIGURES .......................................................................... ix
KEY TO SYMBOLS AND ABBREVIATIONS ............................................. x
INTRODUCTION ........................................................................... 1
Chapter I: REVIEW OF RELATED LITERATURE
I.A. Discourse Analysis And Language Context:
A Review Of The Literature .......................... 6
LA. 1. Nature of discourse .................................... 6
I.A.2. Attempts to integrate linguistic and contextual
approaches .............................................. 9
I.A.3. Fundamental approaches to theories of
discourse .............................................. 10
I.A.4. The essentials of context ............................ 11
I.A.S. Culture as a signiﬁcant determiner in context . . .. 13
LB. Anaphora In Discourse Analysis ................... 14
LB. 1 . An overview .......................................... 14
LB .2. Referential selection among coding devices ...... 15
I.B.3. Topicality, availability, and referential gap ....... 16
1.3.4. Anaphora in syntax .................................. 17
LB .5. Anaphora in Arabic and English ................... 18
LC. Summary ............................................. 23
Chapter II: METHODOLOGY & RESEARCH DESIGN
ILA. Introduction .......................................... 25
II. B. Design Of The Study ................................ 25
II. B. 1. Participants and data collection ..................... 25
II.B.2. Material ............................................... 28
II.C. Analysis ............................................... 3O
II.C. 1. Major anaphoric patterns ............................ 31
II.C.1.a. Sequentially distributed patterns ................... 31
II.C.1.b. Retum-pop anaphoric pattern ...................... 33
II.C.2. Frequency of distribution: referential gap and
referential choice ..................................... 36
II.C.2.a. Frequency of distribution ........................... 36
II.C.2.b. Referential gap ...................................... 36
ND. Referential Choice (RC) ............................ 38
II.E. Topicworthiness Of Participants ................... 39

iv

Chapter III:

Chapter IV:

II.F.

Discourse Hierarchy: Episode as a Unit of
Analysis ............................................... 42

ANALYSIS OF ANAPHORIC PATTERNS IN ARABIC AND

ENGLISH NARRATIVES
III.A. Introduction .......................................... 45
III.B. Anaphoric Patterns In Discourse ................... 45
III.B. 1. Sequentially distributed anaphora .................. 45
III.B.2. Retum-pops .......................................... 51
III.B.2.a. Definition of return-pop ............................. 51
III.B.2.b. Return-pop: A neglected anaphoric relation ...... 53
III.B.2.a. Constraints revisited ................................. 54
III.C. Anaphoric Patterns In Arabic And English

Narratives ............................................. 58
III.C.l Frequency of sequential and return-pop

patterns ................................................ 58
III.C.2. Length of patterns ................................... 59
III.C.3. Why do differences exist ........................... 6O
III.C.4. Episode as an inﬂuencing factor in the

distribution of anaphoric patterns .................. 62
III.C.4.a. Episode and distribution of anaphoric patterns... 63
III.C.4.b. Episode length in Arabic and English ............. 64
III.C.4.c. Cultural background and episode .................. 64
III.C.4.d. Discourse markers of episode ...................... 66
III.C.4.c. Episode and anaphoric patterns distribution ..... 69
III.D. Conclusion ........................................... 7O

REFERENTIAL CHOICES IN ARABIC AND ENGLISH

NARRATIVES
IV.A. Introduction .......................................... 71
IV.B. Referential Choices In Arabic And English
IV.B. 1. English zero-anaphors and Arabic subject
pronouns .............................................. 72
IV.B. 1 .a. Arabic subject pronouns
IV.B.1.b. English zero-anaphora .............................. 74
IV.C. Referential Choices And Topicworthiness ........ 79
IV.C. 1. Influence of topicworthiness on referential
choices in English and Arabic narratives .......... 81
IV.C. 1.a Topicworthiness and referential choices in
Arabic narratives ..................................... 83
IV.C.1.b Topicworthiness and referential choices in
English narratives .................................... 89
IV.D. Episode And Referential Choices .................. 91
IV.D. 1. Topicworthiness, reference and sensitivity of
episode boundaries .................................. 94
IV.D.2. Topicworthiness and discourse boundaries in
Arabic and English narratives ...................... 95
IV.B. Conclusions .......................................... 98

Chapter V: REFERENTIAL GAPS IN ARABIC AND ENGLISH

NARRATIVES
V.A. Introduction .......................................... 99
V.B . Quantitative And Qualitative Aspects Of

Referential Gap ..................................... 100
V.C. Referential Gap In Arabic And English

Narratives ............................................ 100
V.C. 1. Referential distance of retum-pops in Arabic

and English narratives .............................. 103
V.C.1.a. Cognitive capacity represented by referential

gap ................................................... 104
V.C.2. Referential gap and episode boundaries in

Arabic and English narratives ..................... 106
V.C.3. Inﬂuence of topicworthiness on referential gaps

in Arabic and English narratives .................. 109
V.C.4. Inﬂuence of anaphoric coding devices on

referential gap in Arabic

and English narratives .............................. 111

. V.D. Referential Gap And Continuity .................. 113
SUMMARY .................................................................................... 118
BIBLIOGRAPHY ............................................................................. 126
APPENDIX ................................................................................... 131
English Narrative (2) ............................................... 131

Arabic Narrative (10) .............................................. 138

Table

3-4

3-5

4-1

4-2

4-4

4-5

LIST OF TABLES

Page
A summary of information about American participants ........................... 26
A summary of information about Arabic participants .............................. 27
A summary of the topicworthiness status of characters
in the Pear Stories as produced by Arabic and English
participants .............................................................................. 41
Distribution of sequential and return-pop patterns in
Arabic and English narratives ......................................................... 59
Mean length of sequential and return-pop patterns in
Arabic and English narratives ......................................................... 61
Mean length of episode in Arabic and English narratives
according to number of clauses and anaphoric entries ............................. 64
Distribution of sequential and return-pop anaphoric
pattern in Arabic and English narratives within and
across episode .......................................................................... 70
Referential choices (coding devices) in Arabic and English
narratives ................................................................................ 73
Distribution of attenuated coding devices in English
according to their grammatical functions ............................................ 76
Distribution of pronouns and zero-pronouns in Arabic
according to their grammatical functions ............................................ 76
Distribution of different coding devices in Arabic
narratives according to their topicworthiness ....................................... 84
Frequencies of referential choices according to the
topicworthiness of their referents in English ........................................ 90
Frequency of NPs and attenuated anaphors at episode
initial positions in Arabic and English nanatives ................................... 92

vii

Table
4—8

5-1

5-2

5-3

574

5-5

5-6

5-7

Distribution of referents according to their

topicworthiness in and across episodes in English .....................

Distribution of referents according to their

topicworthiness in and across episode in Arabic .......................

Frequencies and percentages of the different anaphoric
patterns (sequential and pop) in relation to the
referential gap for each pattern in Arabic and in

English .......................................................................

Mean referential gap of sequential and retum-
pop anaphors as used in and across episode to

establish reference ..........................................................

Mean referential gaps as represented in and across

episodes in Arabic narratives ............................................

Mean referential gaps of sequential and return-pop
pronouns in relation to topicworthiness of characters

involved in the narratives in Arabic ......................................

Mean ultimate referential distances of sequential and return-
pop anaphors in relation to t0picworthiness of characters

involved in the narratives in English .....................................

Mean of referential gaps of sequential and return-pop
anaphors as used by pronominals and zero-anaphors

in English ....................................................................

Mean ultimate referential distances of sequential and
return-pop anaphors in relation to the type of anaphora

used in Arabic narratives ..................................................

viii

Page

............ 96

............ 97

........... 102

........... 107

............. 108

........... 110

........... 111

........... 112

.......... 112

LIST OF FIGURES

Figure Page
4-1 Illustrates of the across episode anaphoric relations ................................ 95

5-1 The two way relationship between the length of RGs and
persistence of referent in the minds of interlocutors ............................... 114

ix

KEY TO SYMBOLS AND ABBREVIATIONS

 

RC referential choice

URD ultimate referential distance

RD referential distance

RG referential gap

NP . noun phrase

( ) unable to transcribe

" " unable to give proper meaning

// interruption

CL clause

0 a return-pop anaphor

svo subject-verb-object

vso verb-subject-object

" 1* boldface, underline, and italic fonts are used in different places in the
transcriptions to highlight certain parts of the text relevant to the issue
discussed then.

Transcription symbols peculiar to some Arabic sounds

? voiceless glottal stop

9 voiced pharyngeal stop

dh voiced retroﬂexed alveolar stop

zh voiced dental fricative

h voiced pharyngeal fricative

s voiced uvular stop

1; voiceless retroﬂexed alveolar stop

gh voiced uvular fricative

kh voiceless velar fricative

INTRODUCTION

My interest in the study of anaphora in spoken Arabic and English narratives
emerged from several sources:

First of all, through my review of the literature of discourse analysis in Arabic, I
was able to trace no previous research on the topic of my dissertation, not even in its
broader context of anaphora as a cohesive strategy.

Secondly, the role anaphora plays in cohesion has emerged as a major issue in
discourse analysis theory in general. Therefore, pursuing research on anaphora within the
framework proposed by Halliday and I-lassan (1978), Givon (1983), Clancy (1980), and
Fox (1986; 1987) could improve our understanding of the way texts are tied together, how
they are comprehended as units and how they are interpreted (Clancy, 1980; Fox, 1987;
Payne, 1988). This study will add to our knowledge about the role cognitive and discourse
factors play in anaphoric cohesive ties.

Thirdly, since no previous research has dealt with comparing anaphoric patterns in
spoken Arabic and English narratives, such a study would provide the ﬁeld of discourse
analysis with information about cross-linguistic ﬁndings by testing theories about anaphora
in a broader linguistic context, outside Indo—European languages that have previously
attracted attention of discourse analysts (e.g., Fox, 1987; Clancy, 1980; Givon, 1985).

The present research deals with anaphora in oral narratives in Arabic and English
using an analysis which integrates the major strategies in use today for approaching
anaphora in discourse. A detailed description of the system will be given in the

methodology chapter. Close attention will revolve around the following issues:

1. Quantitative approaches to anaphora similar to those used by Givon (1983),

Fox (1987), and Payne (1988) in dealing with the following categories:

a. Major anaphoric patterns in oral narratives produced by Arabic and
English native speakers.
b. Distributional frequencies of referential choices (referent NPs and

their anaphoric pronouns) in oral narratives produced by Arabic and English native
speakers.

2. Cognitive and discourse constraints that seem to be responsible for
anaphoric selection. Clancy (1980) has stressed the importance of discourse and cognitive
constraints in determining the referential options in English and Japanese narratives. The
cognitive constraints include limits on the number of clauses between an anaphor and its
prior mention (referential gap).

Discourse constraints, on the other hand, refer to the referential choices (RC) that
are guided by a range of discourse factors, including the narrator’s need to introduce new
characters into the story and establish them as old information, to avoid ambiguity when
switching references. Within the category of “discourse constraints” I will treat two issues.

First, how does the status of a referent as a major character or merely a secondary
or a minor one affect the referential gap (RG) and referential choice (RC) in Arabic and
English narratives? Payne (1988) referred to this issue as “topicworthiness,” a term that I
will be using from now on. In his study of Yugua narrative discourse, Payne mentioned
that the topicworthiness of a character inﬂuences the length of RD between pronouns and
their referents.

Secondly, in her criticism of traditional approaches to anaphora in discourse, Fox
(1986, 1987) argued that hierarchical discourse structures such as episodes exert significant
inﬂuence on anaphoric relations in discourse. Such a claim does not harmonize with the
traditional views of Givon (1983), where anaphora is merely dealt with in the realm of

linear structure of discourse. For this study, I will investigate the inﬂuence of episode, as a

major discourse unit, on anaphora in narrative discourse. Fillmore (1975; cited in Tannen
1979) referred to this type of discourse frame of narratives as “thematic frames and
scenarios.”

In summary, the main contribution this study brings to discourse analysis and more
speciﬁcally to the theory of anaphora in discourse is that there is cross-linguistic evidence
that establishment of reference in discourse is highly inﬂuenced by contextual factors and
discourse boundaries, rather than linguistic relations alone. The aim of this research is not
only to investigate anaphora in Arabic discourse using modern approaches that have not
been applied previously, and to provide a comparison of two unrelated languages, Arabic
and English, but also to demonstrate that although interlocutors involved in our research
come from a diverse cultural backgrounds the principles they follow in using contextual
information and their treatment of discourse boundaries in narratives to establish anaphoric
relations seem to be cross-linguistically similar.

In the ﬁrst two chapters, I discuss a variety of approaches to the analysis of
discourse, and argue that in our stage of knowledge it is important to use a combined
approach of different strategies previously used for the study of anaphora in discourse.
Any one approach alone would not be adequate to provide explanatory account of major
aspects of anaphora with which I will be dealing in Arabic and English narratives.

Chapter I summarizes issues regarding the nature of discourse and discourse
analysis, with special attention to approaches dealing with anaphora. Examining the related
issues in previous studies of anaphora assisted me in developing a combined
multidimensional system for the study of anaphora in Arabic and English spoken
narratives.

Chapter II focuses on this approach of studying the phenomena of anaphora in
Arabic and English narratives. I combined Clancy’s (1980) technique of studying
cognitive constraints, Givon’s ( 1983) typological approach of discussing referential

distance and referential choices, Fox’s (1987, 1988) patterning of anaphora and their

distribution and her approaches of analyzing the interaction of discourse units with
anaphoric relations, and ﬁnally, Payne’s (1988) method of measuring the inﬂuence of
contextual factors on anaphoric relations.

In Chapter III, I compare the distribution of sequential and return-pop anaphoric
patterns in Arabic and English oral narratives. In the analysis, some theoretical claims
regarding anaphoric patterns were discussed based on evidence from Arabic and English.
In that phase of the study, I also attempt to shed light on anaphoric patterns in terms of how
referential gap is affected by the contextual factor of topicworthiness and discourse
boundaries in Arabic and English spoken narratives. In other words, the major issue here
is the nature of the abstract role that discourse boundaries and topicworthiness play in
human discourse and more speciﬁcally their effect on the way anaphoric relations are
maintained through narratives.

Chapter IV provides an illustrated discussion of how selection of anaphoric devices
by two unrelated languages, English and Arabic, is inﬂuenced by the same contextual and
discourse parameters. Discussion of episode as an inﬂuencing discourse unit on anaphoric
relations in Arabic and English oral narrative discourse took two phases. The ﬁrst phase
was a discussion of the way the boundaries of episodes in narratives exert a signiﬁcant
inﬂuence on the selection of different anaphoric coding devices available in the narrators’
linguistic competence. The second phase was a more detailed analysis of an underlying
contextual factor that seems to affect the kind of referential choices that perforate the rigid
boundaries of episodes.

In the ﬁnal chapter, I show that cognitive constraints represented by referential gaps
seem to be consistent cross-linguistically. Evidence was not limited to results of our
analysis in Arabic and English narratives; ﬁndings from other unrelated languages are
included. A principle that seems to hold cross-linguistically regarding RC is that, although
long distance reference exists, establishment of anaphoric relations tends to employ the

shortest referential gap possible to avoid any misinterpretation.

Apart from the theoretical signiﬁcance discussed above, this research has
implications for pedagogy. Cross-linguistic research in anaphora is very important to the
study of repertoire of cohesive devices, and thus important for educational practices such as
the teaching of writing to native speakers and equally as important to non-native speakers.
Widdowson (1979) indicated that cohesive strategies used in a given text are excellent

guides to coherence. Thus, teaching cohesion devices can be a crucial issue in second

language pedagogy.

CHAPTER I
REVIEW OF RELATED LITERATURE

LA. Discourse Analysis and Language Context:
A Review of the Literature

During the last quarter of this century interest in discourse analysis has
mushroomed. There has been heated debate over theoretical assumptions about the status
of discourse analysis as a subﬁeld of linguistics, especially regarding the nature of its
corpus, “discourse.” This chapter will review contemporary debated issues in discourse
analysis, touching upon the important role the communicative context plays alongside the
linguistic one and how different discourse analysts with different backgrounds evaluated

such roles.

I.A.l. Nature of discourse

Brown and Yule (1983) mentioned that the term “discourse analysis” has been
widely used by scholars from different backgrounds, for instance, linguists, sociolinguists
and psycholinguists, to accommodate a wide range of semantic connotations.

Consequently these different disciplines seem to focus on different domains and provide
different manifestations of the term.

In linguistics, Givon (1983) considered the clause as the basic information unit in
human discourse, and used the term “discourse” to refer to any chain of clauses that “hang
together” by discourse linking devices such as “continuity” (p. 7). These devices play a
major role in discourse analysis. They have been labeled differently by various discourse
analysts. Halliday and Hassan (1976) speak of “coherence” as a notion that marks
discourse as a distinguished linguistic unit of analysis. Givon coined the term “continuity,”

which is more or less synonymous to Halliday and Hassan's “coherence.” So, Givon

perceives discourse as the unit where continuity (or coherence) from one minimal

6

communicative unit (clause) to the following one is the most expected and unmarked.
Consequently, discontinuity is unexpected, and marked in accounting for discourse and
human communication by and large. Givon was most interested in analyzing topic
continuity in discourse and approaching the question by testing discourse for information
that continues to be available to hearers or “still activat .”

Linguists like Brown and Yule (1983) approach discourse analysis from a purely
linguistic perspective in the sense that they investigate the way people use language in
communication, and, in particular, how linguistic messages are constructed and delivered
by addressers and how addressees receive and interpret them. Brown and Yule are
primarily interested in providing an account of “how forms of language are used in
communication” (p. ix). They integrated some of the traditionally considered domains of
semantics and pragmatics like tense, aspect and modality within the domain of discourse
analysis. Brown and Yule characterize discourse analysis as the analysis of language by
what they term its two major functional aspects (transactional and interactional):

a) In the transactional aspect, language functions as an expression of
content. They claim that the written form of discourse, in general, fulﬁlls this category.

b) In the interactional aspect, language serves to express social relations and
personal attitudes.

They sum up their approach to studying discourse by stating, “We have adopted a
compromise position which suggests that discourse analysis on the one hand includes the
study of linguistic forms and the regularities of their distribution and, on the other hand,
involves a consideration of the general principles of interpretation by which people
normally make sense of what they hear and read” (p.10). This division between referential
content and social meaning would be challenged by some researchers in sociology and
anthropology.

Many sociologists are concerned with the social interaction patterns manifested

through conversations. A substantial number of discourse studies have been launched by

ethno-methodological sociologists like Sacks and Schegloff (1974) who were concerned
with the dynamics of conversations and therefore developed painstaking strategies of
transcribing most of the phonological and interactional features of conversations. The
interest of the researchers was not to investigate the principles of formal linguistic structure
of discourse, but rather the organization of social interaction. Coulthard (1985) remarked
on that fact by stating that, “. . . they [Sacks and Schegloff] work with conversational
materials, not because of special interest in language or any theoretical primacy they offered
to conversation, but because they see conversational analysis as a ﬁrst step towards
achieving a naturalistic observational discipline to deal with details of social interaction in a
rigorous, empirical and formal way” (p. 52). However, Coulthard ( 1985) considers
conversational analysis as a part of discourse analysis. In this view social meaning is
always conveyed in linguistic interaction; referential content cannot be divorced from
context. V

It is important that the study of structures both within and outside the text must
continue. Each type of study supports and illuminates the other, and the relevance for this
study is that cohesion needs to be studied in both ways. Any separation of “within” from
“outside of” the text is strictly for the purpose of restricting analysis and does not imply
autonomy of the text. Hierarchical structures of discourse have implemented direct
inﬂuence on the anaphoric relations in Arabic and English narratives.

Findings presented in III.C.4 provide strong evidence for the inﬂuence episode, as
a narrative discourse higher structure, has on referential relations in the different anaphoric
patterns. Chapter IV, presents results that provide strong evidence between discourse
initial slots and the selection of the marked members of anaphoric coding devices. These
ﬁndings lend support to claims of previous researchers who touched on this issue, such as

Chafe (1979), Fox (1987) and Payne (1988).

I.A.2. Attempts to integrate linguistic and contextual approaches
For a large portion of this century, linguistics has concerned itself with sentence
grammar. Earlier in the century linguists focus was on phonology and morphology . It
was Chomsky‘s (1957) revolutionary work W that brought attention to
syntax. Nevertheless, consideration of any factors outside the pure linguistic context has
been excluded. Such premises prevailed until a subdiscipline of linguistics known as
discourse analysis emerged through demands to study language in its communicative and
social contexts. Regardless of such development in linguistics, many linguists have found
it hard to diverge from focusing on the core of traditional linguistic studies that has
persisted for several decades. Exceptions are text-grammarians like Prince (1982) and van
Dijk (1985), who have attempted to develop some kind of grammar, emerging out of
assumptions based on sentence grammar, to account for well-formed texts. The texts are
based on well-formed sentences which are generated by what are called textual rules, some
of which are transformational in nature. Hopper (1989) categorized such an approach as a
pure structuralist movement in discourse analysis. It looks at language as an autonomous
system that can be studied in the absence of any contextual variables. Hopper argued that
such attempts degrade discourse to a simple text that would be included within the
structural theory of language. Hopper alluded to this approach by saying:
It is precisely these limitations, I would like to suggest, that provide the enabling
conceptual framework for grammatical theory and explains why it is unlikely that
grammatical theory as it is now envisaged can be extended beyond such simple
texts, whose simplicity, as I have noted, is in fact illusionary. Indeed the very
restriction of discourse analysis to deeply embedded, allegedly simple textual types
fosters and promotes a grammatical theory based on this analysis and guarantees
that grammar will be seen as a set of core categories and core rules that generate
precisely these kinds of texts. It does not matter whether we call these texts
discourse or sentences, for the problems are the same. Discourse analysis of this
kind is simply an echo of sentence grammar which likewise privileges precisely

those word meanings and sentence structures that can allegedly stand alone. (p.

19)

These text-grammarians (e.g., Prince [1982] and van Dijk [1985]) have attempted

to approach discourse equipped merely with Chomsky’s (1957, 3) idea of accounting for

10

“all the grammatical and none of the un grammatical” discourses. Prince adopted such an
assumption in his study of narration when he stated that, “A theory of narration is therefore
not so much concerned with the history of particular novels or tales, or with their meaning,
or with their esthetic value, but rather with the traits which distinguish narrative from
another signifying system and with the modalities of these traits. Its corpus consists not
only of all extant narrative, but of all possible ones” (p. 5).

So, for these grammarians basic assumptions about the nature of language and the
goal of language studies have not changed regardless of the quality or quantity of language
under analysis. What they did is to transfer sentential grammar over to text grammar, still
regarding language as separate from sender and receiver and situations that accommodate

the text, without realizing the role of language use as a system for intersubjectivity.

I.A.3. Fundamental approaches to theories of discourse

Tannen (1989, 3 ff.) views discourse analysis as an interdisciplinary ﬁeld of study
which interacts with the other disciplines of linguistics. Tannen suggests that discourse
analysis is distinguished for its diversity and novelty in the sense that it crosses the
boundaries of several ﬁelds of knowledge: linguistics, anthropology, psychology,
sociology, and ethnography of communication. She also suggests that discourse analysis
is different from other subﬁelds of linguistics in not being limited to a single theory or a
combined set of theories. Consequently, it does not tie itself to a particular methodological
system of analysis. She clariﬁed this point by adding that the term discourse analysis
speciﬁcally does not refer to any single theoretical view or methodological framework. It is
simply concerned with studying language beyond the sentence level. So, we can assume
that Tannen thinks of the term discourse as language beyond the sentential level where most
works in linguistics have stopped.

Data beyond the boundary of sentence level is extremely heterogeneous from both

linguistic and nonlinguistic perspectives. It includes any language units occurring in any

11

communicative context; linguistic contexts happen to be one of them. Therefore, as Tannen
explained, discourse can encompass spoken data such as tape recordings of conversations,
meetings, story-telling, interviews, or written data including novels, stories, letters,
recipes, and many other spoken and written registers. Considering the large scope of data
that discourse can include in its domain, Tannen insists that discourse analysis is simply no
different from what the term “linguistics” refers to, namely, “the study of language.” She
also emphasized that necessity was behind the invention of the term “discourse analysis,”
since discourse analysts found themselves obliged to distinguish their ﬁeld from other sub-
disciplines of linguistics that are nan'owly focused on limited data.

In responding to some criticism of discourse analysis for not having a monolithic
theory and methodology, van Dijk (1985), Widdowson (1988), and Tannen (1989) labeled
such criticism as immature, since it contradicts the genuine nature of discourse analysis as
an interdisciplinary ﬁeld and contributes toward defeating the main purpose of discourse
analysis in widely opening the doors of language study to invite various theoretical
methods to be able to account for language variation. Any attempt to study such complex

data with a single approach is doomed to failure.

I.A.4. The essential role of context

During the previous discussion, we have seen that scholars such as Longacre
( 1983), Hopper (1989) and Tannen (1988) ﬁnd context to be a source of discourse
meaning. Stubbs (1983) agrees, pointing out that discourse analysis obviously presents
linguistic patterns which can be an outcome of some nonlinguistic factors, as well as the
linguistic context. Stubbs raises the question of whether concepts that linguists have
nourished and developed for decades in their research in phonology, morphology, and
syntax are artiﬁcial and highly idealized. Such a deﬁcit can be noticed in two ways: “by

being restricted to examples which are both smaller than sentence and also studied out of

12

context. Alternatively, it might imply that discourse is not linguistically organized at all”
(Stubbs, 1983, 86).

I agree with the arguments of these scholars I just cited in holding the position that
discourse as a larger linguistic unit is inﬂuenced by social cognitive factors for which
discourse analysts have to account. Discourse analysis, in dealing with natural languages,
is not based on fragmented units of language which are detached from their social and
linguistic contexts. Brown and Yule (1983) viewed “discourse” as the process of
communication between speaker and hearer, or between writer and reader. Such interest in
the process rather than the product of discourse distinguishes Brown and Yule from other
scholars like Chafe (1982) and Olson (1979), who focused on the product of discourse
with little attention to the process. Such an approach led the latter to come up with
inaccurate generalizations about the nature of speech and writing.

Part of the debate around delineating “discourse analysis” centers on attempts to
differentiate between “text” and “discourse.” Stubbs (1983) suggested that the term
“discourse analysis” is ambiguous. He related such ambiguity to the unclearly deﬁned
senses the terms “text” and “discourse” seem to denote. Though he did not provide any
theoretical distinction between the two terms, Stubbs would only conﬁrm that “text” often
implies written discourse, while “discourse” entails spoken discourse. However, Stubbs
provided a proposal to solve this ambiguity by limiting the term “discourse analysis” to
“linguistic analysis of naturally occurring, connected spoken and written data” (p. 9).
Halliday and Hassan (1976) tend to imply that “discourse” refers to lengthy linguistic units,
whereas “text" may refer to very short ones. Van Dijk (1977), on the other hand, uses the
term text to indicate an abstract theoretical process which is realized in discourse. As for
Stubbs (1983), he seems to view discourse analysis and linguistics as referring to different
domains: “whereas linguists study language, discourse analysis can study the actual

mechanisms by which communication, understanding, and interaction are maintained”

(p. 10).

13

The role that contextual factors play in resolving anaphoric ambiguity was very
important in Arabic and English narratives, especially, in situations where potential
ambiguities emerge where syntactic or semantic information can only provide little

assistance to resolve ambiguity.

I.A.S. Culture as a signiﬁcant determiner in context

It is not news any more to report that discourse studies have been very signiﬁcant in
revealing evidence about speech events which are culturally determined, as shown in
studies by Tannen (1980), Ochs (1979), Chafe (1980) and many others. In Tannen's
comparative work on oral narration strategies by American and Greek women in telling
about the pear ﬁlm, interesting cultural and conventional information was revealed during
the process of discourse. Tannen reported that American and Greek subjects showed a
striking difference in their application of coherence principles in narration. American
narrators built the coherence of their narratives on “cinematic jargon” through commenting
upon and criticizing technical aspects of the ﬁlm production (p. 54). Greek subjects, on
the other hand, followed a process of narration concentrated on the events presented in the
ﬁlm as a real experience of narration; they did not comment about the ﬁlm as a ﬁlm. The
present study will provide information on Arabic and English patterns. This comparative
study of speciﬁc discourse features based on controlled discourse situation (genres) and
external factors can contribute greatly to our understanding of cognitive and cultural
conventions and restrictions. Such studies can also provide very useful information for
enhancing cross-cultural communication and language teaching and, most importantly,
provide a basis for generalizations of discourse analysis ﬁndings that can claim some
validity cross-linguistically.

In section III.B.4.c., cultural inﬂuence proved to be a determining factor behind the
absence of a speciﬁc episode during some English narrations. Therefore, it had a

consequent effect on the distribution of certain referents.

14

LB. Anaphora in Discourse Analysis

1.8.1. An overview

The study of anaphora in linguistics has had a long tradition, especially in the realm
of syntactic theory. Langacker (1969), Ross (1969) Karttunen (1967), Kuno (1972, 75)
and Delisle (1973) were pioneers in studying pronominalization rules and introducing them
to the textbooks in syntax. In discourse analysis, studies of anaphora have their own
history, in the works of Li and Thompson (1979), Clancy (1980), Givon ( 1983), Fox
(1987) and Payne (1988). Nevertheless, the novelty of the ﬁeld and the numerous types of
texts create a need for continued research across texts and languages.

Levinson (1983, 67) referred to anaphoric usage as the situation “where some term
picks out as a referent the same identity.” Thus, anaphoric relation takes place when a
pronoun is interpreted in terms of its relation to a referent noun phrase (NP) carrying the
same syntactic and semantic information in the discourse. Anaphora have the feature that
they cannot be interpreted in isolation; their interpretation involves the whole text through
mapping them to their antecedents that are mentioned in during the prior discourse (Tyler
and Marslen-Wilson, 1982). Therefore, anaphora is a prime example of text cohesion and
dependency. It should be noted however, that cohesiveness is a general property of texts;
it is hard to imagine a lexical item that can be interpreted or analyzed properly in isolation.

Halliday and Hasan (1976: 18) referred to the type of relation that holds between
anaphoric pronouns and their referent NP's as that of “cohesiveness.” They proposed two
types of anaphoric relations: exophoric, referring to cases where the pronoun relies on
information outside the boundaries of the given text for interpretation, and endophoric,
where interpretation of anaphoric pronouns is dependent on contextual information within

the text.

15

Analogous to anaphora, Halliday and Hasan introduced another type, termed
cataphora, which involves pronominal anaphora where their interpretation depends on a NP
referent which is mentioned later in the text, as in:

“In addition to his work in linguistics, Chomsky has become well known for his
writings in other areas as in politics and mathematics.”

However, cataphora falls outside the scope of the present study. I will be
concerned with what Camden (1982) referred to as "backward anaphora" as illustrated in the
following sentence:

“Chomsky is such a well known linguist that many modern trends

in linguistics are associated with his name.”

Additional studies have dealt with various aspects of anaphora. Fox (1987), for
instance, studied distributional frequencies of pronominal anaphora in spoken and written
English discourse. Hirst and Brill (1980) and Sanford and Garrod (1981) concerned
themselves with the role of general knowledge in the interpretation of anaphoric pronouns.
Clancy (1980), utilizing data gathered by a methodology similar to that used in the present
study, investigated the cognitive and discourse constraints that governed the selection of
referential choices in Japanese and English.

Givon (1983) and Payne (1988) stated that the structuring of information within
texts affects the choice among the different devices of continuity. Among such discourse
unit are hierarchical thematic or episodic structure of a text and issues of grounding in

categorizing information.

I.B.2. Referential selection among coding devices
The choice in English between nominal or pronominal reference is not entirely
optional (e.g. Hinds [1977] paragraph structure). Chafe ( 1976) has investigated how
speakers select between nominal and pronominal references in relation to the distinction

between “given” and “new information.”

16

Chafe referred to “given information” as the kind of knowledge “which the speaker
assumes to be in the consciousness of the addressee at the time of the utterance”. “New
information” on the other hand is referred to by Chafe as information that the speaker
assumes s/he currently is “activating” or “re-activating” in the hearers' consciousness.
Establishment of givenness is not limited to linguistic context, it can be established by
paralinguistic context as well. The major linguistic effect of the distinction between given
and new information in English, and possibly all languages, as Chafe seemed to suggest, is
that “given information is conveyed in a weak and more attenuated manner than new
information.” In English “given” information is usually weakly stressed and consequently
is easily subjected to pronominalization. Nevertheless, speakers will take care to avoid the

use of pronominals if it would result in ambiguity.

I.B.3. Topicality, availability, and referential distance

Van Dijk and Kintsch (1983) investigated interpretation of anaphora in discourse.
They suggest that topicality was a very inﬂuential factor. Topicality (topic continuity) is
particularly determined by various factors, one of which is the status of the referent as
“new” or “given”, as Chafe (1976) has stated.

Two other major factors were identiﬁed by Givon (1984) as inﬂuencing the
availability of referents in the minds of listeners. One additional factor is the referential gap
between the pronominals and previously mentioned referent. The larger the gap between
the two the higher the probability of difﬁculty in coreferential interpretation. Consequently,
a larger gap requires the speaker to establish reference via stronger linguistic coding.(i.e.
full NP, rather than an anaphoric pronoun). The second factor is the potential of
interference by other referents. Even if the gap (referential distance as termed by Fox,
1987) splitting two mentions of the same referent is relatively short, the appearance of other
referents in the text interfering between the two mentions may complicate the proper

referent interpretation.

17

To study this referential gap and measure potential interference, Givon ( 1983) has
put forth a quantitative topic continuity system. According to Givon's analytical system,
referential distance between the current occurrence of a previously mentioned referent in the
discourse is measured by the number of clauses that divide them,where it is marked by a
speciﬁc grammatical coding device (i.e. NP or pronoun). I will discuss in depth during
discussion about referential patterns in Chapter III that this measurement of RD is adequate
only when accounting for the basic sequentially distributed patterns of anaphora. This is a
shortcoming of the tradition theory of discourse anaphora in this regard. Fox ( 1985)
initiated a more advanced approach that accommodates the difference in nature between the
major two anaphoric patterns. Interference (the introduction of new referents between a
pronoun and its referent) is measured by the disruptive effect that introduced referents may
have on the topic interpretation within the immediately preceding discourse. Fox (1987)
went further in discussing this point by investigating the different effect interfering
referents from the same vs. the different gender may have on anaphora interpretation and
referential choice.

This framework of topic continuity has been implemented by several discourse
analysts working on topic continuity in English: Givon (1983); Brown (1983) in written
English; (Hinds ( 1983) in Japanese; and Sun and Givon (1985) in Mandarin Chinese. All
these scholars have expressed in their ﬁndings that the topic continuity continuum has

proven to be in harmony with their results.

I.B.4. Anaphora in syntax
A major hypothesis in the study of anaphora in syntax is the Forward Only
Hypothesis. Carden(1982) argued that “A grammar that limits its domains to sentences
must necessarily have a rule or rules for Backwards anaphora. If we expand the domain of
the grammar from the sentence to the discourse, we need to consider the hypothesis that

there is no Backwards anaphora rule as such, and that the apparent cases of Backwards

18

anaphora are all really Forwards anaphora in discourse structure” (p. 384). In his
analysis, Carden questions the validity of the Forwards-Only Hypothesis, since he was
able to collect numerous counter-examples from actual language use. The Forwards-Only

Hypothesis carries the claim that anaphoric relationships work only forwards which entails

that the antecedent always precedes its referential anaphor. Therefore, it only accounts for

anaphora interpretation in such structures such as:

NPi [ ........ PROi ........ ] s

 

In this sense, it does not include anaphoric relationships where anaphors precede

the referent NP such as in the following structure;
PROi [ ........ NPi ......... ] s

Such counterexamples, as Carden claims, were as frequent as other types of
discourse, contradictory to what has been predicted by the Forwards-Only Hypothesis. He
went further to argue that such contradictions can be related to the fact that the “architects of
the Forwards-Only Hypothesis based their analysis on the same sort of data that has long
been standard in theoretical work in syntax and semantics: they relied on their personal
introspection.” (p.385). Realizing the shortcomings of limiting the study of anaphora to
the sentence boundaries, I will also discuss in the following section the call for analyzing

anaphora in a broader context.

l9

I.B.S. Anaphora in Arabic and English
Al-Fihri (1982) mentioned that anaphors in Arabic referential relations can have two
forms: Dhahirah, a pronominal appearing in the surface structure or Ghayer dhahirah

elliptical (zero-pronoun), The former takes two shapes:

(a) (munfassil) free morpheme (independent pronominal) as in;
hum qabaal-uu ll-mushrifa

they met the advisor

(b) (mutassil) a bound morpheme (an attached pronominal) as in;
..... qabala-hum ll-mushrifu

met them the advisor

The independent pronominals appear as syntactic constituents and have their
presence at the phonological level (surface structure) as well. Attached pronominals have a
surface existence morphologically as sufﬁxes. Elliptical anaphoric elements (zero
anaphors) have speciﬁc underlying functions just like the previous pronouns, but they do
not appear at the (surface) phonological level. In other words, they have gone through
some form of deletion at the underlying structure (as proposed by Bakir [1979] below).
Therefore, in this view all Arabic pronominals can be said to share the function of being
anaphoric, when bearing a grammatical referential function. Ellipses may be
phonologically empty but present grammatically (at least on the UL structure).

free
Apparent Pronominals {
bound

Pronominals

Elliptical Pronominals

.17

20

Abdo (1973) disagrees with several Arab grammarians who consider the suffixes as

[-uu] , [—aa] and [-na] in the following sentences as pronominals.

a. ?ar-rijal-uu ghadar[-uu] l-mattaar.

the men left the airport

b. ?ar-rajul-an ghadar[-aa] l-mattaar.
the men two left the airport

c. ?n-nisaa?-uu ghadar[-na] l-mattaar.

the women left the airport

. Abdo deals with them merely as morphological markers of number and gender of
the subject. Among several arguments, he mentioned that if these sufﬁxes are considered
as subject pronouns, we contradict the rule that states that a sentence cannot have more than
one subject. Therefore, such verbal sufﬁxes have no syntactic functions. They are
morphological indicators of number and gender of the subject, apparent or deleted (pp. 71-
74).

Contemporary Arab linguists like Ayoub (1957) and As-Samirraa’i (1966) in
addition to Abdo have reached the same conclusion that what traditional grammarians called
an attached pronominal (to the verb) and given the function of the subject, is only one way
of marking number and gender. The slot of the subject is usually ﬁlled with a NP (N or

PRO) as in the following examples;

(a) ?al-?awlaad-u (noun) thabab-uu ?ilaa l-maddrasah
the boys went to the school
(b) hum (pronoun) thahab-uu ?ilaa l-madrasah

they went to the school

21

(c) thahab-uu ( zero anaphora) ?ilaa l-madrasah

went to the school

In (c) above there is no surface noun or pronoun ﬁlling the subject slot. Instead,
there is a NP trace (a zero-pronoun) when enough information to recollect its reference is
available when ﬁlling that slot (Bakir, 1979). In discourse analysis it is referred to as
ellipsis (Clancy, 1980) or zero anaphora (Givon, 1983).

Bakir (1979) coincides with the previous consensus of considering sufﬁxes
attached to verbs as merely carrying certain morphological information about the subject.
He accounted for the absence of the subject pronouns in Topic-Comment structures in

Arabic by the following deletion rule:

NPi ........ @[ ........... NPi { + pro. + nom.} ] ---—> 1 O

l 2
@=S,NP

The application of this deletion rule as described by its SD is restricted to T-
Comment structures where the topic NP C-commands the trace (Chomsky, 1980).
Nevertheless, the application of the syntactic rule is limited to the above sentential context;
zero anaphora in discourse seem to go through a much more complicated process, so that
any approach to account for it has to take several factors beyond the linguistic context into
consideration.

Bloch (1986) considered verb sufﬁxes, which provide information about the
number and gender in Arabic, as the subject (pronoun). Bloch analyzed this as pronoun

reduplication in Arabic discourse, as in the following example;

22

“Maa zhalam-na -hum wa-lakin kaan-uu humu zh-zhaalimiin”
not were unjust them and but were a they the unjust
“We have never been unjust to them, but they were the unjust”

(Bloch, 1986: 1; from Qura’an 43:76)

In this case Bloch did not consider the subject of the second clause as a zero-
pronoun, the sufﬁx -uu is a bound pronoun instead. Therefore he claimed the existence of
two pronouns ﬁlling the subject position of the second clause (this is contradictory to the
consensus of contemporary Arab linguists as I have discussed above, pp. 19-20). Bloch
claimed that the reduplication of pronouns has a discourse function as a “focusing device”:

thus marking this item to stand out against the rest of the sentence, which

contains the presupposed or known (given) information. An item so marked is the
focus of the sentence (p.1, 2).

Such an argument could also be considered if we can determine the unmarked
element of zero-[pronoun opposition, even if we take the ﬁrst theory of considering the
sufﬁxes like the ones mentioned above as merely donating information about the subject of
the clause. The theory that accounts for a single slot for the subject seems to be more
convincing, especially when the provided evidence is compared.

Zero-anaphors occur in Arabic, English, Chinese and many other languages (Fox
1987, Li and Thompson 1979, Al-Fihri 1982 and Givon 1983); however, they seem to be
diverse in their distribution and function. Zero-anaphora in Arabic and English are
restricted to the subject position syntactically and require coreferentiality for their
interpretation. On the other hand, zero-anaphora in Chinese (as reported by Li and
Thompson [1979] depend for their interpretation solely on pragmatic information, in the
sense of native speakers' knowledge about the world. What also makes zero-anaphora in

Chinese so unique is that it is not limited in its coreferentiality or distribution to any

23

grammatical function or syntactic slot. Furthermore, Li and Thompson have pointed out

that:
massive non-speciﬁcation of arguments occurs in Chinese discourse, and that
the interpretation of theSe unspeciﬁed arguments is neither a function of the
syntactic roles of the referents nor a function of the distance referents from the
unspeciﬁed arguments” (p.317).

They added that:

The interpretation of zero-pronouns can not be explained on the basis of
structural factors involving grammatical relations or grammatical functions.
Instead, the referent for a zero-pronoun is inferred from the information provided
by th; discourse context and the knowledge shared by the speakers of the language’
(pp. 21-22).

9

LC. Summary

In this chapter, I have reviewed the literature related to the topic of this study. The
review has focused on the nature of discourse and some of the methods discourse analysis
adapted to meet the demands of its data. Finally, some preliminary issues about anaphora
in Arabic and English were discussed.

The ﬁrst section of this review chapter focuses on several works by prominent
discourse analysts. These studies have highlighted the complex nature of discourse and
broad spectrum of approaches to discourse analysis. It has been repeatedly suggested that
it is a matter of necessity for researchers who share assets in discourse analysis to work out
innovative methods to meet the demands required by the type of data they handle.
Houghton and Hoey ( 1982) have eloquently emphasized the urgency of such collaborative
efforts, by stating, in regard to discourse analysts and second language acquisition
researchers, that:

the linguists must build bridges between their various theories and studies

before the students can build their bridges between the rhetorics of their first

language and the language they seek to acquire. . . (p.14 quoted in Ouaouicha,
1986, 92).

24

The second part of the chapter sheds light on contemporary approaches to
anaphora. Examples are cognitive constraints implemented on anaphora interpretation, as
in Clancy's (1980) study of anaphora in English and Japanese narrative discourse, and the
research on topic continuity in discourse as demonstrated in Givon ( 1983) by several
scholars. Special attention was given to Fox's (1986, 1987) studies of patterns of
anaphora and their distribution in the light of textual inﬂuence of different spoken and
written English genres. Finally, the inﬂuence of discourse factors such as topicworthiness
and discourse structure on referential choices (as studied by Payne, 1988) was also
carefully examined. As I have indicated, these researchers approached anaphora in
discourse from different perspectives to achieve certain goals.

The last section of this literature review chapter aims to explicitly discuss some
preliminary issues regarding the nature of anaphors in Arabic and English with extra focus
on the resolution of some of the conﬂicting theories regarding the nature of zero anaphors
in Arabic. Some recent syntactic approaches to the study of anaphora within the sentence
boundaries were also discussed to show the progress reached in that domain. Urgency for
study of anaphora in discourse is increasing because of its increasing importance in
providing input to ﬁelds like artiﬁcial intelligence, second language acquisition, translation
and psycholinguistics. My present research will deal with anaphora in oral narratives in
Arabic and English with a new system of analysis which will combine the major strategies
of approaching anaphora, particularly those approaches of Clancy (1980), Givon (1983),
Fox (1986; 1987) and Payne (1988) with some modiﬁcations of my own. The following
chapter will present in detail the complex nature of such an approach and how it is expected

to meet the goals of this study.

CHAPTER II
METHODOLOGY AND RESEARCH DESIGN

ILA. Introduction

In this chapter of the dissertation techniques and instruments I adopted in executing
the aims of this study will be deﬁned and explained.

To brieﬂy summarize from Chapter I: discourse analysis is an interdisciplinary ﬁeld
deriving its resources from several theoretical and methodological systems to meet the
difﬁculties of analyzing linguistically and culturally diverse data. Discourse analysts have
established premises for their ﬁeld based on research done in related disciplines. For
instance, techniques for data collection and transcription are borrowed from conversation
analysis and the ethnography of communication. Elicitation and techniques of tape
recording are borrowed from sociolinguistics, and so on. So, I see discourse analysis as
adapting these theories, rather than being tied to a single theory and a single way of
thinking with respect to the diverse nature of its resources. Therefore, in my study of
anaphora in Arabic and English oral narrative, I have made use of applicable previous
techniques drawn from sociolinguistics, cognitive science, linguistics and ethnography of

speech, in my data collection, data transcription, and analysis.

II.B. Design of The Study

II.B.l. Participants and data collection
The data collected for this study are 20 audio-taped oral narratives. Forty people
participated in narration. Twenty of them were native speakers of Arabic (Arabic spoken in
the Southern Province in Saudi Arabia, more speciﬁcally Abha and the surrounding area) in
10 dyads of an addresser and an addressee (see Table 2.2). Twenty were native speakers

of American English (mid-westem dialect), also in 10 dyads of an addresser and an

25

26

addressee (see Table 2.1). All participants were between the ages of '18-28, male, college
students or recent graduates. Subjects participating in each pair of “addresser” and
“addressee” were friends. So, the process of the speech event was natural and not
hindered with complicated edicts as usually happens between strangers.

Subjects were requested to participate in the data collection process in groups of
four to six. After the group had assembled, they were divided into two subgroups, with
each “friend” dyad divided between the two subgroups. Group (a), the addressers,
watched a video presentation of “The Pear Story.” After watching the ﬁlm (which usually
takes about 13 minutes), each then rejoined his colleague in the original “ﬁiend” dyads to
retell what he had watched. Group (b) the addressees, at that time, were waiting in
different separate rooms with a pre-set tape recorder to tape the narration and interaction
that took place. After interactions between participants came to an end, the researcher
requested the addressers to write down what they remembered after seeing the ﬁlm.
Addressees also were requested to write down what they remembered from their
colleagues’ retelling.1

The data collection for Arabic oral narratives took place on the campus of King
Saud University, Abha branch, Abha, Saudi Arabia. English oral narratives were collected

on Michigan State University campus, East Lansing, Michigan.

27

 

 

 

Table 2.1 Summary of information about American participants
PAIR ADR AGE EDU ADSE AGE HJU CL ANP
r 28 us 27 110 106 135
2 23 no 24 RG 254 215
3 21 us 21 US 86 137
4 23 RG 19 UG 90 118
5 26 RG 23 US 157 167
6 18 tn 18 US 72 102
7 22 us 23 110 84 85
8 21 US 21 US 92 113
9 23 m 23 UG 156 205
10 21 us 21 us 84 145
Total 1181 1422

 

Table 2.2 Summary of information about Arabic participants and narratives

 

 

 

PAIR ADRAGE E'DU ADSEAGE EDU CL AN?
1 26 RG 26 RG 86 121

2 20 {I} 21 II} 84 149

3 28 RG 28 RG 192 248

4 28 In 22 In 86 138

5 21 [X3 20 II} 122 131

6 23 In 18 II} 139 192

7 19 I13 18 II} 54 96

8 22 In 28 no 81 132

9 22 II} 18 [Ii 112 181
10 20 IX} 26 IX} 83 114
Total 1039 1502

 

PAIR: the participants. ADR AGE: age of the addresser. EDU1= age of the addresser.
ADSE AGE: age of the addressee. EDU2= education of the addressee. CL: the number of clauses. ANP= anaphoric
patterns.2

There are signiﬁcant differences between the data collection process of this study
and previous studies of discourse elements using the "pear ﬁlm". Studies by Tannen,
Clancy, Chafe, and Du Bois in Chafe (1980) were based on one-sided narratives, in which
a subject watches the ﬁlm and narrates what he has seen to the researchers or another
person who may have watched the ﬁlm for several tirrres; the speech event is not

“authentic” in that the events are not new to the audience any more. A close look at some

28

of the previously transcribed data of the Pear Stories in Chafe (1980: p9) suggests that they
may lack the kind of interpersonal interaction that is central in everyday face to face
communication.

In the current study participants narrate what they have seen in the movie to
colleagues of their choice who have not seen the ﬁlm. This approach assists the
enhancement of pseudo-natural narration. Participants involved in each narration are male
and close friends, who share the same social and cultural background. These shared
characteristics among participants should increase the mutual interaction between each
speaker and recipient during narrations. Indeed, recipients in this study were not merely
passive listeners; they were active in asking for clariﬁcation and commenting on events,

similar to any conversational narration speech event (Polanyi, 1979).

II.B.2. Materials

In both groups of participants involved in this study, the Arabic speakers (AS) and
the American English speakers (AES), the taped narration speech event centers on a seven
minute video presentation of " The Pear Film". It was originally produced as a sound and
color 16mm ﬁlm with out dialogue, which was designed to be used for academic research
in discourse analysis. This speciﬁc task is selected as the elicitation tool for narration for
the following reasons:

1) Chafe (1980), a major researcher in discourse analysis, designed the ﬁlm
for research purposes; to provide insight into how people talk about things they have
experienced and later recalled (see Tannen, 1984).

2) The ﬁlm was designed to be easily interpreted by people from different
cultural backgrounds through the use of an uncomplicated plot.

3) The ﬁlm depicts a group of people and objects participating in the events in

various modes, thus providing good raw material for anaphoric operations (Chafe, 1980).

 

29

4) There is no dialogue and therefore, the ﬁlm is a candidate for retelling by
people who speak different languages.

5) The ﬁlm has proven useful in the research of others in studies involving
anaphora, among them Clancy’s (1980) study of referential choices in English and
Japanese.

6) Genre is consistent. Tannen (1984) suggests that some of the
generalizations about the dichotomy of speech and writing can be blamed on differences in
genre in the texts involved.

Since speakers of both languages, Arabic and English, are telling the same story
they are dealing with the same referential tasks during the course of the nanation process.
However, the indexical forms of reference commonly used by speakers of each language
incorporate differing amounts of information about their referents. Arabic has grammatical
gender and therefore it is required that all referents have gender markings, male or female,
regardless of whether they are animate or inanimate. English on the other hand, does not
have grammatical gender; inanimate objects are not marked for gender.

Chafe (1980) provides this description of the events that took place in the ﬁlm:

The ﬁlm begins with a man picking pears on a ladder in a tree. He descends
the ladder, kneels, and dumps the pears from the pocket of an apron he is
wearing into one of the three baskets below the tree. He removes a bandanna
from around his neck and wipes off one of the pears. Then he returns to the
ladder and climbs back into the tree.

Toward the end of this sequence we hear the sound of a goat, and when the
picker is back in the tree a man approaches with a goat on a leash. As they
pass by the baskets of pears, the goat strains toward them, but is pulled past
by the man and the two of them disappear in the distance.

We see another close up of the picker at his work, and then we see a boy
approaching on a bicycle. He coasts in toward the baskets, gets off his bike,
looks up at the picker, puts down his bike, walks towards the baskets, again
looking at the picker, picks up a basket full of pears. He puts the basket
down near his bike, lifts up the bike and straddles it, picks up the basket and
places it on the rack in front of his handlebars, and rides off. We again see
the man continuing to pick pears.

The boy is now riding down the road, and we see a pear fall from the basket
on his bike. Then we see a girl on a bicycle approaching from the other

30

direction. As they pass, the boy turns to look at the girl, his hat ﬂies off, and
the front wheel of his bike hits a rock. The bike falls over, the basket falls
off, and the pears spill out onto the ground. The boy extricates himself from
under the bike, and brushes off his leg.

In the meantime we hear what turns out to be the sound of a paddle ball, and
then we see three boys standing there, looking at the bike boy on the ground.
The three pick up the scattered pears and put them back in the basket. The
bike boy sets his bike upright, and two of the other boys lift the basket of
pears back onto it. The bike boy begins walking his bike in the direction he
was going, while the three other boys begin walking off in the other direction.

As they walk by the bike boy's hat on the road, the boy with the paddle ball
sees it, picks it up, turns around, and we hear a loud whistle as he signals to
the bike boy. Then the bike boy stops, takes three pears out of the basket,
and holds them out as the other boy approaches with the hat. They exchange
the pears and the hat, and the bike boy keeps going while the boy with the
paddle ball runs back to his two companions, to each of whom he hands a
pear. They continue on, eating their pears.

The scene now changes back to the tree, where we see the picker again
descending the ladder. He looks at the two baskets, where earlier there were
three, points at them, backs up against the ladder, shakes his head, and tips
up his hat. The three boys now are seen approaching, eating their pears. The
picker watches them pass by, and they walk off into the distance.

II.C. Analysis

The current study attempts to provide a complete account of the distribution of a
anaphoric references to third person in narratives by Arabic and English speakers, in order
to bring out the relationship between discourse and anaphora in cross-linguistic texts.
Therefore, the scope of this study is limited to third person in narratives only. The domain
of this study has been narrowed in this way in order to concentrate on anaphora within its
prototypical use (Du Bois, 1980), that is, its function of tracing participants through
discourse, or what Prince (1982) labeled contextually “evoked references”. Because the
objective of this study is focused on the relationship between high level discourse
considerations and anaphora, the patterns of anaphora I will examine represent what has
been called “discourse anaphora”; in other words, the instances analyzed here are not
syntactically controlled. Reinhart (1983) discusses this difference in detail.

Indexically interpreted diectic elements such as I, we, you which depend on other

non-linguistic information for interpretation (Fillmore, 1975) are considered outside the

31

scope of this study. Other uses of anaphora where reference is linked to previous

utterances, rather than to participants in the discourse (e.g. this in the following examples)

are also excluded.

A: wu mashaw wu maruu bi-nafs 'ar-rajul 'ilithi kaan
and walk (they) and passed (they) by the same the man who was
yaqtif ?il-kimithraa - 'ah..'ah - ilithi?akhad min 9ind-ah
(he) picks the pears who (he) took from him
s-salah. 'wukaan-at 'amaam-ah salah mulq-ah wa-l-
the basket and was fem. in front of-him basket thrown- it and the.
'ukhraa fadiy-ah --farigh-ah fas-taghrab 'ar-rajul min
other empty-fem empty-fem. confused the man from
hadaa 'il-hadath- wa 9inda-thalik - ja9al yandur
this the incident and then started (he) looking (he)
'ilay-.yhim

to them

{N.18 A}

A : "and they walked past the same man who was picking the pears - ah..ah-, whom [the
boy] took the basket from. And a basket was thrown in front of him [the man] and the

other was empty.. empty. He was confused from this incident. Therefore, he started

looking at them [the passing by boys]."

Another example from the English data;

A: OK. I see !

B: so, this is it - it wasn't very long.

{NE.10}

32

II.C.l. Major anaphoric patterns
In this study, I will focus on the occurrence of the two major anaphoric patterns
which have been identiﬁed in the literature - sequentially distributed anaphora and retum-

pops. Each type is deﬁned brieﬂy below:

II.C.1.a. Sequentially distributed anaphora.

This pattern has the form of a mentioned NP followed by a sequence of one or
more coreferential pronominals until the sequence is closed, i.e. no reference to the NP is
activated and continuation of pronominalization is terminated. Fox (1986: 29) referred to
this pattern with the following quote:

The ﬁrst mention of a referent in a sequence is done with a full NP. After that, by

using a pronoun the speaker displays an understanding that the preceding sequence
has not been closed down (p.18).

I tried to formalize this pattern as follows:

CLIINP; ..] CL2[..pron; ...]{(CL3[...proni...])(CLn[...proni..])}

The following excerpt will illustrate this anaphoric relation;

S: CL 1: uhm- OK. It starts out with - uh a picture of like..a.. a hill side.
2: you see this guy upon a tree

3: an- and he is like a Spanish looking guy, big mustache and big bushy side
beard.

4: he 's got an a..apron.
6: -- and uh-- he climbs down..down the ladder from the tree.
7: .. an..and he puts them in his apron.

8: .. and he puts them on the basket...

33

9: and he - yeah he,s got a..red neck band.
10: he takes the band off.. and clean the pears.
11: an- then- uh- he puts them in the.. basket.
12: then he climbs back to the tree --
13: .and uhm- - and mm walks by - leading a goat- on a rope
‘ [NA 10]

The sequence of anaphoric relations starts with the NP this guy at 2 as the ﬁrst
mention of a referent to which a sequence of pronouns refers. The continued use of
anaphoric pronouns referring to the same entity from CL3 to CL12 displays the
understanding that the anaphoric sequence is not yet closed. The introduction of a new
'character (underlined) described with an NP at CL13, marking the shift to a new episode in
the narration, is an indication that the previous anaphoric sequence has been closed.

Findings about the distributional aspects of sequential patterns are discussed in III.A.l.

II.C.1.b. Return-pop anaphoric pattern.

Fox (1986: 36) described this pattern as having pronouns that have previously
mentioned referent NPs in the discourse and were lastly referred to by another pronoun.
The term "pop" had been introduced previously in computer science literature by Grosz
(1977) and used by Linde and Goguen (1978), and Marios (1981), as reported by
Veronous and Gass (1985). Nevertheless, Fox was the ﬁrst to use the term "return pop"
referring to a speciﬁc anaphoric pattern in discourse analysis. She observed the existence
of two conditions on the “popped over” clauses (gaps):

1) They should contain information relevant to the referent NP.

2) The 'popped over' material should not include complex structures.3

cll[...NPi....] cln[...proni...] ....(gap >2) ..... c|n[.... proni..]

34

The following example will illustrate the anaphoric relation in this pattern;

81: -- bi-ssudfah fii-yh thalath-ah 'awlaad.. ya9nii wahid min-

by chance there three (male) boys it means one from

hum yimkin 9umr-uh khamisa9ish! .. sitta9sh 1.. with-thaanii
them probably age his ﬁfteen sixteen and the second
thalata9ash! ?rba9ta9ash..! with-thalith ﬁi hiduud l-ihda?ash.
thirteen! fourteen! and the third in boundaries the- eleven
R2: ?aah .. uh.
S3: ?il-lwased hatha ma9-ah.. shif-t l-lmadhrab haq-l..l..l .......
the middle one this with him see I the racket for the the..
R4: //?it-tinss ?
the tennis ?
SS: ?it-tinss ?iwah..! wa-ﬁi-yh kuurah ka-nna-haa marbuttah.. wi-
the tennis yes! and there is aball like it (female) tied with
l-lhabil hathaa blastiik... ya9nii yharrik tadhrib ﬁi-yh wu-
the rope this plastic it means he moves it hits on it and
t-harrik-- ?il-muhim jaa'-uu .9 li-l-walad wa... wa. uh..uh

it moves the important thing come- -they to the boy and and
laqatt-uu l-kimithraah.//
gather they the pears
R6: //‘il-bint raah-at?
the girl went (fem.) ?

S7: 'il-bint raah-at...//

the girl went (fem.)

R8: // mashiiy-ah 9alaa ?aqddama-haa?

walking (fem.) on feet her ?

35

S9: rah-at maa shif-naa-haa ?llaa lamaa...( )//
went she neg. see we her except when ..
R10: // 9alaa rijlay-haa hii mashiiy—ah?
on feet her she walking ?
81 1: la 9alaa d-darrajah.
no on the bike
R12: darrajah thanyah?
bike another one?
813: 9alaa d-darrajah na9am- wu-mash-at ttab9an — 'akhir
on the bike yes and walked she of course the last
mulahadhah la-naa ﬁ-l—felm ?nna-haa mash-at hiiyah
notice for us in the ﬁlm that she walked she
..ya9nii 'athnaa twazzi -hum wu- ma-9aad shahad-naa-haa...
it means while parallel they and neg. see we her
?l-muhim - humah laqattuu ma9-ah l-kimithraa.

the important ~they gather with him the pears
[NA 10]

[Free translation]
S1: -- by chance ..there were three boys one of them is ﬁfteen..! sixteen ..l
.and the second is thirteen..! fourteen...! and the third is around eleven...
R2: ah..uh.
S3: the middle one he has like .. have you seen the racket for..//
R4: // tennis?
SS: tennis! yes! .. and there is a ball as if it were tied with a plastic rope... he
moves it then it (the ball) on it (the racket) .. and then it moves-- the important

thing is that, ethey came to the boy.. w..w uh..uh..they gathered the pears.

36

R6: the girl went away?

S7: the girl went away.

R8: walking on her feet?

S9: we didn't see her except when //( )...

R10: //on her feet, she was walking?

81 1: no! on the bike.

R12: another bike?

$13: on the bike! yes! and she walked away of course-- the last thing we've
have seen in the ﬁlm was that she walked away when they (the boy and the girl)
went side by side then they disappeared... the important thing is that -they

gathered the pears with him.

Although the episode above is rich with several anaphoric patterns of all types, for
our purpose here, I will turn the attention to how the anaphor "0humah" (they) in the last
sentence is a pop returning to its referent NP "thalathah awlaad" (three boys) in the ﬁrst
sentence. This return-pop was established after a single sequence with an anaphor in 5
“jaa'-uu..” (they came). Then, the anaphoric sequence was terminated until the last
sentence, where the reference was re-established with a return-pop. As Fox (1987) has
observed, the popped over material did not include complex structures that could have
hindered the proper interpretation. The popped over material included relevant mentioning
of the referent as in 5, in addition to the fact that the narration process was in the
boundaries of the same episode of the narrative, which is an early observation that requires
further testing before considering it a valid factor in determining retum-pops.

For each anaphoric pattern, distributional frequency, referential gap, and referential
choices will be measured and compared to other patterns and cross-linguistically (Arabic
and English). My aim of studying retum-pops in Arabic and English narratives is to ﬁnd

out if there is any correlation between the use of pops and the inﬂuence of inherent and

 

37

discourse features of referents. I am also interested in testing the constraint Fox (1987) has
suggested on the appearance of return-pops on a cross linguistic basis and on a more
harmonious type of data. Findings in regard to the distribution of retum-pops are reported

in III.A.2.

II.C.2. Frequency of distribution: referential gap and referential choice

II.C.2.a. Frequency of distribution.
The frequency of each pattern will be measured relative to the opposing patterns, to

the text as a whole, as well as to the two languages involved.

II.C.2.b. Referential gaps (RG).

Givon ( 1983) and Fox (1986, 1987) used the term “referential distance” (RD) to
[refer to the "gap" between a previously mentioned referent NP and its coreferential
pronouns. Givon has proposed the clause as an instrumental unit for measuring the
referential distance between referent NPs and their coreferential pronorrrinals, based on the
assumption that “the clause is the basic information processing unit in human discourse”
(p.7) Thus a referential distance of one indicates that the participant was mentioned with a
full NP in the irrrrnediately preceding clause and therefore maximally continues to ﬁgure in
the discourse. In his study, Givon held referential distance to a maximum of 20 clauses.
This value of maximum discontinuity was arbitrary, imposed because theoretically RD is
presumably inﬁnite, and we can not deal satisfactorily with inﬁnite values. I have chosen
to adopt Givon's procedure for measuring referential distance with the exception of
imposing any limit to the number of clauses in RD, since in practice it is feasible to count
the actual number of clauses“. I also introduced other modiﬁcations in terminology as I
will discuss right below. Sections III.B. and III.C. donate explicit analysis referential
gaps in Arabic and English narratives and the distribution of sequential and return-pop

patterns.

38

In this study of anaphoric relations in Arabic and English spoken narratives, I
introduced certain terms and modiﬁed others to help account for the complexity different
anaphoric patterns produce. Referential gap (RG) is utilized as a general term to refer to
any kind of distance that exists between an anaphor and its previous referential antecedent,
regardless of its morpho-syntactic type ( i.e. Full NP, pronoun). Ultimate referential
distance (U RD), as used in this study is a term reserved for sequential anaphoric pattern. It
is different from Givon’s notion of RD in the sense that Givon treats the distance between a
referent and its anaphors in a proximate referential sense. URD as discussed in this study
accounts for the distance in number of clauses between anaphors and their ultimate NP
reference; even if it is not the nearest anaphoric occurrence in discourse. Givon’s RD
refers to “the number of clauses from the last occurrence in the preceding discourse.”
(Givon, 1989; p. 216) Givon does not distinguish between the type of the preceding
anaphoric device, therefore, RD is proximal in this regard.

Although, there is great similarity between URD and Givon’s ( 1983) t0pic
persistence (TP) in the sense of measuring the distance in number of clauses back to the
ultimate topic NP. In my approach to URD, I differ with Givon in two ways, Firstly, I do
not consider the gap of two clauses to terminate the referential persistence between an
attenuated coding device and its full NP referent in sequentially distributed anaphoric
patterns. Secondly, Givon uses TP to refer to “the number of reoccurrences of the referent
in the subsequent ten clauses;” as he mentioned in his most recent publication ( Givon,
1989; p. 216) While, I do not see a reason to set such a limit. I also reserve the use of the
term “referential distance” RD to the referential gap (RG) that exists within return-pop
anaphoric patterns, in the same way Givon used the term to refer to “the number of clauses
from the last occurrence in the preceding discourse.” (Givon, 1989; p. 216) However as I
mentioned earlier I do not impose any maximum number of clauses for RD.

Unquestionably, any inclusive theory of anaphora in discourse has to go beyond

merely calculating the numbers of mentions and referential gaps to encompass discourse

39

factors that may play a role in deterrrrining RG. Two of such factors were proposed by
different discourse analysts:
a). Topic worthiness of participants (Reinhart, 1982)
b). Structure and ﬂow of discourse (Clancy, 1980 and Fox, 1987)
These factors will be discussed in more detail in II.D. Findings about the
interaction between the referential gaps and discourse and contextual factors are presented

in Chapter V.

II.D. Referential Choice (RC)

Referential choice, or choice between a noun phrase, a pronoun or a zero-anaphora
when referring to a previously mentioned participant in the narrative, will be discussed and
'compared in Arabic and English narratives. I will also study the inﬂuence of factors like
topicworthiness, which has been reported by Payne (1988) as signiﬁcant in determining the
RC of referents in Yagua. The inﬂuence of episode, a discourse unit, on the RC3 will be
also discussed in both Arabic and English narratives. Clancy (1980) reported that
‘episodic’ boundaries were signiﬁcant in determining RCs. Findings about the cognitive
and contextual constrains the choice of referents are discussed in the fourth chapter.
Referential gap, the constraints it attracts and overlapping relations will be the focus of

Chapter IV.

II.E. Topicworthiness of Participants

Payne (1988) found that the role that a participant plays in the narrative affects the
RC (choice of a full NP as opposed to a pronoun or a zero-anaphora). Payne referred to
the RC as, the “coding device” a speaker uses when referring to a participant within the
ﬂow of discourse. Payne (1988) and Givon ( 1983) perceive a full NP as a “strong” coding
device, and other choices as attenuated coding devices. Payne observed that central and
major characters are referred to with greater RD; therefore, it appears that narrators think

their audiences can endure a longer distance of discontinuity between referents and their

40

coreferential anaphors than is the case for weak or secondary participants in narratives. In
other words, central and major characters showed a higher level of continuity
(topicworthiness) than weak and secondary characters, and therefore, allowed more weak
coding devices (e.g., pronouns) than were used for weak and secondary characters.

The present study will shed light on the effect of topicworthiness as a contextual
factor in determining the RG as implemented through the anaphoric patterns in narratives
produced by Arabic and English native speakers and on the selection of the different
anaphoric devices. Lichenberk (1988) recognizes the role topicworthiness plays in
discourse in suggesting that;

The importance of a participant in a situation is not a semantic matter: it is a
matter of context. Furthermore, it is relative concept. It can be judged only in

relation to the situation the participants are involved in and in relation to the
importance of the other participants in the situation (p. 342).

In the current study the techniques proposed by Payne (1988) will be taken into
consideration with some modiﬁcation to capture because of the differences of the data used
in both studies. Payne proposed these character statuses to ﬁt his scale of topicworthiness:

1. Central characters - characters that the narrative is built around, and who are
“normally present through the text” (360) . Their position as central characters is not
challenged even if they are not referred to in an entire episode.

2. Major characters - those who are mentioned four times or more in 20
clauses of the same episode. Major characters may not hold onto their status in other
episodes.

3. Minor characters- other participants that do not meet the above criteria. As
for the scale of topicworthiness adapted here, a scale of three degrees of topicworthiness
(major, secondary, and minor) depending in the role they play in the narratives.

Payne (1988) used one long narrative in which he based his topicworthiness scale.

He found himself obliged to distinguish between central and major characters, although the

41

distinction was not very clear. Later during analysis, he dealt with them as a single
category in relation to their anaphoric relations.

Simply because the data of this study includes 20 narratives (ten in Arabic and ten
in English), I had to ﬁnd a technique to generalize the scale of t0pic worthiness adapted
over the 20 narratives. Therefore, frequency of appearance is used to account for the
topicworthiness of referents used in our data in addition to other criteria.

1. Major characters: are characters who play a prominent role in the narratives
and events in the narratives revolve around them. Their role as major characters is not
challenged even if they were not mention in an entire episode. They also scored the highest
percentage through all narratives. Only two characters in "The Pear Story" as told by
Arabic and English speakers ﬁt the above criteria; The Pear Picker and The Boy on The
Bike. The Boy was mentioned by Arabic narrators 404 times making 26.9% of the total
number of characters mentioned. English narrators mentioned The Boy 433 times with a
percentage of 30.5 of the number of characters mentioned. The other major character, The
Pear Picker was mentioned 253 times in English narratives, composing 17.8% of the
number of characters used . Arabic speakers on the other hand, mentioned The Pear Picker
303 times, forming 20.2 % of the number of characters mentioned. Table 2.3 below will
summarize their frequency in comparison to other characters.

2. Secondary characters: are characters in “The Pear Story” that were used by
the narrators in a frequency that is not as high as the major characters and not as low as the
minor characters. These kind characters also play a secondary role in directing the event in
the narratives. Secondary topicworthy characters were treated similarly by Arabic and
English narrators as in Pears and The Three Kids. They differed in their treatment of The
Basket. Only Arabic narrators treated it as a secondary character. It was treated as rrrinor
in English narratives (as shown in the table below)

3. Minor Characters: all other characters were treated as having minor

t0picworthiness role.

 

42

In addition to the quantitative criterion for determining topicworthiness of referent
in the Pear Stories as summarized in Table 2.3, referential persistence and role in building
plot of the narratives are taken into consideration. For a character to be considered major in
the topicworthiness scale it has to persist over 3 episodes including the one in which it was
ﬁrstly mentioned. Only “the bike boy and the pear picker” met this criterion. Secondary

and minor characters were decided solely on their distribution.

Table 2.3 A summary of the topicworthiness status of characters in The Pear Stories
as produced by Arabic and English participants

 

 

 

 

 

 

 

 

 

 

 

Arabic En lish
. character # o #
I Major The boy 404 26.9 433 30.5

The pear 303 20.2 253 17.8
picker _

Secondary The kids 170 11.3 149 10.5
The pears 132 8.8 143 10.1
The basket 120 8.0 - 0

Minor all others <8.0 i < 8.0 ]

0 does not apply.

As the frequency of the different characters might imply that the relation among the
various characters with different topicworthiness status forrrr a continuum from the most
topicworthy to the least topic worthy. In Chapters III and IV, I will discuss how such a
continuum of topicworthiness affects the topicality of such characters in the sense of the
anaphoric RG and the kind of coding device they take. Tables 4-1 ,5-4 and 5-5 summarize

these relationship.

II.F. Discourse Hierarchy: Episode as a Unit of Analysis
Fox (1987) shows that the choice between a pronorrrinal or a full NP in English is
inﬂuenced by the hierarchical structure of the contents of the texts. It also has been noted

by Clancy (1980) that discourse boundaries tend to elicit stronger coding devices than

43

would be expected given a strictly linear view of continuity based on the number of
mentions or the number of clauses since the last mention of a referent. Thus, interaction
between discourse and grammar has been clearly demonstrated in relation to anaphora.

Clancy (1980) found out that “episode” boundary was signiﬁcant in determining
RG and RC (coding devices) involved in anaphora in English and Japanese narratives.
However, she did not deal with retum-pops that I expect will show interesting behavior
across discourse boundaries. Givon (1983) referred to episodic structure of the text as a
discourse unit that inﬂuences the “thematic continuity” and relates to higher level discourse
boundaries that affect the speakers‘ choice of reference (or coding device).

Different speech genres tend to have their own higher level discourse hierarchical
units. The paragraph has been proposed by Longacre (1983) as higher level unit in
expository texts. As for narratives, episode has been used by different discourse analysts
as a major discourse unit .

Clancy (1980), Fox (1986), Payne (1988) and Tannen (1979) referred to episodes
as thematic frames and scenarios. Mandler and Johnson (1977) and Fox (1987 ART)
referred to episode as the major complex unit in narratives and “which dominates the
highest node of the story [structure] which dominates the entire text” (Fox, 1987 ART;

p. 362). According to their categorization each episode consists of two units: (a) an
initiating event in which some sort of conﬂict is developed through the introduction and the
interaction of one or more characters; (b) a development unit which reveals the character(s)’
response to the initiated event and any plans or actions that they take as a reaction.

Chafe (1979) provided a more detailed description of an episode, which depends on
pauses in oral narration. He mentioned that episode tends to begin with hesitations of
longer than two seconds. He added that an episode illustrates uniﬁcation of “character
conﬁguration, spatial location, and coherent temporal and event sequences.”

After examining the data of this study which are based on watching The Pear Story,

episodes as the immediate discourse units for our purpose in this section are based on the

above categorization which consists of 7 scenes that the ﬁlm is based on (pages 18, 19 for
more details). Narrative episodes as presented by Arabic and English narrators were
consistent with the sequence in which they were presented in the pear ﬁlm. However,
some of the participants would skip some episodes all together. Such behavior as I found
out is inﬂuenced by the speakers cultural background.

Episode will be used in this study as the discourse unit against which anaphoric
patterns RC and RG of anaphoric relations will be analyzed. Results will be compared
with other variables in the same language context and cross-linguistically, in narratives
delivered by Arabic and English native speakers.

Episode as an inﬂuencing discourse unit has been discussed in relation to the
distribution of anaphoric patterns within and across episodes, as in III.C.4. The interaction
of episode boundaries and selection of coding devices have been thoroughly analyzed in
IV.D.

 

1The resulting written data will be analyzed in future studies, but not in the present
study.

2This computer assisted research used The Statistical Package for Social Sciences (
SPSS PC +) to assist the qualitative analysis of anaphora in spoken Arabic and English
narratives.

3 This constraint has been challenged by Cornish (1987). I will also provide
evidence from our data to refute the generalizability of this constraints when I focus on the
analysis of the distribution of anaphoric patterns in Arabic and English narratives.

4 Givon’s claim that RD can have an inﬁnite value in the sense that there can be any
number of clauses with maintained continuity between an anaphor and its referent has no
psychological evidence. Instead there seems to be cross-linguistic evidence that RD is
limited because of the limit of the human short term memory, as I will discuss on The RD
chapter. Another evidence against considering the RD as inﬁnite, that we can not deal with
inﬁnite values. Mathematically any value in comparison to inﬁnity equals zero.

CHAPTER III
ANALYSIS OF AN APHORIC PATTERNS IN ARABIC
AND ENGLISH NARRATIVES

III.A. Introduction

In this part of my analysis attention will be directed to anaphoric patterns and their
distribution across the two languages under study, Arabic and English. As I have indicated
in Chapter II, scholars involved in the study of anaphora in discourse have realized the
existence of two types of anaphoric relationships that Fox (1987) categorized into two
types of anaphoric patterns: sequential and return-pop. Nevertheless, I will argue that only
Grosz (1977) and Fox (1987) had treated return-pop as a separate pattern, it was either

neglected or accounted for but not as a separate pattern.

III.B. Anaphoric Patterns in Discourse

III.B.l. Sequentially distributed anaphora

Fox (1986) referred to this type of anaphoric relation as a sequence of anaphoric
chains of pronouns established after an antecedent in prior discourse, which was ﬁrstly
introduced by a full noun phrase referent until another sequence is closed with a full NP
referent‘. Givon (1983) referred to this pattern as “major juncture” of anaphoric relations,
where there is “a continuous sequence of actions without a break in the thematic
continuity.” (p.350).

As long as the sequence of pronouns or zero anaphors referring to a speciﬁc
antecedent is still activated and continues to persist in the interlocutors' minds or until
another character is introduced or a new discourse unit emerges, a sequential pattern of
anaphoric relations is still in effect. Fox (1988) also referred to this form of anaphoric

relation as “The Basic Pattern”.

45

46

I tried to formalize this pattern as follows:
CL1[NP| ..] CL2[...pron...] {(CL3[...pron....])(CLn[...pron,...])}

According to this analysis sequentially distributed patterns were determined on
several bases within the realm of the above formula:

a. An anaphoric pattern is considered sequential when it involves a ﬁrst
mention of a full NP followed by at least one anaphor.

b. A gap of two clauses between an NP and its anaphor or between an anaphor

and following one is the maximum for a sequential pattern to continue.

{ E X . l A}
A: 1 ?al.. 1. lqabl qaliil shahadt .. 2 ya9ni.. hadhaa l-ﬁlm -- # 3
before a few moments I saw (means) this the ﬁlm
kaan rajul fi-mazra9ah .. 4 wa huu waqif 9alaa shajarah - 5
wasaman in afarm and he standing on atrec
yalqutt ¢ ba9dh ?a-thimaar..6 wa hiya l-k.mithraa- 7 wa-
picks he some the fruits and it the pears and
hadha a-shaa?ib yadha9 a-thimaar ﬁ kumm-ah ..
this the old man puts the pears in pocket his
8 thumma yajma9 -¢ -haa ..9 thuma yanzil a bi-haa.. 10 ?ilaa
then collects he them then descend he with them until
?ann yasil o ?ilaa sallah .. 11thuma yadha9 {5 haa haa..//
“ ” gets he to a basket then puts be it it
B: 12 // ?is-sallah 9alaa l-ardh?
the basket on the earth

A: 13 yadha9 a hadhi-hi ?a-thimaar ﬁi s-salah.. 14 wa ka?an

47

puts he this them the fruits in the basket and as if
ma9-ah thalatlr sallaat. 15 _ wa qad malaa? a thintayn ?aw
with him three baskets and already ﬁlled he two (mm) or
ithnataan.. 16 wa- wa sa9ada t! maratan ?ukhraa ?ilaa ash-
two (acc.) and and climbed up he time another to the
shajarah..l7# wa ﬁi ?athnaa?i dhalik ..18 wa huwa yaqttifu 1-
tree and in during this and he picks the
kimithraa ..19 jaa? walad 20 _ yarkab e darajah.. 21.. wa hadha
pears came a boy rides he a bike and this
?a-shaab yablugh 9umr- uh taqriban thalathat 9ashar sanah..
the young man reaches age his approximately thirteen year
22 wa kaan yalbas ¢ quba9ah .. 23wa 9alaa raqabat-ih khayt.. 24
and was wearing he hat and on neck his “ ”
?aw kama yuqaal karafatah hamraa .. 25 thuma jaa?a ¢ ?ilaa a-
or as said a tie red then came he to the
shajarah .. 26 wa nadhara a ?ilaa r-rajul ..27 Thuma ?akhadh
tree and looked he to the man then took
ti as-salah..28 wa wadha9a- a - ha 9alaa ad-darrajah .. 29
he the basket and put he her on the bike
thuma nnsaraf ¢..//
then left he

B: //.. 30 sallah wahidah?
basket one

A: 31 na9am, sallah wahidah..32 Thuma ?akhad-o-ha hadhi-hi

yes basket one then took he her this

48

s-salah .. 33 wa dhahab o .. 34 wa r-rajul hadha laa yadrii..
the basket and went he and the man this not know
[N A 1 5]

[free translation]

A: a.a a.. a few moments ago, I saw I mean in this ﬁlm -- a man was in his

farm .. and he was up on a tree- (he) was collecting some fruits .. and they were

pears - and this old man puts the fruit in his apron .. then he collects them .. then

gets down with them .. until (he) reaches the basket .. then puts it.. //

B: // the basket on the ground ?

A: puts these fruits in the basket .. and it seems that he had three baskets .. _ and

(he) ﬁlled out two of them.. and .. and he climbed back to the tree .. and during

that .. and while he was picking the pears .. a boy came riding a bike .. and this

young man is about 13 years of age .. and (he) was wearing a piece of cloth

around his neck .. or as its called a red tie.. then (he) came to the tree .. and (he)

looked at the man.. then (he) took the basket .. and put it on the bike .. then (he)

left... //

B: // .. one basket ?

A: Yes! one basket .. then (he) took it .. this basket .. and went away.. and this

man does not know

The excerpt above from (NA 15A) included several anaphoric sequential patterns.
Some of them overlap with one another. In clause (3) a new referent was ﬁrst introduced
with an indeﬁnite NP lrajul/ [a man] followed by a sequence of two anaphors in clauses (4)
and (5). Two clauses later the same referent was introduced again but with a deﬁnite NP
this time, since it was established as old information at clause (3). However, the referent
was introduced with an explicit NP /?sh-shaayb/ [the old man] as in clause (7). This

anaphoric sequential pattern involved ten anaphors and persisted over 11 clauses, from

49

clause (7) to (18). This sequence was discontinued at clause (19) by the introduction of a
new referent /walad/. After an anaphoric sequence over two clauses, the sequence was
terminated by a deﬁnite NP referring to /walaad/ [a boy] as an old referent, but with some
modiﬁcation /?sh-shaab/ [the young man] as the narrator has done to the previous referent
/?r-rajul/.
Afterwards, a new sequential pattern emerged containing 8 anaphors and persisted
with an URD of 10 clauses (from cls 19 to 29). However, continuity of reference was not .1
terminated by the end of the sequential pattern; a return-pop pattern took over after a gap of ’
3 clauses at clause (32). ,
When a sequence of attenuated anaphoric coding devices (zero or pronominal
anaphors) is used, it indicates that the ﬁrst mentioned referent with a full NP (strongest
coding device available to the narrators) is still activated in the minds of both addresser and
addressee. The moment such a referent loses its continuity, the sequence has to be
terminated with a more explicit referential choice (RC); otherwise ambiguity and
misinterpretation may occur. Since the use of pronominals is common in natural languages
to achieve anaphoric relations, it is usually the addressee who carries the burden of
interpretation to determine the right referent for zeros or pronorrrinals, assuming that the
addresser is following Grice's (1977) maxims.
Givon (1979, 1983) has proposed that the appearance of a second referent in the
discourse can cause the ﬁrst referent to be referred to with a stronger coding device (a full
NP); reference with a pronoun or a zero-pronoun is not adequate unless the language has
other devices for keeping the two distinct. An example of that is the introduction of /?ar-
rajul/ in clause (34) since the sequence was closed at clause (26) with a full NP. However,
based on the numerous occasions of anaphoric relations in Arabic and English narratives,
that the choice of a full NP or a pronoun to establish reference involves more factors than
just the introduction of another referent. Both characters mentioned in the excerpt (A1) are

of equal topicworthiness status given the deﬁnition on II.E. I expect that referential choice

50

is inﬂuenced by the status of the referents involved in regard to topicworthiness. It may
also have to do with the role a referent plays in the discourse in addition to the structure of
the discourse and the kind of pragmatic information involved to assist interpretation. I will
discuss these factors in greater detail in the section on referential choices (RC) .

In the following example of sequentially distributed anaphoric patterns, a long
sequence of reference to the ﬁrstly introduced referent [the man in the tree] has been
terrrrinated at clause (30) when a new episode was introduced along with a newly
established referent. This new referent has the same gender and number as that of the
previous referent, actually with the exact reference [a man], but as an indeﬁnite referent.
The use of the pronoun [him] in clause (30) is ambiguous, since it could refer to both men.
Nevertheless, the recently introduced [man] has initiated a new sequence and thus replaced

the previous referent as the focus.

{EX.1E}
A: 1.. uhm . what I've seen .. uhm.. 2 there was a. . a man .. 3 a man was
picking .. fruit .. uh .. 4 I could n't identify what the fruit was .. 5 only what
I saw .. 6 what the fruit was.. 7 I saw the name on the tape .. 8 like uh ..( ) ..
9 I thought .. 10 it was pears.. 11 but it wasn't pears .. uh ..12 then he was up
on the tree .. 13 o picking 'm.. 14 ¢ putin' 'm uh .. on .. uh .. 15 Wa' looks
like.. a.. a pron .. ya know.. uh .. 16 iii dropped them in the ground .. uh 17
..and then ..I think he got enough.. 18 then he came down .. uh .. 19 iii put
those in the basket ..( ) .. uh .. 20 then he had this scarf.. 21 that he had
around ..22 he took that off .. 23 then.. he laid it out.. 24 then he put the
piece of fruit on that.. 25 a whipped it up .. 26 then .. a put it on the
basket .. 27 then he took .. uh .. the .. uh .. handkerchief .. he had .. 28
then .. a put it around his neck.. uh.. 29 then you see him going back up the

ladder 30..# uh.. at this time .. uh -- a.. a a man .. came by with a goat .. 31

51

and .. see him walking .. uh .. you know ..32 right by the fruit .. that was in
the basket..33# in the meantime. he passed ( ) _ uh 34 then a little boy

COIDCS ..

[NE 1]
III.B.2. Return-Pops

III.B.2.a. Deﬁnition of return-pop.

I think that Givon (1983) referred to return-pops with the terrrr “major junctures” as
opposed to “minor junctures”, which are “sequential pattern”. However, he did not go on
to test their distribution on the basis of the unique role they play in discourse. In other
words, no theoretical or methodological techniques were assigned or discussed by Givon
(1983) or Clancy (1980). In summary, the above two approaches to return-pops are
lacking in two ways. First, they do not acknowledge the difference between ultimate
referential distance which occurs within a sequential anaphoric pattern and referential
distance which occurs within a return-pop. Secondly, little was mentioned about the nature
of this pattern and the role it plays in discourse or the problems it may pose for
interpretation because of its complex nature.

Fox (1986: 36) deﬁned the return-pop pattern as involving pronouns that have a
previously mentioned referent NPs in the discourse and were lastly referred to by another
pronoun. The term “pop” had been drawn from computer sciences literature by Grosz
(1977) and'used by Linde and Goguen (1978), and Marios (1981), as reported in
Veronous and Gass (1985). Nevertheless, Fox was the ﬁrst to use the term “return-pop”
referring to a speciﬁc anaphoric pattern in discourse analysis. She observed the existence
of two conditions on the 'popped over' clauses (gaps):

1) They should contain information relevant to the referent NP.

2) The ‘popped over’ material should not include complex structures.2

52

What differentiates the return-pop pattern from the sequential one is that in the latter
the ultimate referential distance between the anaphor and its referent is continuous through a
chain of explicit referential devices. In the former, the referential distance takes the form of
no linguistically explicit reference. In other words, there is no pronorrrinal or @-
pronominal link between the pop and its previously mentioned referent.

In my treatment of the distribution of return pops in Arabic and English narratives I
assign a gap of 3 clauses nrinimum for an anaphor (as illustrated in the formula below).

I formalize the return pop pattern as follows;

cll[....NP,..cln[.... pron(n), ..... ].. (gap>2cls)...cln[.... pop....]

The following excerpt from English and {Ex.E1 } above provide an example for
retum-pops from Arabic.

{EX. 2E}
Al: I wasn't sure .. I couldn't tell .. I must be imagining it .. and uhm.. the kid
hits .. after he hits the stone..this big STONE .. looks like .. uhm .. what do you
call ? ..What do you call a_ volcanic stone
0?
Bl: ( )
A2: Looks like a big chunk of [a black ....( ) }
B2: [ black thing is( )in shape. ]
A3: Yeah! .. uhm .. and uh .. there he falls down .. and .. he . he pulls up
his pants .. like .. ¢ pulled up his sock .. and uh ..
B3: Did he hurt his leg ?
A4: Well ! apparently .. because he rubs it .. there is no wounds .. or

anything wrong with his leg .. it must ache a little bit..

53

B4: {( )1
A5: {( )} .. but of course all the pears were scattered on the road ..
B5: Yeah!

A6: an .. uh .. and .. uh uh it turns out .. standing near by are ..a.. aa three ..
of ..uh .. young boys ..uhm .. and two of them walk out into the road .. and
begin helping ohim .. pick up the pears .. ¢ put them in the basket/l

B6: ”Anything that was said during the course of any of this?

A7: No! .. no words .. no words although there was .. one whistle.

B7: O.K.
[NE.2]

The return-pop anaphor him in A6 above is a return-pop referring back to the NP
“the kid” in Al which has been followed by a sequence of pronouns. The sequential
pattern was terminated at A4. Five clauses later a return-pop emerged to resume an

activated anaphoric relation.

III.B.2.b. Return-pop: a neglected anaphoric relation.

Certainly, return-pop as a major type of anaphoric relation in discourse has received
very little attention in traditional works of anaphora. As a matter of fact, return-pops were
excluded from analysis because of their complexity. However, their investigation provides
valuable information about coherence in discourse and reveal a lot about how humans map
information in their minds and how they perceive it (Grosz, 1977; Fox, 1987). Payne
(1988) mentioned that;

.. certainly there are times when a speaker needs to refer to participants that have
not been brought onto the discourse stage have no topicality what so ever.

Therefore, in order to introduce a participant, or to introduce a participant after a

signiﬁcant period of absence, more marked morpho-syntactic coding devices are
called for. (p.356).

 

54

Payne’s claim is based on the assumption of the traditional view of topic continuity
in discourse which is strictly linked to morpho-syntactic markers as that of Givon (1979;
1983). Continuity of anaphoric relations in discourse in such a framework is viewed as a
one-way relation, in the sense that only when there is a continuous chain of reference to a
certain referent is continuity observed. Payne (1988) in a foot note acknowledges that such
a perspective is extreme. He also mentioned that the anaphoric relation be was discussing
in the above quote (return-pop in this analysis) is very difﬁcult to study. Therefore, the
best way to deal with it is to avoid it, or in his terms “to control for them.” (P.389; fn 6)

This analysis and previous studies of anaphora by Grosz (1977) and Fox (1987 and
1987 ART) suggest that the return-pop anaphoric pattern is so prominent in human

discourse that it can not be left unaccounted for because of its complexity alone. -

III.B.2.c. Constraints revisited.

Cornish (1987) rightly challenged Fox's constraint on the distribution of return
pops, which states that the popped over text should include some related information about
the referent. Cornish (1987) was able to refute Fox’s claim regarding this constraint using
Fox’s own data. (Cornish, 1987; p.238). In the analyzed 20 Arabic and English spoken
narratives for this study, several incidents of return-pops appeared with anaphors which
occurred after a long RD (gap), and yet the text popped over did not contain information
pertaining to the previously mentioned referent. Nevertheless, proper interpretation was

not hindered.

{EX 3E}
A1: # uh..m .. It looks like this foreigner he .. was an a a --- in this tree
picking pears O.K. . he was.. he gets a few .. and it puts them on the
basket ya know ! -- o kept doing it.. ya know! .. he put them in the

basket .. he went back to the tree -- ¢ collects mor -- # meanwhile --- this -

55

this white boy .. uh - this boy is driving by in his bike... he comes by .. and
sneakily it takes a basket full of pears .. he just (laugh )'.. looks .. if the guy is
looking .. he wasn't there. he drove up .. ya know!..O.K. ! .. so, then he is
going down like this dirt path - he is going one way balancing the pears. and t3
keeping on track .. ya know! ..( ) .//

B1: // was ITon a car .. ora bike?

A2: on a bike.

B3: O.K.!

A3: # so th- then this.. and here is a girl -- this other foreigner girl looks
like .. she's conring on the - other way on another bike...and ( ) --- uh .. the was
waggling the bike back and forth .. and he's losing his balance

[NE 6]

The example above also contradicts Fox‘s argument that return pops can occur only
when the popped over material is “structurally non-complex.” It can be extremely hard to
test whether a structure is complex or not, especially in regard to discourse structure, where
“complex” is only a relative measure that does not necessarily contrast with simple.
Nevertheless, Fox did not analyze closely this felicity in terminology. The example above
contains a return-pop anaphor in the penultimate clause of A3, “he”. Such a pop appeared
over a complex discourse unit (episode). The referent was ﬁrst introduced with a full NP
in Al, then followed by a sequence of anaphors which was terminated at the ﬁnal clause of
A1. Anaphoric relation was not terminated because of the introduction of a new episode at
A3 in accordance with Fox’s (1986) claim. A new anaphoric pattern emerged with a pop.

Fox (1987) suggested that return-pops can only be done with pronouns in context
with different gender referents. Their appearance in a same gender environment can hinder
proper interpretation. In fact Fox strongly claimed that “same gender environment is the

most restricted environment for the conversational context”. (p.146). Such a restriction

56

could be logically understood to require more effort for interpretation at the addressees
side, but there are arguments against maldng this generalization. In the ﬁrst place,
interlocutors do not depend solely on linguistic information provided in the text to interpret
referents, they also use their background knowledge of the previously mentioned referents,
as well as their ability to reason. I assume that interlocutors make anaphoric relations and
textual ties in general to develop coherence (Morgan and Sellmann, 1980). Also, there are
other factors at work, such as RG, discourse boundaries, shared background knowledge,
and saliency of a referent in the text. The data of this study provide a number of incidents
where pops occur in a same gender environment and only a single case of misinterpretation

took place.

{ E X . 4 E }

A: ----- (pause) a.. and this KID comes by on a BIKE .. an .. and _ uhm _ he is
coming from the other way _ a .. and _ he looks up on the tree .. and 16 sees the
guy busy picking he does n't see him .. then he grabs the basket _ then o puts
it on his BIKE __ in front of the..//
B: // w..w.wait a minute? the goat drags .. drags the basket?
A: NO .. NO.. the goat just looked at it.. it didn't..it didn't actually grab it.. uh _
I don't know why.. uh _ so _ _ he picks up the basket .. a . a. and o puts it on
his bike. and {6 takes off with this basket of pears ..

[NE 10]

In the example above, there is a clear case of an anaphoric overlap of two referents
of the same gender, “the kid” and “the guy”. A sequence of anaphors was established after
the ﬁrst mention of the NP "the kid" over 5 clauses. It was followed then by a return-pop
(in bold) after a gap of 6 clauses. Another sequence was resumed after the pop. Although

the pronoun “he” could have referred to either the guy or the kid, the addressee did not

57

have a problem linking the anaphor to its proper antecedent. This interpretation was made
on the information the addressee has built around the produced text; he knows that the
referent for “he” cannot be the guy since he cannot be picking pears at the tree and at the
same time picking the basket up ﬁom the ground. In addition, the following clause
provides information: “a. a. and {6 puts it on his bike”. Such subsequent information also
plays a role in clarifying any confusion regarding the appropriate referent of the return -pop
anaphor “he”.

As I will discuss in detail, a same gender environment is not the only determinant
for the distribution of return-pops; other factors like contextual information and background
knowledge play a major role in anaphora interpretation, as Li and Thompson (1979) have
suggested. The example above would argue against the restriction put on the distribution
of pops in a same gender environment. It suggests that pops can appear in a same gender
environment in the presence of contextual information to assist proper interpretation.

I think that Fox (1987) has probably reached her contradictory conclusions
regarding “a constraint” on return pop distribution because of two factors she did not
account for in her analysis. First, she was dealing with several discourse genres and text
types (conversations, narratives, and expository prose; in addition to spoken and written
data) at the same time. Such data types have quite different discourse hierarchical units.
For example, conversations have adjacency pairs; expository prose has rhetorical
structures, and narratives have episodes. These discourse units may not be compatible in
either length or structural complexity. Therefore, what is considered complex or non-
complex is only based on a conventional judgment that can not be transferred from one
discourse type to another. Tannen (1984 , 1989) has emphasized and discussed this point
in detail, especially in regard to contrast between spoken and written language. So, it is not
only the hierarchical structure of discourse alone that determine long distance anaphoric
relations as Fox claims, it is also the genre type, precisely because each discourse type has

its own hierarchical units3.

58

Secondly, as Cornish (1987) has stated, Fox did not seem to have an explicitly
deﬁned methodology with respect to her analysis. In addition to the severe restriction Fox
has put on her references where it became “ ...... difﬁcult to reach any really solid
generalizations about discourse anaphora as a whole on the extremely narrow basis
provided by third-person references via referential pronouns and (in the rrrain) proper
nouns to single, human individuals, which is Fox's choice in this work..” ( Cornish, 1987:

242).

III.C. Anaphoric Patterns in Arabic and English Narratives

In the attempt to compare anaphoric patterns in Arabic and. English, providing
distributional information about the two patterns is an obvious option. Anaphoric patterns
(ANP) in Arabic and English narratives will be discussed in relation to their frequencies,

their referential gap, and their distribution in correlation with the discourse unit of episode.

III.C.l. Frequency of sequential and return-pop patterns

The division of anaphoric patterns between sequential patterns and retum-pops was
of interest. The ten Arabic narratives produced 350 sequentially distributed patterns,
71.3% of the total number of anaphoric patterns produced in Arabic. Because of the
number of constraints implemented in the distribution of return-pops, Arabic narratives had
101 return-pops composing 28.7% of the number of patterns produced in Arabic. The
frequency of the sequential and retum-pops in English narratives was strikingly sirrrilar
with sequential patterns composing 70.00% of the number of anaphoric patterns produced.
Return-pops compose the remaining 30.00%. Such similarities among the distribution of
anaphoric patterns in Arabic and English narratives could be attributed to two factors.

Firstly, since both Arabic and English narratives were produced by narrators who
were telling a narrative about the same events and under similar circumstances, it is not
surprising that they produced very similar percentages of anaphoric patterns. This may be

related to the same experience that Arabic and English narrators had in watching the pear

59

story. Secondly, American and Arabic narrators were able to map the anaphoric relations
presented in the ﬁlm in the sequence they were presented. So later they are retrieved and
produced accordingly. Therefore, similar percentages of the two anaphoric patterns were
used. Difference in the number of patterns produced by Arabic and English narrators (31
sequential patterns and 7 pops) can be attributed to the different amount of data presented
by Arabic and English narrators as shown in Tables 2.1 and 2.2 in the previous chapter.
Another interpretation can be attributed to the difference in the length of anaphoric patterns
as produced in English and Arabic narratives, as summarized in Table 3-2. American
narrators produced sequential patterns with an average of 0.41 clauses longer than those

produced by Arabic narrators.

Table 3-1 Distribution of sequential and return-pop patterns in Arabic and English

 

 

 

 

 

 

 

 

 

narratives
anaphoric Arabic English
attem # % # % ‘
sequential 250 71.3 219 70.0
pop 101 28.7 94 30.0

III.C.2. Length of patterns
The mean length for sequential patterns in Arabic narratives was 3.68 clauses.
Return-pop patterns had the mean length of 6.18 clauses per pattern. 2.5 clauses longer
per pattern than sequentially distributed patterns in Arabic. Applying the one way AN OVA
analysis of variance, the variance of means for sequential and return-pops in Arabic had the
signiﬁcance of F Prob. = .0000. Sequentially distributed patterns ranged from a minimum
of one clause long pattern to a maximum of 26 clauses long pattern. Return-pop patterns

ranged from 3 to 48 clause anaphoric patterns.

60

On the other hand, sequentially distributed patterns in English narratives produced a
mean of 4.09 clauses per sequence of anaphors. The mean length of gaps between
anaphors and their referents was 5.61 clauses per gap. The variance of means was at the
signiﬁcant level of F Prob.= .0011. Continuity of sequentially distributed patterns in
English started by l clause as a nrinimum sequence to a maximum of 25 clauses. As for
return-pops , they started at the minimum (already assigned) gap of 3 clauses to a

maximum of 17 clauses.

III.C.3. Why do differences exist?

Sequentially distributed patterns of anaphora in English narratives are slightly
longer than those in Arabic, with an average difference of 0.41 clause for each anaphoric
pattern. Arabic narratives on the other hand allow for a longer gap of syntactic anaphoric
discontinuity of .57 per each anaphoric pattern per average. These types of differences in
the mean length of RG in the two patterns can only be interpreted on different bases,
because of the different nature of the syntactic continuity involved in the two patterns that
traditional work in anaphora overlooked (Clancy 1980, Givon 1983, Fox 1987). For
instance, URD can be very long in sequentially distributed patterns but this thematic
continuity is assisted with syntactic continuity, by a sequence of anaphors (if not in every
clause, then in every other clause). However, the gap of the number of clauses between
the pop and its preceding reference in return-pops can be thematically coherent, but
syntactically disconnected with the lack of anaphorically connected sequence. This is why
its distribution is more restricted to speciﬁc type of referents that have a strong thematic

continuity in the narratives.

61

Table 3-2 Mean length of sequential and return-pop patterns in Arabic and English

narratives

 

 

 

 

 

 

Arabic English
anaphoric mean of mean of
pattern length length‘
sequential 3.68 422
return-pop 6.18 5.61

 

 

English narratives produced 0.41 longer anaphoric sequential patterns than the
Arabic ones. This difference can be attributed to two factors. Arabic narratives included
more clauses and anaphoric entries than English narratives (see Tables 2.1 and 2.2) .
Therefore, Arabic narrators produced a higher number of anaphoric patterns than their
American counterparts (Table 3-1), which might have affected the length of the sequentially
distributed anaphoric patterns. Another interpretation which seems more probable is that
sequentially distributed patterns in English narratives have longer sequences in average than
in Arabic, because Arabic sequential patterns are distributed on the URD of 1-17 clause
long sequences. English narratives on the other hand extend beyond that to include several
anaphoric patterns distributed in the URD range of 18 -26 clause long sequences.

A detailed analysis of factors that might have inﬂuenced the difference in the mean
suggest that the gap between an anaphor and its referent in a return-pop pattern in Arabic
and English narratives can be attributed to the following reason. First of all, the difference
in the mean length of Arabic and English return-pops is 0.41 clauses longer in Arabic.
Surveying the length of retum-pops anaphoric patterns in Arabic, an extremely long pattern
with the length of 48 clauses emerged; the longest pattern (a single pattern) other than this
48 clause gap was one with a 22 clause long gap. This extreme incident of a 48 clause gap
has caused ambiguity, so that the listener misinterpreted the correct referent. This case is

not even conceived to be possible in natural language processing unless strong assisting

62‘

factors were involved. Therefore, if this extreme incident is taken from the calculation of
the mean for retum-pops in Arabic, the mean length will be 5.71. This result is amazingly
similar to that of English retum—pops, 5.61.

The results above can provide evidence that Arabic and English narrators in average
abide by a certain range of continuity. This nright be related to factors among which borh
narrators initiated their narratives based on the same experience (Pear Story). The form and
length of an anaphoric pattern as Fox (1987) and Stevenson and Vitkovitch (1986) have
suggested, are inﬂuenced by our cognitive capacity as humans to form a network of mental
models that link zero and pronominal anaphors with their antecedents in discourse. Fox
(1987 ART) argued that there also seem to be other conventional factors related to the type
of the speech situation which have a role in inﬂuencing anaphoric relations. As this net of
anaphoric relations expands, the more complicated the process of mapping and the more
difﬁcult interpretation will be. Therefore, more elaborate devices of reference are required.
Nevertheless, there are several factors that sustain appropriate interpretation and adapt a
lengthier sequence of reference. Among these are the topicworthiness of the referents
mentioned in an anaphoric pattern and the role of non-linguistic as well as linguistic

information for assisting comprehension of zero and pronorrrinal reference.

III.C.4. Episode as an influencing factor in the distribution of
anaphoric patterns
It has often been assumed that different discourse anaphoric patterns have different

structures and probably implement different discourse functions. However, previous
research has revealed very little about the inﬂuence higher level discourse units like episode
implement on the distribution of these anaphoric patterns in discourse. For the purpose of
this study, “episode” is the higher discourse level meant whenever the expression
“discourse unit” is referred to. Chafe (1979:162) mentioned, based on evidence from

several psycholinguistic studies, that there seems “to be universal tendency for people to

63

express their thoughts in linguistic segments with highly distinctive phonological and
syntactic properties.” The classiﬁcation and evidence for episode will be discussed below.

Clancy (1980) suggests that boundaries between discourse units not only inﬂuence
the choice of different anaphoric coding devices, but also have a great effect on activating
the addressees' consciousness about cohesive relations in discourse. Tannen also alludes
to the importance of such discourse units in determining expectations of speakers of a
language regarding textual and contextual relations by stating that;

. one organizes his knowledge about the world and uses his knowledge to
predict interpretations and relationships regarding new information, events, and
expectations.” (1979, pp. 138-139).

III.C.4.a. Episode and distribution of anaphoric patterns. .

The present computer assisted analysis for Arabic and English narratives includes a
calculation of whether an anaphoric pattern has its anaphoric relation (referent and anaphors
of the same pattern) in the same episode, or the anaphoric relation has gone over episode
boundaries.

My ﬁndings provide evidence that episode as a discourse unit in narrative
implements certain constraints on the distribution of anaphoric patterns across its
boundaries. English is more sensitive to discourse boundaries, but both English and
Arabic recognize episode as a unit of discourse. Episode plays a major role in anaphora
distribution and therefore inﬂuences interpretation. Anaphoric patterns are not, however,
equally sensitive to episode. This can be related to the structural and functional apparatus
of each pattern.

Prior to discussing the distribution of discourse patterns in Arabic and English
narratives, I will ﬁrstly discuss how episodes are realized as discourse units and what kind
of discourse markers they involve in both Arabic and English narratives. There may be
cross-cultural inﬂuences on these units. They probably are conventionally determined, in

addition to the abstract role they play as units of information in human discourse.

Process of narration production by Arabic and English narrators was consistent
with the ﬂow of information as it was presented in the ﬁlm. Such consistency of
information perception and production is evidence for the existence of episode as a
psychologically real discourse unit in narratives. Another kind of supporting evidence
from our data is the fact that in some cases, a speciﬁc episode (e.g., the episode about the

goat man) is skipped in its entirety.

III.C.4.b. Episode length in Arabic and English.

As summarized in Table 3-4, the average length of episode in English is 16.43
clauses for an episode, a difference of 1.67 clauses more than the mean length of episode in
Arabic. The mean number of anaphoric entries (ANE) is 21.19 ANE per episode in
English narratives. Arabic on the other hand has 22.26 ANE per episode. The difference
in length of episode and the number of anaphoric entries of episode in the two languages
can be related to the difference in the number of clauses and anaphoric entries by which

each language is represented as summarized at Tables (2.1) and (2.2).

Table 3-4 Mean length of episodes in Arabic and English Narratives according to
number of clauses and anaphoric entries

 

 

 

 

 

 

Length of Arabic English
episode
In clauses 14.76 16.43
In ANE 22.76 21.19

 

 

III.C.4.c. Cultural background and episode.
A speciﬁc episode in the pear story (the goat man episode) has been treated in
signiﬁcantly different ways by English and Arabic narrators, which can be related to the

difference between the two groups’ cultural backgrounds. Tannen (1980) suggested that

65

narratives provide some aspects of discourse like telling about events that are culturally
determined and conventionalized. Therefore, such conventions in turn contribute to the
impression made by narrators as members of different cultures or subcultures.

Both Arabic and English have used the lowest number of clauses (Arabic 12.33 cls,
English 8.74 cls) and the fewest number of ANEs in their narration of events in reference
to the goat man episode (English 11.62 ANEs, Arabic 18.00 ANEs). Such low
representation of this episode in both languages can be related to two factors. Firstly,
actions in the goat man episode play a very minor role in the development of events in the
narratives. However, there is a signiﬁcant difference between the mean length and number
of anaphoric entries in this episode in comparison with the over all means between the two
languages. In the English narratives, the goat man episode is 7.69 clauses 46.80 % shorter
than the mean length of episodes. In comparison to only 2.43 clauses 16.46% less than
average in Arabic. As for the number of anaphoric entries this episode is represented with
9.57 ANEs 45.16% lower than average episode length in English. Arabic narrative
produced this episode with 4.76 ANEs, 20.91% fewer than the overall mean. One reason
why this episode has a low representation is because it is very low in topicworthiness
representation. It includes minor characters only, (the goat man and the goat).

Arabic narrators on average used 3.59 cls and 6.3 ANEs more than English
narrators when referring to the goat man episode. Moreover, The goat man episode was
missing two times in English narratives“ (narratives 6 and 7). On the other hand, this
episode was missing only once in the 20 Arabic narratives5. Such differences in the
treatment of this pattern in Arabic and English may be due to cross-cultural differences
between speakers of the two languages. All Arabic narrators are from the Southwestern
mountainous province of Saudi Arabia, where goats are a very familiar part of the culture.
This in not true for English participants from mid-Michigan. Therefore, the narration
repertoire of Arabic and American narrators may have been inﬂuenced by their daily

cultural experiences.

66

III.C.4.d. Discourse markers of episodes.

Another major piece of evidence for the realization of episode as a principal
discourse unit in narrative is the frequent use of discourse markers in Arabic and English
narratives with the introduction of a new episode. About 70% of episodes in Arabic
involve the use of some discourse markers. English narratives also used discourse markers
at the beginning of each new episode 81.35% of the times. English narrators made use of
both phonological and syntactic forms of discourse markers. Phonological markers include
using long pauses of more than 3 second pauses, hesitations, repetitions and false starts.
Such markers are either used individually or integrated with syntactic markers, such as
adverbials and conjunctions, where the syntactic form is preceded by a pause. Examples of
these discourse markers are listed below, in examples 5E, 6E, and 7E.

Fox (1988) mentioned that episode-initial positions are strong candidates for the use
of marked morpho—syntactic devices. She went further to propose that this claim will hold
cross-linguistically. Fox (1988) mentioned that such a claim does not restrict the existence
of a marking device to episode positions only. “Rather it is meant to indicate that there
tends cross-linguistically to be a statistical correlation between these two items” (p. 360),
the existence of marked structures and the emergence of new discourse units. An example
that Fox has referred to as evidence is the existence of marked inverted clause types at
episode-initial slot at Tagalog with frequency of 60%.

In the present study, as I will discuss in the following section, there seem to be a
correlation between the distribution of marker anaphoric choices and episode-initial slots in
Arabic and English.

Examples of Discourse Markers in English:

{EX SE}
A: ..uh.. you see him .. going backup up the ladder .. #2 .. uh _ _ at this time ..

uh _ _ a.. a man .. came by with a goat .. and .. you see him walking around ..

67

uh .. you know.. right by the fruit .. that was on the basket ! #3.. in the mean
time .. he passed ( ).. then a little boy comes..
B: pass what?
[beginning of episodes 2 and 3, NEI]
{EX 6E}
A: .. he put them in the basket .. he went both to the tree _ collects more... #3
meanwhile _ _ _ this _ this white boy .. uh _ this boy is driving by in his bike.
[beginning of episode 3, NE 6]
{EX 7E}
A: .. they picked his hat that ﬂew off .. so, the guy brings this hat up to this kid..

¢ give it to him .. OK! #6 _____ theeen then .. the guy .. the fruit

picker (laugh) the fruit picker .. uh .. uh the fruit picker was up on the tree.

[ beginning of episode 6, NE 3]

On the other hand, episodes in Arabic are marked with a set of adverbials. Other
forms of episode initial markers included conjunctions. Especially, long distance

conjunction /thumma/.

{EX 2A}
A:..marr min 9ind-ah shaxss muqaarb la-h ﬁi-s-sinn .. wa ma9-ah
passed from him a person close to-him in the-age and with-him
ma9iz _wa marr min 9ind-ah ..wa raah o #3_ ba9dfatrah qalilah..
a goat and pass from besides-him and left he after period ashort
jaa? tiﬁl ssaghir.. la-hu min-l9umr hawaalii 9ashr sanawat .. ?aw
came a child small has-him from age about ten years or
?ihhdaa 9asher sanah 9alla darrajah hawa?iyah ........................... #4

eleven year on abike air

 

68

waﬁi ?athnaa?i t-tariiq .. wajahat-hu fatah 9alaa darrajah

and in during the-road met him a girl on a bike

?uxraa.. uh.. uh..uh ?thnaa? murruura-ha min 9ind-ah .. waqa9-at
another during pass -her from beside-him fell down
?al-quba9ah haqat t-tiﬁl .. mimaa ?addaa bi-hii .............................
the hat possession of the-boy which led to -him

#5 ba9da hadhaa l-hadith bi-lahattaat .. kaan ﬁih majmu9at

after this the-incidentwith-moments was there a group

?attfaal qariibin min-h

 

children near from-him
[beginning of episodes 3,4,5. NAl]
There are some incidents where narrators used the same discourse marker again and
again each time they start a new episode. These occurrences provide a strong evidence for
the underlying realizau'on for discourse boundaries. This could not occur randomly. An

illustration is the ﬁrst clause occurring at episode-initial slots in Arabic narrative 14.

{EX 3A}
#2 .. ﬁi ?thnaa? 9amli-hi .. ?thnaa? 9arnal-ih...
in during work his during work his
#3 .. wa ?thnaa? l-jannii..
and during the harvesting
#4 _ _ﬁi ?athnaa? maa hu maashii..

in during while he walking

#5 ﬁi athnaa? haadhii..
in during this
#6 .. ﬁi ?athnaa? dhalik ..

in during that

69

#7 .. ﬁi ?athnna? murrur-hum..

in during passing by their
[NA 14]

III.C.4.c. Episode and anaphoric patterns distribution.

After providing evidence for the existence of episode in Arabic and English
narratives, we will now focus on the way anaphoric patterns (sequential and return-pop)
react to episode boundaries. As summarized in Table 3-5, both anaphoric patterns in both
languages are affected by episode boundary in their distribution. The majority of anaphoric
relations tend to take place within episode boundaries. However, as summarized in Table
3-5, anaphoric patterns in English narratives seem to be more sensitive to episode
boundaries than in Arabic. Only 3.55% of sequentially distributed patterns are permitted to
continue their anaphoric relations across episode boundaries in the English narratives,
while 8.95% of sequential patterns in Arabic continued their anaphoric relation across
episode boundaries.

Because of the inherent nature of retum-pops, which establish anaphoric relations
across a gap of syntactic referential discontinuity, return-pops tend to be less restricted than
sequential anaphoric relations. In English 13.3% of retum-pops have established
anaphoric relations across episode boundaries. Arabic narratives seem to have less
restriction than English for return-pop anaphoric relation to occur across episode, with
33.34% of retum-pops in Arabic narratives establishing their referentiality across episode

boundaries, much higher percentage than that of English retum-pops.

70

Table 3-5 Distribution of sequential and return-pop anaphoric patterns in Arabic and
English narratives with in and across episodes

 

 

 

 

, ARABIC ENGLISH
ANP IN A CROSS IN
# % # % # %
sequence 228 91.5 22 8 95 211 96.45

 

 

 

 

 

 

 

return pop 67 66.34 34 33.66 8—1 86.7

 

III.D. Conclusion

In this chapter, I devoted my effort to discuss the recently proposed anaphoric
patterns (sequential and return-pop) that were not rigorously discussed by traditional theory
of anaphora in discourse. I modiﬁed some of the conditions that were prescribed on the
distribution of return-pop patterns. In addition to analyzing the distribution of sequential
and return-pop patterns in Arabic and English narrative discourse, I provided evidence for
episode as a discourse inﬂuencing factor of the distribution of anaphoric patterns in Arabic

and English spoken narratives.

 

1Fox (1987) did not include zero anaphora in her study although she has
acknowledge its existence.

2The second constraint has been challenged by Cornish (1987) using Fox's own
examples; I will provide some evidence for Comish's counter arguments later on the
chapter.

3 It is also important to mention here, that these different discourse situations
require different social and communicative task. In narratives, for example, the narrators
are in control of the communicative task, since they are the source of information for
addressees. This is way interaction in terms of two way communication take place only
when there is need for clariﬁcation from the addressees side or conformation from the
narrators side.

4Ten narratives only.
51 only used 10 of these narratives in this analysis.

5 Starting episodes are not included in the calculations.

 

CHAPTER IV
REFERENTIAL CHOICES IN ARABIC AND ENGLISH NARRATIVES

IV.A. Introduction

In this chapter, I will focus on the way lexico—grammatical choices of anaphoric
coding devices reﬂect the inﬂuence of discourse and contextual factors such as topic
worthiness, grammatical function, and episode in Arabic and English narratives.

While pronouns play many similar roles through out the world’s languages. The
way in which they are interpreted and used within different linguistic systems is highly
dependent on the speech repertoire of the available anaphoric coding devices that they
contrast with (Mithun, 1990). English for example, has a choice between pronouns and
full NPs. Zero-anaphors can be used under more restricted conditions and to achieve
different discourse functions. In many other languages, Arabic as an example, pronouns
occur to indicate emphasis when used in subject position where zero- anaphora are the
unmarked anaphoric choice. Pronouns are used to establish discourse and stylistic
functions other than referentiality except under certain syntactic requirements. Anaphoric
choices otherwise involve full NPs and pronouns.

When anaphoric usage is compared within the context of the full range of choices, it
will be obvious that such selections are not haphazardly executed but rather were highly
systematic and governed by cognitive and discourse constraints as mentioned by Clancy
(1980); Fox (1987,1988); Givon (1979,1983); Payne (1988) and Mithun (1990). Such
high levels of systematicity have great inﬂuence on the process of interpretation of
anaphoric relations. A general principle for retrieving referential information is: the more
available a referent in the memory of interlocutors, the less complicated the process to
retrieve it. Thereafter, bringing a referent to the attention of the bearer with the least

marked anaphoric elements does not affect interpretation.

71

72

This chapter focuses on factors that inﬂuence the choice among the coding devices
available to interlocutors in Arabic and English narratives including the topicworthiness of a
referent in the narrative, the different functions anaphors play in discourse besides

referentiality, and the effect discourse boundaries have on referential choices (RC).

IV.B. Referential Choices in Arabic and English

Instead of direct involvement in the comparative analysis of reference in Arabic and
English narratives, one appropriate starting point is through providing the frequencies with
which referential options available to the narrators in Arabic and English were used during
the narratives.

Indeﬁnite Full NPs are often used to introduce new entities in the discourse process
to the consciousness of the addressee. Full NP as a referential device is the most explicit
coding device among the choices a language allows as cross-linguistic research of Fox
(1987), Payne (1988), and Mithun (1990) has revealed. NPs are normally selected in
environments where their antecedents are presumably outside the addressees’ capacity of
tracing certain referents in the discourse. As I have discussed previously, analysis will be
limited to third person referents, although during the narration process ﬁrst and second
person anaphors were mentioned when narrators refer to themselves or to their addressees
and vice versa.

Speakers of Arabic have three options of referential choices (coding devices): noun
phrases, and more attenuated devices like pronouns, and elliptical elements (in the subject
position) in the form of zero-anaphora. English is also represented by a similar set of
coding devices. However, as will be discussed in detail, some of these coding devices
carry in addition to or in exclusion of their referentiality other discourse functions. As an
example, zero-anaphora exists in Arabic and English, but in different ranges of
frequencies. The distribution of zero anaphora in Arabic and English is restricted to subject

position only; other grammatical slots require more explicit RCs, as Table 4-1 illustrates.

73

Table 4-1 Referential choices (coding devices) in Arabic and English narratives

Arabic En

 

Table 4-1 provides a summary of frequencies that represent different referential
choices in Arabic and English narratives. Arabic narratives provide strikingly similar
percentages of the distribution of their referential choices, 33.9% for full NP’s and zero-
anaphors, and 32.3 % for pronouns. When compared with the distribution of such
referential choices in English narratives, English pronouns ranked highest in their
distribution with 47.1 % of the referential choices used. Full NP’s composed 41.7% of
RCs. The percentage of Zero-anaphora coding devices was signiﬁcantly low. It composed
only 9.4% of the RCs employed in English. Numeral values for each coding device are
summarized in Table 4-1 above.

Comparing the frequencies of the attenuated RC’s with the strongest coding devices
(NPs) in Arabic and English narratives, Arabic provided more freedom to use pronouns
and zero-anaphors; 66.1% of the number of RCs were distributed in these domains.
English narratives on the other hand allowed 58.3 % of RCs to take anaphorically
attenuated forms (pronouns and zero’s).

A ﬁrst look at Table 4-1 above would give the impression that English narrators use
of full NPs with higher percentage to establish reference compared with Arabic narrators,
since full NPs in English exceed those in Arabic by about 7.8%. Nevertheless, a more
inﬂuencing factor to the present hierarchy of referential choices is affected not by whether a
language allows a certain coding device or not, it is rather how much distribution a speciﬁc
coding device is allowed and under what conditions in relation to other RCs. Referential

choices by Arabic narrators are distributed almost equally in terms of their frequencies.

74

Although English narrators have access to a similar array of choices, establishing reference
by using zero anaphora is restricted to certain syntactic contexts.

Interpretation of such statistical information can be misleading because of the
complex relations involved among the different RC8, and because of the constraints each
language implements on the distribution and selection of these coding devices.

Although both languages have referentiality as a part of their discom'se functions,
the comparison between Arabic and English zero-anaphors has little to do with their
anaphoric distribution. Besides referentiality, zero-anaphora in Arabic and in English
narratives are inﬂuenced by the different discourse functions they play. Zero—anaphora in
Arabic is the unmarked coding device to establish reference in most subject positions.
Zero-anaphora in English, as I will discuss in section IV.A.1.b, plays an additional role
besides referentiality.

On the other hand, Arabic pronorrrinals occurring in subject position often do not
have referentiality as their main function since they do not provide additional information to
referentiality. Either they are syntactically motivated as in copulative clauses where they
can have anaphoric function, or they have other stylistic roles to play, as I will present in

the next section.

IV.B.l. English zero-anaphors and Arabic subject pronouns

Von Stutterheim (1989) pointed out that it is extremely difﬁcult to decide if
restriction on the distribution of zero anaphora is attributed to word order or to the
difference in other syntactic properties (subject, object ,etc.). During his analysis of zero
anaphora in German and English, Von Stutterheim found that zero anaphora are restricted
to clause-initial position and not to certain grammatical slots. German, for example, is a
language with a free word order that allows zero—anaphora to appear in object as well as
subject position, but always conﬁned to clause initial position. So, in German, zero

anaphora distribution is inﬂuenced by organization of information through word order.

 

75

Arabic has a somewhat free word order that can include an array of several word order
choicesl, as Thalji (1988) has illustrated. In contrast to German zero-anaphors, Arabic
zero-anaphora are restricted to subject position assuming that they follow the verb.
According to the theory of basic word order of VSO in Arabic (as proposed by Bakir
(1979) and Thalji (1988)), in English, with its rigid SVO word order, zero anaphora is
restricted to subject position only. Although the case can be partially settled in regard to the
appearance of zero anaphora in Arabic and English as restricted to subject position, the
issue seems more complicated, to be decided by reference to a limited number of
languages, and it requires devoted cross-linguistic studies.

Regardless of the complications around the restrictions of zero anaphors, their
choice represents two different discourse functions dictated by the two different linguistic
systems of Arabic and English. In Arabic, it is the unmarked form that attenuated third
person references select zero-anaphora assisted by agreement markings of number and
gender attached to the verb. Use of less attenuated devices, as pronouns, is restricted to
certain functions governed by the clause type or the speciﬁc discourse function intended.

In English reference is established with pronouns in all grammatical slots (subject,
object, object of a preposition, etc.; though in a limited number of cases, zero-anaphora are
used in subject position). Such marked usage of certain coding devices establishes some
meaning beyond their inherent meaning of referentiality as suggested by Clancy (1980) and
O'Connor (1990). In this section, inﬂuences on the appearance of low frequencies of zero-
anaphora in English and subject pronouns in Arabic will be discussed.

As summarized in Table 4-2, zero-anaphora in English represent 23.63% of RCs
occurring in subject position, in comparison to 76.37% pronominals. Moreover, zero-
anaphors compose only 9.4% of the anaphoric choices involved in English narratives.
Arabic subject pronouns on the other hand, represent 19.36%, a minority compared to

80.64% the zero-anaphors (T able 4-3).

76

Table 4-2 Distribution of attenuated coding devices in English according to their
grammatical functions

 

Table 4-3 Distribution of pronouns and zero-pronouns in Arabic according to their
grammatical functions

 

IV.B.1.a. Arabic subject pronouns.

To account for the appearance of Arabic subject pronouns and English zero-
anaphors, a detailed analysis of the distribution and contextual function of each coding
device was executed. Arabic pronominals occurred in subject position for one of the
following reasons. Firstly, there are certain syntactic constraints in Arabic, as in copulative

structures where a surface structure verb is not available (examples 1A and 2A).

{EX 1A}
?al-muhim jaa? rajuul ?ammrikii.. Ka?ann-ah rrrin ha?ulaa? lfursaan

the important came a man American similar he from those the knights

“The important thing is that an American man came .. as if he were from those
knights..”
[NA 1 5]

77

{EX 2A}
kaan rajul ﬁi mazra9ah .. wa huu waqif 9allaa shajarah - yaltaqitt
was a man in a farm and he standing on a tree collecting
ba9dh ?a-thimaar .. wa hadhi- hii ?al-thimaar.. hiyaa l-kimithraaa..

some the fruits and this the pear she tlrepears

“ The man was in his farm he was standing on a tree .. collecting some

fruits .. and these fruits .. they were pears.”

[NA19]

In such a context, the use of pronouns instead of zero-anaphors is obligatory in
Arabic grammar. These situations represent 37.77% of subject pronouns used. Secondly,
in non-copulative structures subject pronouns have discourse functions completely different
from referentiality. They occur by themselves or are often preceded by a certain set of
discourse particles to establish emphasis on certain referents during the narration. 51.11 %
of subject pronouns acted to establish emphasis. Bloch (1983) referred to these pronouns
in Arabic as “focusing devices”. They “stand out against the rest of the sentence, which
contains the proposed or known information” (pp. 1-2). Such occasions are illustrated by

the following examples.

{EX 3A}
A: ...?il-muhim Hummah laqatt-uu ma9-ah?il-kimithraa..?alkabiir..
the important they collected with him the pears the big
?akbar wahid .. wa as-saghiir ?akthar man laqatt.. ?illthaa ?a ?a -
biggest one and the small more who collected this a .. a ..
hathak saa9ad-hum l-wassitt ?illii 9umrah( ) .. ?ilghariib ?inn-
that helped them the middle one whosc age the strange that

ah hu maa 9amil shayy? .. ﬁi ?inn-ah yijjnii l-kimithraa hathaa llii

78

he he did not doathing in that he harvest thepears this which
hathaa. Hum jama99-uu-h .. wa-ba9da maa uhmm- wa- hum ..
that they collected it and after and they
qaam yanfudh banttalun-ah .. 9alla ?ill ..?ill .. 9allaa ad-darajah ..
started clean pants his on the the on the bike
haqat-ah.. ?il-mihim jama9u -uh la-h

belong him the important collected it to him

[NE 13]

In the above excerpts, there are situations in bold face where pronorrrinals are used
to establish emphasis through focusing on certain participants. They clearly have little to
do with accomplishing anaphoric relations. In the ﬁrst case, the pronoun [hummah] was
used after the word [?al—muhim] “the important thing” which has been frequently used to
emphasize the attention of the addressee to a speciﬁc event or character. The use of the
pronoun instead of the zero-pronoun, which was possible grammatically, took place after
digression. This had the effect of establishing focus by using the pronoun in a context
were a zero-anaphora would have been adequate to establish reference. If a zero-anaphora
had been used, the sentence would still be syntactically and semantically well formed. The
use of [hu] in the second occasion is also not to establish referentiality, but rather to
emphasize that he (the middle boy) did not provide any help. The second use of the
pronoun [hum] referring to the other boys who helped pick up the pears is also to establish
emphasis through repetition, that it was the other two boys (the small and the big boy) who
helped and not (the middle one). What has been said for the grammatical possibility of not
using an overt pronominal for the former case is true for the latter ones.

Mithun ( 1990) suggested that “when emphatic pronouns are used, they do not
represent entities that are already within the immediate focal consciousness of speaker and

hearer; in those cases their is no overt reference at all. Their primary function is to draw

79

already active referents back into focal consciousness after some disruption or

discontinuing” (p.375).

IV.B. 1.b. English zero-anaphora.

Though English zero-anaphors carry anaphoric function, they seem to play other
discourse roles as well. The majority, (80.16%), refer to major or secondary topicworthy
characters in the narratives. Only 19.84% of them refer to minor characters. They mostly
occur as demonstrated by the examples below, in an environment of vivid actions where a
sequence of rapidly occurring events are taking place. Probably the shorter the clauses, the

more vivid the narration.

{EX 1E}

A : O.K.! . and he's wearing a red handkerChief around his neck and it's ( ) down
.. uh .. in a sort of a triangle .. the kid was also wearing jeans.. the bike was uh
.. uh . a men's bike .. you know! .. and uh ..( ) just typical kid's bike .. you
know! .. and .. he rode up .. and it got off his bike .. then he .. he looked up ..
and uh.. a picked up .. uh one of the .. one of the - baskets full of pears .. and u
got back on his bike .. and it sets the basket .. uh ..down on his front . uh..
what would you call it?.... a hoop fender?

B: OK.

A: It's fender .. and a put it there .. and ti drove away....

NE 2]

The excerpt above is one of the few instances where zero-anaphors were employed
with some frequency. The episode that describes the bike boy’s maneuvers to steal the
basket of pears was the most vivid and thrilling episode through all narrations. Therefore,

it was not surprising that this episode attracted most cases of zero-anaphora. The last part

80

of the pear story, where the pear picker realized the absence of one of his baskets and at the
same time the three kids were passing by eating pears. At this episode the suspicion and
curiosity of the listener to know and the rush of the addressers to tell is reﬂected in the
choice of anaphoric elements. Zero-anaphors in English, although very few in comparison
to pronominals, occurred more frequently in these two episodes. The excerpt [1E] above
was taken from the former episode and excerpt [2E] was taken from the latter. I do not
think the 133 zero-anaphors, composing only 9.4% out of 1410 RCs occurring in English
narrative, are an adequate basis for a generalization about what determines their
distribution. However, they appear to be associated with the most vivid and exciting
events in the narratives, where attention to the rapidity of actions may have generated short

clauses, causing some pronouns to be deleted.

{EX 2E}
A: _ _ uh _ uh and the three kids walked down the road.. eating the pears
you know! .. and they go back to the Mexican. uh .. in the tree... he comes
down the ladder .. an.. o looks at the pears. and .. uh.. ¢ frowns. ( ) ..
scratching his head .. () what’s going on here..?.. then he looks .. there are
three kids .. walking by his tree .. looks at them .. shakes his head .. I don’t
Know what’s going on.. the kids walked away .. it’s the end of the story..

[NE 4]

The existence of zero-anaphora in contexts that describe rapid and vivid actions can
be related to the close relationship between the organization of thoughts in the narrators
minds and the instantaneous pressure they undergo during speech production, as Chafe
(1979) has suggested. More evidence was provided by Stevenston and Vitkovitch (1986)
during their experimental work on the role sentences with zero-anaphors and sentences

with explicit anaphors play in discourse. Sentences with zero—pronominals provide faster

 

81

response time than pronoun sentences. They also mentioned that their results give evidence
to the view “that the reference assignment is delayed rather than immediate” when a
pronoun is used to establish reference, because of the mental process it requires.
Particularly this holds true for the elliptical sentences, where such evidence supports the
view that the integration of subsequent information affects the speed of pronoun
assignment.

These results of experimental work by Stevenston and Vitkovitch (1986) provide
some support to our view that the assignment of zero-pronoun in English is pragmatically
manipulated to execute some discourse functions beside referentiality. Therefore, we can
claim that discourse functions play a role in facilitating the process of anaphoric relations

which was the underlying hypothesis the above researches were investigating.

IV.C. Referential Choices and Topicworthiness

As previously revealed, English and Arabic speakers’ choice of attenuated versus
fully speciﬁed coding devices is inﬂuenced by factors such as RG and the emergence of
discourse boundaries. In this phase of the study, attention will be given to the role
topicworthiness of referents plays in the choice among the coding devices available to
Arabic and English narrators.

Research on topicworthiness by discourse analysts such as Payne (1988) and
Lichtenberk (1988) is an indication of their commitment to expand the boundaries of
anaphora research beyond the limited traditional focus on the effect of the number of
clauses since previously mentioned referents. The importance of the role a participant plays
in a situation and the importance given to such a character in the narrative is not simply a
semantic matter; but rather a contextual one.

This importance of topicworthiness, as Payne (1988) called it, is a relative concept.
It can be determined only in relation to the roles other participants play in a speech

situation. Therefore, the topicworthiness of each character is determined by the importance

82

it has in the discourse, in relation to other characters. The topicworthiness of a referent is
reﬂected in the type of anaphoric coding devices the speakers choose and at the same time
be anaphorically coherent to the hearers.

Reinhart (1982) has referred to the topicworthiness of certain participants in
discourse in terms of how often they are “talked about”. Payne (1988) categorized
topicworthiness of discourse participants in two ways. There are inherently topicworthy
participants, e.g., humans are inherently more topical than non-human, animates are more
rcpical than inanimate. The other type of topicworthiness which is in the focus of this part
of the research is “context imported topicworthiness” (p.357). In all speech acts
interlocutors like addressers or addressees are highly tapicworthy. They are conscious of
themselves all the time. Interaction between them is taking place and therefore is available
for reference with almost no ambiguity. This status of speech interlocutors explains why
they are referred to universally with attenuated referential choices, e.g., ﬁrst and second
person pronouns, regardless of whether they were previously mentioned in the text or how
many times they were mentioned (Payne, 1988). I

Clancy (1980) recognized a decade ago that topicworthiness plays a role in RC.
She suggested in general terms that the status of a participant as a main character in
narration affects the addressers’ selection among coding devices available in their
languages. However, Clancy did not pursue any rigorous analysis to test her hypothesis or
go beyond that to provide examples and how much inﬂuence involved.

Topicworthiness, as I perceive it, is an active process which emerges out of the
interactive relationship among participants within a speech act. Categorization of
characters’ topicworthiness in Arabic and English narratives was decided on the basis of

the density of involvement each character had in the narratives, as discussed in II.E.

83

IV.C.l. Inﬂuence of topicworthiness on referential choices in English
and Arabic narratives

In this portion of the current study, I will analyze the effect that topicworthiness
hierarchy has on the choice among referential coding devices available to Arabic and
English speakers in their narrative repertoire.

The hierarchy of topicworthiness2 has exerted a signiﬁcant inﬂuence on the
distribution of referential choices in Arabic and English narratives. The most salient effect
of topicworthiness is the signiﬁcant correlation between the number of coding devices
involved in anaphoric relations and the hierarchy of topicworthiness. The more
topicworthy referents are, the higher number of anaphoric relations in which they are
expected to appear. Therefore, they attract a higher number of coding devices. Tables 4-4
and 4-5 provide a summary of results regarding the correlation between topicworthiness

and referential choices in Arabic and English.

IV.C.1.a. Topicworthiness and RC in Arabic narratives.

Major referents in Arabic narratives attracted 707 anaphoric coding devices, 47.1%
of the total number of RCs used in the whole body of data. Secondary topicworthy
characters were involved in 528 anaphoric operations, with 35.2% of the total number of
coding devices used in Arabic narratives. The most restricted distribution is that of the least
topicworthy characters. Minor referents composed 17.8 % of the anaphoric relations in
Arabic. Table 4—4 provides more details. There is a remarkable relation between the
number of referents and their distribution. Major referents in Arabic narratives are only
two characters, but they account for about half of the anaphoric relations and therefore a
high percentage of RCs in the narratives was attracted. Secondary and minor characters
included 13 of the 15 characters mentioned in the narrations. Regardless of their larger
number in comparison to major characters, secondary and minor characters shared the other

half of references and coding devices involved.

84

These ﬁndings are strong evidence that the topicworthiness of a character is a
signiﬁcant determiner of the frequency of its anaphoric relations. The following chapter
will provide an in depth analysis of the effect that topicworthiness has on referential gaps
and how powerful highly topical characters are in penetrating discourse boundaries.

While the most topicworthy characters are involved in the highest number of
anaphoric operations, they attract the fewest number of the strongest referential coding
devices (full NPs). The least t0picworthy characters in contrast are mostly referred to with
fully eXplicit referential choices. 61.42 % of reference involving minor topicworthy
referents and 38.25% of anaphoric relations involving secondary referents were established
with full noun phrases. Only 17.68% of referential relations involving major topicworthy

referents used the strongest RC of full NPs (Table 4-4).

Table 4-4 Distribution of different coding devices in Arabic narratives according to

 

their topicworthiness
RC major secondary mhror
# % # % # %

 

 

 

 

 

 

 

 

E 12—5 17.68 __"3_220 8.25 164 61.42"
PRO 257 36.35 163 30.87 65 24.34
ZERO _325 45.97 145 4—27.46 38 14.23
[_TO—TAL# _707 100 —528 100 267 100

 

 

 

 

 

% 47.1 35.2 17.8

 

These ﬁndings suggest the existence of a strong correlation between the choice
among the available coding devices in Arabic (full NPs, pronouns, zero—anaphora) and the
topicworthiness value of a referent. The more topicworthy a referent is, the higher number
of attenuated coding devices it attracts, and therefore the fewer strong coding devices it
attracts. Major characters in the Arabic narratives were refereed to with 582 attenuated

coding devices, 82.32% of the total number of coding devices used. Minor characters,

85

because of their low topicality, depended heavily on strong coding devices (61.42% of
RCs) to establish reference. Attenuated referential choices implemented only 38.57% of
the number of RCs used to establish reference to minor characters. Secondary topicworthy
referents, on the other hand, fell in the middle of the continuum of major and minor
referents in terms of attracting more attenuated coding devices and the number of RCs they
exerted, as Table 4—4 has summarized.

The following example from narrative nineteen clearly illustrates how sh-shab [the
young man], a major character in the Pear Story was, referred to by the narrator with a
continuous usage of attenuated coding devices (pronouns and zero-pronouns) over 16
clauses, since the ﬁrst time he was introduced—in line 2 of the excerpt below. This is an
indication of the continuity major topicworthy characters maintain over long stretches of
discourse. All RCs are represented in bold face.

Even the introduction of more than four referents (indicated by italics) did not
disturb the continuous usage of ﬁfteen attenuated referential devices or create any ambiguity
in interpretation. Minor topicworthy referents like [Id-darrajah] [the bike] and /?s-salah/
[the basket] were referred to with the strongest coding devices available. Regardless of the
introduction of a same gender referent as in lines ﬁve and nine, which theoretically is
expected to cause difﬁculty for interpretation (Fox 1986), reference with pronouns and zero
pronouns to the bike boy was maintained until the sequential pattern of referentiality was

closed at the line with a full NP reference.

{EX 4A}

A: Wa.. ﬁi ?amnaa?thalik.. wa huwa yaqttifu l—kimithraa.. Jaa?a
and in during that and he picking the pears came

waladun yarkabuu darrajah .. wa hatha sh-shaab .. yablughu
aboy riding abike and this young man reaches

9umur- uh taqriban thalathaat9ashara sannah .. wa . kaana yalbas o

86

age his approximately thirteen year and was wearing (he)
quba9atan .. wa 9alaa raqabat-ah khayt ..?aw kamaa yuqaal

a hat and on neck his a cloth or as said

karaﬁtah hammra?.. thumma jaa? a ?ilaa ash-shajarah .. wa nazhara
a tie red then came (he) to the tree and looked
ti ?ilaa ar—rajul .. thumma ?akhatha a a s-salah .. wa wadha9a-haa a
(he) at the man then took (he) the basket and put it (he)
9alaa ad-daraajah .. thuma nsarraf o ../l

on the bike then left (he)
B: // salah wahidah?
basket one
A: na9am sallah wahidah .. thuma ?akhatha a hathihii s-sallah .. wa

yes basket one then took (he) this the basket and

thahaba a .. wa r-rajul hathaa laa yadrii.. # wa baynmaa huwa ﬁ- t-

left (he) and the man this does not know and while he in

tarriq .. ?akhathat. ?akhathat d-darrajah tanharif min-h that l-

the way started started the bike to diverge from him to the

yarrriin wa thata sh-shimaal .. thumma wajahat-ah fatah saghirah
right and to the left then faced him a girl small

bi- darrajah ukhraa .. wa ﬁi tariqa-ha 9indama taqabalaa .. ?akhathat
with a bike another and in way her when they met took

a quba9atu-h .. thumma nazhara t! ?ilay-ha .. wa n-nharaf-at ?ad-

(she) hat his then looked (he) towards her and diverged the
darraajah 9an1-tariiq .. wa dharab-at ﬁihajar.. fa-ta9athar-at ?d-

bike from the road and hit at a stone and stumbled the

87

darrajah .. wa n-nqalaba sh-shab wa .. wa .. juriha a ﬁ rijli- hi

bike and flipped over the youngman and and was heart (he) on leg his
[NA 19]

Emﬂmnslaticm
A: and .. during that time .. while he was picking pears. a boy came riding a bike
.. and this young man is approximately thirteen years of age. and (he) was
wearing a hat .. and a cloth around his neck .. or a red tie as they say.. then (he)
came to the tree .. and looked at the man .. then (he) took the basket .. and put it
on the bike .. then (he) left ..//
B: [lone basket?
A: yes. one basket .. then (he) took this basket .. and left .. and this man did n’t
know.. # and while he was on the road .. he lost control over the bike towards
right and left .. then a young girl met him with another bike .. on her way when
they met .. (she) took his hat .. then (he) looked at her .. and the bike diverged
from the road .. and hit a stone .. therefor it lost control .. and ﬂipped over the

young man .. and (he) was wounded on his leg ..

Excerpt 5A below illustrates how extensive the use of NPs becomes to establish
reference to minor characters. [The goat man] saahib l-ma9iz (in bold face) was referred to
in three consecutive clauses with the strongest coding device possible, because of its weak

t0picality of the addressers mind and the minor role he played in the narrative.

{EX 5A}
A: ba9d hatha .. jaa?a tiﬂl saahib darrajah .. darrajah 9addiyah ..
after that came a boy owner a bike a bike regular
maa hiyii narriyah .. wa d-darrajah thii yukhayl lii ?akbar min-h _

is not it motorcycle and the bike this seems to me larger than him

88

li- ?anna-h yassuqa - haa.. //
because he rides it
B: // yajiid su9ubah ?
ﬁnds difﬁculty

A: na9am..! jaa?a ¢ min l-itijah .. ?illii thahab ?ilay-h saahib I-

yes came (he) from the direction that went to it owner of the
ma9iz .. ya9nii saahib l-ma9iz min ?itijah ..lakin jaa- ¢-h ba9ad
goat means owner of the goat from a side but came (he) him after

fattrah .. ba9ad maa- raahjahib l-ma9iz .. bi- fattrah .. jaa? (25 min

sometime after went owner of the goat with sometime came he from
l-ttijaah ..//
the side
B: // ?il-m9aakis ?
the opposite

A: ?il- m9aakis..

the opposite
[NA 13]

[Free translation]
A: .. after that .. a child came on a bike .. regular bike it’s not a motorcycle .. and
it seems to me that the bike is larger than him _ because he rides it ..//
B: // ﬁnds difﬁculty?
A: yes! .. (he) came from the side where the goat man had left ..
I mean .. the goat man came from that side .. but (he) came after a while .. after
the goat man had left .. after a while (he) came from the side //
B: // the opposite side ?

A: the opposite side ..

89

IV.C.l.b. Topicworthiness and referential choices in English narratives.

Similar to the behavior of referents in Arabic, in English there is a correlation
between referent topicworthiness and distribution of anaphoric coding devices. English
narratives have almost identical correlations; differences exist only in the density of such
relations. Minor topicworthy referents attract the highest number of the strongest coding
devices available. Minor characters employed 265 NPs, 76.59% of all coding devices to
establish reference. Only 23.41% of reference for minor topicworthy referents was
established by attenuated anaphoric devices, as summarized in Table 4-5. Secondary
t0picworthy characters manifested more attenuated RCs
than minor referents, but they are still far below the number of pronouns and zero-
pronouns used for major characters, 81.3% of the total number of RCs used.

The strong correlation between the topicworthiness of referents and the type of
referential references they take holds both in English and in Arabic . The more topicworthy
characters in Arabic and English narratives, the higher the number of attenuated coding
devices used to establish anaphoric relations, and therefore the fewer number of strong
coding devices they employ. This expresses the lack of need by addressers to refer to
major characters with strong coding devices continuously, because of their high topicality
in the minds of interlocutors, and therefore in the process of continuity (coherence). Minor
characters, however, need more elaborated focus of reference in order for them to be salient

in the minds of interlocutors, and therefore, avoid ambiguity.

90

Table 4-5 Frequencies of referential choices according to the topicworthiness of their
referents in English.

 

 

 

 

 

 

 

IRE major secondary minor
# % # % # %
TN UN W 18.7 W 51. 4 2 5 7 .5
—PRONOUN 466 68.63 145—18—70 70 0.2
—ZERO 86 12.67 36 (Ts—3'5 11 3.18
_TOTALS # 6—79 100 ——385 100 "346 100—
% 48.2 27.3 24.5

 

 

 

 

 

 

The following excerpt from English narrative 10 provides an excellent example of
how major topicworthy referents persist on employing attenuated anaphoric devices, and at
the same time maintain referentiality. “This ldd” is a major referent (in bold face) which
was introduced with a full NP at line one. From there on, a sequence of pronouns and
zero-pronouns has followed and persisted over seven clauses with no need to reintroduce
the referent with a stronger coding device (NP). Regardless of the introduction of another
referent in the story line with the same number and gender “the guy” (in italics). There is
an extremely interesting use of the return pop “be”, referring to the kid after an interruption
by the addressee to negotiate about who dragged the basket. The narrator was expected to
use a Full NP referring to the kid to solve any misinterpretation; However, he employed a
pronoun instead, as a return-pop over a gap of ﬁve clauses, with no ambiguity resulting. I
think that the addresser did so because of the high topicality that the referent possesses in

the narration that minor and in most times secondary characters can not have.

{EX 3E}
A: # _ _ __ _ and this KID comes by on a BIKE .. an .. and uhm _ he is coming
from the other way _ a .. and _ he looks up on the tree .. a sees the guy busy
picking he does n’t see him .. then he grabs the basket _ then it puts it on his
BIKE _ in front of the.. //

B: // WAIT a minute? the goat drags .. drags the basket?

91

A: NO .. NO .. the goat just looked at it .. It didn’t .. It didn’t actually
grab it .. uh _ I don’t know why .. uh _ so ohe picks up the basket .. a. a. and

.. ¢ puts it on his bike .. and it took off with this basket of pears.

[NE 10]

IV.D. Episode and Referential Choices

Several scholars have provided evidence that certain grammatical phenomena can be
clearly understood when studied in relation to their hierarchical place in discourse (Grosz
1977; Payne 1985, 1988; Fox 1987,1988). In relation to the study of anaphora in
discourse, Fox (1987) has studied the relation between anaphoric morpho—syntactic
markedness and discourse structure in Tagalog and English. Her analysis was based on
the assumption that;

There tends to be a correlation between the marked member of morpho—syntactic

oppositions involving particular tracking and the slot at the beginning of a discourse
unit (p.360).

This principle that Payne (1985) has discussed and was supported with evidence by
Fox (1987 ART) does not suggest that the single function marked members of a morpho-
syntactic opposition is to indicate the start of every discourse unit. Neither does it imply
that the beginning of every discourse unit ought to be occupied by the marked member of
the morpho-syntactic opposition. Instead the principle indicates that there is a tendency
cross-linguistically for statistical correlation between them.

Fox (1987) considered Full NP as the marked member of the assumed opposition
of the coding anaphoric devices Full NPs and pronouns (zero-anaphora was not included
because of their limited distribution in English expository prose). Full NPs were chosen as
the marked members of the opposition because of the phonological and morphological
complexity they involve in comparison to pronouns. I would add that pronouns are

statistically more frequent than noun phrases which makes them more predictable, and

 

92

therefore unmarked. “ The more predictable the information, the less coding it receives”
(Givon, 1983).

In this part of the current study attention will be directed towards the distribution of
the different anaphoric coding devices in correlation to episode boundaries in Arabic and
English narratives. Fox (1987 ART) found that the presence of full NPs is correlated with
the beginning of discourse units. Pronouns on the other hand, tend to be linked to noun-
initial positions of discourse units. This does not imply that pronouns never appear in

initial position of discourse units, but they do with low frequency.

Table 4-6 Frequency of NPs and attenuated anaphors at episode initial positions in
Arabic and English narratives.

 

 

 

 

Although this study is concerned with spoken narratives in Arabic and English,
results about the distribution of NPs and pronouns in episode initial position were
consistent with that of Fox (1987 ART) in regard to expository prose in English. There is
a strong tendency for full NP, the marked member of the coding devices opposition, to
occur at the beginning of episode. Table 4-6 indicates that 82.60% of episode initial slots
in Arabic were occupied by NPs to establish anaphoric relations (with in the initial two
clauses of the episode). Only 23 episodes 32.85% of the 69 episodes in Arabic started
with a pronoun or a zero— anaphor. On the other hand, 60 episode initial positions in
English 88.23% of the 68 episodes started with a NP. Only 22 episodes started with a
pronoun.

These ﬁndings support previous claims by Payne ( 1985), Von Stuttarheim and

Klein (1989), and Fox (1987) that discourse units tend to attract the marked members of

93

morpho-syntactic markedness opposition. So, Von Stuttarheim and Klein during their
study of anaphora in English and German suggested that;
the beginning of a new episode -or of a comparable segment in descriptive texts- is

marked by an explicit reference .......... in most cases also by a new mention of
dominant referential link in the text. ( 1989: 61)

Clancy (1980) has also reported that while pronouns are the most frequent form of
anaphora in spoken narratives, nouns are used at major structural junctures such as episode
initial positions, or at locations in text where change in t0picality was intended. However,
in narratives, episodes usually provide new characters to the narration, so the use of NPs to
carry that task is expected.

Following are episode initial slots with explicit use of NPs.

{EX 4E}

3. _ _ _ and uh.. um _ and this guy walks by _ leading a goat

4. _ __ _ _ and this KID comes by on a BIKE .. an .. and .. uhm _

he is coming from the other way ...............................

5. Right __ an _ uh _ uh the guy in the tree does not see ..........

6. .. and _ _ _ there is three other kids come up to him an _

uh .. 95 start helping him pick up the pears ..............................

[NE10]

{EX SE}

3. OK. so _ _ _ a a meanwhile .. this guy comes around with a

goat (laugh) .....................
4. .. uh .. the guy is up in the tree .. and this KID rides up

on a bike .....................................

 

94
5. _ __ _ _ and uh .. this girl comes the other way on a bike _
she’s got long brown pony tail .................................
[NE 9]
{EX 6A}
3. _ _ ba9d fattrah qaliilah .. jaa?a ttiffl saghiir ...........
“After a short period, a young boy came ..”
5. _ _ ba9d hathaa l-haadith bilahadhaat.. Kaan ﬁih ma jmu9at attfaal
qarribiin min-hu ...........
“A few moments after this accident, there was a group of children near him ....”
7. _ _ _ ?aysh saar ba9d kithah ? .. rajaa9 l-?awlaad jihat ?ashakhs ?illii ﬁ-
shaj arah ....................

“What happened after this? the boys returned towards the man on the tree

[NA 11]

This behavior of referential choices in Arabic and English provides more cross-
linguistic evidence for interaction between grammar and discourse as postulated by Grimes
(1975), Duncan (1982), Longacre (1983), Grosz (1977), Payne (1988), Fox (1987) and
Muthin (1990). Fox (1988) showed that the choice between a pronoun and a full noun
phrase in English is inﬂuenced by the hierarchical structure of the content of the text.
Studying the type of anaphors that cross discourse boundaries to establish anaphoric
relation, they seem to be highly restricted to a speciﬁc class of referents depending on the

role these referents play in the narrative.

IV.D.l. Topicworthiness, reference and sensitivity of episode
boundaries
The section above shed some light on the role episodes as discourse units play in

the selection of different coding devices in Arabic and English narratives. At this phase of

95

the present study attention will be focused on the role the contextual factor of
topicworthiness plays in establishing anaphoric relations across discourse boundaries, as
illustrated in 4- 1. Anaphoric reference within discourse boundaries will be mentioned for
the sake of comparison. My major interest here is to ﬁnd out what kind of characters have

their reference established across episode boundaries.

 

#x {NPi ....(proi(n)).. } #Y{.....proi....

 

# = episode beginnings. ( ) = optional occurrence.

Figure 4-1 Illustrates of the across episode anaphoric relations

Anaphoric relations across episode boundaries refer to the anaphoric relations where
anaphors in episode (Y) have their antecedents in the previous episode(X). Example 4A
above provides an illustration of anaphoric relations that go across episode

boundaries. In example 4A the pronoun [huwa/ [he] referred to a referent flash-shaab/ [the
young man] which was introduced with a full NP 15 clauses away in the previous episode.
Such referents that go beyond discourse boundaries to establish reference, ought to be
salient in the consciousness of interlocutors. Otherwise, they may be misinterpreted.
Results of studying anaphoric relations across episode boundaries in Arabic and English
narratives, as summarized in Tables 4-8 and 4-9, indicate that it is the unmarked and most
expected case for an anaphor to hold reference within the boundaries of the episode. There

are very few anaphoric relations across discourse boundaries.

96

IV.D.2. Topicworthiness and discourse boundaries in Arabic and
English narratives

Anaphoric relations in both languages (Arabic and English) usually are restricted to
the boundaries of episodes in narratives. Only a minority of anaphors establish reference
beyond the episode boundaries of their referents. During English narrations, narrators only
involved 41 anaphors, 2.9% of the total number of anaphors, in anaphoric relations that
carry reference across the boundaries of the episodes of their referents, as shown in Table
4-8. Similarly, Arabic narratives employed a small number of anaphors in anaphoric
relations across episode boundaries. As Table 4-9 reveals, only 7.7% of anaphors used in
Arabic narratives had their antecedents outside the boundaries of their episodes. English in
this regard seems to apply more restrictions on anaphoric relations across episodes. This
may be related to the rigid boundaries that English episodes have, especially in terms of the
long pauses that mark beginnings and ends of episodes. These pauses often go beyond
ﬁve seconds.

Anaphors that hold anaphoric relations across episode boundaries seem to be
signiﬁcantly inﬂuenced by the topicworthiness of their referents. 70.73% of these
anaphors had reference to the major characters mentioned in the narratives (the pear picker
and the bike boy). Secondary topicworthy characters were able to attract 24.39% of across
episode reference. Referents of minor t0picworthiness involved 2 anaphoric relations
across episode boundaries, composing 4.88% of cross episode anaphoric relations. The
relation between topicworthiness of a referent and its distribution and continuity has been
evident through my discussion of TW and RC. It will be more evident when I will discuss
the effect TW has on the persistency of anaphoric relations in terms of RG during the next

chapter.

97

Table 4-8 Distribution of referents according to their tOpicworthiness in and across
episodes in English. - .

 

Arabic narratives maintain the same scale for distribution of topicworthy referents
across episode boundaries. Referents with highest t0picworthiness were more likely to
achieve anaphoric relations across episode boundaries. Major characters were the referent
for 90 anaphors, 78.26% of reference across episode boundaries in Arabic. Secondary
topicworthy characters penetrated episode boundaries with only 25 anaphors, 21.74% of
the total number of anaphors with referents in a preceding episode. Minor topicworthy

referents were not topical enough to go beyond the boundaries of their episode to employ

anaphoric relations.
Table 4-9 Distribution of referents according to their topicworthiness in and across
episode in Arabic.

 

More interesting results emerged out of the three way comparison of the frequencies
of anaphors establishing reference across episode, in episode and in the whole data. There
seems to be striking consistency among these relations in the narratives from English and
Arabic. The frequency of major characters of narratives in both languages is much higher
in establishing across episode reference than in in-episode reference and in the whole body

of data in terms of the over all percentages. This comparison was reversed for minor

98

characters. They established minimum reference across episodes, 2 anaphors in English
and none in Arabic. Minor characters’ anaphoric relations within the episode boundaries
were signiﬁcantly higher, in English they were referents for in-episode 25.13% of all
anaphoric relations and 24.5% in the whole data; in Arabic 19.26% within episode and
17.8 in the Arabic narratives3.

Our results provide clear evidence for the major role contextual and conventional
factors play not only in detemrining the choice of coding devices to establish anaphoric
relations, as suggested by Fox (1986; 1987) and Payne (1988), but also in affecting the

type of anaphoric devices that penetrate discourse boundaries as discussed above.

IV.B. Conclusion

' This analysis and discussion of episode as an inﬂuencing discourse unit on
anaphoric relations in Arabic and English oral narrative discourse took two embedded
phases. The ﬁrst phase was a discussion of the way the boundaries of episodes in
narratives exert a signiﬁcant inﬂuence on the selection of different anaphoric coding devices
available in the narrators’ linguistic competence. The second phase was a more detailed
analysis of an underlying contextual factor that seems to affect the kind of referential
choices that perforate the rigid boundaries of episodes.

Two major points can be drawn from the discussion as presented above.
Topicworthiness as a contextual factor not only affects the distribution of anaphoric
relations in general, but more importantly topicworthiness is most powerful in deciding the
form of reference across discourse boundaries. Secondly, it seems that it is easier for
interlocutors to form anaphoric schemata within the boundaries of an episode in narratives,
rather than across episodes. This is why such occurrences are low in distribution and only
reference to referents with supporting contextual information or with salient function in

discourse employed attenuated coding devices and do not cause misinterpretation.

99

 

1This presumably free word order in Arabic has gone through historical change
exempliﬁed by the loss of case markings on most modern Arabic dialects. The consequence
was (I think) is a developing rigid word order to compensate for the vanishing case system.

2 This hierarchy has been discussed in details in II.C.l.

3 The reasons behind the relatively lower frequency of minor characters in Arabic
rather than English have been discussed above in the section about TW and RC in Arabic
and English narratives.

CHAPTER V
REFERENTIAL GAPS IN ARABIC AND ENGLISH NARRATIVES

V.A. Introduction

In this portion of the study, analysis will be focused on ﬁndings about referential
gap that anaphoric devices (anaphora and zero-anaphora) employed in term of their
reference in Arabic and English oral narratives. The analysis is not merely concerned with
linear aspects of referential gaps. It includes description of the hierarchical nature of
narratives. This is done by studying variance of RG within and across episodes. RG will
also be investigated in relation to topicworthiness.

Referential gap1 is a quantitative measure that has been used by several discourse
analysts and psycholinguists to account for the domain of topicality in regard to coherence
of anaphoric relations in discourse in terms of production and interpretation. Most studies
of RG aimed at understanding the capacity of human short memory to recall and interweave
anaphoric relations within different texts. The concept of RG (referential gap between an
anaphor and its referent) was initiated on the theoretical assumption that our cognitive
abilities as humans have certain limits. RD is assumed to be one way to measure the span
of this ability in regard to comprehension of anaphoric relations. Givon (1983) referred to
RD as the distance the most recently mentioned referent has to the following anaphoric
device with which the anaphoric relation is established. Clause as the smallest
communication unit was used as a measuring device for RD and has been widely used by
several researchers such as Clancy (1980), Givon (1983), Fox (1987) and Payne (1988).

Traditional theory as implemented in the work of Givon (1983;1985), Clancy
(1980) and Payne (1988) on anaphora in discourse have dealt only with RD in sequential
patterns. However, in the present analysis two kinds of RD (URD and RD) are

recognized, depending on the anaphoric pattern type - sequential or return-pop.2

100

101

V.B. Quantitative and Qualitative Aspects of Referential Gaps

In this section of the study of anaphora in Arabic and English oral narratives, focus
will not be merely on the quantitative aspect of RG but on the qualitative side as well. RG
provided the theory of anaphora with a quantitative measure of the cognitive ability people
have in terms of short term memory exempliﬁed by recognition of reference across text.

Qualitative consideration of RG in anaphora went beyond the mere consideration
for the number of clauses that separate the ﬁrstly mentioned NP referent and its pronominal
reference. For example, The inﬂuence of genre on anaphora has been dealt with great
caution. Fox (1988) studied RG in different genre types (expository, narrative and
conversational texts). Referential gap was largest in conversational texts and shortest in
expository written texts. Based on such ﬁndings, Fox concludes that the least constrained
the discourse type the largest referential gap it allows. She related this difference in RG to
major structures a text is composed of. “ Narratives are centered on action, the structural
units have actions as a central feature, whereas since expository prose is centered on claims
and support for claims, the structural units have static propositions as a central feature.”
(P.43)

In this study, I will add to these qualitative factors the inﬂuence the hierarchical
aspect of narrative discourse has on RG . Another qualitative factor is the effect of the
topicworthiness value of a referent on referential gaps.

Both quantitative and qualitative factors will be discussed. Both have been proven
to have signiﬁcant effects on the selection of different coding devices and on the

distribution of anaphoric patterns.

V.C. Referential Gaps in Arabic and English Oral Narratives
Table 5—1 provides a detailed account of RG for pronorrrinal and zero-pronominal
coding devices involved in anaphoric relations in Arabic and English oral narratives. Mean

RG values for each category have been calculated at the bottom of Table 5-1. The Table

102

provides the total number of incidents where anaphors are presented with their frequency of
occurrence with each RG value. The percentage of each RG value and frequency are also
indicated.

The mean length of URD in Arabic sequential anaphoric patterns was 3.68 clauses,
the mean length of URD in English was 4.09 clauses, as shown on the bottom of Table 5-
1. As presented in Table 5-1, we observe that the URD of one clause value is the most
frequent RG for sequentially distributed anaphoric relations. This means that a pronoun or
a zero-pronoun referred to an antecedent referent in the immediately prior clause(URD=1).
There is striking similarity between Arabic and English in terms of this distribution: in
Arabic this URD 25.87% of all sequential anaphoric relations, and in English it is 25.95%.
Furthermore, the majority of instances of URD (66.65% in Arabic and 69.01% in English)
fall in the range between 1 to 4 clauses. This may have to do with the limited capacity of
humans short term memory to accommodate referential gaps between the previously
mentioned full NP and its reference.

Speakers in both languages seem to have a close range within which topicality of a
speciﬁc referent is kept intact. The number of anaphors involved in sequential anaphoric
relations decreases as their URD increases in value, until they reach a level where the
frequency of anaphors equals zero. In Arabic there is no anaphoric relations beyond the
17th clause line (U RD: 17). English employed only 6.92% of anaphors to establish
reference beyond the 11111 clause line (U RD=1 1), slightly above Arabic which had 2.12%
of anaphors beyond eleven clause URD. Fox (1986) suggested that their is some abstract
cognitive constraints inﬂuencing the distribution of RG on a hierarchical mode.
Nevertheless, as I will discuss below, there is clear evidence for the involvement of

conventional factors as well.

103

Table 5-1 Frequencies and percentages of the different anaphoric patterns (sequential
and pop) in relation to the referential gaps for each pattern in Arabic and in
En gl i sh

ARABIC ENGLISH

# % # # %

Totals

893 100 101 100 729 100 94 100
”6898 3.68 cls 6.18 cls 4.09 cls 5.61 cls

 

104

It is not surprising to ﬁnd that shorter RG is much more frequent than longer RG.
There is substantial cross-linguistic and psycholinguistic evidence (Clancy, 1980; and
Lichenberk, 1988) that shorter RG is more available for quick and less ambiguous retrieval
by interlocutors. This provides an explanation for the high frequency of shorter URD one
to four clauses long in sequential patterns, and (three to seven) in pops as will be discussed

below.

V.C.l. Referential distance of return-pops in Arabic and English
narratives

Table 5-1 presents referential distance counts for anaphors involved in return-pop
anaphoric patterns, as well as counts (just discussed) for those involved in sequential
patterns. In III.B, I showed that the type of referential distance employed by retum-pops is
of a completely different nature from that employed in sequential patterns. Although
thematic continuity is maintained, structural continuity is absent. This feature of RD in
retum-pops has been either undetermined or ignored in many studies of anaphora,
especially in traditional paradigms of anaphora as in Clancy (1980), Givon (1983) and
Payne (1988). This issue was investigated in Chapter III.

Arabic and English narratives employed similar frequencies regarding RD within
return-pop situations. The highest frequency of RD was three clause. RD for pronorrrinals
and tel-pronominal involved in pops: in Arabic, 38 anaphors, 37.62% of all anaphoric pops,
had an RD of three (RD=3). In English, again with striking similarity between the two
languages, there were 32 anaphoric pops with an RD of three clauses, or 34.04% of all
pops.

The majority of pops in Arabic (56.43%) employed the shortest range of the scale
of three to four clause gap to establish return-pop anaphoric relations. The frequency of
pops becomes less as the number of clauses ascends on the scale of referential distance.

Referential distances of ﬁve to ten clauses compose 30.69% of the anaphors used as to

105

establish return-pop anaphoric relations. Long distance anaphoric pops, with RD of 11 to
48, compose only 12.88% of all cases of reference involving pops. Table 5-1 shows that
there are only two cases of reference beyond the 17 clause line in Arabic, merely 1.98% of
the number of anaphors involved in this form of referentiality.

Similarly, English pops were mostly distributed on the RG range from three to four
clause gap, composing 61.7% of the anaphoric return-pops used. The second highest
percentage of pops 30.69% were scattered on the range from ﬁve to ten clauses of RD.
Beyond the tenth clause line, there were only a very limited number of scattered pops that
had established referentiality. English narrators devoted only 7.61% of their pops to

execute longer distance references, with RD between ten and twenty-ﬁve clauses.

V.C.1.a. Cognitive capacity represented by ultimate referential distance.

Interpretations of the results in Table 5-1 about the RG of anaphors in Arabic
sequential and return-pop anaphoric relations seem to have a maximum ceiling of 17
clauses, regardless of the nature of RC involved.

There is also the tendency for referential gaps of both sequential and retum-pop
anaphors to acquire higher frequency at the low end of the RG scale. This may be related
to the cognitive restraints applied to RG. The longer RG employed the more difﬁcult it
became to connect anaphoric ties except under very restricted circumstances.

It seems that Arabic allows a longer gap for RD as in pops. However, these are
only exceptional occasions that were discussed previously, and they should be excluded
from the interpretation. English on the other hand executed a long distance RD of 25
clauses. Arabic exceeded the 17 clause line with an exceptional case. One of them resulted
in misinterpretation. Although English narratives allowed a wide range of distribution of
RG, there is a tendency for the majority of referential gaps to concentrate at the bottom of

the RD scale as illustrated in Table 5—1.

 

106

Givon (1983) suggested that a maximum length of 20 clauses be count as the RD
even when the actual RD was greater than that. Givon made that suggestion because he
believes that RD is potentially inﬁnite. Results of the present study in addition to others
presented in the RD literature from unrelated languages seem to argue against Givon’s
claim that referential distance can be inﬁnite.

I do not accept Givon’s account for RD (my RG) as having inﬁnite values, because
this claim would entail that humans’ cognitive capacity to trace referents in discourse is
inﬁnite, which I do not think is true. I believe Givon’s choice of an arbitrary maximum(20
clauses) is problematic for two reasons. First, If Givon is right and RD is inﬁnite, an RD
of 20 clauses does not (and cannot) account for inﬁnity, just by hypothesizing empirically
invalid values. Logically, any value in comparison to inﬁnity equals zero.

Secondly, there is evidence from four unrelated languages that the large majority of
anaphoric relations employ referential gap values between zero and ﬁve clauses. In Givon
(1983), there was only one anaphoric relation that involved 20 or more clauses of
referential distance in spoken English, while 94.72% of pronoun anaphors had an RD of
between 1 and 4 clauses. On the other hand, All zero-anaphors in the same data involved
RD of 1 or 2 clauses.3 In Japanese narratives, Clancy (1980) found that 97% of anaphoric
relations employed a ultimate referential distance range of zero to four clauses. Only
.001% of the number of anaphoric relations established reference on the 20 or more clause
referential gap range. 4 In Ta’aba’ita5, Linchtenberk (1988) reported that 97.2% of
anaphoric relations take place at the range of one to four clauses, and there are no anaphoric
relations beyond the eighth clause referential ran ge5

In the present study, 77.37% of Arabic sequential anaphors involved in a ultimate
referential distance range of zero to ﬁve clauses. No anaphoric relation took place beyond
the 17 clause RG line. Only 4.8% of anaphoric operations functioned beyond the ten
clause line in Arabic. As for return- pop anaphoric relations, 70.29% of them established

reference on the range from three to six clauses on the RD scale. Only two cases of return-

 

107

pops went beyond the referential distance 20 clause line. English sequential anaphoric
relations on the ultimate referential distance span of 0 to 5 clauses were almost identical in
percentage to the Arabic one (77.38%). 76.59% of English retum-pops employed
referential distances on the range of 3 to 6 clauses.

In this section, I demonstrated with cross-linguistic evidence from four unrelated
languages that referential gap is not hypothetically an inﬁnite phenomenon as Givon (1983)
would claim. But rather, a ﬁnite and a highly restricted parameter in the theory of anaphora
in discourse.

Based on the logic behind the results collected about RG length, I do not expect the
existence of numerous anaphoric relations beyond the 20 clause . If such relations
occurred, the prediction is probably that most of them will result in misinterpretation;
unless they are supported with strong contextual or pragmatic information, as in the case of

major t0picworthiness of referents in Arabic and English narratives.

V.C.2. Referential distance and episode boundaries in Arabic and
English
Several studies have revealed that discomse boundaries exert inﬂuence on the ﬂow

of information on texts. The structural aspects of cohesive patterns in texts are affected as
well. Previous discussion of referential choices (Chapter IV) and anaphoric patterns
(Chapter III) in Arabic and English narratives illustrated the role episodes play in
determining the selection of anaphoric coding devices and anaphoric relations. At this
phase, the current study focuses on how referential gaps of sequential and return-pop
anaphors react to episode boundaries in Arabic and English narratives. I will exarrrine this
relation by comparing the variance in the mean length of RG of both types of anaphors in
and across episodes. Variance among mean length of referential gaps was determined by

using the one way ANOVA tests. Such relations are summarized in Tables 5-2 and 5-3.

 

108

Table 5-2 Mean referential gaps of sequential and return-pop anaphors as used in and
across episode to establish reference

[Episode |MeanTJW Mean W |
[In |3_—T—__[_—T.97 7 3 5.41 1 |
I

[Across |7.32 25 I684 13

 

 

 

 

 

Tables 5-2 and 5-3 present the distribution of sequential and return-pop anaphors
along with referential gaps means for each anaphoric group, within and across episode
boundaries. Variance between the means of sequential anaphors in and across episode in
English falls on F. P. = .0001 range of signiﬁcance. The average ultimate referential
distance for in-episode sequential anaphor was 3.97 clauses. As for across episode URD,
it was 7.32 clauses. Variance in means of return-pop anaphors (RR = .3161) was not as
signiﬁcant as that of sequential anaphors, for some reasons that will be discussed latter.
In-episode mean RD for return-pop anaphors was 5.41. Across episode ones had an
average RD of 6.84 clauses. Table 5-3 provides more details regarding RGs of anaphors
in English narratives.

Variance analyses of mean referential gaps between sequential anaphors and retum-
pop anaphors in Arabic narratives were highly signiﬁcant . Variance of means among
sequential anaphors in relation to their reference with episode boundaries was on the
signiﬁcant scope of F. Prob. = .0000. Variance among RD means of return-pop
anaphors in correlation with episode boundaries resided on RR = .0060 level of
signiﬁcance. These results indicate that episode boundaries exert signiﬁcant inﬂuence on

determining the mean length of R08 in anaphoric relations.

109

Table 5—3 Mean ultimate referential gaps as represented in and across episodes in
Arabic narratives

 

 

 

 

 

[Episode [Mean URD(SEQ) # |MeaT7‘GF——|n r) _P )#
|In |3.609_——883 1 5.08 67 |
| Across | 6.521 115 I835 34 I

 

Mean URD for sequential anaphors within episode limits was 3.60 clauses. Across
episode anaphors scored a mean URD of 6.52 clauses. It was not surprising to ﬁnd
signiﬁcant differences among the mean ultimate referential distance of anaphors in and
across episode boundaries of narratives. Mean RD of retum-pops in Arabic and English
narratives, as shown in Table 5-2 and 5-3, is clearly affected by episodic boundaries, since
this type of reference in average establishes referentiality over a larger text than sequential
anaphors. Return-pop anaphors in Arabic employed a mean RD of 5.08 clauses ‘with in
episodes. Their across episode mean RD was 8.35 clauses.

Variance between mean referential distance of p0ps in and across episode
boundaries in English is not as statistically signiﬁcant in English as it is in Arabic. The
tendency for referential distance mean length for across episode pops to be larger than
within episode ones is salient. Across episode anaphoric pops had an RD mean of 6. 84
clauses. In episode pops employed an average RD of 5.41 clauses.

Episode boundaries, as we have discussed in the earlier chapter, work as a ﬁltration
mechanism where only referential relations with marked roles in the majority of times
penetrate these boundaries. Such boundaries implement different strength of rigidity from
one language to another. This issue has been also discussed by Fox (1986; 1988).

Comparing these results of variance among means of referential distance in Arabic
and English narratives, Arabic anaphoric pops had a longer RD (1.51) clauses than English
pops when establishing reference across episode boundaries. Such ﬁndings may provide
extra evidence for our earlier claim in Chapter III that episode boundaries in English are

more rigidly structured than Arabic. Across episode English sequential anaphors employed

110

a 0.80 clause of URD than Arabic sequential anaphors. This may be related to the very low
distribution of sequential anaphors establishing reference across episode (25 anaphors).
Arabic on the other hand established more than four times of the English sequential across
episode anaphors, 115. This treatment of reference across discourse units in English
narratives is additional evidence for the strict constraints episode boundaries employ on

across-episode referentiality.

V.C.3. Inﬂuence of tapicworthiness on referential gaps in Arabic and
English narratives

It was evident that topicworthiness functioned as a major contextual factor in
determining the choice among the different anaphoric coding devices in Arabic and English,
as I have discussed in the previous chapter. As it is the case for referential choices,
topicworthiness exerts signiﬁcant inﬂuence on determining the topicality of a referent in
terms of its continuity on the consciousness of interlocutors. Clancy (1980) mentioned that
main characters seem to have inﬂuence on anaphoric relations. However, Clancy did not
go beyond that step to come up with an explicit mechanism to account for such inﬂuence.

It was Payne (1988), as I have previously mentioned, who initiated a scale of
topicworthiness for referents during his study of Yugoua narratives.

During this part of the present research, attention will be concentrated on how the
t0picworthiness of a referent affects the continuity of its referentiality when attenuated
coding devices (pronouns and zero-pronouns) are used during narrations. Such a scale of
topicworthiness does not represent discrete values; instead it describes a continuum,
starting from the most topicworthy referent (major) to the least topicworthy (minor). Our
theory in this regard is that the more topicworthy is the referent, the longer referential gap it
will tolerate. Therefore, a stronger level of persistence it attained in the short-term memory

of the interlocutors.

lll

Means of ultimate referential distance for sequential anaphors in Arabic and English
narratives followed a strikingly similar pattern of representations on the topicworthiness
scale. As shown on Table 5-4 and 5-5, English and Arabic major characters implemented
the largest URD on the topicworthiness scale with a mean URD of 4.57 clauses for Arabic
major referents, and 4.89 clauses in English. Mean URD for anaphors referring to
secondary and minor referents in Arabic and English narratives were almost identical in
their interaction with topicworthiness. Sequential anaphors referring to secondary
topicworthy referents had a mean ultimate referential distance of 2.43 clauses in English
and 2.50 clauses in Arabic. As for sequential anaphors refening to minor referents, they

had an identical URD mean of 1.98 clauses.

Table 5-4 Mean referential gaps of sequential and return-pop pronouns in relation to
t0picworthiness of characters involved in Arabic spoken narratives

r

8C0" ar
DOT

 

Variance among the ultimate referential distance means in Arabic and English
sequential anaphors to referents with different levels of topicworthiness proved to be highly
signiﬁcant, F. P. = .0000. These results express the close correlation between the
topicworthiness of referents and the amount of persistence in the mind of interlocutors
during the narration process. Referents on the higher level of topicworthiness scale seemed
to have a longer presence on the short memories of the narrative partners and vice versa for

referents on the least topicworthy referents.

 

112

Table 5-5 Mean referential gaps of sequential and return-pop anaphors in relation to
topicworthiness of characters involved in Arabic spoken narratives

 

The mean referential distances of return-pop anaphors in Arabic and English, as
shown on Tables 5-4 and 5-5, do not seem to completely coincide with the scale of
topicworthiness as sequential anaphors do. This is not surprising, especially if we realize
that the variance among the distribution of pops in relation to the topicworthiness scale in
Arabic and English is not statistically signiﬁcant, In Arabic F. P.= .3896 ; in English F.
P. i—--- .7390. These results can be a consequence of the limited distribution they have in
comparison of sequential anaphors. Pops compose only 10.16% of the total number of
anaphors in Arabic, 1 pop to every 8.84 sequential anaphors. In English pops occupied
11.42% of anaphors of the total number of anaphors used, 1 pop to every 7.75 sequential

anaphor.

V.C.4. Inﬂuence of anaphoric coding devices on referential gaps in
Arabic and English narratives
During the previous discussion, the focus was on the effect that factors not
inherently or structurally a part of the anaphoric relation had on RGs such as
topicworthiness and episode. In this section, I will shed some light on how the choice of a
pronoun or a zero-pronoun could have affected their RGs when used by narrators in Arabic
and English narratives. I think it is appropriate to note that the opposition of pronouns and
zero—pronouns differs signiﬁcantly in relation to their distribution and markedness in Arabic
and English; as it has been previously discussed in Chapter IV, (IV.A.l).
Mean referential gap values for pronoun and zero-pronoun (sequential and p0p)

anaphors in Arabic and English are summarized in Table 5-6 and 5-7. English sequential

113

pronouns had an ultimate referential distant mean of 3.98 clauses. For sequential zero-
pronouns, 4.59 clauses was the mean length of the ultimate referential distance, 0.91
clauses longer than URD of sequential pronouns. Variance among the means fall in F.

Prob. = .1361 level of signiﬁcance. Table 5-6 shows these results.

Table 5-6 Mean of referential gaps of sequential and return-pop anaphors as used by
pronominals and zero- anaphors in English

RC |Mean WWI-Tr [Mean R‘_6")"""'_|D (P P #
Pro | 3.89 597 I 5.79 88 |
|

f Zero |4.59—130 I366 5

 

 

 

 

Variance among pronouns and zero-pronouns in Arabic was more signiﬁcant,
although follow the same pattern of relation between anaphoric coding devices and RGs.
Mean ultimate referential distance of sequential anaphoric pronouns in Arabic was 3.58
clauses. Sequential zero-anaphors persisted over a larger URD mean of 4.27 clauses.
Variance between URD means of the referential sequence coding devices in Arabic was F.

Prob. = .0024 on the scale of signiﬁcance. These relations are summarized on Table 5-7.

Table 5-7 Mean referential gaps of sequential and return-pop anaphors in relation to the
type of anaphora used in Arabic narratives

 

 

r0 | 3.58 484 | 6.19 57 |
ro |4._27 508 | 6.18 44 |

E: |Mean UR—W ( E IMean R'ﬁT'—_'|D P P) #
P _ _—
e

Mean RD of pronoun and zero-pronoun return-pop anaphors in English and in
Arabic behaved in different ways as presented in Tables 5-6 and 5-7. Such a difference in
RD means reveals how these coding devices are utilized in the grammatical system of each
language. Pronominal pops in English narratives scored a mean referential distance of 5.79

clauses. The mean RD of zero-pronominal pops was 3.66 clauses. The variance among

 

 

114

the mean referential distance of the coding devices was expectedly insigniﬁcant ( F. Prob.
= .4505) because of the very unequal distribution of return-pop zero-pronouns and
pronouns. Pronominal pops composed 94.68% of retum-pops in English narratives,
leaving zero pops with 5.32%. These results can be interpreted in relation to the discourse
role that zero-pronominals play in addition to the role of referentiality they play in English
narratives, as has been discussed in detail in IV.A.1.b.

Variance between the mean referential gaps of zero-pronominals and prononrinals
was insigniﬁcant (F. Prob.= .9923). Probably that was because of the almost identical
referential ranges zero and pronoun pops had in Arabic narratives. Mean RD of
pronominal pops was 6.19 clauses. Mean referential distance for zero-pops was 6.18

clauses, as shown in Table 5-8.

V.D. Referential Gaps and Continuity

Givon (1983) suggested that continuity of a referent in the minds of interlocutors in
a discourse situation is related to the coding devices used. A referent (a topic) is more
continuant when referred to with a more attenuated referential choice. When a full noun
phrase is used in reference, then it indicates referentiality is least continuous. Here, I
related continuity to RGs believing that the correlation between continuity and referential
choices as discussed by Givon was only a consequence of referential relations in terms of
RD between a referent and its anaphor. Longer distance reference beyond the average often
requires explicit referential choices (NPs). Anaphoric relations with in the average
referential gaps as we have discussed above, tend to be realized as pronouns and zero-
pronouns. Therefore, the shorter the referential gap is, the less effort required from

addressees to link anaphors to their referents and vice versa as illustrated in Figure 5—1.

Short RG Long RG

Most continuous Least continuous

Figure 5-1 The two way relationship between the length of ROS and persistence of
referent in the minds of interlocutors

As discussion about RG has revealed, the majority of anaphoric relations was
established with a short referential ran ge7, Therefor, preserves high continuity during the
narration process and also provides easier referent interpretation.

Nevertheless, as I have discussed earlier in this chapter, reference over a long
referential distance (in most cases, a lot above the average RGs, whether in Arabic or
English) exists in exceptional or low percentage cases. Extra long RGs may cause
ambiguity or improper referent interpretation. With the few cases where continuity was
preserved over long referential gap, there usually were some assisting elements that we
have discussed, such as topicworthiness or discourse boundaries. Topicworthiness of a
referent exerts signiﬁcant inﬂuence; more topicworthy referents in narratives can pursue
anaphoric relations using attenuated referential coding devices over long referential gaps
without sacriﬁcing the proper referential link (Tables 5-4, 5-7 for Arabic; 5-5, 5-6 for
English). Episode boundaries proved to be signiﬁcant as well; they pause restrictions on
long distant anaphoric relations across episode boundaries. Such constraints were more
noticed in English where only 4.62% of anaphoric relations operated across episode.
Anaphoric relations in Arabic although were abiding by the episode constraint, more
anaphoric relations take place across episode boundaries than in English, (13.55%).

The choice of a pronoun or a zero-pronoun in Arabic and English narratives seems
also to have some effect on referential gaps, especially in relation to sequential anaphoric
relations where the majority of references occurred. Zero-pronouns tend to attract
anaphoric relations with longer referential gaps than pronouns as summarized in Tables 5-6

and 5-7. Although these ﬁndings represent coding devices that do not necessarily have

116

parallel discourse functions (as discussed in IV.A.l) seem to differ from the scale of
topicality in relation to coding devices as proposed by Givon (1983)8 in the case of
sequential anaphoric relations in Arabic and English. Such contradictions between our
ﬁndings and that of Givon can be related to the different approaches we have towards
categorizing anaphoric patterns as discussed in III.A.2.b.

Our results indicate that it is not only the inherent nature of coding devices alone
that affects the RGs between referents and their anaphors, but rather other factors are
involved as well. This portion of our research about referential gaps has emphasized that a
complex multidimensional process of several factors such as topicworthiness, episode
boundaries, referential choice, referential gaps and grammatical functions interact to

inﬂuence the form and function of referential relations involved.

 

1The difference in using terms in this study in contrast to others is discussed in
II.C.2.a.

2 This issue has been discussed at length in chapter III.

 

 

 

3

RD d-anaphors pro-anaphors
ll % # %

l 115 98 336 79

2 2 02 46 ll

3 12

4 8 06

5 6

6 3

7 l

8 1

9 2

10 2

1 l 2

l2 2 04

13 1

14 l

15 l

16 0

l7 0

l8 0

1 9 0

20 1

total 117 100 423 100

 

 

 

 

 

A summary of referential distance values and distribution in spoken English. (In Givon (1983),
Table 11, p.353).

117

 

 

 

 

4

RD Ellipsis

it %
0 0 0
1 486 72
2-4 168 25
5-10 22 03
1 1-20 1 001
Total

677 100

 

 

 

 

Referential distance length in Japanese narratives. (Figure (2). Clancy (1980)).

5 To’aba’ita is a language spoken on Malaita in the southeast Solomon Islands. It is
a member of the Oceanic branch of the Austronesian family (Lichtenberk, 1988).

 

 

 

6

RD Pronouns
ll %

l 86 61.9
2 34 24.5
3 1 1 7.9
4 4 2.9
5 3 1.4
6 0 0
7 1 0.7
8 1 0.7
Total 139 100

 

 

 

 

Referential distance of pronouns in Ta’aba'ita. (Table 5 p. 336 Linchtenberk (1988)).

7 I am using short RD here in a relative sense, referring to the majority of
anaphoric relations taking Rd on the range of 0 to 4 of RD scale, as shown in detail in
Table 5-1.

3 More accessible
I Zero-anaphora
| Bound pronouns
l Independent pronouns
lFull noun phrases

Least accessible

CHAPTER VI
SUMMARY

Discourse analysis is an interdisciplinary ﬁeld deriving its resources from several
theoretical and methodological systems to meet the difﬁculties of analyzing linguistically
and culturally diverse data. Discourse analysts have established premises for their ﬁeld
based on research done in related disciplines. For instance, techniques for data collection
and transcription are borrowed from conversation analysis and the ethnography of
communication. Elicitation and techniques of tape recording are borrowed from
sociolinguistics, and so on. So, I see discourse analysis as adapting these theories, rather
than being tied to a single theory and a single way of thinking with respect to the diverse
nature of its resources. Therefore, in my study of anaphora in Arabic and English oral
narrative, I have made use of applicable previous techniques drawn from sociolinguistics,
cognitive science, linguistics and ethnography of speech, in my data collection, data
transcription, and analysis.

In my study of anaphora in Arabic and English oral narratives, I have developed a
combined multidimensional system that included, within its domain, a combination of
major approaches previously applied by Clancy (1980), Givon (1983), Fox (1987, 1988)
and Payne (1988). In this approach, I combined and modiﬁed Clancy‘s method of
studying cognitive constraints, Givon's technique of studying referential distance and
referential choices, Fox's patterning of anaphora and their distribution and approaches to
studying interaction between discourse units and anaphoric relations, and Payne's
methodology of measuring the inﬂuence of contextual factors. .

I will try to summarize my ﬁndings in relation to the following four major questions
that were raised during the early stages of this research. A question will be presented then

the related ﬁndings are presented.

118

119

no . ...1‘ 21211-1 o-.1‘11°1.‘. «.1 0.1.... 3'1or-. 12m " i us. '0 um
I ”In.” A”... H. H .171, -1- 1.1”” “I" .m- .' 1.”.
1° 1 1 . 2

In Chapter III, the discussion was directed at the type of anaphoric patterns used in
Arabic and English narrative discourse. I argued against some of the constraints1 proposed
by Fox (1987) on linguistic contexts in which return-pop anaphoric pattern could take
place.

In the present study, distribution of anaphoric patterns (sequential and return-pops)
was very similar in Arabic and English spoken narratives. This similarity may be
attributable to the factors that the Arabic and English narratives were produced by narrators
who were narrating the same sequence of events and under similar circumstances. ‘
Therefore, it is not surprising that they produced a similar percentage of anaphoric patterns.
Both, American and Arabic narrators were able to map the anaphoric relations presented in
the ﬁlm in the sequence they were presented. Later they are retrieved and produced
accordingly, and therefore, similar percentages of the two anaphoric patterns were used.
These ﬁndings lend support to Chafe’s (1979) hypothesis that there is close consistency
between the ﬂow of thought and the ﬂow of speech.

Difference in the number of patterns produced by Arabic and English narrators
(Arabic narrators produced 31 sequential patterns and 7 pops more than English narrators)
can be attributed to the different amount of data (length of narratives) presented by Arabic
and English narrators as shown in Chapter II, Tables 2-1 and 2-2. Another interpretation
can be attributed to the difference in the length of anaphoric patterns as produced in English
and Arabic narratives, as summarized in Table 3-2.

My ﬁndings in Chapter III provide support for the claim that episode, as a
discourse unit in narrative, implements certain constraints on the distribution of anaphoric
patterns across its boundaries. Hence, English was more sensitive to discourse

boundaries, but both English and Arabic recognize epiSode as a unit of discourse that plays

 

 

 

120

a major role in anaphora distribution. The unequal sensitivity of Arabic and English to
episode boundaries was related to the structural and functional apparatus of each patterns
and to the level of rigidity episode boundaries have in Arabic and in English narratives.
Although both Arabic and English narrators consistently used discourse markers of episode
boundaries, English narrators executed a more complex use of phonological and syntactic
discourse markers Arabic narrators used syntactic episode markers more extensively, as

discussed in III.B.4.d.

2 ‘11 111.. 0.0 M .01-11' 111.1 0.1.“ 110.111“ 'n°V1-._1._1' th‘or

 

Findings from Chapter IV provide evidence for the strong correlation between the
topicworthiness of referents and the type of referential references they take. This inﬂuence
holds both for English and Arabic spoken narratives. The more topicworthy characters in
Arabic and English narratives are, the higher the number of attenuated coding devices used
to establish anaphoric relations, and therefore the fewer number of strong coding devices
they require. This expresses the lack of need by addressers to refer to major characters
with strong coding devices continuously, because of their high topicality in the minds of
interlocutors, and therefore in the process of continuity (coherence). Minor characters,
however, need more elaborated focus of reference in order for them to be salient in the
minds of interlocutors, and therefore, avoid ambiguity.

Although, Arabic and English function similarly in this regard, there are some
differences that are related to the way different characters were treated according to their
topicworthiness as perceived by Arabic and English narrators. Within Arabic and English
spoken narratives, Arabic and English pronouns and zero- anaphors seem to have different
discourse functions besides referentiality in certain grammatical slots which played a major
role in the difference of selecting attenuated referential choices in Arabic and English

narratives.

 

 

121

These ﬁndings lend support to Payne’s (1988) claim about the inﬂuence of
topicworthiness in the selection of anaphoric coding devices in Yugoua. In a more general
sense, these ﬁndings also support Blass’s (1990) assumption that contextual factors exert
strong effect on the relations between parts of human discourse regardless of difference in
cultural and linguistic backgrounds of speakers.

In Chapter IV evidence is also presented for the assumption that the selection of an
anaphoric coding device is inﬂuenced by episode boundaries in Arabic and English oral
narratives. These ﬁndings support Fox’s (1988) claim that episode initial slots in
narratives are usually associated with full NPs, as the marked member of anaphoric coding
devices.

This analysis of episode as a discourse unit which inﬂuences anaphoric relations in
Arabic and English oral narrative discourse took two embedded phases. The ﬁrst phase
was a discussion of the way the boundaries of episodes in narratives exert a signiﬁcant
inﬂuence on the selection of different anaphoric coding devices available in the narrators’
linguistic competence. The second phase was a more detailed analysis of an underlying
contextual factor (topicworthiness) that seems to affect the kind of referential choices that
perforate the rigid boundaries of episodes.

Two major points can be drawn from the discussion as presented in Chapter III.
Topicworthiness as a contextual factor not only affects the distribution of anaphoric
relations in general, but more importantly topicworthiness is most powerful in deciding the
form of reference across discourse boundaries. Secondly, it seems that it is easier for
interlocutors to form anaphoric schemata within the boundaries of an episode in narratives,
rather than across episodes. This is why such occurrences are low in distribution and only
reference to highly topicworthy referents can employ attenuated coding devices and do not

result in misinterpretation.

 

 

122

.0 . ,[1‘1‘ '1'1q., we a '1_‘1.. .10 f _n-gct ,1,010; 1‘ gas. ,l‘, i" I

'1‘. .o. 1.10 . J.!~' ._ 0 9 All: .! DI 21.01 0.0 .150 134. . IIQI‘

f r h h n h ' ' ?

Our ﬁndings about referential gap as a measurement of cognitive constraints on
anaphoric relations in Arabic and English oral nanatives demonstrated that RG is sensitive
to contextual and discourse factors, regardless of linguistic and cultural backgrounds of the
narrators. Arabic and English speakers established anaphoric relations using similar ranges
of RG. The majority of sequential and return-pop anaphors established reference at the
lower range of the RG scale (1 to 4 clauses for sequential anaphors and 3-4 clauses for
return-pops).

It was not surprising to ﬁnd signiﬁcant differences among the mean referential gaps
of anaphors in and across episode boundaries of narratives. Mean RD of return-pops in
Arabic and English narratives is clearly affected by episodic boundaries, since this type of
reference in average establishes referentiality over a larger text than sequential anaphors.
Return-pop anaphors in Arabic employed a mean RD of 5.08 clauses with in episodes.
Their across episode mean RD was 8.35 clauses.

Variance between mean referential distance of pops in and across episode
boundaries in English is not as statistically signiﬁcant in English as it is in Arabic. There is
a tendency for referential distance mean length for across episode pops to be larger than
within episode ones. Across episode anaphoric pops had an RD mean of 6.84 clauses. In
episode pops employed an average RD of 5.41 clauses. Episode boundaries work as a
ﬁltration mechanism where, in the majority of times, only referential relations with marked
roles penetrate these boundaries. Such boundaries suggest a different strength of rigidity
from one language to another.

Comparing these results of variance among means of referential distance in Arabic
and English narratives, Arabic anaphoric pops had a longer RD ( 1.51) clauses than English

pops when establishing reference across episode boundaries. Such ﬁndings provided extra

123

evidence for our earlier claim in Chapter III that episode boundaries in English are more
rigidly structured than Arabic. Across episode English sequential anaphors employed a
0.80 clause of URD, less than Arabic sequential anaphors. This may be related to the very
low distribution of sequential anaphors establishing reference across episode (25
anaphors). Arabic on the other hand established more than four times of the English
sequential across episode anaphors, 115. This treatment of reference across discourse
units in English narratives is an additional evidence for the strict constraints episode
boundaries employ on across-episode referentiality.

Means of ultimate referential distance for sequential anaphors in Arabic and English
narratives followed a strikingly similar pattern of representations on the topicworthiness
scale. As shown on Tables 5-5 and 5-6, English and Arabic major characters implemented
the largest URD on the topicworthiness scale with a mean URD of 4.57 clauses for Arabic
major topicworthy referents, and 4.89 clauses in English. Mean URD for anaphors
refening to secondary and minor referents in Arabic and English narratives were almost
identical in their interaction with topicworthiness. Variance among the ultimate referential
distance means in Arabic and English sequential anaphors to referents with different levels
of topicworthiness proved to be highly signiﬁcant, F. P. = .0000. These results express
the close correlation between the topicworthiness of referents and the amount of persistence
in the mind of interlocutors during the narration process. Referents on the higher level of
topicworthiness scale seemed to have a longer presence on the short memories of the
narrative partners and vice versa for referents on the least t0picworthy referents.

The mean referential distances of return-pop anaphors in Arabic and English, as
shown on Tables 5-5 and 5-6, do not seem to completely coincide with the scale of
topicworthiness as sequential anaphors do. This is not surprising, especially if we realize
that the variance among the distribution of pops in relation to the topicworthiness scale in
Arabic and English is not statistically signiﬁcant (in Arabic F. P.= .3896 ; in English F.

P. = .7390). These results could be a consequence of the limited distribution retum-pops

 

124

have in comparison to sequential anaphors. Pops compose only 10.16% of the total
number of anaphors in Arabic, one pop to every 8.84 sequential anaphors. In English
pops occupied 11.42% of the total number of anaphors used, one pop to every 7.75
sequential anaphor.

Mean RD of pronoun and zero-pronoun return-pop anaphors in English and in
Arabic behaved in different ways as presented in Tables 5-7 and 5-8. Such a difference in
RD means reveals how these coding devices are utilized in the grammatical system of each
language. The variance among the mean referential distance of the coding devices was
expectedly insigniﬁcant ( F. Prob. = .4505) because of very unequal distribution of
return-pop zero-pronouns and pronouns. Pronominal pops composed 94.68% of retum-
pops in English narratives, leaving zero pops with 5.32%. These results can be interpreted
in relation to the discourse role that zero-pronominals play in addition to the role of

referentiality they play in English narratives, as has been discussed in detail in IV.A.1.b.

. .- ..A_ . ... ”dy‘rfu . .-'., 1.1.1 1- -..
l . 1' 9

Since both Arabic and English spoken narratives were based on the same
communicative task and under similarly controlled situations (age, educational level,
linguistic variety, and genre), they were very similar ﬁndings in regard to the way
anaphoric relations were established in Arabic and English. Such ﬁndings seem to be
consistent with the claim raised by Chafe ( 1976) that there is consistency between the ﬂow
of thought and the ﬂow of speech.

Our study of anaphoric relations in Arabic and English spoken narratives supports
the assumptions made by Payne (1988) and Blass (1990) that contextual factors exert
strong effect on the relations among parts of discourse. This effect was illustrated through
the way referents with different t0picworthiness values have interacted with anaphoric

relations in terms of referential gap and referential choices. In regard to Fox’s (1988)

 

125

assumptions about the role discourse boundaries play in determining the selection of
members of morpho-syntactic markedness oppositions, my ﬁndings revealed that it is not
only discourse boundaries that determine the selection of marked members of the morpho-
syntactic anaphoric devices, but rather both discourse boundaries (represented in narratives
by episodes) and topicworthiness of referents involved in the narration process as a
contextual factor. This interaction between topicworthiness of referents and episode
boundaries plays signiﬁcant roles in determining the kind of anaphoric patterns involved,
the average referential gap exerted and the type of referential coding devices being selected.
A major difference I have with Givon’s (1983) approach to anaphora is related to
his claim that RG as an anaphoric phenomena is inﬁnite in its value. I demonstrated with
cross-linguistic evidence from four unrelated languages (V .C.1.a) that referential gap as I
have deﬁned it is not hypothetically an inﬁnite phenomenon as Givon (1983) would claims,
but rather, a ﬁnite and a highly restricted parameter in the theory of anaphora in discourse

that is inﬂuenced by numerous factors; I discussed some of them in this study.

 

1 The two constraints I argued against were (a) the anaphoric relations in the
popped over text should not involve same gender referents, and (b) the popped over
material should not involve complex structures.

BIBLIOGRAPHY

 

BIBLIOGRAPHY

Abdo,Dawud. 1973. Abhaathen ﬁi lg-ghahg-g‘il-9mbiyyeh. Beirut: MaktabatLubnaan.

Akinnaso, RN. 1981. "The Consequence of Literacy in Pragmatic and Theoretical
Perspectives." An r olo n E uca ion art rl , 12, 163-200.

Akinnaso, EN. 1982. "On the differences between spoken and written language."
ngage and Speech, 25, 97-125.

Al-Fihri, A. A. 1982. Alli ni w lu- h hl- i -n m hi'
Q'leliyeh, Ad-daaru l-baydhaa?: Daar Tubgaal.

As-saamiraa?ii, I. 1966. Al-ﬁ211g;zamaanghg we ?ebniyetghg. Baghdad: Matba9at ?al-

9anni.
Ayoub. A. 1957. Duaammdduahfmmahulﬂmhu Cairo: The Anglo-Egyptian
Library.

Bakir, M. J. 1979. Aspeets ef £21 at; se Sﬂetgee 1n Arabie: A Study 1n Werd met

WPK D. dissertation. Indiana University, Bloorrrington:
Indiana.

Bauman, Richard. 1986. $1955, Perfermenee, and Event; Centextual Smdies ef Qeel
Narrative. Cambridge: Cambridge University Press.

Beaman, K. 1984. "Coordination and subordination revisited: syntactic complexity in
spoken and written narrative discourse." In D. Tannen (ed.), oh no in ken
and Written Diseoursg (PP. 45-80). Norwood, New Jersey: Ablex.

Biber, D. 1986. "Spoken and written dimensions in English: resolving the contradictory
ﬁndings." Langgage, 62, 384-414.

Blass, R. 1990. Relevence Relations in Discourse. Cambridge: Cambridge University
Press.

Bloch, A. A. 1986. Studies in Arabic Syntax and Semantics. Otto Harrassowitz:
Wiesbaden.

Brown, Gillian and George Yule. 1983. Discourse Analysis. New York: Cambridge
University Press.

Caden, G. 1982. "Backwards anaphora in discourse context". Journal of Linguistics 18,
361- 397.

126

 

 

 

 

127

Chafe, W. L. 1979. "The ﬂow of thought and the ﬂow of language". In T. Givon (Ed)
121W, New York: Academic Press.

Chafe, W. L. (ed.). 1980. ' - ' '
W211. Norwood, New Jersey: Ablex.

Chafe, W. L. 1982. "Integration and involvement in speaking, writing, and oral
literature." In D. Tannen (ed)
Literacy (pp. 35- 53). Norwood, New Jersey: Ablex.

Chomsky, Noam. 1957. Wme Hague: Mouton.

Chomsky, Noam. 1980. "On binding". Wm 11, 1-46.

Clancy, P. M. 1980. " Referential Choices In English and Japanese narratives". In W.L.
Chafe (Ed.), (pp. 127-202), Norwood, New Jersey: Ablex.

Coulthard, Malcolm. 1985. WESSCX Longman Group
Ltd.

Cornish, F. 1989. "Review article of Discourse Structure and Anaphora: Written or
Conversational English". Liam 79, 229- 243.

Delisle, G. A. 1973. “Discourse and backward pronominalization’. [gdjaga unjygm'm
UM. Bloomington: Indiana.

Duranti, Alessandro. 1986. "The audience as co-author: an introduction." Text, 6(3),
239-247.

Du Bois, J. W. 1980. "Beyond deﬁniteness: the trace of identity in discourse". In W.
Chafe (Ed.), (pp. 203-273), Norwood, New Jersey: Ablex.

Filmore, C. J. 1975. W15. Bloomington: Indiana University
Club.

Fox, B. A. 1987. '
En glish. Cambridge: Cambridge University Press.

Fox, B. A. 1987. “Local patterns and General principles in cognitive: Anaphora in
Written and conversational English”. Text 6 (1), 25-51.

Fox, B. A. 1988. “Morpho-syntactic markedness and discourse structure”. mm
Emgmatjes 11, 359-375.

Givon, T. (ed.). 1983. W. Amsterdam: Benjamins.

Goodwin, C. 1986. “Audience diversity, participation, and interpretation.” jlfext, 63, 282-
316.

Grosz,B. 1977. l'i'u"1o.101 an! '0 or. '1... 0" nc‘r -0n°' SRI
Technical Notes 151, Menlo Park, California.

Halliday, M.A.K. and R. Hassan. 1976. Cohesion in English. London: Longman.

 

128

Hinds, J. 1977. “Paragraph structure and pronominalization”. Eager}; in [,jngm'sﬁgs
10.77-99.

Hirst, W. andG. Brill. 1980. “Contextual aspects of pronoun assignment”. lenmaLef
leaninunﬂerhaLEehaxim 19: 168- 175

Hopper, Paul J. 1988. "Discourse analysis: grammar and critical theory in the 1980s."
(pp. 18-24).

Horowitz, Rosalind. 1987. "Rhetorical structure in discourse processing. " In R.

Horowitz and S J Samuel (edsu) WW (PP
117- 155). New York: Academic Press.

Houghton, D. and M. Hoey. 1982. “Linguistics and written discourse: contrastive

rhetoriCS” AnnuaLRsximtAnuliesLinsuistics 3: 2- 22

Hymcs. Dell. 1974. EgundaticnscLSmiclineuisticsLAn WW
Philadelphia: University of Pennsylvania Press.

Karttunen, L. 1968. “Coreferential discourse”. LSAAmuelMeeﬁng, New York.

Kuno, S. 1975. “Three perspectives in the functional approach to syntax”. Bapemfrem
W167336. Chicago: Chicago Linguistic Society.

Labov, W. 1972. “The transformation of experience in narrative syntax”. In W. Labov

(Bd.) W. Philadelphia: University of Pennsylvania.

Labov, W. 1980. Locating Language in Time and Space. New York: Academic Press.

Langacker, R. W. 1969. “On prominalization and the chain of command”. In D. A.
Reinbel and 8. Shane (Eds. ) Wlmdw. Englewood
Cliffs, New Jersey: Prentice- Hall.

Levinson, S. 1983. WCambridge: Cambridge University Press.

Li, C. and S. Thompson. 1979. “ Third person pronouns and zero-anaphora in Chinese

discourse”. W 12, 311-335.

Lichtenberk, F. 1988. “Nominal anaphora in To’aba’ita”. Wage, 12(2): x-
344.
Longacre, Robert. 1976. "Mystery particles and afﬁxes. " P 1
April, 23— 25.

Longacre, Robert. 1983. The Grammar of Discourse. New York: Plenum Press.

Mandler, J. and N. Johnson. 1977. “Remembrance of things parsed: strong structure

and recall”. CegnjtilLEsychQLng 9,111-151.
Milroy, L. 1987. Wage. Oxford: Basil Blackwell.

Mithun, M. 1990. “Third person reference and the function of pronouns in Central Pomo

Speech” IntematmnallcumaLcLAmLcanDngmstrcsﬁ (3). 361376

129

O’Connor, M. C. 1990. “Third person reference rn Northern Pomo conversation: the
indexing of discourse genre and social relations”. WWW
linguistics 56 (3), 377-409.

Olson, D. 1977. “From utterance to text: the bias of language 1n speech and writing.”

limrdﬂusationauicxim. 47. 257- 281

Ouaouicha, D. 1986. -. _ .
WWW Ph D dissertation The University of

Texas: Austin, Texas.

 

Payne, Thomas E. 1988. “Referential distance and discourse structure in Yaguba”.

W 12(2) 245- 392

Piaget. J. 1962. WWWMW Routlcdge and
Kegan Paul.

Prince. G. 1982. W The Hague:

Mouton.

Reinhart. T. 1983 ArranboraandﬁsmantrslmmtmnsChicagm University of
. Chicago.

Revero, M. 1980. “On left-dislocation and topicalization in Spanish”. Linguisjjelnguim.

Sacks, H. and E. Schegloff. 1974. “Opening up closings.” In R. Turner (ed.),
Ethmmsthodologx (Pp. 233-264). Harmondsworth and Baltimore: Penguin.

Sanford, A. J. and S. C. Garrod. 1981. MW. Chichester:
Wiley.

Shillcock, R. 1982. “The on-line resolution of pronorrrinal anaphora”. Languagejnd
Stretch. 385-398.

Stevenson, R. and M. Vitkhovitch. 1980. “The comprehension of anaphoric relations”.

Winch. 29:
Stubbs, Michael. 1983. W. Chicago: University of Chicago Press.

Tannen, Deborah. 1980. “A comparative analysis of oral narrative strategies.” In W.
Chafe (ed.), mmp. 51-87). Norwood, New Jersey: Ablex.

Tannen, Deborah (ed..) 1982. Analyzing Discourse: Text and Talk. We
Wags Washington. D C
Georgetown University Press.

Tannen. D. 1984. (Ed) WNW Jersey:
Ablex.

Tannen, Deborah. 1989. ' ' i 1
Cam. nCambridge: Cambridge University Press.

Thalji, A. I. 1986. “Marked vs. unmarked structures in modern written Arabic”. A];
whim 19: 109-126.

 

130

Tyler, L. K. and W. Marslen-Wilson. 1982. “The resolution of discourse anaphora:
Some on-line studies”. Iext 2(1-3): 263-291.

van Dijk,T.A. 1977. W. London: Longman.
van Dijk, T..A (ed..) 1985. W. New York: Academic

Press.

Varonis, E. M. and S. Gass. 1985. “Non-native/non-native conversations: negotiation

of meaning”. Applicdhnenistiss. 6(1): 71-90.

von Stutterheim, C. 1989. “Referential movement in descriptive and narrative discourse’.

In R Dietrich and G F Graumann (Eds) WWW
Elsevier Science Publishers: B. V. North Holland.

Vygotsky, L. 1978. 'n ' ' ° v
MCambridge, Massachusetts: Harvard University Press.

Watkins, L. 1990. “Noun phrase verses zero in Kiowa discourse. Imematigmeﬂeumel
Wags, 56(3). 410-426.

 

APPENDIX

 

APPENDIX

English Narrative (2)

A:

E???

.3399}???

2’52”}???

All right ! .. well .. I'll tell you about it .. barry.. there was this little
about uh .. about uh.. OK. it opened with a. a . roaster .. uhm .. really ..
uhm ---
was this .. was this a real roaster and ..?
we didn't see the roaster .. we only heard the sound.. of the roaster to give the
indication that .. it was produced in the morning .. but it..//
0. K..
but it was bright sun light..//
// O.k. is this alive action on the ﬁlm?
It's all live action .. Yeah..! .. and -- O.K. -- there was a man with a. . a satchel
on his belt .. O.K. he was on a ladder .. a wooden ladder .. uh .. and uh .. a fruit
dropped down and .. uh ..//
// from where?
Dropped down from a tree .../l
// OK.
that he was climbing ( ) ﬁrst of all, they turn out to be pears .. and
apparently, .. that he was gathering from for quite some time. because there
were two or three bushels of them.. right under the base of the tree and .. uh --
lets see --- oh wait .. wait - that .. lets see .. one .. one man walked up to
him and .. had a goat on a leash .. a brown and white goat .. uh .. ( ) I

don't I don't actually know .. what this guy with the .. with the goat was

131

 

 

 

52°

2’99???

132

doing because he just kind of walked up and .. you know! .. looked up at
the tree...//

// Did he look up at the guy?
Yeah! an .. and just walked away the goat trailed after him .. wherever he
goes ( ) I'm not sure -
All right !
but .. uh --- lets see .. What els -- the .. uh .. and after .. and after the the goat
person went away .. uh .. a . a little boy .. rode upon a bike .. he was
wearing ah .. uh.. a kind of a ﬂoppy shappy hat .. uh .. almost like a
cowboy hat .. uh ..but .. but uh .. not as stiff ..
O.k. [ laughter]
OK... and he's wearing a red handkerchief around his neck and it's ( ) down .. uh
.. in a sort of a triangle .. the kid was also wearing jeans.. the bike was uh .. uh
. a men's bike .. you know .. and uh ..( ) just typical kid's bike .. you know! ..
and .. he rode up and .. got off his bike .. then he .. he looked up .. and uh.. he
picked up .. uh one of the .. one of the - baskets full of pears .. and got back on
his bike .. and sets the basket .. uh ..down on his front . uh.. what would you call
it ? .. a hoop fender?
O.K.
It's fender .. and put it there .. and drove away....
Was there somebody to help him .. or something?
No! .. he just held it with his hands .. uhm .. the whole thing took place .. or uh
..setting ( ) although there were more trees and the road was .. uh .. I don't
actually see how he steers on the road .. I imagine .. it was a drive way I think it
was a dirt path...
O.K. .. O.K. .. is that essentially the whole story?
NO! .. no..

P? 2’99???

.3???

PP

PP???”

133

Oh . a.ah! O.K.! ..there is more!
There is more //

// Lots more. alright O.K.!
Uh huh any way, so he is riding on the road .. and ..uh .. uh ..It is like
a typical case of a boy meets a girl .. uhm ..he he hits a rock with his bike ..
and falls down .. uh .. uh [laughter] male f female relationship all .. in a
little piece of ﬁlm but .. uh .uhm [cough] .. lets see .. aaah -- this little girl
comes riding from the other direction .. he looked at her .. and his hat blows
off .. on the road .. as she rides by .. there maybe that .. she reaches out ..
pulls ..( )...
You couldn't tell?
I wasn't sure .. I couldn't tell .. I must be imagining it .. and uhm.. the kid hits ..
after he hits this big STONE .. looks like .. uhm .. what do you call ? ..What do
you call a volcanic stone ( )?
( )
Looks like a big chunk of [a black ....( ) }

{ black thing( ) in shape. }

Yeah! .. uhm .. and uh .. there he falls down .. and .. he . he pulls up his pants ..
like .. pulled up his sock .. and uh ..
Did he hurt his leg ?
Well ! apparently .. because he rubs it .. there is no wounds .. or any thing wrong
with his leg .. it must ache a little bit..
[( )1
[( )] .. but of course all the pears were scattered on the road ..

Yeah!

 

A:

.3???

P?

?PP?P???PP??FP

PP???

134

an .. uh .. and .. uh uh it turns out .. standing near by are ..a.. aa three .. of ..uh
.. young boys ..uhm .. and two of them walk out into the road .. and begin
helping him .. pick up the pears .. put them in the basket//

//Any thing that was said during the course of any of this?
No!.. no words .. no words although there was .. one whistle.
OK.
But I've not come to that yet.!
There's sound track .. but . but there is no natural voice.
Right! you hear the wheel rolling...
Yeah!
.. and ..uh and you hear the pears falling .. and nobody is .. nobody is saying
anything.
It would be terriﬁc ( ) . [laughing] go on .. go on!
( ) [laughing] and .. uh ..( ) and uh ..o.k. so they get all the .. all the pears ..
and put them back into the basket..
Did the girl keep riding away?
Yeah! you'll never see the girl any way..
Alright!
-- uh .. but apparently the others saw what she 'd done to him .. took piety of him
.. whatever .. maybe .. ( ) uh ..uh ..but there is this other kid who is dressed on
like a.a.. a blue dress shirt with a loose collar.. instead of button down all buttons
{( )..}
{Oh . is this the third ] of the petrio .. is this..//
This is the third petrio .. uh .. and he's just standing there .. he's got .. uh he's
got [laughing] .uh.a peddle ball thing in his back pocket .. he takes it ..and

plays. you hear the ball going .. tick .. tick.. [laughing] and he misses up .. you

EFF???

PP

?PE’EPP???

135

know .. and goes on .. and starts over again .. tick .. tick .. tick... while
everything is picking the pears ..when they got every thing picked up...
OK.
..and they relocated it back on the bike -- and ..uh.. and ..uh -- uh..
How old are these kids?
Oh.. they're .. uh .. I'd say twelve .. or so ..
All of them 're about the same age?
All of them about the same age... yeah! ( ) .. the girl is the same age .. more or
less. uhm .. --uhm ---
He starts riding right..?
No.! he does n't start riding. he walks away..//
//walkin g his bike?
walking his bike. and .. I just now... was trying to remember .. what happened
.. with the pears .. because it seems to me .. it would be hard to to get the .. uh
.. the from him.. to have been carrying the pears. because he's also... he had
to guide his bike. uhm
Did he set it on the seat of the bike or something?
It could be .. it could be .. I can't remember exactly .. what he did -- uhm --
any way ..he's walking away in the other direction.. down the path .. O.K.?
.. the .. the .. ya.. the guy the kid who took the pears in the ﬁrst place is
walking .. uhm ..uhm .. walking across .. what .. looks pretty ( ) .. looks pretty
( ) .. low .. uhm .. low kind of shrubs that grow .. all over the places .
Yeah! { ( ) }
{ ( ) } .. uh .. and .. uh .. uh .. but on the road .. the kid with the
peddle ball .. toy .. uhm.. sees. sees .. the other kid's .. the kid's hat .. he

picks it up .. and looks back and whistles .. to get his attention..( )

P???

99????

Pl

9?

99????

136

and .. uh .. and ..uh .. then he walks back .. and he gives it to him .. and he
comes back .. carrying three pears .. uh--
The kid who picked up the hat?
Yeah!
He gets three pears and exchanges with him the hat back/l
”Because .....

// Is n't interesting to see that exchange with in the scene and see the kid coming
back with pears and not with the hat?
Yes! .. Right!
Alright!
O.k.! and .. uh.. --- and.. O.k. .. so the little girl ..and the little boy
disappeared we will never see them again nobody knows what happens
to them ( ) ..w'll never see them.
O.K. its now with the.. the three boys?
The three boys and they walk by ---
So.. they're walking back .. down the path.. the direction that the kid on the
bike came from?
Right! yeah! ..so .. in other words walking towards the man ..who's picking pears
in the tree .. Right! .. and - - one of them bits on his pear and we don't know
..if he ﬁnishes. his pear or throws it ..( ) and that ( ) because -- (laugh) they
walk past the guy .. who's picking pears .. and he's now descending from his
ladder .. and .. he's regarding .. the .. the bushel baskets. uhm...
How many bushel baskets are they?
Well! .. there is only one missing .. and he's displeased...
Oh ...also he did..//

// and he's puzzled.
Oh. he didn't see the other kid carrying the....//

 

.99

PP

?FP?PP??PP??FP?PPP?PP

137

// apparently not... I didn't think
that there was anyway ..he would've missed it .. but ..uhm..( )
alright!
Uh .. but uh .. there was..{ ( ) }
[ ( ) } .. you can't see things with your
eyes..unless your eyes are clean. ( )
What did the kids look like ? .. are they Caucasian... or some ( ) ethnic group ( )?

Well! ---- It's a good question.

( )
African but I wasn't sure.
O.K.

..... uh uhm the girl was blond

( )

The girl was blond .. an .. uh .. the a.. ah .. and .. the three little boys
( ) It was instinct a durational composition of the cast..

Uh.. uh ..they're mostly dark haired ..

O.K. .. go on !

They're .. all of the kids walked by.. and .. the other one was still playing with his
paddle ball ........ and he watches them go by .. but they are not caring .. ( )..
he does not say any thing . . or do anything during the whole period

Go on!

THAT'S THE WHOLE THING.

O.K.

 

Arabic Narrative (10)

0.6.. WMJLaicila'igdsx r191! 1.15", ...:Lgsl 1:13.: "1:13.45 IJﬂ..C:I.L.1 We???“ .1
all-.311 oh: "6er “J: .163”; cl“, 9.1 14416:" 14.11.3163," J)L.Jl 5,.6 “59‘ 6.1.11: .. 6,913.:
mrsBJchging..JJli.6ciL-.§3AJ ..11..Ml¢%J#E6J+ﬂI 9.9th
69;}. Y, "gel.“ 63.41% )'C'C 12.3.; «..if: 6:53.: 41 LesbgeLll" 1.1.6 jelll H" 5gb

01.9
"asserts, ...116,5.=u16,._~....,( )..gsa, ...mgrfﬂrl .. 653133.44] .1
//..._1Lal gags-:2...

86).-.4319”
3143,51! 4.2L ”as", claim). dime; 11.36 .1.13..6J.-.s..::l| Osaka, JIJLaraJ .i
// 61.291 4.3:...

( )..loatys // ~24
// 9.2.1... ..I 1“": $51.!

....en ciﬁlw' // ' ....

// 64:30.... .1

mews» «sz ....
.. raj gl ..a..: ”or! .1
1..”; .6...
.31.... (in, .. d...“ “.12 has”: ..a..,” gs 9111b: “Masai-‘9' 5,. ...rsyl c... .1
.411! .671
4:1th 6?ALAJDL 191111.111... .. 6M] “.1! .1ch ".1131“ng "$516)..er .l
Maw 411.11 git .. 411.11 0.11... gust 9.! as, ,5; 91,111, 4,,1; 94.1. ..
“as“ 4+4” ~2-

138

 

139

3.111.213. g... .. 36111 a.....1....t,11 1.43 gm .1939: but...” 1'...- .1
//.Ls:91;,..1_.....,;r,;.6u1 est... (:1, use...;srr,.1_.;,s1 ..tglgsgts
10.51.11 // ..,.

1.5 6,1. "tel: glide ..4.-=.l_,.11| Jerri Gﬁeilsigdejl‘, ..a..-1:, 3.0.5611 .1
..ulisgajsslelldi'

..ul ...}.
gamin(LuJ1t?6l.AUJAJ"ﬁtSJbe§ﬂ3uLi§”uLak‘flﬁbmllej‘ngldj
451$ 1.0.1., ”4.an 4614.9 .. cards-.1 111 "1.3.1.191 L. 1.13.12,“ "1.21.64le
.. 153,515,243?ng .. 1.3541114? Labs, 14:13 .. wahid 41...“ .. dean.
//4§1JJ.11I=‘54A1.13J.6L=.¢1L 1,..6, 41....“ 41.3.,

l..4,_.l4_3.l_,.111// 4,.

84552.64]le 21,696.,” ”4,311,... WJMJ..JBQI&,3@J6M@.1
"$1,255.41. ..9191 w..w-win ..wa... 13.5..“de

...«tal H

31.“ ...Cwlsll Lflejasssté 1.1.14.1le g... uh..-.11 Mugging“ .. 1.1.6 .11le L5...':... .1
{ssmlﬁjs‘sl a..-:91 wa.-..a... 9.361.

{441...} ' ....

.....queelseu, 9111 5.31.11 wri 1.. .. 91 11.31.4619“ gang-Legh: .....st .1
91:, u..Us.,,-,.a,.,,al,slsgrs,..tr..lssdalsllesus 1,11...“ ...srtms-JI
Jest ,3 .. 6.3.6.111:ij cast..-541w 95 we. us... gm “2,1 w.” L. .. 4:51” 1.151,
..a..... till ,u....,.1_,.rn altars-swore 11 ”591.11 «1.01:. Hangers, .. 1.1g
guru-ihﬁﬁhligidmgtjgﬂl 139,152,,er ..dsjwtrsetxsg
53.1: gs 9.11:1“, ..uiehrwi ”may: gm“, gist... .Lﬁzme,“ 5&6: (+6 .1le
.. Uiclssﬁl

1..“1 .9

 

140

//..91 (,4 9,211.43 "44.11.. L..I,11.1

t,_,.::JI// .4,

..4,. 40.2.25 ..aﬁgrs, ”4,251, 1.1.. 4,311, ..41.,,,.. unsung, 1 m1 ”.221! .1
Gem 1,1.21..1.,., 11,11 131*,“ 41,53,

s:...1_,.::,11..,.

//..-...I,.:.-.,11.1

styldalulaesu // ..,.

43,1, 1191 outset-.21, .1

“sibglﬁleoglﬁ-e

4,1,1“ 9.1.9.1

$4,384.41} .4,

swig..gw qulrltillgi magnet, tsetse-....“ ..raj..-4:?1J.l.11“.lc .1
”gem“. lwmraw )lhdhlsllulebaﬁll ..4._._i6L°..1.6L:"..11e L., f.,._;.1,:
()1:th 1231,11 [...an 1..-11:11.. 111111.31 a..;srxml, 3.41,,591 ..,,,§11
5,111,... tray-:11 lia..,5,isﬂl%ulgi4.sl_x§1¢ae L. Lapel 418,,eoJiJI
H1!..ci.6.4_-..l_,.11lulcY|Y|UJ:4r,lldr,Uai.u, "I.BW‘LA, .. [£4.12wa
4:th berm-3149515111044 1--...-5345ts.2-‘-1v41-J' g1" Likes-416.1%
//.J . .11 5.43.1.4 .. rany .. Hr. L...I,11.. 1,11

..,_,s.,..111// ..,.

at... est-disuse, 6,311 6116:1343” .. 4,1,. 1,,1..,1.,=s .. “1'.er 1.11.1. 6111 .1
1,.2.,.J§t,,4,:_,,rt,s.sl...,in11s..1,s-.....11.,,,:r91..11;,..

( )«a
t”fill lhdsjlmwﬁéglogalwﬂll ..u,'il|6Llse1L6.1.1, ..éﬁrALlnelY .1
”6&3.” 1:6"

46" ~s-.4

 

141

.13....1.,.~_-..1.1.5;,,_-..,..g1111:1..1.,( )g. 1.5,.,J,-.;..4,11 1,1A,,1,3,1.,11.1
.4'c_,l§1¢.,59..1.;,3..nc()..¢1L,1.ga

4%“ .1...

...YpJubvlce9Edt‘Jl 9.39.6.ng11 1.143,,"u.,u,.;,111 .1
91”,”.24354515l35m44juzg" mdlﬁlsglﬁuﬁlglcah‘itﬁglwﬁdﬁ
c,_,§,.:.r.r¢11..1;1§ﬂ,.c1u. 14.19., .. 6,391 11.1,. g. .111le 5,151 4.11, .1.1 1.1.1. ..12.-
$111,391.11 £091.”, .. 13,19,411 “.251, 4,4,11,29,13 1.,L,u.1,.1.:,1,..
.1,11411..1,.g111, ..4..,,,1.11.,...1..4,.,,.:;.

$131M -- QWJLAJHJ ~e-1

 

"7'1 11111111111111.1711