Dutch-Moroccan Code Switching among Maroccans in the Netherlands [Reprint 2019 ed.] 9783110877182, 9783110131024

Citation preview

Dutch-Moroccan Arabic Code Switching

Jacomine Nortier

Dutch-Moroccan Arabic Code Switching among Moroccans in the Netherlands

¥

1990 FORIS PUBLICATIONS Dordrecht - Holland/Providence RI - U.S.A.

Published by: Foris Publications H o l l a n d P . O . B o x 509 3300 A M D o r d r e c h t , T h e N e t h e r l a n d s Distributor for the U.S.A.

and Canada:

Foris Publications U S A , Inc. P . O . B o x 5904 P r o v i d e n c e RI 02903 U.S.A. CIP-DATA Nortier, Jacomine D u t c h - M o r o c c a n A r a b i c : C o d e S w i t c h i n g a m o n g M o r o c c a n s in the N e t h e r l a n d s / J a c o m i n e N o r t i e r . - D o r d r e c h t [etc.] : Foris. W i t h bibliogr., index I S B N 90 6765 431 0 SISO 803.2

U D C 800.73:[809.27+803.931](492) ( = 1 . 6 4 - 8 2 )

Subject heading: c o d e s w i t c h i n g ; D u t c h M o r o c c a n A r a b i c .

I S B N 90 6765 431 0 © 1990 Foris Publications - D o r d r e c h t N o p a r t of this publication m a y be r e p r o d u c e d or t r a n s m i t t e d in any f o r m or b y any m e a n s , e l e c t r o n i c or m e c h a n i c a l , including p h o t o c o p y , r e c o r d i n g , or any i n f o r m a t i o n storage and r e t r i e v a l system, w i t h o u t permission f r o m the c o p y r i g h t o w n e r .

Contents Acknowledgements

xiii

1

Introduction

1

2

Moroccan Arabic Grammar: An Overview

7

2.1 2.2

2.3

2.4

3

Notation and Phonology Morphology and Basic Categories 2.2.1 Verbs 2.2.1.1 Auxiliaries 2.2.2 Personal pronouns 2.2.3 Nouns and adjectives 2.2.4 Prepositions 2.2.5 Adverbs Syntax 2.3.1 Basic word order 2.3.2 Clause types 2.3.2.1 Nominal sentences 2.3.2.2 Raising 2.3.2.3 Simple verbal clauses 2.3.2.4 Complex sentences without a conjunction 2.3.2.5 Relative clauses 2.3.2.6 Complex sentences with u (and) 2.3.2.7 Subordinate clauses introduced by an interrogative 2.3.2.8 Conditionals 2.3.2.9 Negation Moroccan, Classical Arabic and Berber

.

8 12 13 16 19 21 29 30 31 31 33 33 34 35 36 36 39 39 40 40 42

The Linguistic Situation in Morocco

45

3.1 3.2

45 46 46 48 50

Morocco Individual Languages 3.2.1 Berber 3.2.2 Arabic 3.2.3 French

vi

CONTENTS

3.3 3.4 3.5 4

The 4.1 4.2 4.3

4.4

3.2.4 Other languages 3.2.5 The Jewish community in Morocco Language Attitudes in Morocco Education Diglossia

51 51 52 54 56

Moroccan Community in the Netherlands Moroccans in the Netherlands, an Overview Moroccans in Utrecht Social and Geographical Aspects 4.3.1 Religion 4.3.2 Shops 4.3.3 Restaurants, cafés and coffee houses 4.3.4 Groups and networks Language Proficiency

59 59 60 61 68 69 70 70 73

5

The Present Study 5.1 Selection of Speakers 5.1.1 Criteria 5.1.2 Looking for subjects 5.1.3 The informants 5.2 The Recorded Conversations 5.3 Transcriptions 5.4 A Preliminary Classification Model

77 77 77 78 81 85 91 91

6

Bilingual Language Proficiency 95 6.1 Introduction 95 6.2 Methods for Measuring Bilingual Ability 96 6.2.1 Self-report 97 6.2.2 Actual use of Dutch, Moroccan Arabic and a mixed variety . . 98 6.2.3 Judgement by native speakers 102 6.2.4 Error analysis of spoken Dutch in the corpus 104 6.2.5 Comparison between methods 108 6.3 Characteristics of the Moroccan Arabic used Ill 6.4 Bilingualism and Code Switching 113 6.5 Socio-Psychological Aspects 116 6.6 Conclusion 118

7

Classification of the Switched Material 7.1 Intersentential Code Switching 7.2 Extrasentential Switches 7.3 Intrasentential Switches 7.3.1 Intrasentential switches between coordinated clauses

121 121 123 124 126

CONTENTS

7.4

8

vii

7.3.2 7.3.3 7.3.4 7.3.5 7.3.6 7.3.7 7.3.8 7.3.9 7.3.10 Single 7.4.1

Intrasentential switches around adverbs 126 Switches inside VP (between V and NP) 130 Switches between main and subordinate clauses 131 Switched appositions and dislocated elements 133 Switching of PP 134 Switching after subject 135 Switching inside NP 136 Switching between P and N(P) 139 Problems with the classification of intrasentential data . . . . 139 Word Switches 140 Word classes to which single switched words belong 141 7.4.1.1 Nouns 141 7.4.1.2 Adverbs 141 7.4.1.3 Adjectives 142 7.4.1.4 Conjunctions 142 7.4.1.5 Verbs 143 7.4.1.6 Prepositions 144 7.4.1.7 Pronouns 145 7.4.1.8 Numerals .145 7.4.2 Syntactic classification of single word switches 146 7.4.2.1 Single word switches inside NP 146 7.4.2.2 Single word switches between preposition and noun . 148 7.4.2.3 Single word switches inside VP 149 7.4.2.4 Single word switches after subject 151 7.4.2.5 Switching of single adverbs 152 7.4.2.6 Switching of single conjunctions 153 7.4.2.7 Single word switches functioning as adverbial PP . . 154 7.4.2.8 Switched subjects after a conjunction 154 7.4.2.9 Word-internal switches 155

Constraints on Code Switching 8.1 The Notion of Matrix vs. Embedded Language 8.2 Constraints on Particular Constructions 8.3 Constraints Claimed to be Universally Valid 8.3.1 The equivalence constraint 8.3.2 The government constraint 8.3.3 The subcategorization constraint 8.3.4 Constraints formulated by Joshi 8.3.5 The free morpheme constraint 8.3.6 The size-of-constituent constraint 8.4 Conclusions

157 157 161 169 169 175 175 178 179 180 181

viii 9

CONTENTS Single Word Switches 9.1 Definitions 9.2 Single Words in the Present Study 9.2.1 Borrowing from French and Spanish 9.2.2 Single Moroccan Arabic words in Dutch sentences 9.2.3 Dutch single nouns in Moroccan Arabic sentences 9.2.4 Functional aspects of switched single Dutch nouns 9.2.4.1 Moroccan Arabic equivalents or not? 9.2.4.2 Semantic fields 9.2.5 Formal aspects: the use of articles 9.2.5.1 Explanations for the patterns found 9.3 Nonce Borrowing 9.4 Constituent Insertion

183 183 187 187 187 188 191 191 193 197 202 206 207

9.5

209

Summary and Conclusion

10 Conclusion

211

A Transcriptions

215

B Borrowings from Romance Languages

223

Bibliography

225

Author Index

233

Subject Index

235

List of Figures 2.1 2.2 2.3

Indefinite NP Moroccan Arabic definite NP Dutch and English definite NP

25 27 27

3.1 3.2

Morocco Languages attitudes in Morocco

47 53

4.1 4.2 4.3 4.4 4.5

Utrecht divided into neighborhoods Utrecht: concentration of immigrants by neighborhood Eight neighborhoods with the highest percentage of immigrants Percentages of immigrants in different parts of Utrecht Moroccans in Utrecht: Age and Sex

63 64 65 66 67

5.1 5.2

Morocco - Home towns of the speakers Model for classification

6.1 6.2

Formal representation of possible language use Actual language use ordered by percentage of Dutch

99 101

7.1

Intrasentential switch sites

125

...

81 92

List of Tables 2.1 2.2 2.3 2.4

Table of Moroccan Arabic consonants (based on Harrell (1962:3) . . . The verb: perfect and imperfect tense Personal pronouns and suffixes A comparison between Moroccan Arabic, Classical Arabic and Berber

11 14 20 43

5.1 5.2 5.3

The speakers, by age, length of stay, place of birth and occupation . . 86 Portions transcribed of the recorded conversations 88 Topics discussed in the recorded conversations 89

6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 6.11 6.12

Languages preferred by the speakers 98 Actual language use in the conversations, in percentages 100 Actual speech time of the informants (in minutes) 101 Average judgements by native speaker judges on a five point scale. . . 104 Errors made in Dutch 106 Cross-classification of self reported preference and proficiency 108 Difference between D and MA scores compared to error indices . . . . 109 Rank orders 110 Rank order correlations between four measures of bilingualism . . . . 1 1 1 Places of origin of the speakers 113 Code switches: main types 114 The relation between age at arrival and proficiency in D or MA . . . 1 1 8

7.1 7.2 7.3 7.4 7.5 7.6 7.7 7.8 7.9 7.10 7.11 7.12

Intersentential code switches 122 Distribution coordinate conjunctions in intersentential switches . . . . 122 Extrasentential code switches 123 Intrasentential code switches 126 Intrasentential switching between coordinate clauses 127 Distribution coordinate conjunctions in intrasentential switches . . . . 127 Intrasentential switching around adverbs 129 Types of adverbs 129 Switches inside VP 130 Code switching between main and subordinate clauses 132 Switched appositions and dislocated elements 133 Switching of P P 135

LIST OF

TABLES

7.13 7.14 7.15 7.16 7.17 7.18 7.19 7.20 7.21

Switching after the Moroccan Arabic subject Switching inside NP Word classes to which single word switches belong Single switched conjunctions Syntactic classification of single word switches Single word switches inside NP Single word switches inside VP Single word switches after subject Types of single adverbs

136 137 141 143 146 147 149 151 153

8.1 8.2

Constraints on code switching valid in particular constructions . . . .162 Conjunctions involved in code switching 164

9.1 9.2 9.3 9.4 9.5 9.6

Single Dutch nouns vs. Dutch nouns in multi-word switches 192 Semantic fields 193 Dutch single nouns divided among semantic fields 194 Dutch nouns in multi-word switches divided among semantic fields . . 195 The use of articles with Dutch nouns in Moroccan Arabic sentences . 198 Contrasts between MA-Fr and MA-D switches 201

Acknowledgements The book at present is a report on an investigation carried out from the end of 1985 until the middle of 1989. The project was financially made possible by the Foundation for Linguistic Research which is funded by the Netherlands Organization for Research (NWO). A slightly different version of this book was defended as a doctoral dissertation at the University of Amsterdam. Of course, the book could never have been written without the solid help and support of many friends and colleagues, of whom I will mention here Pieter Muysken who supervised me, Henk Penning who did all the work in which computers were involved, Henriétte Schatz who corrected my English and Mustapha Lkoundi who collected the material. Furthermore, I would like to thank those who have commented on (parts of) earlier versions of the book: Rob Schoonen, Jan Hoogland, Jeanine Treffers and many other colleagues working in the field of linguistics. Finally, support at home from Henk and Loura was indispensable. The book is typeset using IATgX (Lamport, 1986).

Chapter 1 Introduction Over the last decades, scholars have become interested in language contact, its social context, and its consequences. Code switching, the use of more than one language in the same conversational context by bilinguals, is one of the phenomena that have attracted special attention. In code switching research a distinction is usually made between switches within sentence boundaries (intrasentential switches), switches involving complete sentences (intersentential switches), and semi-independent stretches of discourse (extrasentential switches, such as tags). (1) and (2) axe examples of intrasentential code switching: (1)

You can it ^ zondag doen. ['You can do it on Sunday.'] (English-Dutch; Crama and Van Gelderen, 1984)

(2)

Les femmes et le vin 4 ne ponimayu. ['I don't know anything about women and wine.'] (French-Russian; Timm, 1978)

Intersentential code switching is shown in (3): (3)

Ik ben kritisch, he? 4 ka-nqul: 9las ka-yxess-ni neqra? ['I am critical, am I not? I say: Why do I have to study?'] (Dutch-Moroccan Arabic; kalb477, this study)

An example of extrasentential code switching, in which a quote is switched, is (4): (4)

Le dísi: • "Ma aní asé?" ['I said to her: "What am I to do?'"] (Spanish-Hebrew; Berk-Seligson, 1986)

Ideally, the field of code switching is the field in which several linguistic disciplines meet. Sociolinguists investigate why members of a certain speech community switch codes. Therefore, in this study, quite some attention is paid to a description of the speakers' social and linguistic background. A psycholinguistic orientation is required

2

CHAPTER 1.

INTRODUCTION

to investigate the organization of speech ability in a speaker's mind. An attempt to determine bilingual proficiency should be seen in this light. Grammatical research is needed to establish the linguistic facts that constitute code switching. Many people outside the field view code switching as incomplete and careless language use. Poplack (1980), however, has shown that only balanced bilingual Puerto Ricans are able to use code switching effectively. McClure (1977) has made clear that children do not use intrasentential code switching intensively before the age of eight. Both findings indicate that a high degree of bilingual proficiency is required in code switching. Berk-Seligson (1986) has found, on the other hand, that balanced bilinguals, Ll-dominant bilinguals, and L2-dominant bilinguals switch codes in very much the same way. The related phenomenon of borrowing is often considered a degeneration of a language. Bentahila and Davies (1983), like many other authors, have pointed out that it is important to draw a distinction between code switching and borrowing: "The latter refers to the use in one language of items which originate in another language, but which are currently felt to form an integral part of the borrowing language." (1983:302) Phonological and morphological integration are sometimes viewed as a criterion in the distinction between code switching and borrowing, a view that is not held by Pfaff (1979) and Bentahila and Davies (1983). More reliable criteria are functional in nature: borrowing occurs, for example, in the speech of both monolinguals and bilinguals, while code switching requires proficiency in both languages under consideration (Sridhar and Sridhar, 1980; Kachru, 1982). Shaffer (1975) has remarked that "... switching is confined almost exclusively to informal speech." (1975:489) As an example of borrowing Bentahila and Davies (1983:302) mention the Moroccan Arabic word bisri, which originates from French "epicerie".

In a number of studies on code switching syntactic constraints on possible switch sites have been formulated. The first major constraints have been formulated by Poplack (in several publications, in cooperation with others), mainly on the basis of Spanish-English language material. Many constraints have been developed since. Some proposals for universally valid constraints are: Equivalence constraint, Poplack (1980) Code switching is only possible at points where the two languages have the same word order. Free m o r p h e m e constraint, Poplack (1980) Code switching is not possible within words. G o v e r n m e n t constraint, DiSciullo et al. (1986) Code switching is not possible between a governing and a governed element. Subcategorization constraint, Bentahila and Davies (1983)

3 All items must be used in such a way as to satisfy the (language-particular) subcategorization restrictions imposed on them. Closed class constraint, Joshi (1985) Only constituents can be switched, with the exception of closed class items. Size-of-constituent constraint, Berk-Seligson (1986) Higher-level constituents (sentences, clauses) tend to be switched more frequently than lower-level constituents or smaller ones (i.e. one-word categories) with the exception of nouns which comprise the greatest number of switches. The number of language pairs involved in code switching research is still too limited to allow statements about the universal validity of these constraints. I do not intend to add to the available list of investigated language pairs with this study. Neither do I think that it is necessary to investigate all possible language pairs. However, evidence from typologically distant languages is required to further determine the factors conditioning code switching (see also Clyne, 1987) and this book should be viewed as a contribution towards this end. French-Moroccan Arabic code switching has received a fair amount of attention in several publications of which I will mention the especially valuable work by Bentahila and Davies (1983), Abbassi (1977), Ennaji (1988), Saib (1988) (all based on Moroccan material), and Nait M'Barek and Sankoff (1988) (based on material collected among bilingual Moroccan students in Canada). In my own work I hope to be able to contribute to this research tradition. However, the setting in the present study differs from that in earlier studies: Moroccans in the Netherlands belong to an ethnic minority group that generally has low social status. The pressure of Dutch society is enormous. Moroccan Arabic is a language that has received relatively little attention compared to other languages spoken in the Netherlands. There is no such thing as a "Standard" Moroccan Arabic and it is not automatically the native language of all Moroccans. Fortunately, over the last decade scholars have shown a growing interest in Moroccan Arabic as an object of research. Apart from the existing language courses and (mostly) normative grammars, studies on Moroccan Arabic syntax, phonology and sociolinguistics are now appearing in Morocco as well as in other countries. Generative syntactic and phonological studies on Moroccan Arabic have been carried out by scholars like Ennaji (1980, 1982), Boudali (1984), and Heath (1987). Recently I have also received some work by Moroccan students and scholars from Prof. Condon in the United States. Unfortunately, I have not had time to incorporate these studies in the present research, but in the bibliography some of them have been included (i.e. Ben Jilali (1984), Drissi (1986), El Ayadi (1984), Belahnech (1986) and El Medkouri (1986)). Work on sociolinguistics in Morocco has been carried out by Abbassi (1977), Gravel (1979), Bentahila (1983a, 1983b, 1987), Bentahila and Davies (1983, 1988), Saib (1988), Heath (1989), and Ennaji (1988).

4

CHAPTER

1.

INTRODUCTION

One of the papers sent to me by Prof. Condon also concerns Moroccan Arabic sociolinguistics (Mahi, 1986). In the Netherlands Otten has done work on Moroccan Arabic; he has written a Moroccan Arabic-Dutch dictionary that I have used frequently in preparing this book (Otten, 1983). De Ruiter has recently completed a dissertation about young Moroccans' acquisition of Dutch in the Netherlands (De Ruiter, 1989). There are certainly other, and more recent publications concerned with Moroccan Arabic linguistics, but they cue not available in the Netherlands. The reason for my concern with Moroccan Arabic is connected with the growing interest in Moroccan Arabic in general. Compared to the amount of work on other ethnic minority languages in the Netherlands, Moroccan Arabic has been underrepresented in scholarly investigations. It is in close contact with Dutch, which will undoubtedly have consequences for the development of Moroccan Arabic in the Netherlands. This book is an attempt to describe what happens in the Dutch-Moroccan Arabic contact situation. It is not the place to predict the future of Moroccan Arabic in the Netherlands. The contact situation between Moroccan Arabic and Dutch is far too young, not much older than twenty years. Those who view code switching as a first step to (future) language loss have reasons to be pessimistic about the future of Moroccan Arabic in the Netherlands, since code switching between Dutch and Moroccan Arabic occurs frequently in the speech of bilingual Moroccans. On the other hand, frequent code switching may also be viewed as a phenomenon that helps to sustain bilingual skills (Pedrasa et al., 1980). Moreover, in Morocco, where code switching between French and Moroccan Arabic is part of normal everyday conversation, neither Moroccan Arabic, nor French has disappe'ared (see chapter 3) from the speech community. This study deals with code switching between Dutch and Moroccan Arabic in a corpus of tape recorded conversations. It represents an attempt to combine the two topics that have been outlined here: code switching between two typologically distant languages on the one hand, and code switching behavior among speakers whose native tongue is Moroccan Arabic, on the other. Many readers of this book will not be familiar with Moroccan Arabic, so chapter 2 provides a rough outline of Moroccan Arabic grammar. This outline is not meant to be exhaustive. It only presents some general facts and syntactic descriptions, in particular concerning syntactic constituents that will eventually be involved in my analysis of code switching. I will not present a grammatical overview of Dutch, but relevant facts about Dutch grammar will be presented throughout the text. Those who want to know more about Dutch are referred to Donaldson (1981). Chapters 3 and 4 deal with the background of the speakers who have been included in the present study: chapter 4 sketches the Moroccan community in the Netherlands, and in particular Utrecht, where the study was carried out, while chapter 3 deals with the highly complicated linguistic situation in Morocco, where the speakers were born and have lived for at least a few years. People who switch codes frequently are often not aware of doing so. In order to observe code switching, it is important to keep the setting as informal as possible,

5 since code switching is highly informal behavior. Chapter 5 explains how the best possible situation for eliciting code switching was created. In chapter 6 matters concerning bilingual proficiency will be discussed. It has been shown in several studies (Poplack (1980), among others) that speakers should be balanced bilinguals in order to switch codes effectively. Problems in determining bilingual ability will be discussed and I will suggest a solution to these problems. Bilingual proficiency will also be related to the use of certain code switching patterns. In the last section, an attempt will be made to relate bilingual proficiency to sociopsychological factors. In this study this was only possible to a limited extent, though, mostly due to the homogeneity of the group of informants. The language proficiency of each individual speaker will also be considered in chapter 6. In chapters 7, 8 and 9, where code switching phenomena are discussed, the group of speakers will be treated as a whole, because the corpus is not large enough to differentiate for each speaker and each type of code switch and it may often be a matter of coincidence that some speakers did not use certain types of code switching, while others did so frequently, because there is a high variety in speech time and, accordingly, the number of switches used. In chapter 7 the code switches that were found in the corpus axe classified according to switch types. The analysis of the switches is presented in chapters 8 and 9. Chapter 8 deals with the constraints on code switching as they have been formulated in the literature. A distinction is made between specific and universal constraints. Also in chapter 8 these constraints are confronted with the Dutch-Moroccan Arabic material. In chapter 9 attention is focused on single word switches which constitute almost half of all recorded code switches. The question whether we axe dealing with borrowing here, or perhaps with other language contact phenomena will be discussed in depth. This book should by no means be viewed as an exhaustive treatment of code switching. Many issues could not be taken up. I have already mentioned the sociopsychological factors that influence language behavior. Functional aspects of code switching axe also related to the socio-psychological background of the code switchers, but I will not go into the functions of code switching in detail in this book. In discussing the single word switches (chapter 9), however, their function will be touched upon briefly in an attempt to explain the patterns that were found. In Schatz (1989) a functional model for the analysis of code switching and borrowing is presented in which (grammatical) constraints and motives for code switching are related to each other. A combination of a functional analysis and an analysis of grammatical constraints seems a promising avenue for future research. In this study, I will concentrate on a description of the speakers' social and linguistic background, a detailed analysis of bilingual proficiency, and on the grammatical aspects of the Dutch-Moroccan Arabic switched material. I have limited myself to Dutch-Moroccan Arabic code switching. Berber-Moroccan Arabic and Classical-Moroccan Arabic code switching will have to wait for a later study. In the literature the terms code switching, code mixing and code alternation are used frequently, but there is no unanimity as to which term should be used for which

6

CHAPTER 1.

INTRODUCTION

phenomenon. For instance, Bokamba (1988) uses code switching to indicate switches between sentences and code mixing for switches within sentence boundaries. I will follow Clyne (1987:740) who employs the term code switching in the sense of "the alternative use of two languages either within a sentence or between sentences." and Bentahila and Davies (1983:307) who use the term code switching to refer to the use of two languages within a single conversation, exchange or utterance. The result is in either case an utterance or interaction of which some parts are clearly in one of the bilingual's languages while other parts axe in the other language.

Chapter 2 Moroccan Arabic Grammar: A n Overview The literature on Moroccan Arabic covers only a limited set of titles. This may be a consequence of the fact that there does not exist a broadly accepted written variant. Furthermore, the low social status of Moroccan Arabic may play a role in this matter. Linguistic studies on Moroccan languages are mostly concerned with (Classical) Arabic and in recent years Berber languages have also moved to some degree within the scope of scientific interest. The most important classical studies on Moroccan Arabic are generally descriptive: Brunot (1950), Harrell (1962), Sa'id (1950), Abdel-Massih (1973), Mercier (1956), among others. Dutch contributions to the literature on Moroccan Arabic are among others, Hoogland (1983), Otten (1983), Van Mol (1983). In recent years some detailed theoretical studies have been carried out, mainly on syntactic aspects of Moroccan Arabic: Ennaji (1980, 1982), Boudali (1984), Condon et al. (1987). Heath (1986) has concentrated on phonology and more recently he has written extensively on borrowing, in which all grammatical levels are represented (Heath, 1989). The orthography of Moroccan Arabic as it is used in this dissertation will be discussed and explained in section 2.1; in the same section some phonological phenomena will be treated briefly. In section 2.2 Moroccan Arabic morphology will be the central theme, while in section 2.3 syntax will be discussed. I do not intend to give tin in-depth and exhaustive description of Moroccan Arabic: this short grammatical overview is meant to give the reader who is not familiar with Moroccan Arabic a rough idea about the language. Only those constructions in which frequent code switching occurred in the corpus will be discussed in more detail. Examples are, whenever possible, taken from the Dutch-Moroccan Arabic corpus used in the present study. In the glosses, the following abbreviations will be used:

CHAPTER 2. MOROCCAN ARABIC GRAMMAR: AN OVERVIEW

8

perfect tense PF IMP F imperfect tense I M P E R A T imperative mode singular SG PL plural M masculine feminine F DU durative PC pronoun clitic participle PTC first part of discontinuous negation NI N2 second part of discontinuous negation DUA dual demonstrative DEM DEF definite article FU future particle When the meaning of the sentence is not directly interpretable from the gloss translation, a free translation is given as well.

2.1

Notation and Phonology

I will use the Latin alphabet to represent Moroccan Arabic words and sentences, as has been done by many of the authors cited. Although in Morocco the Arabic alphabet is sometimes used for this purpose (Hoogland, 1983), I have decided not to do so in this dissertation for the following reasons: • Moroccan Arabic has more phonemes than Classical Arabic. Besides, in Classical Arabic short vowels are usually not written, while in the Latin alphabet they are. • When the Arabic alphabet is used, a word-for-word translation to English will be problematic, since Arabic is written from right to left, while English is written in the opposite direction. This could easily cause misunderstandings. • In order to achieve maximum accessibility, also for those who are not familiar with Arabic, the Latin alphabet is preferred over the Arabic script. The notation is based on the one Otten has used in his dictionary (Otten, 1983). The main difference with Otten's notation is that, where Otten uses a dash to indicate pharyngalized sounds, I will use a dot. Besides, I will use 9 to refer to the pharyngal consonant. I have also adjusted the examples taken from the literature to conform to this convention. In order to avoid redundancy and to accomplish maximum adequacy, only semantically distinctive sounds (phonemes) are used. In other

2.1. NOTATION

AND

PHONOLOGY

9

words, I have adopted a phonological notation. Before the symbols and their pronunciation will be presented (in alphabetical order) some remarks about Moroccan Arabic are in order. To people who axe not familiar with Arabic languages, the presence of guttural sounds is probably most striking. These are uvulars and pharyngals. Also typical are the so called emphatic consonants. In fact, they are pharyngalized, which gives the impression of emphasis, so the consonants axe often called emphatic. I will use the term pharyngalized consonant, since it refers to their phonetic quality. About the way they should be pronounced Harrell (1962) writes: There are eight emphatic consonants, b, m, t, d, s, z, I, and r. These sounds are lower in pitch than their non-emphatic counterparts. They are pronounced with greater muscular tension in the mouth and throat and with a raising of the back and root of the tongue toward the roof of the mouth. The English speaker can notice this contraction of the throat very easily by prolonging the "1" in "full", since this English "1" is exactly like the Moroccan I. (1962:8) Vowels axe "darkened" when they are accompanied by one or more pharyngalized consonants, or by q, or 9. The symbols which have been used to represent the sounds of Moroccan Arabic are presented here with a rough approximation of how they are pronounced: a Full vowel; the pronunciation varies from a in the English word man to a in father. Its quality depends on the surrounding consonants: in the neighborhood of pharyngalized sounds it is approximately pronounced like in English father. b Not different from English or Dutch. b Pharyngalized b, rarely used. d About the same pronunciation as English d. d Pharyngalized d. e Schwa, pronounced as the second e in he&ven. f Not different from English or Dutch. g As in English good. g No English or Dutch equivalent; it is pronounced as the French uvular r in Paris. h As in English and Dutch.

10

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

h Pharyngalized h, the sound of someone breathing on a mirror in order to clean it. i Full vowel; the pronunciation varies from the sound in English beast or Dutch pier to the one in English beer. The latter only in the neighborhood of pharyngalized consonants. k As in English or Dutch. 1 As in (British) English or Dutch. 1 Pharyngalized I; it sounds like I in American English or in the dialect of Amsterdam (the velarized 1). m As in English or Dutch. m Pharyngalized m, rarely occurring in Moroccan Arabic. n As in English or Dutch. q No equivalent in Dutch or English; Harrell (1962) about the pronunciation: The q is similar to the k except that contact between the upper surface of the tongue and the roof of the mouth is further forward for k than for q. (1962:5) r Pronounced as Dutch (or Scottish) trilled r. r Pharyngalized r. s As in English or Dutch. s Pharyngalized s. s Pronounced like the initial sound in show. t About the same as in English, usually pronounced with a noisy release which sounds similar to s: ts. t Pharyngalized counterpart of t, but unlike t pronounced without the s-sound. u Full vowel; pronounced as in English or French tour, or Dutch goed. Before or after a pharyngalized consonant, 9 or q it sounds like o in French tot or Dutch boot. u Unstable, short vowel; sounds like English u in put when syllabic or as the onset to a syllabifying sonorant, as in xubz (bread). w Pronounced with rounded lips as in English, unlike Dutch.

2.1. NOTATION

AND

PHONOLOGY

11

Glottal Pharyngal Uvular Velar Palatal Apical Labio dental Bilabial Voiceless Stops: - t t - k q ' Voiced Stops: bb - dd - g Voiceless Fricatives: f ss s - x h h Voiced Fricatives: zz z - g 9 Nasals: m m n Lateral: H r r Flap-Trill: w Semi-Vowels: y Table 2.1: Table of Moroccan Arabic consonants (based on Harrell (1962:3) x As ch in Scottish loch or in Dutch/German Bach. y As the initial sound in English you or Dutch jongen. z As in English Zappa or Dutch zee. z Pharyngalized counterpart of z. z As in French or Dutch jus. 9 Pharyngal consonant. Its pronunciation is close to a in English father, unlike a it is a voiced fricative.

but

' Glottal stop. It rarely occurs, except in some originally Classical Arabic words and is often omitted in Moroccan Arabic. It sounds like the break between vowels as heard in the exclamation "uh oh" (Harrell, 1962:7). The glottal stop (or hamza in Arabic) is considered a consonant. Schematically the consonants are organized as in table 2.1. There are three full vowels, a, i, u, and two unstable short vowels: rounded u and unrounded schwa, (e). The full vowels are articulated clearly, and if there are severed of them in the same word, the stress pattern tends towards even and staccato. (Heath, 1989:31)

12

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

There are also a few sounds which appear only in loanwords: [e], [e], [o], [o], [p], [u]. To conclude this section, some phonological phenomena will be discussed briefly. Similar to English and unlike Dutch, Moroccan Arabic does not have voiced stops that change into voiceless stops when they occur word-finally: wad (river) sounds like [wed], not like *[wet]. Generally speaking, a consonant cluster in Moroccan Arabic consists of no more than two different consonants; more consonants in a cluster are only allowed as a consequence of the process of doubling (cf. 2.2.1: tbeddlat she changed). The word-internal order of consonants is fixed, whereas vowels may change place, a process to which the unstable vowels e (schwa) or u axe senstitive. In fact, it is one of their main characteristics. The vowels in question can only occur before a consonant cluster or a final consonant, not in open syllables (i.e. before a single non-final consonant). If verb conjugation or suffixation would necessitate such clusters, one of the following patterns occurs (which can be analyzed in the model of McCaxthey (1981)): • The variable vowel is absent. When sifet (he sent) is followed by the suffix -u, the resulting form is not *sifet-u, but sift-u (he sent him). • The variable vowel occurs elsewhere. If zber (he found) is made plural, -u is suffixed. The resulting form is not *zbeni, but zebrv. (they found).

2.2

Morphology and Basic Categories

Characteristic for the morphology of Arabic languages is the great number of regularities with relatively few exceptions. Some of the regularities with respect to Moroccan Arabic verbs and nouns will be discussed here. In Moroccan Arabic, most words are built on a basic consonantal skeleton called the "root". The consonants in this root are also called "radicals". The root has a basic meaning which can be modified by combining the radicals with vowels and other consonants according to fixed patterns. The root k-t-b, for example, always has something to do with writing. By using different patterns, different words can be formed like ktab (book), kteb (he wrote), ka-nketbu (we are writing) and mektaba (library). Word classes and root types are closely related: Verbs usually consist of three radicals (sometimes four), although there are exceptions. The same can be said about adjectives and nouns. Members of most other word classes consist of fewer than three radicals.

2.2. MORPHOLOGY AND BASIC

2.2.1

CATEGORIES

13

Verbs

The Moroccan Arabic verb is very complex, which means that it is capable of carrying a great deal of semantic information. There are four different verb types, characterized by the quality of the radicals: Simple (or sound) The three radicals are different from each other, like kteb (he wrote). Hollow The second radical, -w- or -y-, has become invisible; its place is taken by a vowel, as in l a / ( h e saw). W e a k The third radical, originally -w or -y, has been replaced with a vowel: bga (he wanted). Doubled or Geminate The second and third radical are equal: hell (he opened). In addition, there is a small group of irregular verbs which cannot be classified according to the listed verb types. From the verb types, stems or measures can be derived. This is a highly productive process in Classical Arabic, which can form nine measures from the basic stem, while Moroccan Arabic still uses five of them. In traditional grammars the measures are indicated by Roman numerals: I is the basic stem, measure II (in Moroccan Arabic) makes the verb causative by doubling the second radical. For example: xrez(h.e left), xerrez (he caused to leave, i.e. he sent away). In measure IX -a- is inserted between the second and third radical, meaning "to be colored or defective": hmar (become red). This is not the place to give complete paradigms of the possible measures in Moroccan Arabic. More details axe to be found in the literature. Person, number and gender of the verb axe indicated by prefixes and suffixes. There axe two possible tenses, traditionally called the perfect and imperfect tense. The perfect tense refers to the past and to the fact that the action has been completed while its result may still continue. The imperfect tense is used to refer to both the present and the future. The perfect can be recognized by its suffix, indicating person, number and gender, while the imperfect is characterized by its prefix, indicating person and gender, and a suffix for number, as table 2.2 illustrates. In the imperfect, when progressivity or habituality is referred to, the durative particle ta- or ka- (in Northern and North Eastern parts of Morocco) is used. Reference to the future is expressed by means of the prefix ga-, which is, in fact, a short form of the paxticiple gadi (he will), gada (she will) or gadin (they will) (the longer forms axe used as well). For example: (5)

kayekteb DU- write:IMPF:3:SG ['He is writing.']

14

CHAPTER 2. MOROCCAN ARABIC GRAMMAR: AN OVERVIEW

Number sing -

plur -

Person 1 2 2 3 3 1 2 3

Gender m-f m f m f m-f m-f m-f

Perfect xrez-t xrez- ti xrez- ti xrez xerz- at xrez- na xrez-tu/-tiw xerz- u

Imperfect ne- xrez t e-xrez t-xerz-i y e-xrez t e-xrez n-zerz-u t-xerz-u y-xerz-u

Table 2.2: The verb: perfect and imperfect tense (6)

kayketbu DU- write:IMPF:3:PL ['They are writing.']

(7)

ganekteb F U - write:IMPF:l:SG ['I am going to write']

(8)

gadin nketbu will:PL write:IMPF:3:PL ['We are going to write']

The imperative is formed by an imperfect without prefixes, and occurs only in second person masculine and feminine singular and plural (one form for both masculine and feminine): (9)

xelliha! (Lkoundi, 1) let:IMPERAT:2:SG- PC:3:F:SG ['Let her!']

Negative imperatives do not exist in Moroccan Arabic. Other modes, like adhortatives or desideratives, are indicated by the imperfect form without ta- or ka-: (10)

nemsiw (Harrell, 1982:174) go:IMPF:l:PL ['Let us go.']

2.2. MORPHOLOGY AND BASIC (11)

CATEGORIES

15

llah yrehmu (Otten, 1983:921) Allah have mercy:IMPF:3:SG- PC:3:M:SG ['May Allah have mercy upon him.']

Following fixed patterns, active participles can be derived from verbs. They can be used predicatively as finite verbs, like saddin in the following example: (12)

huma saddin dakssi Hi qdim bezzaf they keep:PTC:PL that:M:SG DEF- thing:M:SG which old:M:SG much (Lkoundi, 2) ['They keep to what is very old.']

Boudali (1984), who deals with differences and similarities between finite and infinite verb forms, suggests that "a number of facts distinguish the participles from their verbs: 1. Participles, unlike verbs, do not inflect for person. They are adjective-like in this respect, since adjectives always agree in gender and number but not in person. 2. Participles cannot be in the imperfective or perfective form, in fact, they do not show such a contrast. 3. Paxticiples do not carry any time reference, they rather denote states or activities and seem to require a copula in constructions which are marked for time reference." (1984:15) When reference is made to the past, a form of the verb kan is used, illustrated by the following examples: (13)

daba xeddama / - Amersfoort (kala041) now work:PTC:F:SG in['Now I am working in Amersfoort']

(14)

kunt xeddam fwahed bejaardenhuis (ahmb072) be:PF:l:SG work:PTC:M:SG in- a old people's home ['I was working in an old people's home.']

Infinitives as they are known in English and Dutch do not exist in Moroccan Arabic, since fully unmarked verb forms do not occur. Moroccan Arabic verbs are always inflected for number and gender, and except in the case of participles, also for person. However, imperfect forms without ka- or ta- have some characteristics in common with Dutch or English infinitives: e.g. the form is used after a modal or in a verbal complement. This is illustrated in (15) and (16):

16

CHAPTER

(15)

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

bgit ndirhum. fe- ttumubil. (soua007) want:PF:l:SG do:IMPF:l:SG- PC:3:PL in- DEF- car:M:SG ['I want to put them in the car.']

(16)

ana muste9dd bas nebda 9aw-tani. I ready:PTC:M:SG in order to start:IMPF:l:SG again

(mora092)

['I am ready to start again.'] Note that bga, used in (15), while showing perfective declination, is used with imperfect meaning. Ennaji (1982:22) has remarked about the difference between finite and non-finite in Moroccan Arabic: "The distinction between finite and these 'non-finite' forms is that the latter do not show any opposition of tense form." Boudali (1984:61) has shown that Moroccan Arabic has an infinitive form that corresponds to the infinitive in English in some respects, although there is a major difference between the two: English infinitives are not inflected and the Moroccan Arabic counterparts exhibit person, number and gender inflection. In both languages infinitives lack tense. More details axe to be found in Boudali's work. Just like English and Dutch infinitives, Moroccan Arabic imperfect forms without a durative or future particle cannot form an independent affirmative sentence: (17)

* yemsiw. go:IMPF:3:PL

Most verbs have a corresponding verbal noun: e.g., dreb (to hit), verbal noun drib (hitting, or the action of hitting). The regular meaning of the verbal noun is the activity or state indicated by the verb from which it is derived. Verbal nouns behave as nouns; they take regular nominal affixes. 2.2.1.1

Auxiliaries

Again, I will use Harrell (1962) as the main source, this time for a description of auxiliaries: A verb phrase is a sequence of two or more verb forms, all of which have the same subject (...). Typically, a verb phrase consists of two verb forms, the first of which is an auxiliary which modifies the following verb form as to time (...) or manner (...) of action. Phrases of three or more verb forms are relatively rare and consist entirely of a compounding of auxiliaries. (1962:179)

2.2. MORPHOLOGY

AND BASIC

CATEGORIES

17

Next, Harrell discusses the most commonly used auxiliaries. Here I will only mention the auxiliaries that were frequently used in the data corpus: kan (to be), bda (to begin), bqa (to remain), bga (to want), msa (to go), za (to come), gadi (indicating future), xess (must) and qder (can). kan It usually occurs in the perfect tense in combination with a following durative or an active participle. The meaning is that of past habitual, repetitive, durative or progressive action: (18)

maNl-

künts ka- ne9ref hadak sbe:PF:l:SG- N2 DU- know:IMPF:l:SG that:M:SG DEF-

si (hasb281) thing:M:SG ['I did not know that.'] Together with an 9end- construction, kan is used to indicate possession in the past tense, as in (19): (19)

kanet 9endbe:PF:3:F:SG with-

i rügba (mora048) PC:1:SG desire:F:SG

['I had a desire'] bda Forms of bda (to begin) are followed by the durative, and in some dialects the imperfect is used. (According to Harrell (1962:181), who only mentions the region of Fez.) The occurrences of bda I have found in the data corpus are followed by an imperfect without durative article, as illustrated in (20): (20)

kayebdaw yhed.ru. (nada022) DU- begin:IMPF:3:PL talk:IMPF:3:PL ['They are beginning to talk.']

bqa (remain) is used frequently in the data corpus. According to Harrell (1962:181) it is followed by a durative or by an active participle: (21)

bqit tanetsenna (moral72) remain:PF:l:SG DU- wait:IMPF:l:SG ['I kept on waiting.']

18

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

When negated, bqa in verb phrases is equivalent to 'no longer', with either past or present reference (Harrell, 1962:182). The only example of a negated form of bqa in my corpus, is in combination with an 9end- construction. However, it clearly carries the meaning 'no longer': (22)

maNl-

bqats 9endremain:PF:3:F:SG N2 with-

ek rugba. (mora049) PC:2:SG desire:F:SG.

['You did not have the desire anymore.'] bga (want) is always followed by an imperfect verb. The meaning of perfect bga usually is equivalent to English present tense: (23)

bgit ndirhum fe- ttumubil. want:PF:l:SG do:IMPF:l:SG- PC:3:PL in- DEF- car:M:SG. (soua008) ['I want to put them in my car.']

msa, za These are two frequently used auxiliaries of motion. When followed by the perfect, they signify completed action, as in (24). When an imperfect follows, the action is not necessary completed, as in (25): (24)

msa nzel fe- blad xra. go:PF:3:M:SG settle:PF:3:M:SG in- country:F:SG other:F:SG. (Harrell, 1962:183) ['He went and settled in another country.']

(25)

msaw I- Iwad ysebbnu. go:PF:3:PL to- DEF- river:M:SG wash:IMPF:3:PL. (Harrell, 1962:183) ['They went to the river to do their laundry.']

gadi is always followed by an imperfect verb form and indicates future (Harrell, 1962:183-4). For most speakers gadi as an auxiliary is invariable in number and gender. As a participle, the feminine is gad(y)a and the plural is gad(y)in: (26)

u gadin nsufu. (kara038) and will:PL see:IMPF:l:PL. ['and we will see.']

Often gadi is shortened to ga- (cf. section 2.2.1):

2.2. MORPHOLOGY (27)

AND BASIC

CATEGORIES

19

ganesriw hadik. (mora408) will- buy:IMPF:l:PL this:F:SG. ['We wiU buy this.']

xess (must) takes an object pronoun ending which serves as the subject of a following imperfect verb form (Harrell, 1962:185). This auxiliary was commonly used in my data corpus. Its form is perfective, while its interpretation is imperfective. (28)

xessni ndewwez 9am. (drial44) must:PF:3:M:SG- PC:1:SG spend:IMPF:l:SG year:M:SG. ['I have to spend a year.']

(29)

xesshum yebqaw hna. (soua087) must:PF:3:M:SG- PC:3:PL stay:IMPF:3:PL here. ['They have to stay here.']

qder (can, to be able) is usually followed by an imperfect verb form: (30)

maNl-

neqders nelbes zellaba. (mohal44) can:IMPF:l:SG- N2 wear:IMPF:l:SG gown:F:SG.

['I cannot wear a zellaba (Moroccan gown).'] Of course, there are more auxiliaries than those presented here, but these were the ones used most frequently in my corpus.

2.2.2

Personal pronouns

Moroccan Arabic has independent personal pronouns and dependent clitic pronouns. They both express person, number and gender. Personal pronouns and suffixes are listed in table 2.3: Subject pronouns occur only when the verb is absent, or, since Moroccan Arabic is a 0-subject language (Rizzi, 1982), in cases of emphasis. This will be discussed in more detail in the section about Moroccan Arabic syntax. Pronominal elements are expressed by the use of suffixes. The clitics can be attached to verbs, nouns and prepositions. The paradigm for the pronominal suffixes is also given in table 2.3. A clitic pronoun attached to a noun expresses a relation of possession or connectedness. I will use the term possession for this group. More details are to be found in the section about nouns. Direct and indirect pronominal objects are attached to the verb, as can be seen in (31) and (32), respectively:

20

CHAPTER

sing -

plur -

1 2 2 3 3 1 2 3

2. MOROCCAN

m-f m f m f m-f m-f m-f

ARABIC

independent personal pronouns ana nta nti huwa hiya lina ntuma huma

GRAMMAR:

AN

OVERVIEW

pronominal suffixes -(n)i/-ya (after vowel) -k/-ek (after consonant) -k/-ek (after consonant) -u/-h (after vowel) -ha -na -kum -hum

Table 2.3: Personal pronouns and suffixes (31)

seftha (kara060) see:PF:l:SG- PC:3:F:SG ['I saw her.']

(32)

9ezbatek (kalb410) please:PF:3:F:SG- PC:2:SG ['It pleased you.']

The pronoun can also be suffixed to a preposition thus creating a prepositional phrase: (33)

m9a- ya with- PC:1:SG ['with me']

The preposition l(i)- (to) occurs in prepositional phrases attached to a verb: (34)

ka- ygululi- k (soua283) DU- tell:IMPF:3:PL- to- PC:2:SG ['They tell (to) you...']

The combination I- -j-clitic can be attached to a verb which already has a suffixed object: (35)

sifethum.I- i (Harrell, 1962:139) send:PF:3:M:SG PC:3:PL- to- PC:1:SG ['He sent them to me.']

2.2. MORPHOLOGY

AND BASIC

CATEGORIES

21

However, when a verb already has a first or second person object suffixed to it the combination /- +clitic follows the verb independently: (36)

qeddmek li- hum. (Harrell, 1962:140) introduce:3:M:SG- PC:2:SG to- PC:3:PL ['He introduced you to them.']

Possessive sentences axe verbless, and contain the preposition 9end- (with) and a clitic: (37)

9endwith-

ek 'asdiqa' (Lkoundi, 3) PC:2:SG friend:M:PL

['You have friends.'] This prepositional phrase behaves as a verb in many respects. Ennaji (1982) writes about this phenomenon, to which I will return when I discuss negation: The fact that PPs can occur with negative and interrogative markers which normally appear with verb stems might suggest that they are verb substitutes but as they cannot be conjugated (i.e. carry tense markers) it is difficult to decide whether they belong to the verb group. (1982:22)

2.2.3

Nouns and adjectives

Moroccan Arabic nouns can be either masculine or feminine. Usually, nouns ending in -a axe feminine while most other nouns axe masculine. Nouns have fixed gender, whereas adjectives can be masculine or feminine, by adding feminine a-. Sometimes feminine nouns can be derived from masculine nouns by adding -a, as in malik (king), malika (queen). In general, Moroccan Arabic nouns and adjectives show singular and plural marking, a few nouns also have a dual form. There axe severed possible rules to turn a singular into a plural form, but the choice which rule applies to which singulax form cannot be predicted. The two major plural forms for nouns and adjectives are the sound and the broken plural. The sound plural is formed by adding -in, -a or -(a)t to the singulax form, e.g. haza pi. hazat (thing-s), ferhan pi. ferhanin (happy). Plurals formed by an internal change in the stem, with or without the addition of an ending, are called broken plurals: e.g. bermil pi. bramel (barrels), bab pi. biban (doors). Moroccan Arabic adjectives always follow the noun and they correspond in number and gender, as can be seen in (38) and (39): (38)

ktab mezyan book:M:SG good:M:SG ['A good book']

22 (39)

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

ktub mezyanin book:M:PL good:M:PL ['Good books']

The dual form, used only for a few nouns that denote time or quantity, is accompanied by a plural adjective: (40)

9amayn mezyanin year:M:DUA good:M:PL ['Two good years']

The definite article I- is invariable for all genders and numbers. It assimilates completely to a following stem-initial alveolar or postalveolax coronal consonant (i.e. a consonant articulated with the tip of the tongue): (41)

Ibab DEF- door:M:SG ['The door']

(42)

Imasaktl DEF- problem:M:PL ['The problems']

(43)

rrial DEF- man:M:PL ['The men']

Unassimilated I- becomes le- when followed by a consonant cluster: (44)

lemra DEF- woman:F:SG ['The woman']

According to some authors, when a noun has the definite article, the accompanying adjectives must have one as well. However, there does not seem to be general agreement on this matter: Otten (1983:923) is very explicit about the use of adjectives and articles. According to him, an attributive adjective must always be accompanied by a definite article if the noun is also definite, or else the resulting construction can be interpreted as a nominal (or equational) sentence. Abdel-Massih (1973:41) is not as explicit as Otten, but he gives examples which make clear that he is of the same opinion. Sa'id

2.2. MORPHOLOGY

AND BASIC

CATEGORIES

23

(1950:75) also writes that if the noun is definite, the adjective is usually definite, too. Vein Mol (1983:12) expresses the same idea: adjectives have the definite article mainly when they are accompanied by a definite noun. On the other hand, Harrell (1962) seems to have a different viewpoint: If the noun it [the attributive adjective] modifies has the definite article, there is considerable variation as to whether the attributive adjective does or does not take the definite axticle. In general, the definite article with the adjective in such a case implies a superlative meaning, e.g. lehwayez l-mezyanin 'the best clothes' (...). In cases where the Moroccan definite article with the noun does not indicate specific previous reference, the attributive adjective usually does not take the definite article. (1962:205) In Mercier's work (1956:192-193) both opinions are represented from another point of view: Adjectives can be either "épithète", or "attributive". In the first case the adjective has the definite article and must be interpreted as in his example "Mon frère malade n'est pas venu". (My sick brother has not come.). However, when the adjective is used attributively, it does not have the definite article, and then its meaning is as in "Mon frère, malade, n'est pas venu". (My brother, sick, has not come.). These different opinions on a subject which in my view is quite essential for Moroccan Arabic grammar are rather confusing. Perhaps they should be viewed in the light of regional variation, although most of the studies cited are based on the Northern urban dialects. Another possible explanation for the difference in opinion is that in some descriptions of Moroccan Arabic the influence of Classical Arabic is more prevalent than in others. Van Mol (1983:51) states about Classical Arabic that the attributive adjective agrees with the noun in definiteness. This would place the analyses of Otten, Sa'id, Abdel-Massih and Van Mol closer to Classical Arabic than Mercier's or Harrell's. It should be borne in mind, however, that Ferguson (1959a) has pointed out that the modern colloquial Arabic dialects do not directly descend from Classical Arabic. This whole matter is a prime example of the fact that there is no unanimity as to what variety of Moroccan Arabic is considered standard. In the data corpus used for this study, the speakers did use the definite article as described by Otten (1983), and others. However, it was not always possible to hear whether or not a definite article was assimilated or absent, when the initial consonant of the relevant noun was a coronal: (45)

1haza Hi ka- tderrni? IDEF- thing:FEM:SG that DU- hurt:IMPF:3:F:SG- PC:1:SG DEFmakla! Ima'kulat food:F:SG DEF- food:PL

(s-) sahiya, l(DEF-) delicious:F:SG DEF-

24

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

mawadd Igeda'iya! (mora221) nutritious matter:PL DEF- daily:F:SG ['The thing that hurts me? The food! The delicious food, the daily nutritious matters!'] In (45), clearly 1-ma'kulat (s-)sahiya is meant to be an NP, as 1-mawadd 1-geda'iya, and not an equational sentence. In the first NP it is not possible to hear whether an (assimilated) definite article is used, but since the corresponding noun is definite (because 1- is used), (s-) should be interpreted as being present. (Note that in both NPs the non-animate plural noun is accompanied by a feminine singular adjective, according to the rules of Classical Arabic.) In another example, (46), a lengthening of the initial consonant s- is heard: (46)

m9a- na s-surinami wahed (hasb025) with- PC:1:PL Surinamese:M:SG one:M:SG ['There is one Surinamese (man) with us.']

The context does not allow a definite article here. It is more appropriate, therefore, to assume that the speaker hesitated. With respect to its use, the definite article in Moroccan Arabic differs somewhat from Dutch or English. In generic expressions Dutch and English usually have plural without an article, whereas Moroccan Arabic only uses a singular form with a definite article: (47)

ssbe9 xtir DEF- lion:M:SG dangerous:M:SG ['The lion is dangerous. (Lions are dangerous.)']

Indefiniteness is expressed by the absence of a determiner, as in (48), or by the presence of the determiners si (one or another, some) or wahed (one; combined with the definite article), as in (49) and (50): (48)

mra woman:F:SG ['(A) woman']

(49)

si mra some woman :F:SG ['Some woman']

2.2. MORPHOLOGY AND BASIC

CATEGORIES

(Moroccan Arabic)

25

(Dutch and English)

N N Det

N

A

Det

A

sgira (little)

een (a)

klein (little)

\

N meisje (girl)

Figure 2.1: Indefinite NP (50)

wahed Itmra one:M:SG D E F - woman:F:SG ['A woman']

Figure (2.1) presents a Moroccan Arabic indefinite NP-construction, with the adjective following the noun: The indefinite article is parenthesized because in many cases it may be absent. If wahed is used as a numeral, it appears postnominally, corresponding to the noun in terms of gender: (51)

mra wehda woman:F:SG one:F:SG ['One woman']

All other numerals invariably appear before the noun, and do not influence congruence between noun and adjective. Demonstrative pronouns occur independently, or in combination with a noun. In the latter case, they correspond with the noun in number and gender, as in examples (52), (53) and (54): (52)

dak rrazel DEM:M:SG D E F - man:M:SG ['That man']

(53)

dik leDEM:F:SG D E F ['That woman']

mra woman:F:SG

26 (54)

CHAPTER 2. MOROCCAN ARABIC

GRAMMAR:

AN

OVERVIEW

duk lewlad DEM:PL D E F - child:M:PL ['Those children']

An exception is had (this) which, like wahed, is invariable for masculine, feminine, singular and plural. Of course, wahed is only used in the singular. Examples are (55), (56) and (57). (55)

had rrazel DEM D E F - man:M:SG ['This man']

(56)

had lemra DEM D E F - woman:F:SG ['This woman']

(57)

had lewlad DEM D E F - child:M:PL ['These children']

W h e n a noun is preceded by a demonstrative, the use of the definite article is obligatory; also t h e accompanying adjective must have the definite article as in (58), or else the resulting construction will be interpreted as a nominal or equational sentence, as in (59): (58)

duk lewlad lekbar DEM:PL D E F - child:M:PL D E F - big:M:PL ['Those big children']

(59)

duk lewlad kbar DEM:PL D E F - child:M:PL big:M:PL ['Those children are big.']

W h e n the demonstrative is used independently it is preceded by ha-: (60)

haduk mezyanin DEM:PL good:M:PL ['Those are good.']

2.2. MORPHOLOGY

AND BASIC

27

CATEGORIES

f wahed 1 \ DEM J

.

.

{ DEM }

(the)

bent

?gl?a

(girl)

(little)

['A (DEM) little girl'] Figure 2.2: Moroccan Arabic definite NP

\

i (the) \ DEM J

(little)

(girl)

'Het kleine meisje' ('Dit/dat kleine meisje') ['The little girl' ('This/that little girl')] Figure 2.3: Dutch and English definite NP

28

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

In figure (2.2) a Moroccan Arabic definite NP-construction is given. A second tree is presented in figure 2.3 in order to visualize the difference between Moroccan Arabic and Dutch or English. In figure (2.2) DEM and wahed axe not completely interchangeable: when wahed is used, the adjective does not have the definite article; when had or any other demonstrative pronoun is used, the adjective must have the definite article. Syntactically, an indefinite construction with wahed behaves like a definite NP construction. Therefore, wahed is given in figure 2.2 (definite NP), although its meaning is indefinite. The definite NP-construction in figure (2.2) shows the most striking difference with Dutch: between a demonstrative (had, (ha)dak, (ha)dik, (ha)duk) or wahed and a noun, the use of a definite article is obligatory.

Possession can be expressed in different ways: First of all, with dyal (of) or the shorter form d-, comparable to English " o f ' and Dutch "van" constructions: (61)

zzmel dyal had rrazel DEF- camel:M:SG/PL of DEM:M:SG DEF- man:M:SG (Harrell, 1962:202) ['That man's camel(s).']

(62)

zznaqi d- lemdina (Harrell, 1962:202) D E F - street:PL of- DEF- city:F:SG ['The streets of the city']

A pronoun directly attached to the noun in order to express possession is only used in representing the relationship between family members, parts of the body, or in some fixed expressions, e.g.: (63)

razelha (Harrell, 1962:203) man:M:SG- PC:3:F:SG ['Her husband']

In the case of alienable possession, personal pronouns are suffixed to the preposition dyal: (64)

ddar dyalD E F - huis:F:SG van['My house.']

i (Harrell, 1962:203) PC:1:SG

2.2. MORPHOLOGY

AND BASIC

CATEGORIES

29

The 'construct state' is a construction that is used relatively rarely, compared to the d- and dyal constructions. It consists of a noun or an adjective followed directly by a noun or a pronoun, e.g. hanut baba-ha (her fathers shop) (Harrell, 1962:191). The first term of the construct phrase is the possessed entity, which almost never takes a pronoun ending or definite article (there are only very few exceptions). The second term is the possessor: (65)

ras I9am (kara053) head:M:SG DEF- year:M:SG ['Newyear']

Between the two terms of the construct phrase only a definite article or demonstrative can occur: (66)

rezlin had Iweld (Harrel:191) foot:M:PL DEM DEF- boy:M:SG ['This boy's feet']

Only a limited set of words can occur as the first term of a construct phrase, while any semantically appropriate noun may occur as the second term. When the first term of the construct phrase is a feminine noun the final -a changes into -1: (67)

sebba (Harrell, 1962:129) reason:F:SG ['Reason']

(68)

sebbti (Harrell, 1962:129) reason:F:SG- PC:1:SG ['My reason']

2.2.4

Prepositions

Prepositions are usually short and invariable. They govern nouns, as in (69), or pronouns, as in (70): (69)

sme9tha fe- rradiu. (soual82) hear:PF:l:SG- PC:3:F:SG on- DEF- radio:M:SG ['I heard it on the radio.']

30 (70)

CHAPTER maNl-

2. MOROCCAN

ARABIC

GRAMMAR:

AN

kayen fiha masihiyin bezzaf. be:PTC:M:SG in- PC:3:F:SG christian:M:PL much

OVERVIEW

(kara071)

['There are not many Christians in it (Egypt).'] In Moroccan Arabic, prepositions cannot be stranded. Ennaji (1982:24) has given the following examples (orthography adjusted to the one used in the present study): (71)

xellat Fatima l-beztam fe-d-dar. (Ennaji, 1982:30i) She left Fatima the wallet in-the-house ['Fatima left the wallet in the house.']

(72)

* d-dar xellat Fatima l-beztam f . (Ennaji, 1982:30ii) The-house she left Fatima the wallet in ['The house Fatima left the wallet in.']

Some frequently used prepositions are f(i)~ (in, at), I- (to), m9a (with) and 91a or (before a vowel) 9el (on), dyal (of) takes an exceptional position because it is only used in possessive constructions. Its syntactic behavior is usually prepositional, though.

2.2.5

Adverbs

Adverbial constituents or invariable adverbs occur in any position. An example of a sentence-initial adverb is given in (73): (73)

merra tatkun rugba bas ndir Sometimes DU- be:IMPF:3:F:SG desire:F:SG in order to do:IMPF:l:SG sihaza. (mora029) some- thing:F:SG ['Sometimes I have the desire to do something.']

Adverbs also occur in final position, as in (74): (74)

kunt hezzam, xeddam berra. (mora068) be:PF:l:SG hairstylist work:PTC:M:SG outside ['I was a hairstyler, working outside the house.']

And adverbs may occur in the middle of a sentence, between other constituents, such as verb and P P , as in (75):

2.3.

SYNTAX

(75)

31

ka- nemsi nhar wahed le- Imedrasa. (kala004) DU- go:IMPF:l:SG day:M:SG one:M:SG to- DEF- school:F:SG ['I go to school one day.']

In the literature on Moroccan Arabic grammar, adverbs or adverbial constituents do not seem to be a topic of interest. In chapter 7 it will become clear, though, that bilinguals favor the direct proximity of an adverb as a switch site. One of the reasons may be that the occurrence of adverbs does not seem to be governed by strict rules.

2.3

Syntax

2.3.1

Basic word order

Classical Arabic is a typical VSO-language. In Moroccan Arabic, several word orders are possible. Ennaji (1982) remarks about word orders in Moroccan Arabic: Unlike English which is strictly an SVO language, Moroccan Arabic has two basic word orders: an unmarked VSO order used in any context, and a marked SVO order used only under special discourse conditions. (1982:14) Harrell (1962) is quite explicit about the basic word order in Moroccan Arabic (1962) According to him the subject usually follows the predicate, as in (76) and (77): (76)

zaw ddyaf (Harrell, 1962:160) come:PF:3:PL DEF- guest:M:PL ['The guests have come.']

(77)

msa huwa (Harrell, 1962:160) go:PF:3:M:SG he ['He went.']

About the occurrence of a prestated topic, a concept frequent in the Arabist literature, Harrell (ibid.) writes the following: (...) A common expansion of the elementary simple sentence (...) is the prestated topic. The prestated topic is a noun or a pronoun which is placed at the beginning of a sentence and which refers to a pronoun, either independent, suffixed or signaled by verb inflection, occurring later in the sentence. (1962:160-161) The notion of prestated topic corresponds to what is often called left-dislocation. An example of this is the following sentence (78):

32 (78)

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

had ssezra, raha fi-ha DEM:SG D E F - tree:F:SG look- PC:3:F:SG in-PC:3:F:SG

OVERVIEW

lDEF-

kenz (Harrell, 1962:161) treasure:M:SG ['There is a treasure in this tree.'] (The literaly meaning of ra is "look" (imperative). Other forms of the verb rarely occur. Usually ra appears with nouns, pronouns, or pronominal endings.) Harrell goes on to say that (...) in many cases a noun which seems to be the subject of a following verb is separated from the verb by a slight pause and is to be interpreted as a prestated topic rather than directly as the subject. as in the following example (79): (79)

lemgarba, kayaklu tlata wella reb9a D E F - Moroccan:M:PL DU- eat:IMPF:3:PL three or four

deof-

lDEF-

merrat fennhar (Harrell, 1962:161) time:F:PL on- DEF- day:M:SG ['(As for) the Moroccans, they eat three or four times a day.'] In Boudali's opinion (1984:56), both VSO and SVO word orders are possible in Moroccan Arabic, without restrictions with respect to the use of one or the other. In my corpus both word orders occur. An example of SVO is given in (80), while (81) is an example of VSO: (80)

ana I

maNl-

kanesrebs Ibira. (kadal51) DU- drink:IMPF:l:SG N2 DEF- beer:FEM:SG

['I do not drink beer.'] (81) is an example of VSO: (81)

weq9u masakil (mora093) happen:PF:3:PL problem:M:PL ['Problems happened (occurred)']

In the data used for this study the pronominal subject was often not expressed, which made it impossible to determine whether a sentence was VSO or SVO, as in (82):

2.3.

33

SYNTAX

(82)

dewwez zuz de- ssimanat (ahma243) spend:PF:3:M:SG two of- DEF week:F:PL ['He spent two weeks']

Unlike in English and Dutch, yes/no question sentences do not show inversion in Moroccan Arabic. There is no difference in word order between interrogative and affirmative sentences. Interrogative sentences can be introduced by an interrogative pronoun or adverb. This pronoun or adverb occurs sentence-initially and its presence does not require a word order different from the one in affirmative sentences: (83)

9las mawhy N l -

ka- yehtaflus bih? (kara071) DU- celebrate:IMPF:3:PL- N2 with- PC:3:M:SG

['Why don't they celebrate it?'] Yes/no questions can be introduced by the interrogative particle was, which is functionally comparable to French est-ce-que. The use of was is not obligatory: because of a different intonation pattern, ambiguity as to whether a sentence is affirmative or interrogative can be avoided: (84)

9endwith-

u leflus? PC:3:M:SG DEF- money:M:PL

['He has the money. / Does he have the money?'] (85)

was 9endINTERROGATIVE with-

u leflus(?) PC:3:M:SG DEF- money:M:PL

['Does he have the money?']

2.3.2

Clause t y p e s

In the present section, several frequently used clause types will be discussed. First the type without a verb form will be considered. In subsequent paragraphs, the verbid clauses (simple and complex) will be discussed. Finally, some remarks on negation will be made, because its form will turn out to be closely linked to the presence or absence of a verb form. 2.3.2.1

N o m i n a l sentences

Moroccan Arabic exhibits a frequent use of nominal sentences, consisting of a subject (noun or pronoun) and a predicate (noun, pronoun, adjective, adverb or prepositional phrase). Characteristic is the absence of a finite verb:

34 (86)

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

s9iba 9li- na (moha266) difficult:F:SG for- PC:1:PL ['It is difficult for us.']

(87)

huma they

I'asdiqa' (moral39) DEF- friend:M:PL

['They are the friends.'] The verb kan is used (in the perfect tense) to indicate the past: (88)

rrazel mrid DEF- man:M:SG ill:M:SG ['The man is ill.']

(89)

rrazel kan mrid DEF- man:M:SG be:PF:3:M:SG ill:M:SG ['The man was ill.']

Sometimes the predicate precedes the subject, e.g. when the predicate is a prepositional phrase with a possessive meaning: (90)

(kanet) 9end(be:PF:3:F:SG) with-

i wahed nnedra mezyana PC:1:SG one DEF- opinion:F:SG good:F:SG

bezzaf (kala435) very ['I have (had) a very good opinion.'] huwa, hiya and huma (personal subject pronouns, 'he', 'she' and 'they', respectively) often occur in stead of a copula, as in (91): (91)

hada huwa Iferq. (kalb271) DEM:M:SG he:3:M:SG DEF- difference ['That is the difference.']

2.3.2.2

Raising

Raising involves the movement of a constituent, in this case the subject NP, from a subordinate (a 'that-' clause) to a matrix clause (Ennaji, 1982:178). In Moroccan Arabic, subject clitic pronouns axe an integral part of the verb, which is the reason why raising of a subject clitic pronoun does not occur in Moroccan Arabic: clitics will

2.3.

SYNTAX

35

occur anyway in their original position. However, it is possible to consider raising as involving copying of a clitic to the verb of the matrix sentence, while maintaining it in the subordinate clause at the same time. A sentence like (92) results after raising the subject in (93) (orthography adjusted to the convention used in the present study): (92)

dher (belli) Moh.am.ed razel mezyan. seems (that) Mohamed man nice

(Ennaji, 1982:455i)

['It seems that Mohamed is a nice man.'] (93)

dher Mohamed (belli) razel mezyan. seems Mohamed (that) man nice

(Ennaji, 1982:455ii)

['Mohamed seems to be a nice man.'] The raised NP (Mohamed) is subject of the main clause. Raising to subject can occur across many clauses, as pointed out by Ennaji (1982:182, for a thorough analysis). According to Ennaji, raising to object exists in Moroccan Arabic as a general grammatical rule, which applies in order to map complement subject NPs into matrix clause object NPs. The following examples are again taken from Ennaji (1982:199). In (94) raising to object has not applied; in (95) it has. (94)

ka-ndenn I think

(belli) nzeh Mohamed. (that) succeeded Mohamed

(Ennaji, 1982:521i)

['I think that Mohamed has succeeded.'] (95)

ka-ndenn I think

Mohamed (belli) nzeh (Ennaji, 1982:521ii) Mohamed (that) succeeded

['I believe Mohamed to have succeeded.'] I will not go deeper into the details of reusing here. Those who are interested in the topic are referred to Ennaji (1982). 2.3.2.3

Simple verbal clauses

A verb, with or without a subject and an object (direct, indirect, adverbial or prepositional) constitutes a simple verbal sentence. Simple sentences are not necessarily short. With the help of coordinating conjunctions, adjectives, etc, the sentence can be expanded. The finite verb is highly inflected and is marked for number, person, gender and tense. Therefore, the pronominal subject is only present in case of emphasis, as was the case in (80) above. Otherwise it is dropped as in (82).

36 2.3.2.4

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

Complex sentences without a conjunction

Complex sentences consist of a main clause and one or more subordinate clauses. This is not the place to discuss all possible kinds of complex sentences, so only some frequently used constructions will be considered. More detailed information can be found in the publications cited in the first part of this chapter. In a complex sentence without a linking conjunction the subject of the main clause functions as a 'prestated topic' for the subordinate clause, in the sense of Harrell (1962:160-161, see section 2.3.1): (96)

ka- nhawel nterzem 9li- h (kala020) DU- try:IMPF:l:SG translate:IMPF:l:SG for- PC:3:SG ['I try to translate for him.']

In this example, the auxiliary is the finite verb and the second verb functions as an infinitive: it is marked for number, gender and person, but ka- or ta- is absent (cf. above in 2.2.1, the section about the verb in Moroccan Arabic). 2.3.2.5

Relative clauses

According to Ennaji (1982) and Boudali (1984), Hi (who), -men (who) and -as (what) are relative pronouns which can introduce a relative clause in Moroccan Arabic, the latter two being used as the complement of a preposition (-men occurs only sporadically). A relative clause is always linked to another Noun Phrase (NP) which is part of another clause. Examples of relative clauses with Hi, -men and -as are the following (taken from Ennaji with some orthographical changes): (97)

Iweld Hi msa (Ennaji, 1982:51iii) DEF- boy:M:SG who leave:PF:3:M:SG ['The boy who left']

(98)

lemra m9a-men seftek (Ennaji, 1982:51i) DEF- woman:F:SG with-whom see:l:SG- PC:2:SG ['The woman with whom I saw you']

(99)

rrazel 9el-as tqelleqti (Ennaji, 1982:51ii) DEF- man:M:SG with-whom be angry:PF:2:SG ['The man with whom you became angry']

When the element in the main clause to which the relative refers functions as the object of a preposition or as the possessor of a noun in a relative clause with Hi, the relativized element is repeated as a pronoun ending, -u in (100) and -ha in (101) are instances of pronoun retention:

2.3. (100)

SYNTAX

37

was sefti rrazel lli srit menn-u INTERROG see:PF:2:SG DEF- man:M:SG that buy:PF:l:SG from-him ddar? (Harrell, 1962:164) DEF- house:F:SG ['Did you see the man that I bought the house from?']

(101)

ha Ibent lli xutha here DEF- girl:F:SG that brother:M:PL- PC:3:F:SG ka-yeqraw fbariz (Harrell, 1962:164) DU-study:IMPF:3:PL in- Paris ['Here's the girl whose brothers are studying in Paris.']

Relative clauses can be restrictive, appositive or free. Restrictive means that there are additional members of the class referred to by the relativized element in the main clause which do not share the same qualifications of the subordinate clause. An example is: (102)

/hulandiyin lli m9a- ya / - groep kaDEF- Dutch:M:PL who with- PC:1:SG in- group DUyhetmu lemgarba (kalal40) disdain:IMPF:3:PL DEF- Moroccan:M:PL ['The Dutch who are in my group disdain the Moroccans.']

Appositive relative clauses often function as parenthetical comments which are marked by a specific intonation pattern that clearly sets them apart from the main sentence as in (103) (Ennaji, 1982): (103)

had rrazel, lli 9ad xrez, sahbi DEM DEF- man:M:SG who just leave:PF:3:M:SG friend:M:SG PC:1:SG (Ennaji 1982:122) ['This man, who just left, is my friend.']

Except for intonational and semantic differences, restrictive and appositive relative clauses are also syntactically distinct from each other. For instance, only restrictive relatives co-occur with indefinite antecedents (Ennaji, 1982). (104) is an example of a restrictive relative clause with an indefinite prestated topic: (104)

ktab lli sra li-ya bba twedder book:M:SG that buy:PF:3:M:SG for-PC:l:SG father:M:SG lose:PF:3:M:SG (Ennaji 1982:125i) ['A book which my father bought for me was lost.']

38

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

(105) is not a possible variant, since appositive relatives cannot have indefinite antecedents: (105)

* ktab, lli sra li-ya bba, twedder book:M:SGthat buy:PF:3:M:SG for-PC:l:SG father:M:SG lose:PF:3:M:SG (Ennaji 1982:125ii) ['*A book, which my father bought for me, was lost.']

In (106) and (107), the antecedent is definite, which makes both restrictive (as in (106)) and appositive relativization (as in (107)) possible: (106)

lektab lli sra li-ya bba DEF- book:M:SG that buy:PF:3:M:SG for-PC:l:SG father:M:SG twedder (Ennaji 1982:125iii) lose:PF:3:M:SG ['The book which my father bought for me was lost.']

(107)

lektab, Hi sra li-ya bba, DEF book:M:SG that buy:PF:3:M:SG for-PC:l:SG father:M:SG twedder (Ennaji 1982:125iv) lose:PF:3:M:SG ['The book, which my father bought for me, was lost.']

Free relatives do not have an overt antecedent: (108)

ne9mel Hi bgit (kala292) do:IMPF:l:SG what want:PF:l:SG ['I do what(ever) I want.']

It is not easy to give just a few simple facts about the use or the nature of relatives in Moroccan Arabic. Harrell (1962:164-166) dedicates a number of pages to 'the relatives Hi and -at?, but Ennaji (1982:37), on the other hand, concludes that Hi is a complementizer, while -as and -men are relative pronouns. Then again, Boudali (1984:9) states that lli, -as and -men all three axe relative pronouns, whereas belli, bas and was are complementizers. In addition, Harrell (1962:164) remarks that "As an alternative to lli and a preposition with a pronoun object, -as can be used after the preposition in those cases where a human being is not referred to." Ennaji (1982:32) also complicates matters by writing that "In M[oroccan] A[rabic], the relative pronoun -men is

2.3.

SYNTAX

39

marked for animacy (...) but as seems to be in free variation since it can relativize humans and non-humans alike." However, Boudali (1984:7) states that -men refers to animate antecedents, -as to inanimates, and Hi to both. These statements make clear that it is not sufficient to name the relative pronouns, since no unanimity seems to exist about what a relative pronoun is, or how exactly it is used. T h e thorough analyses by Ennaji and Boudali throw a different light on the problem than the traditional, purely descriptive approach, but they do not solve the problem. 2.3.2.6

C o m p l e x s e n t e n c e s w i t h u (and)

In Moroccan Arabic u is used to express coordination: (109)

Ihut ssgir ka- ytah u lDEF- fish:M:SG/PL DEF- small:M:SG DU- fall:IMPF:3:M:SG and DEFhut lekbir ka- yebqa. (mora021) fish:M:SG/PL DEF- big:M:SG DU- stay:IMPF:3:M:SG ['The small fish fall through (the nets) and the big fish stay.']

Apart from coordination, u often also indicates a semantic modification of the second clause with respect to the first: (110)

xrezt u wqe9 li- ya hadak sleave:PF:l:SG and happen:PF:3:M:SG to- PC:1:SG DEM:M:SG DEFsabotage (moral 70) sabotage (Dutch) ['When I left, sabotage was committed against me.']

2.3.2.7

S u b o r d i n a t e clauses i n t r o d u c e d b y an i n t e r r o g a t i v e

A sentence introduced by an interrogative pronoun or adverb can function as a subordinate clause: (111)

maNl-

9refts as bgiti te9ni biknow:PF:l:SG- N2 what want:PF:2:SG mean:2:IMPF:SG with-

ha (Lkoundi, 4) PC:3:F:SG ['I don't know what you (want to) intend with it.'] Subordination does not change the form of these interrogative sentences.

40 2.3.2.8

CHAPTER

2. MOROCCAN

ARABIC

GRAMMAR:

AN

OVERVIEW

Conditionals

In Moroccan Arabic, a semantic distinction is made between two types of conditionals: counterfactual and non-counterfactual conditionals. Counter}actual conditional sentences are introduced by kun: (112)

kun kan 9endil be:PF:3:M:SG with-

i patroondiploma, raPC:1:SG master-certificate (Dutch) look-

neqder nhell Imahall dyalh PC:3:M:SG can:IMPF:l:SG open:IMPF:l:SG DEF- shop:M:SG ofi (mora061) PC:1:SG ['If I would have had a master-certificate, then I could have opened my own shop.'] Non-counterfactual (113)

conditional sentences are introduced by ila:

ila seft ana /rasi, kanqul: when see:PF:l:SG I in- head:M:SG- PC:1:SG DU- say:IMPF:l:SG 9las? (kalb051) why ['When I see myself, I say: Why?']

2.3.2.9

Negation

In Moroccan Arabic, negation is expressed by the use of masi or ma...s. The continuous form masi occurs before the word or (verbless) sentence to be negated: (114)

masi not

heqq (kalal76) right:M:SG

['It is not right.'] The discontinuous negation ma...s is mainly used for the negation of verbs. The negation encloses the verb, including all its affixes: (115)

m a - kayebqaws fe- ddyurhum (kala315) N l - DU- stay:IMPF:3:PL- N2 in- DEF- house:F:PL PC:3:PL ['They don't stay in their houses.']

2.3. (116)

41

SYNTAX maNl-

kemmeltfinish:PF:l:SG-

has (youa089) PC:3:F:SG- N2

['I did not finish it.'] In a sequence of verbs, the finite verb, ususally an auxiliary, is negated, as in (117): (117)

ma- neqders nelbes zellaba (mohal44) N l - can:IMPF:l:SG- N2 wear:IMPF:l:SG zellaba:F:SG ['I cannot wear a zellaba (Moroccan gown).']

When the sentence does not contain a finite verb but a participle or a predicatively used adjective, both negation forms can be used. According to some speakers, the use of the continuous form makes the negation stronger when both forms are used: (118)

ma- galis (soua040) N l - expensive:M:SG- N2 ['It is not expensive.']

(119)

masi gali not expensive:M:SG ['It absolutely is not expensive.']

However, the participles derived from the existential verb kan never have a continuous negation: (120)

ma- kayen(s) muskil (heard in Morocco) N l - be:PTC:M:SG- (N2) problem:M:SG ['There is no problem.']

(121)

* masi kayen muskil

The first example is grammatical, because kayen is enclosed by the two parts of the discontinuous negation, whereas the last example is ungrammatical because the continuous negation is used, which is not allowed with forms of the active participle kayen. When prepositions axe used with a noun in a nominal sentence, the negation form has to be continuous, as in (122): (122)

masi fe- ddar (Harrell, 1962:156) not in- DEF- house:F:SG ['Not in the house.']

42

CHAPTER 2. MOROCCAN ARABIC GRAMMAR: AN

OVERVIEW

However, when prepositional constructions are used with pronominal endings in nominal sentences, the use of the discontinuous negation is obligatory, as in (123): (123)

maNl-

9endwith-

is (Harrell,1962:156) PC:1:SG- N2

['I don't have (it).'] Consistently, (124) is not possible: (124)

* masi 9endnot with-

i PC:1:SG

In summary, a discontinuous negation is generally used for the finite verb. If there is no finite verb, the discontinuous negation can be used for a form that has some characteristics in common with a finite verb. In any case, this verb or verblike form should be marked for number and gender, in combination with either person or tense: When the main verb contains a tense marker, the discontinuous negation is used. In all other cases masi, the continuous negation, is required.

2.4

M o r o c c a n , Classical Arabic and B e r b e r

In the previous chapters I have sometimes referred to Classical Arabic and Berber. Although these two languages are beyond the scope of this study, I will briefly present some of the most striking differences with Moroccan Arabic. To illustrate the difference between Moroccan Arabic and Classical Arabic, some examples will be given from Heath (1989). Sentences (125) and (127) axe Classical Arabic, while (126) and (128) are their Moroccan Arabic equivalents. (125)

sa-'a-dhab-u I-will-go

gad-a-n li-d-dar-i li-'a-ra 'ab-T. (Heath la) tomorrow to-the-house in-order-to-see my-father.

['I will go tomorrow to my home to see my father.'] (126)

gadi n-msi gedda le-d-dar bas nsuf bba. (Heath lb) will I-go tomorrow to-the-house in-order-to see father. ['I will go tomorrow to (my) home to see (my) father.']

There are some similarities between (125) and (126), but the following two examples are very different from each other, (127) and (128): (127)

mada ta-f9al-u huna-ka? what you-are-doing there ['What are you doing there?']

(Heath 2a)

2.4. MOROCCAN,

CLASSICAL

feature radicals VSO SVO 0-subj. 0-copula case N/A order def. art. demonstr.

43

ARABIC AND BERBER

Moroccan Arabic

+ + + + +

Classical Arabic

+ + -

+ + + + +

-

+ +

prenom.

prenom.

Berber

+ + + + + -

+ -

postnom.

Table 2.4: A comparison between Moroccan Arabic, Classical Arabic and Berber (128)

as ta-d-dir temma? what you are doing there?

(Heath 2b)

['What are you doing there?'] I will not go into details of the differences between Moroccan and Classical Arabic. My only aim here is to show that Classical Arabic is a language that must be learned by Moroccans, because it differs from their native language. Table 2.4 schematically presents distinctive features for Moroccan Arabic, Classical Arabic and Berber. With respect to some features, e.g. radical structure, VSO word order, 0-subject, 0-copula and noun/adjective order, the three languages show the same characteristics, while Morocan Arabic and Classical Arabic can be grouped together on the basis of the existence of a definite article and prenominal use of the demonstrative pronoun. Berber and Moroccan Arabic, on the other hand, have in common that they both allow SVO word order, while they have no case marking. Berber and Classical Arabic have no characteristics in common. Although the three languages are quite different from each other, they apparently also have much in common. The languages have been spoken side by side for centuries, which implies that extensive lexical borrowing has taken place, sometimes with semantic change. Some examples: the Classical Arabic word for "house" is bait, while in Moroccan Arabic bit is used for "room". Classical Arabic has zadid for "new", which is ¿did in Moroccan Arabic and uzdid in Berber. Other facts about Berber and Classical Arabic will be given throughout the text, whenever relevant.

Chapter 3 The Linguistic Situation in Morocco Linguistically speaking, Morocco is a very complex and dynamic country. The speakers reported on in the present study, were born and raised in Morocco and have lived there for a shorter or longer period of time (between two and twenty years). Their linguistic norms and behavior were, of course, formed by life in Morocco, or, if they were very young when they left, by members of their families in the Netherlands. For this reason it is important to give a brief overview of the sociolinguistic situation in Morocco, since it can clarify the informants' behavior and attitudes in the Netherlands.

3.1

Morocco

Morocco is the Westernmost part of North Africa. Together with other North-West African countries it constitutes the Maghreb, which is Arabic for West. The name "Morocco" is derived from Marrakes, one of the capital cities in South Morocco. Moroccans and other Arabic speakers use the name l-megrib to indicate their country. In Morocco there are some rather inaccessible mountain chains: the Rif in the North, the Atlas mountains in Central and Western Morocco, and the Anti-Atlas in the South. These geographical facts have influenced the linguistic situation in the country immensely. The tribes in the mountains were rather isolated from external influences and could therefore maintain their original Berber languages, while those in the lower areas were subject to Arabization more readily. Berber was originally spoken by all indigenous inhabitants in the Maghreb countries. It is a member of the Hamito-Semitic language family and related to OldEgyptian and to the Kushitic, Chadic and Semitic languages. Arabic belongs to the Semitic language group. Together with Islam, the Arabic language was brought to Morocco in the seventh and eigth centuries. The Arabization process started more

46

CHAPTER

3. THE LINGUISTIC SITUATION IN MOROCCO

than twelve hundred years ago, but about one third of the Moroccan population is still Berber-speaking. Initially, Arabic was only introduced in urban areas. The indigenous population was converted to Islam but did not adopt the Arabic language en masse. Arabic spread from the Arabized cities to the surrounding areas only bit by bit and at a very slow pace. In the Arabic dialects dating from the early period (the so called pre-hilalic mountain dialects) traces of a Berber substrate are noticeable. Berber is still spoken today in large parts of the mountainous areas. In the eleventh century, Morocco underwent another wave of invasions. This time the invaders were the Banu Hilal and other Beduin tribes from the Arabic peninsula who settled in the Atlantic plains and on the East Moroccan plateaus. At present, dialects related to these Beduin dialects are still spoken in large parts of the country. Another period of great importance, particularly linguistically, was the time of the French protectorate, from 1912-1956. While most of Morocco was in French hands, the Northern paxt of the country was a Spanish protectorate, of which the two Spanish enclaves, Ceuta and Melilla are still reminders. Over the centuries, Morocco has thus developed from a primarily Berber-speaking nation to a complex multilingual society through influences which are mainly historically determined. In the following sections the individual languages (Berber, Arabic, French, other) will be discussed in more detail, followed by a section about language attitudes and interaction between the different languages.

3.2 3.2.1

Individual Languages Berber

It is not correct to speak of Berber as if it were one language. In fact, several Berber languages are spoken in Morocco and they can be divided into three major groups, which are separate both linguistically and geographically. Figure 3.1 illustrates the distribution. In the Northern Rif-mountains tarifit is spoken; in the central Atlas tamazight, and in the Southern Atlas and Anti-Atlas the Berber variety is called taselhit or selha. These three language varieties are not entirely mutually intelligible (Bentahila and Davies, 1988). In the Netherlands, in particular in Utrecht, more than half of the Moroccan immigrants is Rifian (Van der Meer, 1984:124). This is partly a consequence of chain migration, but it is also a result of depopulation measures by the Moroccan government. Several times in history Rifians were leading rebellions against the then dominating powers. Rifians were involved in rebellions during the period of the French protectorate, but also in attacks directed against King Hassan II in 1971 and 1972. (Van der Meer, 1984:41-42)). Of the total Moroccan population (25 million, according to Storcken (1986:2)) more than one third uses Berber as an everyday language (Abbassi, 1977). Outside

3.2. INDIVIDUAL

LANGUAGES

47

±

Liiache El Kar cl'Kcbb t Oucujni

CjubUnct A temmour El-Jadlda

Afadlf

F-Fei T T u i G2S3

mountain

Rabat •

town w i t h urban

i Tinger •

town w i t h n m i x t u r e

dlaloct dialect of

urban and b e d u i n / m o u n t a i n dialects [M^

beduin

Safi •

town w i t h b e d u i n

dialect

tar.hclhit

(or

dialect

shellia)

Lainazight ^ ^

MOROCCAN ARABIC

unnDER LANGUAGES

tarific

[based

on

Otten

Figure 3.1: Morocco

(1982:98^

48

CHAPTER

3. THE LINGUISTIC SITUATION IN MOROCCO

Morocco Berber languages are spoken by about 35% of the Algerian population, and in Libya by 20% of the population. In Tunesia and Egypt only a few isolated tribes axe Berber-speaking (Schippers and Versteegh, 1987: 52). Tuareg is also related to the Berber languages and is spoken by the Tuaregs in the countries of the Sahara desert: Southern Algeria, Mauritania, Mali, and Niger (Ennaji, 1982:7). Berber is not a written language. Only the Tuaregs have an old written form of their language which is called tifinagh. Penchoen (1973) makes the following remarks about this alphabet. Although the alphabet has been handed down over a very long period it does not seem ever to have been used for recording history, literature or anything of the sort. (...) Modern use of the alphabet seems to be restricted to writing of formula« of magical significance on rocks, finger writing on the palm of a friend's hand on occasions when silence is required, and, more recently, for writing letters. (1973:3) Berbers are considered illiterate (by Berbers as well as by non-Berbers) as long as they do not have a written command of Arabic. According to Keil (1986:154-155) Berber is a language with low social prestige, but high affective value: it is used as a vehicle for old songs, stories and legends. The use of Berber expresses ethnic solidarity; it is a home language (Gravel, 1979) and it is the language for everyday conversations (Forkel, 1980). According to Bentahila and Davies (1988) Berber is being lost more rapidly than ever before in recent years, in favor of Moroccan Arabic. Berber and non-Berber speakers show little regret about this shift. Bentahila and Davies cite a girl (1988:3) "who remarked that her father was more upset about her wearing European-style clothes than the fact that she had forgotten her Berber." Gravel (1979:Ch 4) also notices a growing decrease in the use of Berber.

3.2.2

Arabic

In the case of Berber I have pointed out that it is not correct to speak of one language. This is even more true of Arabic. The spoken variant is generally referred to as Moroccan Arabic, while for literary, religious and other formal purposes Classical Arabic is used. The official national language is Classical Arabic. According to Ennaji (1988:9) "Morocco decided in early independence to choose Classical Arabic as the official language because it is codified and standardized, unlike Moroccan Arabic, and above all because it represents the language of the Coran, and of literary and scientific tradition." Moroccan Arabic, like Berber, is acquired as a first language, whereas Classical Arabic is learned at school. Classical Arabic derives directly from Koranic Arabic, and is considered highly prestigeous. It is used throughout the Arabic world, which has enormous advantages. From the Atlantic to the Indian Ocean it is used in books, newspapers, and political and other formal speeches, so a speech by PLO-leader

3.2. INDIVIDUAL

LANGUAGES

49

Yassir Arafat, for example, is understood perfectly by educated people in Morocco. Keil (1986) calls Classical Arabic the language that unifies in space and time. For centuries it has remained more or less unchanged, and it is used by many speakers all over the Arabic world. The disadvantage of Classical Arabic is that it is a difficult language which always has to be learned in years of hard work because it is nobody's native tongue. Otten (1982:94) compares the linguistic situation in the Arabic countries with that of medieval Europe: the vulgar languages like French and German were the native languages, but in school Latin had to be learned for formal and literary purposes. In the Arabic world the dialects (i.e. Moroccan, Syrian, Algerian, Jordanese etc.) are as distant from Classical Arabic as are French, Spanish, Italian from Latin. Whereas the European languages have developed into literary languages, the Arabic dialects are still unwritten. Proposals to introduce an official written version of Moroccan Arabic axe not taken seriously or they axe viewed as threats to the Arabic heritage (Hoogland, 1983; Bentahila, 1983). The discussion about the appearance of a Moroccan Arabic-Dutch dictionary by Otten in 1983 was typical. Moroccans living in the Netherlands argued vigorously against introduction of a written form of Moroccan Arabic by means of the Latin alphabet which would help Moroccan children learning Arabic in the Netherlands. (This discussion took place in Utrechts Nieuwsblad, November 12th, December 6 and 19, 1983.) Ferguson (1959a) has explained why, in his view, the modern Arabic dialects do not descend directly from Classical Arabic. Since the modern Arabic dialects have characteristics in common which cannot be attributed to Classical Arabic, he assumes a common ancestor of the dialects which he calls the koine and which differs from Classical Arabic in several respects. The lingua franca in everyday Moroccan life is Moroccan Arabic (Gravel, 1979). Whereas Berber, the other native language in Morocco, is never learned as a second language, Moroccan Arabic is acquired by Berbers as soon as they seek contact with the world outside, as in market-places or in small towns (Van der Meer, 1984:141). In contacts with Moroccans outside the native community, Moroccan Arabic is learned and spoken as a consequence of communicative need. Women are more often illiterate and Berber monolinguals than men. Van der Meer (1984:25) states that in his experience, Berbers speak Moroccan Arabic easily and fluently. The way they express themselves in Moroccan Arabic is not essentially different from the way native speakers of Moroccan Arabic from a comparable social-economic background speak it. There axe some intonational differences, and he also remarks that Berbers produce short sentences in Moroccan Arabic, and with a higher pitch, just like they do in Berber (Van der Meer, 1984:25,141). Abbassi (1977) estimates the number of monolingual Berbers in Morocco at about 45% of the Berber population. In the case of Moroccan Arabic it is not correct to speak of one language or dialect. In fact, Moroccan Arabic is the collective name for all the Arabic varieties spoken colloquially in Morocco. In recent years, under influence of mass media like radio and television, the variety used in the Western cities is spreading throughout

CHAPTER

50

3. THE LINGUISTIC SITUATION IN MOROCCO

the country ¿is a kind of new standard form of Moroccan Arabic (Otten, 1983:12ff; Abbassi, 1977:19). Gravel (1979:Ch.4) predicts a continuation of the important role of Moroccan Arabic, and even an increase in importance for a 'higher' variety (i.e. closer to Classical Arabic). In some studies, a continuum is assumed, ranging from dialectal or vulgar Arabic at one extreme to Koranic Arabic at the other (e.g. Meiseles 1980). Keil (1986) distinguishes Classical Archaic Arabic and Modern Literary Arabic, while Bentahila uses the term Classical Arabic in his publications to designate the literary language of both past and present. Otten (1983:12fF) also uses the term Classical Arabic to indicate the formal forms of Arabic, one of which is Modern Standard Arabic. Since the distinction between the non-vulgar forms of Arabic suffers from vagueness, I will follow Bentahila and Otten and use the term Classical Arabic, as opposed to Moroccan (colloquial) Arabic. The focus of this dissertation will be on the variety used in informed situations by speakers of whom the majority has little or no educational background in Morocco. Their writing skills in Classical Arabic are minimal and as such they are representative for a large group of Moroccans in the Netherlands.

3.2.3

French

Since the period of the French protectorate, French has played an important role in Moroccan society. It is learned in primary school, and it is used on a wide scale in many domains. According to Bentahila (1983:14-15) it is the language often used in radio and television broadcasting, in newspapers and other written publications, in administration, science, technology, and medicine (see also Forkel, 1980; Ennaji, 1988). French functions as the link with the non-Arabic world, but it is also often used alongside Arabic in documents, road signs, names of shops, and so on. It has high social status among Moroccans (Keil, 1986; Ennaji, 1988) and it is often associated with prestige and power (Schippers and Versteegh, 1987:93). According to Heath (1989:21) the use of French peaked during the nineteen sixties and early seventies, and it is probably on the wane now. In Morocco many people with no more than an elementary (oral) knowledge of French use a variety with a heavy Moroccan accent which differs sharply from French as it is spoken in France. Bentahila (1983:120) found that speakers of 'High Moroccan French' are evaluated more highly in a language attitude survey than speakers of (Moroccan) Arabic, with respect to social status, education and general sophistication. This was not necessarily true for speakers of 'Average Moroccan French'. Bentahila (1983) thoroughly investigated the alternate use of Arabic and French and the attitude of the speakers towards the two languages. He concluded that Moroccan Arabic has a purely practical value, while the value of Classical Arabic is largely aesthetic and that of French is clearly instrumental. Each is recognized as useful in Moroccan society, and bilinguals evidently

3.2. INDIVIDUAL LANGUAGES

51

feel that each is necessary to the Moroccan. (1983:165) In a study in which language contact and language use in Casablanca were investigated, Abbassi (1977) distinguishes different varieties of spoken French in Morocco: High Standard French, Moroccan Standard French and Sub-Standard French ('pidgin' or 'petit-negre'). Multilingualism was also investigated by Gravel (1979) in a study among students of English in Rabat. On the basis of his findings, he predicts that the role of French in Morocco will not decrease in the future.

3.2.4

Other languages

In this section, some languages which do not or no longer play dominating roles will be discussed briefly. Spanish, which was only spoken in the North which comprised the Spanish protectorate, has almost entirely been supplanted by French. One of Gravel's findings is that Spanish is losing its importance (Gravel, 1979:Ch.4). Van der Meer (1984:25) writes that the older generation in Northern Morocco prefers Spanish over French, while for the younger generation the opposite is true. He observes that Spanish doctors axe popular among Moroccans who live in the North. On the national level, Spanish plays a minor role, therefore it will not be incorporated in the following analysis. English is of growing importance. It is used in the classroom and is also often heard in the cinema (Bentahila, 1983). Moroccans learn English in secondary school and it is taught at the universities. Gravel (1979) expects that the role of English will become more and more like that of French.

3.2.5

The Jewish community in Morocco

Several groups of Jews live in Morocco: the Toshiban or native Jews, who have been living in Morocco since long before the Arabs came, and the Sephardic Jews, who came from Spain in the fifteenth century. Around 1950, about 250,000 Jews lived in Morocco. Originally, each important city had its own Jewish quarter, called the mellah. Heath (1989) writes about the position of the Jews: The Jews were mostly artisans and merchants; the wealthier ones owned land which was rented out, or (if agricultural) was leased to Muslim peasants who worked the land in exchange for a percentage of the harvest. Intercommunal marriage was almost unheard-of, and social relations with Muslims were primarily limited to workplace and commercial interactions (chiefly involving men). Even in the colonial period there was fairly little interaction, since the Jews had their own schools (often using Fr[ench]). Under these conditions, special Jewish communal dialects developed and stabilized over the centuries. Although most of the Jews were descended

52

CHAPTER

3. THE LINGUISTIC

SITUATION

IN

MOROCCO

from Spanish and Portugese Jews who fled the purges of the Inquisition in the fifteenth and sixteenth centuries, the attested Jewish Arabic (Judeo-Arabic) dialects of Morocco appear to be basically Moroccan in origin (rather than imported from the Iberian peninsula). However, they are archaic in this context, and tend to reflect older MCA [Moroccan Colloquial Arabic] patterns which most Muslim dialects have given up; they also have a number of lexical items of foreign origin (Hebraisms, especially in religious matters, and some Hispanisms going back to the Judeo-Spanish (Ladino) which many of t h e m spoke in Spain). (1989:1112) Since its existence, many Jews have emigrated to Israel. Currently, there are about 20,000 Jews left in Morocco. Among Jews, Bentahila and Davies (1988) have noticed a shift in recent years from Moroccan Arabic to French as a home language.

3.3

Language Attitudes in Morocco

The languages discussed in the previous section are each associated with several often overlapping domains and values. Keil (1986:154-155) provides a systematic comparison of the languages with respect to function, dimension, concretion, extension and connotation. In the literature two or three languages or varieties are generally discussed: Schippers and Versteegh (1987) compare the use of Classical Arabic with dialectal Arabic, and Bentahila (1983) concentrates on the use of Arabic versus French. In figure 3.2 the domains and values axe ordered as to the languages they are associated with, summarizing the discussion in the previous sections. The data have been taken from the literature. It appears from figure 3.2 that the most frequently used languages in Morocco, i.e. Berber (B), Moroccan Arabic (MA), Classical Arabic (CI A), French (F) and Spanish/English/German (other) can be placed in a continuum. Berber and Moroccan Arabic overlap in several domains; they can both be used in many situations. Berber and Classical Arabic cannot be used alternately and neither can Berber and French. Figure 3.2 shows a clustering of Berber and Moroccan Arabic on the one hand, and Classical Arabic and French on the other. Particularly the use of Berber or Moroccan Arabic is often associated with the same values and domains, as Forkel (1980) has also observed. He states that Classical Arabic is most important in writing, as is French. In oral language use, language choice depends on the degree of formality: when the situation is very formal, Classical Arabic is used. In informal situations (everyday use) both colloquial Arabic and Berber are spoken 1 . '"Im schriftlichen Bereich herrscht das Hocharabische (und Französische). Im mündlichen Bereich hängt die Wahl der Sprache von der Formalität ab: Ist die Formalität hoch, (...) so ist das Hocharabische üblich. In informellen Situationen, d.h. vor allem bei Alltagsgesprächen, werden dagegen das Dialektarabische und das Berberische gebraucht."

3.3. LANGUAGE ATTITUDES

IN MOROCCO

Value/domain ethnic solidarity old stories, songs, legends high affective value oral language home language everyday use ignorance, poverty illiterateness and subordination socially low jokes and fun in-group mother tongue (best mastered) intimacy lingua franca unifying practical patriotism official, formal sacred, religious aesthetic official language in education out-group morality and tradition unpleasant in everyday communication panarabic unifying in space and time successful career written language wealth and power formal upper and middle class media (newspapers, radio and tv) informal among the educated (oral) high educational prestige high social prestige instrumental law administration, finance science classroom tourism cinema modernism, future contact with non-Arabic world

53

B°

MA

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

Cl A

F

other

X X X X

X X X X X X X X X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X X X X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

"Of course, Berber values and domains only apply to Berber speaking Moroccans. Figure 3.2: Languages attitudes in Morocco.

54

CHAPTER

3. THE LINGUISTIC SITUATION IN MOROCCO

Figure 3.2 shows that only a limited set of domains and values is restricted to either Berber or Moroccan Arabic, whereas Classical Arabic and French have more separate domains and values, so their overlap is less striking. Bentahila (1987) wrote about the values of the many languages that Moroccans speak: Moroccans cope happily with three, four or five languages in everyday use because each of these languages has its own value for them, its own usefulness, its own identity. If one language merely duplicated the functions of another, it would probably be quite readily abandoned. (1987:2) Bentahila and Davies have thoroughly investigated the loss of Berber (Bentahila and Davies, 1988). One of their findings was that speakers of Berber easily abandon their mother tongue in favor of Moroccan Arabic. Although the languages differ greatly in terms of grammar, they are sociolinguistically more closely related, which is clear from figure 3.2. To the speakers who abandon Berber, the unifying and practical character of Moroccan Arabic seems more important than the ethnic solidarity and folklorism associated with Berber languages. In reading figure 3.2 it must be borne in mind that it does not represent an absolute state. Often the values and domains must be seen from the viewpoint of the speakers. A Berber peasant, for example, will value Moroccan Arabic much more highly than an upper class inhabitant of one of the cities. In other cases, more objective distinctions have been made, as in the domains of finance and administration (French) or contacts with the Arabic world (Classical Arabic). In these cases, the viewpoints of the speaker do not play such a role of importance. Although figure 3.2 has its shortcomings, it illustrates the gradual and overlapping character of the domains and values of the languages used in Morocco.

3.4

Education

Ennaji (1988:9) notices that 78% of the Moroccan population is illiterate. The use of Classical Arabic and French is therefore restricted to a minority of educated people. Schooling is possible from the age of four or five in Kindergartens and especially in Koranic schools, where the children are made familiar with Classical Arabic by learning the Koran by heart and copying passages from it. Moroccan children go to primary school when they are six or seven and they stay there for at least five years (Bentahila, 1987). From the start Classical Arabic is the medium of instruction. The first two grades are taught exclusively in Arabic (Hermans, 1983:28-31), while French is introduced in the third grade. The children may well be more or less familiar with French by that time, under the influence of television programmes or by hearing it spoken around them. It is not easy for the child to learn French, because French is not related to the language(s) it already knows. Bentahila (1987) reports that it is not unusual for children to be rather confused for a shorter or longer period of time:

3.4.

EDUCATION

55

I have seen a child in the early stages of learning Fench attempting to write it from right to left across the page: this is of course because he was used to writing Arabic in this way. Such difficulties axe, however, of the developmental type and they are usually overcome by the end of primary education (...). (1987:2)

Major subjects in primary school are Arabic, French and arithmetic. Whatever the child's first language, it will always become at least bilingual when it attends school. The process of becoming bilingual or multilingual is unidirectional. A Berber always acquires Moroccan Arabic as a second language, and later Classical Arabic and French, while native speakers of Moroccan Arabic acquire Classical Arabic and French. Berber can only be a native language, whereas French and Classical Arabic never are. Figures on school attendance vary in the available literature. According to Hermans (1983:29), 90% of the children in urban areas attend school; in the rural areas this is only 15%. Figures given by Storcken (1986:31) are 80% in the cities and 25% in the country, whereas Wagtendonk (1982:274) states that 95% of the children in the cities attend school, and 35% in rural areas. It is clear, however, despite the fact that there has been compulsory education since 1963, that there are still large groups of children who never receive primary education. This is especially true for girls. According to Hermans (1983:29) only 36% of all primary school children are girls, in rural areas this percentage is even lower. This is mainly due to the fact that in certain areas or communities education is not considered necessary for girls (Hermans, 1983:29; Wagtendonk, 1982:276; Storcken, 1986:31). However, as the infrastructural conditions improve, the degree of education is increasing. Secondary education takes seven years and is meant to prepare for higher education. Of all children attending primary school, one third goes on to secondary education, which amounts to 550,000 children yearly. Of those, only 15,000 pass the final exams ('baccalauréat'). Universities are growing: between 1975 and 1981, the number of university students increased from 25,000 to 80,000. The only full scale university is the one in Rabat. Selected faculties (mainly Law and Arts) have been established in Fes, Marrakes, Casablanca, Oujda, Tetouan and Meknes. Technical education is almost non-existant, except for the Mining School, the Institute Mohamed V, INAV (where technical engineers, agronomists and veterinarians are educated) and the Engineering School in Mohammadia near Casablanca. There are two medical faculties, one in Rabat and one in Casablanca, while the first Dental School was founded only a few years ago. University students receive a moderate grant from the government. In addition, 15,000 Moroccans are studying in France (Storcken, 1986:31-32).

56

3.5

CHAPTER

3. THE LINGUISTIC

SITUATION

IN

MOROCCO

Diglossia

The Moroccan sociolinguistic situation has often been characterized in terms of diglossia. Ferguson (1959b) has defined diglossia as follows: Diglossia is a relatively stable language situation in which, in addition to the primary dialects of the language (which may include a standard or regional standards), there is a very divergent, highly codified (often grammatically more complex) superposed variety, the vehicle of a large and respected body of written literature, either of an earlier period or in another speech community, which is learned largely by formal education and is used for most written and formell spoken purposes, but it is not used by any sector of the community for ordinary conversation. (1959b:336) In this now famous article, Ferguson (1959b) gives examples of diglossic communities, in the Arabic countries and elsewhere. Classical Arabic represents the high variety and the Arabic dialects the low varieties. It is clear from the preceding discussion that Morocco has many features of diglossia, and this is confirmed by several authors. Forkel (1980) characterizes the Moroccan situation as diglossic 2 , because Arabic exists in two forms, Classical or high Arabic and Moroccan dialectal Arabic. The relation between these two forms is a diglossic one. Dialectal Arabic is the natural mother tongue and everyday language, while Classical Arabic is used for written and other formal purposes. Grabe (1979) also describes diglossia in the Moroccan context, elaborating on Ferguson's definition. He concludes that in Morocco, Classical and Moroccan Arabic are in a diglossic relation, on the basis of the following arguments (p.15): 1. Diglossia is a stable language situation that may last for centuries. "Arabic Diglossia seems to reach as far back as our knowledge of Arabic goes." (Ferguson, 1959, p.327) 2. Diglossia relegates specific functions to each language (the High language-superposed variety and the Low language-dialect variety). 3. Diglossia distinguishes a prestigious language from a regional dialect. Speakers of both languages tend to view the dialect as a non-language even though it will be used for most day-to-day purposes. 4. Speakers of both languages tend to believe that the superposed variety is more beautiful. 2 D a s Arabische existiert in zwei Formen, dem Hocharabischen (klassischen Arabisch) und dem marokkanischen Dialektarabisch. Diese beiden Sprachformen stehen in einem diglossischen Verhältnis zueinander, wobei das Dialektarabisch die natürliche Mutter- und Umgangssprache, das Hocharabische die Schrift- und Hochsprache darstellt.

3.5.

DIGLOSSIA

57

5. In the case of Arabic the superposed language has a religious base. 6. The superposed vaxiey has a large body of written material which is held in high regard. 7. In diglossic situations the dialect variety is the mother language while the superposed variety is learned through the educational and religious institutions. 8. "In diglossia there axe always extensive differences between the grammatical structures of H and L" (Ferguson, 1959, p.333). 9. Lexically, many words will be different words in the H and L varieties, with the same meaning. 10. Phonologically, the L variety is similar to the H variety, but the H variety will have more phonemes from either a parasystem or a subsystem (Ferguson, 1959, p.335). Diglossia must be distinguished from a standard vs. dialect situation. Both varieties can be used for ordinary conversation in a standard vs. dialect situation, while in a diglossic situation the speech community never uses the high variety for ordinary conversation (Ferguson, 1959b:337). Diglossia should also be distinguished from bilingualism. According to Ferguson (1973), Bilingualism exists in those situations in which the two languages in question both have relatively complete levels of discourse, whereas diglossia exists in a situation in which the burden is shared by certain levels in the one, and different levels in the other (Ferguson, 1973:83). If the statements in the literature about diglossia are compared with the information presented in the previous section on domains and values, it becomes clear why only Moroccan Arabic and Classical Arabic are considered to be in a diglossic relation: both varieties have clearly distinct domains and values, while other possible language pairs (e.g. French and Classical Arabic) show too much overlap, or do not cover all domains (e.g. Moroccan Arabic and French), while Classical Arabic and Moroccan Arabic axe complementary. Sociolinguistically, Berber and Moroccan Arabic often have the same function, but the relation between Berber and Classical Arabic is not a diglossic one, in the sense meant by Ferguson. As Forkel (1980) remarks, "Sociolinguistically, Berber and dialectal Arabic are related to high Arabic in the same way, although the relation between Berber an high Arabic is not a diglossic one, since Berber and Arabic are linguistically not closely related languages." 3

3 D a s Berberische (...) steht, soziolinguistisch gesehen, zum Hocharabischen in einem ähnlichen Verhältnis wie das Dialektarabische, obwohl es sich hierbei nicht um ein Diglossieverhältnis handelt, da Arabisch und Berberisch nicht näher miteinander verwandt sind.

Chapter 4 The Moroccan Community in the Netherlands In this chapter, the Moroccan community in he Netherlands will be the focus of attention. After a short introduction about the situation in the Netherlands in general, a more thorough description of the Moroccan community in Utrecht will be given. In many ways the situation in Utrecht is typical for the situation in other large cities in the Netherlands. From what follows it will become clear that Utrecht is a very suitable place to carry out a study in which Moroccans take a central position.

4.1

Moroccans in the Netherlands, an Overview

In January of 1987, 14.6 million people were living in the Netherlands (Centraal Bureau voor de Statistiek, 1988). Out of these, 567,900 had a nationality other than Dutch (3.9%). 123,400 were of Moroccan, and 161,700 of Turkish origin. Together they constitute half of the non-native Dutch population. Most people from Surinam and from the Netherlands Antilles are not included in these figures, because many of them axe Dutch citizens. The same applies to stateless Moluccans (estimated at about 20,000). In the nineteen sixties the industrialized countries in Western Europe had an acute shortage of unskilled labor. This desperately needed labor force was not available. In the Mediterranean countries, on the other hand, the labor reserve was enormous and thousands of people, mostly young men, came to work in the Netherlands as well as in other countries of Western Europe. Most of them were Italian, Spanish, Yugoslavian, or Greek. A few years later, Turkey and Morocco became more and more popular as a source for foreign workers. In 1980 more than 10,000 Moroccans settled in the Netherlands, as opposed to fewer than 5,000 in 1984. If the socio-economic situation does not change, expectations are that about 17,000 Moroccans will have migrated to the Netherlands between 1985 and 1990. Not only the number of migrants but also the social make-

60

CHAPTER 4. THE MOROCCAN

COMMUNITY

IN THE

NETHERLANDS

up of the migrant groups has changed. As far as women are concerned, until 1975 mainly young women without children or with only a few small children moved to the Netherlands. Later on the migrant women were mostly older, with completed families. In recent years, most immigrants are the young partners of second generation Moroccans living in the Netherlands. The tendency for young Moroccans to look for (Moroccan) partners in the Netherlands is expected to increase (Buitenlanders Bulletin, November 1985:4). In the nineteen seventies and eighties there was in the large Dutch cities a tendency for immigrants to move from the inner city to the periphery. Some important reasons for this shift were the reunion of the families and the moving of autochtones to newly built satellites. In section 4.3 this will be discussed in more detail with respect to the situation in Utrecht which is typical for the situation in other Dutch cities. The remainder of this chapter presents an outline of the Moroccan community in Utrecht which in some ways —but not in others, as will become clear— is characteristic of other large Dutch cities. The main purpose is to sketch the background of the informants whose speech constitutes the basis of the investigation presented here.

4.2

Moroccans in Utrecht

Since it is not possible here to describe Moroccan life in Utrecht in full detail, the following sections will center around two major themes, the living situation of the Moroccan community and its language use. The information was obtained in three different ways, with the help of Mustapha Lkoundi who is familiar with the Moroccan community in Utrecht: Literature A frequently used source for statistical data was Vreemdelingen in Utrecht (1988), a publication of the Municipal Department of Statistics in Utrecht. It provided information on where Moroccans and other immigrants live in Utrecht and about the stratification in terms of age. Other sources have been used as well: references are given in the text. The literature mainly provided information on geographical aspects of the Utrecht Moroccan community. Community Centers In neighborhoods where many Moroccans live we visited community centers during the summer of 1986. Informal talks with Dutch and Moroccan staff members helped us to gain insight in Moroccan life in the different districts of Utrecht. Topics in those conversations were usually the composition of peer groups, their interests, their mobility and their participation in language classes, as well as their activities in the community center. During these informal conversations notes were made which were worked out

4.3. SOCIAL AND GEOGRAPHICAL

ASPECTS

61

afterwards. They provided us with social and geographical as well as sociolinguistic information. Observations Observations were made in several places in Utrecht: in shopping halls, in cafes, in a mosque, and in the streets. Sometimes these observations led to informal conversations, especially when Mustapha Lkoundi was the observer, as a logical consequence of his role as participant observer. Notes were taken whenever possible and they were worked out afterwards. The observations were used mainly to corroborate our impression about language choice in everyday situations. As could be expected, these different ways of collecting information created quite some overlap, which was exactly what we aimed for to achieve as much objectivity as possible. The chance that the data would be colored by the viewpoints and possible bias of one particular source was minimized in this way and the fact that the same information was often given by different sources makes it more reliable.

4.3

Social and Geographical Aspects

Utrecht is the fourth largest city in the Netherlands with a total of 229,326 inhabitants (in January 1987). In Utrecht 22,624 people are of foreign origin which constitutes 9.9% of all inhabitants. 10,249 have Moroccan nationality, while 6,425 axe Turkish. No figures are availabe about immigrants who live in Utrecht illegally. Utrecht is different from other major Dutch cities in terms of the size of the various immigrant groups. Of all immigrants in Utrecht, 45.3% are Moroccan and 28.4% Turkish, while on the national level these figures are 21.7% for the Moroccan community and 28.5% for the Turkish population. This makes Utrecht somewhat exceptional in the Netherlands, since of all immigrants in the Netherlands 4% live in Utrecht. For the Moroccans, this percentage is 8.3%. Utrecht also holds this special position with respect to the number of Greek immigrants: 26% of all Greeks in the Netherlands live in Utrecht. Utrecht is different from other Dutch cities in still another way: At the beginning of 1985, 6,305 Surinamese and Antillians were living in Utrecht, according to figures from the Municipal Department of Statistics. Bovenkerket al. (1985:39) have noticed that the number of Surinamese in Utrecht is higher than the national average, but below the average in other major cities in the Netherlands. I will not go into possible explanations for the presence in Utrecht of relatively many Greeks and few Surinamese and Antillians, because I want to concentrate on the situation of the Moroccans in Utrecht. One of the reasons for the overrepresentation of Moroccans in Utrecht has to do with recruitment practices. Companies in Utrecht started to employ Turks and Moroccans more recently than businesses in other Dutch cities and because contracting

62

CHAPTER

4. THE MOROCCAN

COMMUNITY

IN THE

NETHERLANDS

with Turkey began earlier than with Morocco these local companies 'missed' the mainstream of Turks. A large group of Moroccan workers did not come on a contract basis, but on their own initiative, and this is typical for the Moroccans in Utrecht (Bovenkerk et al., 1985:43). About 30% of the Moroccans did not migrate directly from Morocco, but came via France or Belgium. Moroccan migration to the Netherlands was mainly a matter of chain migration: the first migrants pulled family and friends along, like a chain. For this reason Moroccans in the Netherlands and in Utrecht mainly originate from three regions: the Rif mountains in the North, the Casablanca region in the West and the Anti-Atlas area in Southern Morocco. Chain migration and origin from specific regions are less evident among Turkish migrants. They were selected from anonymous lists and, as a consequence, they originate from all over Turkey. For them personal contacts played a less important role in recruitment than for Moroccans. Utrecht started to recruit Turks relatively late, and mostly not directly from Turkey: many were already in the Netherlands after having worked in Belgium or Germany. Furthermore, infrastructural differences (e.g. with respect to the labor exchanges) between Turkey and Morocco contributed to the different nature of the Turkish and the Moroccan immigrants. In practice this implies that, during the sixties, the growth of the Turkish population in Utrecht was greater than the growth of the Moroccan community. In the seventies this difference more or less evened out, and in the eighties the Moroccan population increased much more than the Turkish population. In the seventies the recruitment of migrant workers stopped, because they were not needed anymore. The stream of immigrants did not stop, though. The new immigrants were members of the foreign workers' families, who came to join their husbands and fathers in Northern Europe. Within the Turkish group, this process has more or less been completed. Among the Moroccans it is still going on, though not on such a large scale as was expected a few years ago. One of the main reasons may be the government's policy in recent years of discouraging migration to the Netherlands. Figure 4.1 is a plan of Utrecht with the different neighborhoods identified by number. The circled neighborhoods are the ones with more than 10% foreign inhabitants, about which more statistical details are given in figure 4.3. The neighborhoods consist of smaller clusters and the stratification of the individual clusters in each neighborhood may differ considerably. Figure 4.2 shows the concentration of immigrants in Utrecht. Percentages are calculated for the small clusters which together form neighborhoods, which is the reason why some of the areas in figure 4.2 with more than 10% immigrants are not included in figure 4.3. In these cases the average of the whole neighborhood is less than 10%. Between 1985 and 1988, the number of Moroccans in Utrecht increased from 9,160 to 10,985. The neighborhoods in which the Moroccan community grew faster than anywhere else are 17 (Hoograven), 19 (Kanaleneiland) and 5 (Nieuw Zuilen). In order to better understand this seeming coincidence, it is necessary to turn to the recent socio-geographical history of Utrecht with respect to foreign inhabitants.

4.3.

00 01 02 03 04 05 00 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22

SOCIAL

AND

GEOGRAPHICAL

Oude sLacI noord Oudc stad znid AZU, Iloog Catharijnc Amstcrdam.sostraalwrg, Fijlswccrd Amstcrdamsestraatwcg, Ondicp Nicuw Zuilcn Locvcnhoutflcdijk, Jagcrskadc Ovcrvccht znid Ovcrvccht noord Buitcngcbied Overvccht Votulast Tuindorp, Karcl Doormanlaan Duitcn YViltcvrouwcn Duitcn Tolsloog Oudwijk, Williclminapark Buitengcbicd oost Lunetten Tolstecg, Iloogravcn Crocscliian, Rivicrenwijk Kanalenciland Oog in al, Den HomrncI Vleutcnscwcg, Cartesiusweg Buitcngcbicd west

63

ASPECTS

x

.09 08 Pt,

22

loop 02) 20

151

. 01»,

>

v(2r 22

\S

' P i

18

F i g u r e 4 . 1 : U t r e c h t d i v i d e d i n t o n e i g h b o r h o o d s ( t h e d o t t e d lines i n d i c a t e s m a l l e r clusters within neighborhoods)

64

CHAPTER 4. THE MOROCCAN COMMUNITY IN THE

0 -2.4% 2.5-4.9% 5.0-9.9% 10.0-24.9% 25.0-50.0% not included

NETHERLANDS

•

B IS3 11 'rj1- S

•

*

Figure 4.2: Utrecht: concentration of immigrants by neighborhood

4.3. SOCIAL AND GEOGRAPHICAL

Area"

03 21 19 10 17 04 18 05 Total:

%of population is immigrant 25.8 23.2 15.2 13.9 13.2 12.9 11.3 11

%of immigrants is Moroccan 41.9 40.5 53.2 50.4 64.3 36.2 56.7 51.9

65

ASPECTS

Moroccans (absolute) 656 1929 1446 915 1256 631 850 1320 9003

%of immigrants is Turkish 35.4 42.4 28.3 27.5 18.5 41.1 18.3 29.4

Turks (absolute) 554 2016 770 498 361 716 275 748 5938

"Entries refer to neighborhoods as in figure 4.1

Figure 4.3: The eight neighborhoods in Utrecht with the highest percentage of immigrants Bovenkerk et al. (1985:32-33) have divided Utrecht into six parts which differ from each other in historical and socio-economic aspects: 1. The inner city: built within the moats, before the industrial revolution. It contains many privately owned houses. A large variety of social classes can be found here. (00, 01 and 02 in figure 4.1) 2. The old working class quarters: situated around the old inner city and built between 1870 an World Wax I for the industrial workers of those days. They contain houses of poor quality, which have been privately owned since the days they were built, (around Vleutenseweg (21) and Amsterdamsestraatweg (03) and (04)). 3. The old middle class neighborhoods: built also between 1870 and 1917, they contain big, comfortable houses which are still inhabited by the well-to-do or have been turned into student houses or offices (Wilhelminapark, (14) in figure 4.1). 4. The working class districts built between the World Wars: built for the 'better' working class. They are owned for the most part by building societies and are generally of good quality. (Nieuw Zuilen (05), Rivierenwijk (18)). 5. The middle class quarters, built between the World Wars: they consist of excellent houses which are mainly in private hands. (Oog in Al (20) and Tuindorp (11))-

66

CHAPTER

1. 2. 3. 4. 5. 6.

4. THE MOROCCAN

COMMUNITY

inner city old working class quarters middle class quarters working class quarters 1914-1940 middle class quarters 1914-1940 postwar buildings

1971 10.1 7.6 5.7 1.8 0.7 1.4

IN THE

1976 9.0 12.2 5.2 2.8 1.1 1.7

NETHERLANDS

1981 8.3 19.3 5.6 6.5 1.3 4.5

Figure 4.4: Percentages of immigrants in different parts of Utrecht (based on Bovenkerk et al., 1985:40) 6. The apartment buildings around the perimeter of the old city: built after World War II and often owned by the Municipality or by building societies, they house all social classes. Hoograven (17) and Kanaleneiland (19) were built in the fifties; Overvecht (7 and 8) in the sixties and Rijnsweerd (in 15) and Lunetten (16) in the seventies and eighties. In 1971, most immigrants lived in the inner city, while in 1981 most of them lived in the old working class quarters. Figure 4.4 illustrates this shift. When the reunion of the families began in the seventies, many immigrants were urged to look for a bigger place to live. The families would only get permission to live in the Netherlands if they had suitable living accomodations. Since there was not enough rental housing, a large group of Turkish and Moroccan immigrants were forced to buy a house (Kaufman and Verbraeck, 1986:26). Cheap houses were mainly situated in the old working class districts but often these houses were in poor condition and there was no relation between the price and the quality of the houses. Many buyers were not able to meet their high payments and were soon forced to sell their precious house only a short time after they had moved in. In the Amsterdamsestraatweg and Vleutenseweg area (21, 03 and 04 in figure 4.1) houses were cheapest, and, as a result, the Moroccan immigrants were most likely to buy houses there. Because Turkish family reunion had started earlier than the reunion of Moroccan families, the Turks also had a head start on looking for large, inexpensive housing. Even now, the number of Turks in the original immigrant neighborhoods is higher or equally high as the number of Moroccans. Moroccans often had to start looking in less obvious neighborhoods, which explains why in some districts there are more than twice as many Moroccans as Turks. Neighborhoods like Votulast (10), Kanaleneiland (19), Tolsteeg/Hoograven (17), Croeselaan/Rivierenwijk (18), Nieuw Zuilen (05) (figure 4.1) are a case in point. In the nineteen seventies, many native inhabitants of Utrecht moved to newly built satellites, like Maarssenbroek, Nieuwegein or Houten (Kaufman and Verbraeck,

4.3. SOCIAL AND GEOGRAPHICAL

0 , 4% 19,4% 13,3%

32 ni 1 . 54 9 III H08 III

67

ASPECTS

•) r

65 • : 41 p e r s o n s

585 f

45-65: 2.134

658 r

30-44: 1,466

persons

persons

29%

1 , 807 III

1 .382 f

15-29: 3 . 1 8 9 persons

37,8%

2 , 1 1 0 III

2,045 f

0-14: 4,155 persons

Figure 4.5: Moroccans in Utrecht: Age and Sex 1986:26). The apartments they left behind became available to large groups of Turkish and particularly Moroccan immigrants. This caused an enormous increase of foreign families in Tolsteeg, Kanaleneiland, Croeselaan/Rivierenwijk and Nieuw Zuilen. As fax as the distribution of Moroccans across age and sex is concerned, figure 4.5 shows that there are more Moroccan men (6,306) than women (4,679) in Utrecht. Among children younger than 14, the distribution of boys and girls is equal. However, in the high age brackets, the percentage of men increases: among Moroccans aged 45-65, 72.6% axe male and only 27.4% axe female. This is a logical consequence of the fact that not all families have been reunited. In spite of the government's attempts to stimulate remigration, it is not very popular. Remigration among Turks and Moroccans is lower than in any other foreign groups (Bovenkerket al., 1985:40). Most men who are still active in the labor force are working as unskilled laborers, e.g. in factories on assembly-lines, or in cleaning jobs. Only a small group of Moroccan women between 15 and 64 is employed: 3.3%, while 59.1% of the men in the same age group is registered as having a job. In 1983, 65% of the Moroccans in the Netherlands between age 15 and 64 did not have any education beyond primary school, while among the Dutch this was only about 20%. Among unemployed Moroccans (and Turks), 90% only have primary education (Buitenlanders Bulletin, November 1987:24). There is a decreasing demand for unskilled labor in the Netherlands. Although no exact figures axe available, it is clear that a considerable group

68

CHAPTER 4. THE MOROCCAN COMMUNITY IN THE

NETHERLANDS

of immigrants is unemployed and their number is increasing. According to Buitenlanders Bulletin (October 1987:15) unemployment among immigrants is 2.5 times as high as among the Dutch. Buitenlanders Bulletin (September 1987:19) reports that, in general, youth unemployment is decreasing in the Netherlands, while it is growing among foreign youth. Taking all this together, it can be concluded that the economic situation of Moroccans is generally weaker than that of the native Dutch people.

4.3.1

Religion

Almost all Moroccans in the Netherlands are moslem. The total number of moslems in the world is estimated at 610 million (Storcken, 1986). Of the world population, 18% axe Islamic, which implies that Islam is the second largest religion in the world after Christianity (30%). Islam is based on the Koran: texts written down by the prophet Mohamed as revealed to him by Allah. The demands made on every moslem are summarized in the five pillars of Islam: 1. shahada: the confession 2. salat: the duty to pray 3. zakat: the payment of alms to the poor 4. ramedan: the month of fasting, the ninth month of the Islamic year 5. hadj: the once-in-a-lifetime pilgrimage to Mekka Moslems in the Netherlands (Turks, Moroccans and Surinamese) are Sunnites. An important difference between the Sunnites and the Shiites, the other major group of moslems, is that the Shiites consider the descendants of the prophet the only possible leaders of the Islamic community, while the Sunnites accept a democratic leadership and axe convinced that all believers are equal before God. The Shiites mainly live in Iran, but smaller groups of Shiites are to be found in Yemen, Syria, Iraq, Lebanon and India. Of all moslems, 90% axe Sunnites. The Islamic calendar differs from the Christian calendar. It begins in 622 A.D., the yeax that Mohamed settled in Medina, and the Islamic year is eleven days shorter than the Christian yeax. The length of the Islamic yeax is based on the cycles of the moon, while Christian years and months are based on the position of the earth with regard to the sun. This means, for instance, that the month of ramedan occurs eleven days earlier each year. During ramedan men and women who are in good health (in case of illness, other rules apply) are not allowed to eat and drink during the daylight hours. Of course, in Northern countries this is particularly difficult in summertime when days are long and nights are short. In Morocco and other Islamic

4.3. SOCIAL AND GEOGRAPHICAL

ASPECTS

69

countries public life is adjusted to a different life style during ramedan, while in the Netherlands and other parts of Europe this is not the case. In the Moroccan community in Utrecht, ramedan is an important and wellattended religious event. Social control is strict: most people do not want to be seen eating or drinking during the day. Even those who are not very strict in following the Islamic rules pay attention to ramedan. The nightly meals axe important social events, taking place in the circle of family and close friends. After having eaten many people visit a mosque. When ramedan is taking place in summer people often do not return home until well after midnight. This raises problems for those who live far from the mosque, because in Utrecht there is no public transport at night. In Utrecht there are five mosques in the center and North-Western part of the city (1986). Two of them axe Turkish, two are Moroccan and one is used by the Surinamese moslems in Utrecht. In other neighborhoods there are no mosques. The many Moroccans who live there consider this a serious lack of the possibility to practice their religion. A moslem is forbidden to eat pork. Other kinds of meat should come from ritually slaughtered animals. At the beginning of 1985 there were 18 Islamic butchers in Utrecht (Bakker et al., 1985). The exact numbers axe not known, but it is clear that there are more Moroccan than Turkish or Surinamese butchers. The number of slaughter-houses that provide meat to the Islamic community is increasing.

4.3.2

Shops

The Islamic butcher-shops in Utrecht are mainly situated in and around the city center. Shops often change ownership. According to Bakker et al. (1985) Islamic customers seldom buy their meat from Dutch butchers. Turkish shops axe usually patronized by Turks and Moroccan shops by Moroccans, but in both cases members of other ethnic groups are also frequent patrons. One of the reasons for this segregation is the problem of language. The butchers often also sell other products beside meat, like bread, spices and (dried) vegetables, which are typically Moroccan or Turkish. Sometimes they also sell reading material, and now and then the shops also function as meeting-places. Islamic shops sometimes axe regaxded with suspicion by the local Dutch population. Nevertheless, an increasing group of Dutch customers shop there, since mutton or lamb is better and cheaper there than in Dutch butcher shops (Bovenkerk et al., 1985: 285). In the street market, for example at Vredenburg on Wednesdays and Saturdays, Moroccan men with loaded shopping bags are a familiar sight. Moroccans often eat fish, and it is usually bought in the street market. In recent years, market-stalls with Moroccan products like spices, olives and olive oil, but also non-food items like ornaments, clothes and music cassettes have been added to the assortment. Besides, originally Dutch stalls have expanded their assortment with Mediterranean

70

CHAPTER 4. THE MOROCCAN COMMUNITY IN THE

NETHERLANDS

and West Indian products. Sometimes they have ethnic employees in order to attract more customers (Bovenkerk et al., 1985:286). There is only one Moroccan bakery in Utrecht, since Moroccan bread can be bought from butchers and many Moroccan families eat home-made bread, just like in Morocco. As far as their other shopping needs are concerned, Moroccans patronize the same supermarkets and department-stores as the Dutch, because they are relatively inexpensive.

4.3.3

Restaurants, cafés and coffee houses

There axe only about half a dozen Moroccan restaurants in Utrecht, as well as a number of coffee houses and cafés. Kaufman and Verbraeck (1986), who investigated the use of hard drugs among Moroccans, visited Moroccan coffeehouses and cafés in neighborhoods like Lombok (Vleutenseweg, 21 in figure 4.1). In Lombok they found at least three Moroccan coffee houses and at least two Moroccan cafés with a licence for the sale of alcohol. The coffee houses are very diverse with respect to the clientele and the drinks that axe sold. A few are patronized by traditional, elderly men, who mainly drink mint tea and who consider the coffee house as a meeting-place. Other coffee houses function as gathering places for young Moroccans who go out in groups, for instance to Hoog Catharijne, a large shopping center in the middle of Utrecht. Young Moroccans from the neighborhood, but also from other parts of town where there axe no such establishments, spend their evenings in the coffee houses. The ethnic cafés ('etnische kroegen', a term used by Kaufman and Verbraeck (1986:26) to distinguish establishments where alcoholic beverages are served) have a more dubious reputation. Since Lombok can be considered the central area for the Moroccan community in Utrecht, the number of Moroccan coffee houses and cafés there may not be representative for the rest of Utrecht and it is difficult to estimate the total number of Moroccan meeting places, especially because a number of them is not officially licensed.

4.3.4

Groups and networks

Moroccans who have only recently arrived in the Netherlands often have friends and acquaintances in the closed group of people they already knew, directly or via other people, in their place or region of origin. The longer they live in the Netherlands, the more possibilities they have to move outside this closed group. This does not necessarily mean that everyone who has lived in the Netherlands for a long time has many friends from all over Morocco or from outside of Morocco. The degree of contact with people outside the family depends strongly on age, sex and the kind of work someone does. Of course, those who work in shifts do have less time to maintain friendships than those who are not. According to our observations, the patterns were different for men and women and for young and old people. It must be emphasized, though, that these patterns should not be interpreted as more

4.3. SOCIAL AND GEOGRAPHICAL

ASPECTS

71

than an indication. Of course, many people cannot be incorporated in any group or do not find it desirable to do so. Van den Berg-Eldering (1977:64-65) reports that in Moroccan families the mother is usually responsible for the raising of children up to the age of about eight. Until that age, children's lives are mostly concentrated in and around the home. Older girls stay at home until marriage and are under the direct responsibility of the mother, whereas over time boys become more and more the father's responsibility. Boys grow up to become part of the men's world outside the home. This tradition can clearly be recognized in the fact that male Moroccan adolescents are often seen together in groups. The cold and wet climate in the Netherlands, as opposed to Morocco, does not really seem to influence this habit. Besides, there are enough places that are sheltered from wind and rain where youngsters can gather, as we have found in the course of our research. This gathering in groups is quite a normal phenomenon among Moroccans and is not necessarily associated with negative factors like poor housing conditions or problems at home (although, of course, they may play a role). In general, older Moroccans often maintain contacts with family members or with other people they knew from the homeland. This is especially true for Berbers who do not speak Arabic or Dutch. Young Moroccans often have made friends in other places, like community centers. As a consequence, they do not only socialize with Moroccans, but also with the Dutch and with other ethnic groups. For the younger generation the language also plays a crucial role. Someone who only speaks Berber may have friends from the same region in Morocco, but there are few young Moroccans who only speak Berber. Most of them also speak at least Moroccan Arabic, which enables them to make friends among people from other parts of Morocco. When they speak Dutch as well, contacts with other immigrants and the Dutch are, of course, possible. Peer group pressure plays an important role in the acquisition of Dutch: Kaufman and Verbraeck (1986:60) report on a Moroccan youngster who spoke Dutch with a Surinamese accent, which indicates that his contacts with the Surinamese were more frequent than with the Dutch. In the street, in the market or in community centers, first generation Moroccan men and women are seldom seen with members of other ethnic groups, or with Dutch people. On the other hand, adolescents of Moroccan, Surinamese and Dutch descent are often seen together. Contacts between Turks and Moroccans seem to be less frequent, as was also observed by Kaufman and Verbraeck (1986). In 1986, when we made our observations, a disco dance was organized every week in Hoograven (17 in figure 4.1) where half of the 150 visitors was Dutch and the other half Moroccan, Surinamese and Turkish. Again, the Turks did not mix with members of the other groups, while Moroccans, Surinamese and Dutch frequently mixed. There are always groups of Moroccan adolescents in Hoog Catharijne (HC), the shopping mall mentioned earlier. HC functions as a meeting place, sheltered from rain, wind or cold and it is centrally situated. It can be reached by bus and train, so that the young Moroccans who congregate there do not only come from

72

CHAPTER

4. THE MOROCCAN

COMMUNITY

IN THE

NETHERLANDS

the neighborhoods around HC, but also from other parts of town or from outside Utrecht. Many regular visitors of HC are truants or are unemployed. Among the Dutch, HC has a negative reputation and among Moroccans esteem for HC is not much higher. It is often associated with the use of all kinds of drugs, which is not far off the mark, according to the investigation by Kaufman and Verbraeck (1986). However, the majority of Moroccans visiting HC has nothing to do with problematic heroin-users or alcoholics. We were told, and we observed ourselves, that in different community centers groups of adolescent Moroccans of 16 and older often have fixed routes: one day they visit community center A, the next day center B, and the following day they can be found in C. When the weather is good, groups of young Moroccans also get together outside in the parks. There are also soccer clubs which are joined by Moroccans and sports schools where they work out. Moroccan adolescents often get together in community centers. In Utrecht there are about 40 such centers (Bovenkerk et al., 1985:293). There are considerable differences between the various neighborhoods. The type, quality, character and the activities for immigrants of a community center are related to the number of immigrants in a certain neighborhood. It also depends on the length of stay of the majority of the immigrants in an area and even more on the staff's policy and the mentality of the local people. Bovenkerk et al. (1985:295) have noticed that a heterogeneous Dutch population, in which many age groups and social classes are represented is much more tolerant towards immigrants than a homogeneous population. When a neighborhood is not only populated by families but also by students and other 'non-mainstream' groups, immigrants axe frequent visitors of the community center and the atmosphere is more pleasant and tolerant than in more homogeneous neighborhoods. An example of a heterogeneous neighborhood is Votulast (10 in figure 4.1), which houses Dutch and foreign families as well as many students and other single-person households. In 1985, a group was organizing activities in Votulast especially for young Moroccans, and there were several well attended Dutch language classes. Every year on the 5th of May there was a Liberation Day party for the whole neighborhood, organized for and by the people who live there, including the Moroccans. The community center in Votulast functioned very well. In Kanaleneiland, on the other hand, with its majority of Dutch families and (since 1980) a greater number of Moroccan inhabitants than Votulast, activities for Moroccans were minimal. There were two community centers in Kanaleneiland. Sewing lessons for immigrant women were organized, but they had to be ended because of the threats from the local population who objected to the fact that the lessons were free of charge. In 1985 a meeting place for Moroccan teenagers became available, but only for a few hours a week. It is not difficult to understand why many young Moroccans prefer to spend their spare time in Votulast than closer to home in Kanaleneiland. It must be noted that the situation described here is based on observations made

4.4. LANGUAGE

PROFICIENCY

73

in 1986. The situation is very dynamic and rapidly changing and by the time this study will be published it may have changed completely. There are, for example, indications that the conditions for immigrants in Kanaleneiland have become much better as their number has increased.

4.4

Language Proficiency

Van der Meer (1984:141) estimates that about 70% of the Moroccans in the Netherlands are Berber-speaking, while the remaining 30% axe native speakers of Moroccan Arabic. It is not known how many Berbers also speak Moroccan Arabic, but a good many do. Of the people who speak only Berber, many axe older women, because their outside contacts axe limited, and a large proportion of them have never attended school (Van der Meer, 1984:25). Moroccan children who go to school in the Netherlands often have the opportunity to follow a few lessons a week in Arabic language and culture. Of course, the results are not comparable to the results of Arabic education in Morocco, but the lessons in the Netherlands enable the children at least to become familiar with Arabic language and culture. Whatever the children's school results are, it is clear that attending Dutch schools stimulates proficiency in Dutch. All the Moroccans and Dutch we have spoken to during the observation period agreed on one point: Moroccans who speak Dutch well have usually participated in the Dutch educational system. Among Moroccans, usually Moroccan Arabic is spoken, and those who speak Dutch as well often mix their native language with Dutch. One of the Moroccan staff members of a community center in Votulast observed that proficiency in Dutch depends on a combination of factors related to work, length of stay, educational career and personality. Those who arrived around 1976, when they were about 15 or 16 (25 or 26 at the time of observation), immigrated most of all in order to earn money. As soon as they left school they began to look for a job, which at the time was not as difficult as it became later. Members of this group speak Dutch relatively poorly, because they only had minimal Dutch education. Another group arrived at the same age, but a few years later, around 1980. At that time there was considerable unemployment, which forced many young people to go back to school. Looking for work was postponed more easily than a few years earlier, because the chances of finding a job were minimal. As a result, there is a group of Moroccans in their late twenties who have lived here for more than ten years, and who have no more than a moderate proficiency in Dutch. There is also a younger group, who have spent between five and ten years in the Netherlands, with a much greater proficiency in Dutch. This example makes clear that the importance of length of stay should not be overestimated. The same Moroccan staff member also noticed that many Moroccan girls speak better Dutch than boys do: for girls school is considered one of the few acceptable alternatives for staying home all day, while the boys can go out as they please

74

CHAPTER

4. THE MOROCCAN

COMMUNITY

IN THE

NETHERLANDS

whenever they do not attend school or go to work. It is also thought that girls may speak better Dutch than boys because truancy is less pervasive among them, and, as a consequence, they have more opportunity to learn Dutch. Opinions about the status of Dutch vary somewhat in the Moroccan community. Among older Moroccans, proficiency in Dutch is generally highly thought of. It is certainly functional for reading and filling out forms. In a community where only few people read and write Dutch and Arabic their position is important. As I have mentioned before, the largest proportion of Moroccans with a high proficiency in Dutch is under 25. The fact that not all Moroccans, including the older people, take the trouble to learn Dutch must be seen in connection with the deep-seated wish to remigrate which is expressed by many Moroccans. In reality, remigration does not take place on a large scale for a number of reasons, the economic situation in Morocco perhaps being the most important one. At many locations in Utrecht weekend and evening Dutch language classes are given, mostly by Dutch volunteers. Moroccan students attend the lessons because they want to learn how to speak Dutch. Usually they axe not beginners anymore, and often most attention is paid to teaching reading and writing skills. In most cases the volunteers are ill-prepared for their task. Usually they start teaching after no more than a few training-sessions in which only a few general points can be discussed. The existing language methods often do not meet the needs of the students. When the Moroccans start their language classes, their proficiency levels vary greatly. Some have lived in the Netherlands for many years, without speaking a word of Dutch, while others have recently arrived and seem to have learned much in a short period of time. Some have attended school in Morocco and know French, while others are illiterate. Without professional help it is not easy to assess a student's proficiency level, so it often takes weeks or even months for a student to get started. Besides, practice is important: some people have a chance to use Dutch every day, others only use it during the language classes. On the other hand, formalizing of language classes may discourage attendance. In Votulast, for instance (in 1985), the language classes were so popular that not only the students came to the center to socialize, but other Moroccans did too, and together with students and teachers they drank coffee before and after the lessons. This very informal atmosphere was attractive to new students which was one of the main reasons why the classes were so well attended. During these informal gatherings Moroccan Arabic or Berber was usually spoken, since the majority did not speak Dutch well enough to converse, or in some cases because the Dutch who were present were not supposed to understand everything. Similar to what was noticed in other situations, Moroccans and Turks do not mix in these language classes. The classes in Votulast were meant for all immigrants, but in practice more than 90% of the participants was Moroccan. In some neighborhoods, such as Vleutenseweg, language classes have been discontinued when attendance started to drop. Most members of the younger generation by now spoke Dutch very well and did not need language classes while older people with language problems

4.4. LANGUAGE

PROFICIENCY

75

could rely on friends or family for translations. Moreover, many Moroccans believe that their stay in the Netherlands is only temporary, and there is still a sizeable group of women who do not venture outside the house. When, in addition, attendance is not socially motivated, it is evident that there is no future for Dutch language classes as described above. It must be kept in mind, though, that a very dynamic and rapidly changing situation exists in Utrecht and that these observations represent little more than a snapshot taken in the summer of 1986.

Chapter 5 T h e Present Study In the previous chapters I have discussed the background of Moroccans in the Netherlands. A short overview of Moroccan Arabic grammar was presented (chapter 2). The linguistic situation in Morocco (chapter 3) and the social and geographical situation in Utrecht (chapter 4) were sketched briefly. Now I will turn to the study of my corpus of code switching data. In this chapter the selection of speakers (5.1), the recorded conversations (5.2) and the transcriptions (5.3) will be discussed, while in (5.4) a preliminary model for classification of the switches will be presented.

5.1 5.1.1

Selection of Speakers Criteria

In order to obtain data from a group of speakers who might be expected to switch codes, five criteria were formulated which the speakers had to meet: 1. They had to be male. 2. They should be native speakers of Moroccan Arabic. 3. Their age should be between 16 and 25. 4. They had to be proficient in Dutch and had to have been in the Netherlands for more than four years. 5. They had to be proficient in Moroccan Arabic and had to have been in the Netherlands since the age of six or older. These five criteria were used as a guideline, not as a strict rule, as the following discussion will make clear.

CHAPTER 5. THE PRESENT STUDY

78

ad 1: At first it was assumed that speakers would all be male, because the Moroccan research assistent was male and it was feared that female speakers would not talk freely when the conversation was attended by a man. In practice, however, the problems that were foreseen for female speakers did not occur at all. Of the 15 Moroccans who were involved in the research, five were female. The topics discussed with the women did not essentially differ from the topics discussed with the male speakers. Moreover, the contributions of the women axe as useful as the conversations in which only men were involved. ad 2: Most people involved in the study had Moroccan Arabic as their native language. Of the 15 speakers, one was a Berber. More details are given in 5.1.3. ad 3: The average age of the 15 informants is 22.9 (SD=5). Two participants were older than 25 (27 and 38). The reason why these persons were involved in the investigation is that they could not be distinguished from the target group by other factors than age. Neither their looks, nor their use of Moroccan Arabic and Dutch set them apart and they were both usually together with friends who were younger. The women in the sample were slightly older than the men (between 24 and 27). ad 4: The speakers in the sample had all been in the Netherlands for a longer period than four yeaxs, their length of stay ranging from five to 19 years, with an average of 13.3 years. Purely in terms of time, they had all had sufficient opportunity to reach a high degree of proficiency in Dutch. ad 5: As for the age on which the speakers from the sample left Morocco, five were younger than six at the time. In practice, their proficiency in Moroccan Arabic did not seem to deviate from that of the other subjects. Proficiency in Dutch and Moroccan Arabic will further be discussed in the first section of chapter 6.

5.1.2

Looking for subjects

Mustapha Lkoundi, the research assistent for the present study, played a crucial role in the selection of the speakers. He is a native speaker of Moroccan Arabic, born in Casablanca, who has achieved a high degree of proficiency in Dutch since he has moved to the Netherlands in 1972. In 1986, when the data collection took place, he was 37 years old. He was living in Utrecht and knew many Moroccans and the places where they often get together. Many informants were either members of his family or friends and acquaintances or they were brought into contact with him by one of them. He became a member of a body-building club in Utrecht where he was able to convince some Moroccan adolescents to take part in the investigation. An advantage of working with family or friends was that it was not necessary to inquire in orientation meetings about the speakers' proficiency in Dutch and Mo-

5.1. SELECTION

OF

SPEAKERS

79

roccan Arabic, because Lkoundi already knew. On the other hand, it was felt to be disadvantageous that the contact between the interviewer and the informant(s), who sometimes knew each other quite well, suddenly acquired a more formal character in a conversation that had been planned beforehand. Usually, meetings were spontaneous, but now they had to be arranged on purpose for the study. When he did not know the speakers, Lkoundi started recording as soon as possible. It was his experience that appointments with people he knew little or not at all were often not kept. Besides, special arrangements were felt to block spontaneity which was not conducive to a relaxed atmosphere and consequently inhibited frequent code switching. The degree of formality and methods for achieving the least formal situation will be discussed further in 5.2. In the selection procedure the appearance of potential speakers and the way they were dressed did not play a role. The speakers were dressed in the same way as Dutch people of similar age and social class. Only very few young Moroccans wear a zellaba (traditional North African type of robe with a hood). When he was asked whether he would wear a zellaba, Mohammed, one of the speakers, answered 1 : No, oh that's not possible, according to me, then they will, they laugh at you. They think what are you doing now, axe you drunk or something? Everybody would look at you. (...) They would say: Look, what is wrong with him? He comes straight from the country! (moha094-097) (For the sake of clarity, examples from the conversations in this chapter are translated directly into English, since linguistic analysis is not of concern here. The originals axe given in footnotes.) The next step in the investigation was to assess proficiency in Dutch and Moroccan Arabic for each informant. The main reason for this assessment is that several studies have shown that only fluent bilinguals are able to switch codes effectively (e.g., Poplack, 1980). The criteria that were used in establishing whether speakers were more or less fluent bilinguals, were factors such as length of stay, educational career in Morocco and in the Netherlands, actual language use and language choice. This was easy enough for the persons Lkoundi already knew. Usually, he had a fairly accurate impression of their proficiency in both languages. Those he did not know he observed for a while before the interview, paying special attention to their use and choice of Dutch and Moroccan Arabic. Once he had found out that they spoke both Dutch and Moroccan Arabic he initiated an orientational superficial conversation in which he tried to find out what the speakers' native language was: Moroccan Arabic or Berber. This was not difficult, since most native speakers of Berber have a clearly recognizable accent when speaking Moroccan Arabic. Predominantly Berber 'Nee, ah dat kan niet volgens mij, dan gaan ze je- lachen ze je uit- ik denk ze lachen je uit, ze denken van wa- wat ben je nou mee bezig? Ben je bezopen of zo? gulu 1-ek sekran u la. Ja, natuurlijk wel, want iedereen zit te kijken. ta-ygulu I-ek hada ma-l-u hada, yallah za yhett l-kar.

80

CHAPTER 5. THE PRESENT

STUDY

speakers were not to be included in the sample. One Berber, however, knew Moroccan Arabic well enough to be included. In a conversation with other (non-Berber) Moroccans he only used Moroccan Arabic and Dutch. In certain respects he could be considered comparable to a native speaker of Moroccan Arabic, so there was no hesitation to include him in the study. While getting acquainted with speakers he did not know beforehand, Lkoundi asked, among other things, at what age they had come to the Netherlands and whether they had attended a Dutch school for any length of time. As was pointed out in chapter 4, people who have attended at least some years of education in the Netherlands usually have a better command of Dutch than those who never went to a Dutch school at all. In a later stage, after the recordings were made, proficiency in Moroccan Arabic and Dutch was judged intuitively by Lkoundi and myself. It was our impression that the speakers expressed themselves rather easily both in Moroccan Arabic and in Dutch. Only in a few cases was a speaker's pronunciation of Moroccan Arabic not perfect. In Moroccan Arabic minor grammatical mistakes were sometimes made. Of course, such impressions are highly subjective. Therefore, I looked for a more objective measure of the speakers' language proficiency. At first it seemed logical to establish the syntactic complexity by means of frequently used measures such as MLU (Mean Length of Utterance) and a subordination index. I decided not to do so because syntactic complexity cannot provide information about pronunciation, which might indicate a decline of Moroccan Arabic or interference from Dutch, and, as such, is exactly what I was interested in. Furthermore, there is no information available with which the results could be compared, such as data on the syntactic complexity of Moroccan Arabic which has not been exposed to Dutch influence. Eventually, I chose to present monolingual fragments of the speakers' recorded speech to a jury of native speakers of Dutch and Moroccan Arabic. In order to do so, two cassettes were made with stretches of monolingual speech that had been extracted from the recorded conversations with each speaker involved in the study. The cassette with Dutch material was heard and judged by 17 native speakers of Dutch. The Moroccan Arabic fragments were judged in Morocco by four native speakers of Moroccan Arabic. This was arranged by Prof. Moha Ennaji in Fes, who was kind enough to help me out. Analysis of the judgements has generally confirmed our own impressions. In chapter 6 special attention will be given to matters of bilingual competence. It will suffice here to state that, on the whole, the speakers were fluent both in Dutch and Moroccan Arabic. If it is true that frequent code switching is only possible for persons who speak two (or more) languages well, an additional argument for a high degree of bilingual proficiency of the speakers in the sample can be found in the fact that they were all frequent code switchers. In fact, Lkoundi only selected speakers of whom he knew (e.g. through observation) that they were used to switching from one language to the other.

5.1. SELECTION OF SPEAKERS

81

Figure 5.1: Morocco - Home towns of the speakers

5.1.3

The informants

Nine conversations have been analyzed in which 15 speakers were involved. In this section a profile of the individual informants will be given: short descriptions of the speakers at the time of recording in 1986. The speakers who have been excluded from the analysis in a later stage will also be discussed briefly. At the end, in table 5.1, the data about the informants will be summarized. In order to maintain the speakers' privacy, their names have been changcd. Figure 5.1 is a map of Morocco on which the speakers' places of origin have been indicated. Ahmed, 20 year old, was born in Kser, and is a native speaker of Moroccan Arabic. He came to the Netherlands, together with his family, when he was four. He now speaks Moroccan Arabic and Dutch and, according to his own judgements, he is more fluent in Dutch than in Arabic. At home he uses both Dutch and Moroccan Arabic with his brothers and sisters, and mainly Moroccan Arabic when speaking with his parents. With friends he primarily uses Dutch, since most of them do not speak Arabic: they include Dutch, Surinamese and Turkish friends, in addition

82

CHAPTER

5. THE PRESENT

STUDY

to Moroccans. After finishing primary school he completed lower level technical education (LTS). He is now working as a heating engineer. He did not have any education to speak of in Arabic. Since he has acquired Dutch nationality, he will not return to Morocco. Hassan was born in Casablanca and is 19 years old. He also migrated to the Netherlands with his family when he was four. His native language is Moroccan Arabic. In addition to Dutch he also learned English at school. He completed primary school and intermediate level secondary education (MAVO). He goes to evening classes for higher level secondary education (HAVO) and during the day he works through cm temporary agency. When he was 18 he went to a private school in Morocco for a period of three months in order to refine his Arabic. He sometimes uses Arabic for writing and reading, but does so with considerable effort. Dutch is dominant in fill domains. His use of Moroccan Arabic is restricted to the same domains as for Ahmed. They have many friends in common. Hassan sees his future in the Netherlands, not in Morocco. Morgan, 38 years old, is the 'nestor' of the sample. He was born in Casablanca where he acquired Moroccan Arabic as his first language. In Morocco he completed primary and lower secondary education, so he is quite proficient in French and Classical Arabic. He arrived in the Netherlands when he was 20. In either Morocco or the Netherlands he went to a hair dressing school. He has had several jobs, ranging from working on an assembly line to hair dressing. He is currently unemployed and is waiting for a residence permit on which his continued stay in the Netherlands depends. He believes Moroccan Arabic was to be his best language, followed immediately by French and Dutch. He knows how to read and write Dutch, but finds it hard to use and prefers Arabic and French for writing. He has mainly younger Surinamese, Turkish, Moroccan and Dutch friends. Kadi was born in Marrakes 22 years ago. He is a native speaker of Moroccan Arabic, although his parents are of Berber origin. He has lived in the Netherlands since he was nine. In Morocco, he had a few years of primary school which he continued in the Netherlands. He reads and writes Dutch exclusively. Since he has finished MAVO he now takes evening classes in commercial correspondence. He does not work and spends a lot of time studying, which he says he likes very much. Except sometimes with his sisters, he usually speaks Moroccan Arabic at home. At school he learned French and English. His best language is Moroccan Arabic; Dutch is second-best. With friends (Turkish, Surinamese and Moroccan) he speaks Dutch and Moroccan Arabic. He does not have many contacts with Dutch people. He is often taken for Surinamese because he has dark skin. Living in Morocco would only be acceptable to him with a lot of money. As long as he does not have that he will stay in the Netherlands. At first he was not very enthusiastic about getting involved in the study, but eventually he did agree to participate. He was rather quiet and reserved in most of the recorded conversation. Despite his modest involvement, he was not excluded from the analysis, because qualitatively his contributions were very useful indeed.

5.1. SELECTION

OF

SPEAKERS

83

Mohammed came to the Netherlands at the age of six, 14 years ago. He was born in Casablanca and his native language is Moroccan Arabic. In Dutch he is more fluent than in Moroccan Arabic which he only uses occasionally in speaking with his parents or the very few Moroccan friends he has. Most of his friends are Dutch or Turkish. He does not read or write Arabic which he resents very much. After primary school he started MAVO, which he is completing in evening classes. In the daytime he is working as an unqualified nursing assistant. Since he is gaining experience now he would like to become a professional nurse after he has finished school. He intends to stay in the Netherlands. Kaltum is a 24 year-old woman, who was born in Tetouan and has been in the Netherlands since she was nine. She comes from a Moroccan Arabic-speaking family. In Morocco she attended two years of primary education. In the Netherlands she went to a regular and to a Moroccan primary school. After that she attended a series of schools: lower and intermediate domestic science training and intermediate vocational training. Now she is attending the College for Social Work during the evenings, while in the daytime she works as a welfare worker at the Municipal Department of Social Services. She sometimes writes and reads Arabic, but most of the time she uses Dutch, which she finds much easier. With respect to speaking, she reported neither of the two languages as being dominant. She does not have many Dutch friends, most of them are Moroccans. She has married a Moroccan in the Netherlands. She very much liked to participate in the study. Souraya is 27 and left Casablanca when she was 11, a native speaker of Moroccan Arabic. In Morocco she attended primary school for a few years. In the Netherlands she finished primary school, where she attended comparatively many classes in Arabic language and culture. She obtained her certificate in geriatric care, and she is working in an old people's home now. Moroccan Arabic is her dominant language, and in reading and writing she also prefers Arabic. In the recorded conversation she mentions that she is soon going to marry in Morocco where she and her future husband will be living. With her brothers and sisters and her Dutch and Moroccan friends she alternately speaks Dutch and Moroccan Arabic (with Dutch friends only Dutch, of course). Karima, Souraya's sister, is 25 and has been living in the Netherlands since she was nine. Her linguistic and educational history is about the same as Souraya's. Unlike their younger brother Hassan, they both prefer to write and read Arabic. For the time being, Karima has no plans to leave the Netherlands. She is a qualified nursery school teacher, working in an international day care center. She is the only speaker who participated in two conversations. Khadija, 24 years old, was born in Rabat and left when she was five. After primary school she had lower vocational training. She is working as an unqualified nursing assistant. Only with her parents she sometimes speaks Moroccan Arabic, her native language. In the recorded conversation, in which she participated together with Souraya and Karima, she exclusively spoke Dutch, even when she was addressed in Moroccan Arabic (which she seemed to understand perfectly). For this reason, her

84

CHAPTER

5. THE PRESENT

STUDY

contributions were excluded from the linguistic analysis, although she added her bit to the recordings in terms of content. Driss left Morocco when he was nine, and is 24 at present. He was born in Kser and acquired Moroccan Arabic as his first language. He started primary school in Morocco and finished it in the Netherlands. After that he attended LTS, a school for lower level technical training and MAVO. He did not complete his education, which he does not really mind. In the recorded conversation he discussed the mischief he got into at school with a great deal of fervor. He does not use Arabic for writing and reading. In his language use he does not show any particular preference for either Dutch or Moroccan Arabic. His friends are Dutch and Moroccan. He is unemployed and often goes to a body-building club, where he met Mustapha Lkoundi. Youssefwas born in Tetouan, 21 years ago, and has been living in the Netherlands since he was two. At home he acquired Moroccan Arabic as his first language. After primary school he tried different types of education: he started with HAVO and ended with intermediate commercial education, which he did not finish. He prefers to use Dutch in most situations and uses Moroccan Arabic at times with his parents or with Moroccan friends. Most of his friends are Surinamese and Dutch. Just like Driss he is unemployed and a frequent visitor of the body-building center. Nadia is 26. She was 18 when she left Kenitra, where she was born. She went to secondary school in Morocco, which she had no time to complete; she only failed the last year. Her French is very good and she uses it alongside with Dutch and Moroccan Arabic for writing purposes. In the Netherlands she completed intermediate vocational training. She has had several jobs in places like a shelter for runaway girls and women. When the conversation with her was recorded she was working as a nursery school teacher. Her friends are of different nationalities. Although she speaks Dutch fairly well, she feels more at ease when using Arabic. She has a slight distrust with respect to Dutch people. It is not clear whether or not she wants to stay in the Netherlands. Wali is almost 20 and has been living in the Netherlands since he was 14. His first language is Berber (Tarifit), which he uses at home and with Berber-speaking friends. He also speaks Moroccan Arabic, Dutch and Spanish, but he is most fluent in Berber. In the recorded conversation, however, he uses Moroccan Arabic and Dutch, with a lot of code switching between the two, which is the reason why his material has been used in the analyses. In Morocco he only attended a Koranic school. In the Netherlands he started lower technical education, but he did not finish. His friends are mainly Moroccans, who taught him Moroccan Arabic, and Surinamese. He views his f u t u r e as rather hopeless: he is too much of a Moroccan to feel at home in the Netherlands, and at the same time he has been alienated from the Moroccan way of life. He has had a few short jobs and is trying hard to find work, so far without success. He spends much of his time in Hoog Catharijne. Unfortunately, his case is representative for a large group of young Moroccans. The difference between Wali and Mohammed, who expressed some of the same feelings, is that Mohammed has plans for the future and is working hard to realize them, while Wali does not have

5.2. THE RECORDED CONVERSATIONS

85

any diploma that gives him access to a better job and a better living situation. Abderrahman was born in Tetouan. He left Morocco when he was 12 and is now 18. His native language is Moroccan Arabic. In Morocco he had a few months of secondary education after primary school. He has a solid knowledge of French and Arabic. In the Netherlands he continued his educational career at a MAVO school, but he did not finish it. His use of Dutch for written purposes is poor. His oral command of Dutch is better, though not excellent. Most of his friends are Moroccans with whom he usually speaks Moroccan Arabic. He is one of the rare Moroccans who has learned to speak Berber from friends. His contribution to the recorded conversations is only marginal. It was analyzed linguistically, though, because he used both Moroccan Arabic and Dutch. Finally, Abdelhaq participated in the same conversation as Wali and Abderrahman. His linguistic and social situation is not very different from Abderrahman's. He was quiet most of the time, and if he spoke it was only in Moroccan Arabic. He was excluded from the linguistic analysis. Some of the data about the speakers discussed above are schematically summarized in table 5.1. Deatils on the conversations in which they participated will be given in the next section.

5.2

The Recorded Conversations

The recorded conversations provide us not only with linguistic but also with sociolinguistic information on the speakers. It has become clear that code switching in the speech of bilinguals and multilinguals only occurs in informal settings (e.g., Bentahila and Davies, 1983). Poplack (1980) has pointed out that the interlocutor should preferably have the same ethnic background as the informants: The importance of data collecting techniques cannot be overemphasized, particularly in the study of a phenomenon such as code-switching, which cannot be directly elicited. The actual occurrence of a switch is constrained, probably more than by any other factor, by the norms or the perceived norms of the speech situation. The most important of these norms for the balanced bilingual was found (Poplack, 1978a) to be the ethnicity of the interlocutor, once other criteria (appropriateness, formality of speech situation) were met. (1980:595) During the planning stages of the fieldwork, Mustapha Lkoundi established that certain factors axe likely to promote or inhibit informality. These factors will be discussed briefly: N u m b e r of participants When the field worker only has one conversation partner, the conversation will easily have the character of a formal interview, especially when the interlocutors are not familiar with each other. It is therefore

CHAPTER 5. THE PRESENT

86

Age at arrivai 4 4 20 9 6

Years in Neth. 16 15 18 13 14

Place of birth Kser Casablanca Casablanca

9 11

m

24 27 24 24

5 9

15 16 19 15

Tetouan Casablanca Rabat Kser

You

m

21

2

19

Tetouan

7 8,5 9

Nad Kar Wal

f f m

25 25 19

18 9 14

7 16 5

Kenitra Casablanca Nador

9

Abr

m

18

12

6

Tanger

9

Abh

m

17

12

5

Tetouan

Conv.

Name

Sex

Age

1 1 2 3 3

Ahm Has Mor Kad Moh

m m m m m

20 19 38 22 20

4 5 5 6

Kal Sou Kha

f f f

Dri

6

Marrakesh Casablanca

STUDY

Occupation heating engineer ec° HAVO casual work unemployed hairdresser ec; no work ec MAVO; nursing assistent social worker geriatric helper hospital assistent unemployed; school finished unemployed; school finished nursery assistent nursery assistent unemployed; school finished unemployed; school finished unemployed; school finished

unun-

ununun-

"ec=evening classes Table 5.1: T h e speakers, by age, length of stay, place of birth and occupation

5.2. THE RECORDED

CONVERSATIONS

87

advisable to avoid such conversations. In practice, this was not always possible: in four out of nine conversations Lkoundi only had one conversation partner. In all cases, they did already know him beforehand, though. Presence of a cassette recorder When speakers gave the impression that the presence of a cassette recorder might inhibit their spontaneity they were only told that their speech had been taped after the recording had talcen place. They were told afterwards that they had the right to refuse further use of the tape for analysis, but in practice they never did. Spontaneity Recordings should be made as spontaneously as possible and not through special arrangements: appointments axe often not kept and give the setting an unnecessary formal character. Place of Recording Recordings should be made in normal everyday situations: at home, in a café, in a sports club. Formal environments like classrooms and studios should be avoided. In the present study, recordings were actually made in home situations, in coffee houses, in cafés, on terraces, in a sports club and on a parking lot. Confidence There should be a reasonable degree of confidence and familiarity between the field worker and the informant(s). When the speakers do not know each other, confidence can be built by undertaking activities together, like drinking in a café or sporting activities. Planning As was mentioned before, the conversations not only had to provide linguistic material, but also data about the social circumstances of the speakers. This meant that some topics had to be covered in all the conversations. If an unforced conversation is to be achieved, it is not advisable to use a questionnaire. Instead, the topics were presented quasi-spontaneously. It appeared from the conversations that the degree of formality created by the presence of a cassette recorder had no effect on code switching. The use of Arabic, however, seemed to be more sensitive to a formal atmosphere. In a formal setting, the speakers try to use a variety of Arabic which is maximally close to Classical Arabic. The variety depended on the speakers' background and education: the more educated in Arabic, the more closely he or she will approximate Classical Arabic. As the situation becomes more formal, the variety of Arabic used will be one with higher prestige. Kaltum, for example, did not have other conversation partners, except Lkoundi, of course. In terms of formality the number of code switches in her speech should, therefore, be restricted. However, the percentage of transcribed (code switched) material is only slightly lower than that in the other conversations, as is illustrated in table 5.2. She did put considerable effort into approximating Classical Arabic, which is clear from her pronunciation and her choice of words in the following example:

CHAPTER 5. THE PRESENT STUDY

88

Conversation 1 2 3 4 5 6 7 8 9

Length 42 min. 22 min. 30 min. 60 min. 31 min. 31 min. 62 min. 26 min. 31 min.

Mixed 26 min. 13 min. 13 min. 17 min. 15 min. 13 min. 23 min. 12 min. 10 min.

percentage 39.0 46.4 43.0 28.3 48.4 42.0 37.1 46.1 32.3

Table 5.2: Portions transcribed of the recorded conversations (129)

r-razel kbir Hi ka-yexdem /e-n-nadafa (kal) the-man old who he-works in-the-cleaning business ['the old man who works in the cleaning business']

In this sentence the word razel is pronounced according to the rules of Classical Arabic. The word nadafa is also Classical Arabic. In Moroccan Arabic tendif can be used as well. From the recorded conversations only the parts in which code switching occurred (about 35%), and the contexts (about 5%) have been transcribed, which amounts to roughly 40% of the conversations. The exact portion of transcribed material for each conversation is shown in table 5.2. 65% of each conversation was monolingual (standard deviation 8.8). None of the conversations consisted of exclusively monolingual stretches in either Dutch or Moroccan Arabic. Of course, it must be borne in mind that only an average of 40% of the conversations was transcribed (i.e. only the parts in which code switching occurred). Although there are fluctuations, we may conclude that the conversations all were very bilingual. I will now take a closer look at some of the topics that were discussed in the conversations. Although a questionnaire has not been used, it is clear that topics Lkoundi wanted to talk about were brought up in each conversation, as is illustrated in table 5.3. None of the speakers had a purely positive attitude towards the Dutch. They had all had bad experiences in terms of discrimination; and all informants had sad stories to tell, like this one by Wali 2 : 2 J e loopt gewoon in de stad he, gewoon, je komt iemand tegen: "He zwarte kop, wegwezen!" En naar de discotheek natuurlijk, weet je wel, eh hij mag naar binnen, hij mag naar binnen, en jij? Daar! Hoe kan dat nou? Ja, gewoon.

5.2. THE RECORDED

CONVERSATIONS

Topic work/school adjustment/integration discrimination clash of cultures language use/choice jokes, tall tales situation in the Neth. going out, friends (vacation in) Morocco position Mor. women (lack of) future plans religion age documents, permits sports family environment a! pollution food looks/clothes drugs New Year's Eve Utrecht

Conv. 1

2

3

4

5

6

7

8

9

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X

X X

X

X

X

X X

X

X

X

X

X X

X

X

X

X

X X

X

X X X

X

X X

X

X

X

X

X X

X X

X

X

X

X X X X X X X

Table 5.3: Topics discussed in the recorded conversations

90

CHAPTER 5. THE PRESENT

STUDY

You are walking in the city, just like that. You meet someone: "Hey blackface, get lost!" And in the discotheque, of course, you know: "He can come in, he can come in", and you? There! How come? That's how it is. (wala400-405) Kaltum, however, thinks that not only the Dutch are to blame 3 : ...the young people who use drugs. I find that hard to take, because it bothers me. I, as a Moroccan girl, see the young men of my country in that area. I find it a shame. (...) I think that they confirm the inferiority of immigrants. (ka!a456-480) The speakers were well aware of belonging to two cultures. Mohammed expressed this sentiment as follows4: I feel like on a balance, really (...) I don't know where I am, whether I am Moroccan or Dutch. (...) You adjust as much as you can, really a lot, but not really: you are not really Dutch, you are still a Moroccan. (...) You take off your nationality, you replace it, you mix it all up, you forget the Moroccans, (...) but you remain a Moroccan. (moha344-347 and 431-437) As can be seen in table 5.3, four conversations (3, 4, 7 and 9) contain no jokes or stories. The atmosphere was more serious than in the other five conversations. This did not seem to influence the degree of informality. Furthermore, the position of Moroccan women and girls in Dutch and Moroccan society was only discussed by the female participants. It was brought up by the women themselves, not by Lkoundi. Beforehand, the speakers were usually informed about the purpose of the study, for ethical reasons. One of the consequences was that language use and choice were also discussed in some of the recorded conversations, like the one in which Youssef participated 5 : Here you speak Dutch with Moroccans, sometimes Moroccan with Surinamese. For example, you go to Hoog Catharijne, well, let's say 80% 3 s-sabab lit ka-y- eh j a lit ka-yeste9mlu l-muxeddirat, daar heb ik veel moeite mee he, li-anna maka-nerda-s ana kama saba megribiya nsuf le-wlad blad-i, he?, 9 aye sin f-dik eh f-dik l-midan hadak he, ma-ka-nerda-s. (...) ana ka-ndenn duk s-sabab ka-y9emlu bevestiging bi-anna buitenlanders minderwaardig zijn. 4 ia-nhess bhal Ilifuq l-mizan, een weegschaal, echt. (...) Eh ik weet niet waar ik aan toe ben, (...) was ana megribi, was ana hulandi? Je past je zo veel aan, echt heel veel, maar niet echt. Je bent niet echt een Nederlander, je bent nog altijd Marokkaan en zo (...) theyyed tdir zensiya tbeddel-ha u tdir kill-si txerbiq u tensa t-mgarba (...) Maar je blijft een Marokkaan. 5 A l s je hier bent dan praat je Nederlands met Marokkaans, soms Marokkaans met Surinaams. Bijvoorbeeld je komt op Hoog Catharijne, nou, zeg maar tachtig procent van de Marokkaanse jongens die zegt altijd tegen jou in het Surinaams "hoe g a a t ' t ? "

5.3.

TRANSCRIPTIONS

91

of the Moroccan guys will say to you in Surinamese: "How are you?"! (youa277-280) In conversation 3 Mohammed mentioned problems with respect to language choice6: My mother, she speaks very little Dutch, but sometimes, then I want to say something to her and then, really, it just jumps out of my mouth, in the Dutch language. With my brother, I speak a lot of Dutch with him, also Moroccan. But also half-half, just like that: zib li-ya (get me) a glass of water, or something. (moha064-074)

5.3

Transcriptions

For the purpose of the present study it was not necessary to transcribe the conversations completely. Instead, only those parts in which code switching occurred were transcribed, together with the immediate context. Monolingual stretches exceeding three or four sentences (depending on their length) were not transcribed. All switches between Dutch and Moroccan Arabic were classified according to direction and type of switch and syntactic context. Switches between speakers were not coded; only switches in the speech of one speaker were included. Table 5.2 showed how much of the conversations was transcribed. In this section I will discuss the way the transcriptions were made and how they were used in the analysis. After listening to the tapes intensively, the parts which had to be transcribed were selected. At first rough transcriptions were made. Later the orthography for the Moroccan Arabic fragments, as it has been presented in 2.2, was adjusted and standardized. A Dutch morphemic translation was added, and finally a rough translation was made, usually for stretches no longer than the speech turn of one speaker. This was done for every new turn in the conversation. Here, the Dutch translations have been replaced by English ones. The transcriptions were all entered into a computer. In Appendix A parts of conversations 5 and 6 are presented, by way of example.

5.4

A Preliminary Classification Model

For the syntactic analysis the code switches were taken out of context, and classified according to the model given in figure 5.2. In figure 5.2, the possible values for direction of switch are first summed up. These directions have been chosen on linear grounds. D-+MA means that there has been 6 M i j n moeder, die kan heel weinig Nederlands, m a a r weleens, d a n wil ik w a t zeggen tegen 'r en d a n eh echt, eh s p r i n g t ' t zo uit m ' n mond van eh in de Nederlandse taal d a n . m9a xu-ya ik spreek heel veel Nederlands m e t h e m , ook Marokkaans. Maar ook half half, om en o m , echt, d a ' s echt niet n o r m a a l . Gewoon, zib li-ya een glas water of zo.

CHAPTER

92

DIRECTION MA->D — D—>MA = D in MA MA in D -

5. THE PRESENT

STUDY

from Moroccan Arabic to Dutch from Dutch to Moroccan Arabic Dutch word in a Moroccan Arabic sentence Moroccan Arabic word in a Dutch sentence

TYPE 1 2 3 4 5

—

= —

=

SYNTACTIC a — b — c = d = e f = g h i = j k = m — n = o = s t

=

X

=

intersentential switches extrasentential switches intrasentential switches single word switches repetitions and broken sentences at switch sites CONTEXT between sentences with a coordinating conjunction quotation metalinguistic (the word itself is discussed) between separate sentences between main and subordinate clause between verb and object NP after subject adverbial (constituents) whole PP between preposition and NP NP internal left/right dislocation rest code switching single noun, bare single noun, with article and/or plural, and/or diminutive marking single word, not a noun rest single words

Figure 5.2: Model for classification of code switches by direction, type and syntactic context

5.4. A PRELIMINARY

CLASSIFICATION

MODEL

93

a switch in a sentence or between sentences from Dutch to Moroccan Arabic, while MA—>D means the opposite. D in MA and MA in D are reserved for cases where one word or a short idiomatic expression from language A appears in a sentence in language B. This is where the notion base language comes in. For the time being, I will assume that the base language is the language with which the sentence has been started. I will return to this subject in a special section in chapter 8. In fact, D in MA is a special case of MA—>D and MA in D of D—>MA. The difference is that when D in MA or MA in D is used, it is assumed that the speaker switches back to the original language immediately after the element under consideration (except when the element is the last one in a speech turn). In figure 5.2, five types of code switching are listed. Intersentential switches are switches between sentences. A number of switches have been classified as intersentential when they occur immediately before or after a coordinate conjunction. Sometimes such switches were marked by a break or a pause. In those cases the conjunction was used as an element to introduce a new topic. This type of switching was interpreted as intersentential code switching, while cases involving 'true' coordination, without a pause or a break; were classified as intrasentential code switching. The second category is extrasentential code switching. This term has also been used by others, such as Boucherit (1987) and Scotton (1987). Extrasentential switching concerns words or stretches of sentences that axe part of the discourse structure of a sentence but are outside the sentence's syntactic structure, such as "ja" (yes), "nee" (no), "he" (isn't it), "weet je wel" (you know), fhemti? (you understand?). Most 'fillers' axe included in this category to which I will return in chapter 7.2. Reported speech is also classified as extrasentential code switching. The next type is intrasentential switching. Apart from switches between sentences coordinated by a conjunction after a pause, all switches that took place within the sentence belong to this category. The fourth switch type is single word switching. All one-word elements from language A uttered in a stretch of discourse in language B are included in this category. The question whether these axe instances of borrowing, intrasentential code switching, or another phenomenon of language contact wil be discussed in chapter 9. Finally, repetitions and broken sentences were scored, but they were not analyzed further. In figure 5.2 the syntactic contexts in which the code switches took place are presented in more detail, of course only in those cases where there is a need to differentiate. In chapter 7 the syntactic contexts in which switching takes place are analyzed further. Altogether 1110 switches have been coded for direction, type and syntactic context. The different kinds of code switching will be discussed in detail in the following chapters. The following examples show how the classification model applies:

CHAPTER

94 (130)

5. THE PRESENT

STUDY

neqder ne9mel 4 cursus van negen maanden. (youa210) can:IMPF:l:SG do:IMPF:l:SG course of nine months. ['I can do a nine-month course.']

Classified as: MA->D 3 f (V-MA + Obj NP-D). (131)

Dat is wel voldoende, 4 u hada huwa Iferq. that is about enough and that it DEF- difference:M:SG (kalb270) ['That is about enough, and that is the difference.']

Classified as: D-+MA 1 a (D + u-MA). (132)

ta-nehder m9a si 4 buitenlander f DU-talk:IMPF:l:SG with a foreigner:M:SG

axor. (hasb023) other:M:SG

['I talk to another foreigner.'] Classified as: D in MA 4 o.

In addition to the classification model presented in figure 5.2, the switches were later coded in more detail for precise syntactic context. The classified code switches have all been labeled, so that their contexts can easily be traced in the original transcription. Now that the speakers and the conversations in which they participated, and a preliminary sketch of the way in which their switches have been classified have been presented, I will proceed to the analyses of the speakers' bilingual competence and the code switching data, in the following chapters.

Chapter 6 Bilingual Language Proficiency 6.1

Introduction

In 5.1.2. I briefly mentioned how the speakers' proficiency in Dutch and Moroccan Arabic was determined and why it was necessary to do so. In this chapter the matter will be discussed and explained in detail. Some authors have claimed that it is essential for speakers to have a high degree of bilingual proficiency in order to switch codes effectively. Poplack, reporting on an investigation on code switching between Spanish and English among Puerto Ricans in New York, remarked that "while speakers who axe dominant in one language show a strong tendency to switch into L2 from an LI base, more balanced bilinguals often alternate base languages within the same discourse." (Poplack, 1980:597) In addition, she claims that intrasentential switching (from one language to the other within the sentence) requires more bilingual skill than emblematic or tag-switching: (...) most of those who report that they know, feel more comfortable in, and use more Spanish than English, tend to switch into L2 by means of tag-like constructions, sometimes to the exclusion of sentential or intrasentential switches. Those who claim to be bilingual, on the other hand, show a reversal. (...) They favour large amounts of the switches hypothesized to require most knowledge of both languages, sentential and intrasentential switches. The most favoured switch type for bilinguals is clearly intra-sentential, while the least favoured is tag-like switching. (1980:606) Berk-Seligson (1986), on the other hand, reports findings contradictory to Poplack's, in her investigation of code switching between Hebrew and Spanish: In addition, type of code switching (...) is, with the exception of one minor case, unrelated to degree of bilingualism. That is, contradictory to findings to date, balanced bilinguals, Ll-dominant bilinguals, and L2dominant bilinguals code switch in an indistinguishable manner. This

96

CHAPTER

6. BILINGUAL LANGUAGE

PROFICIENCY

finding, therefore, strongly disputes the currently held notion that the ability to code switch, particularly at the intrasentential level, is a mark of a bilingual's high competence in his/her two languages. (1986:334-335) The contradictory findings of Poplack and Berk-Seligson lead me to to a close study of the relationship between bilingual proficiency and code switching behavior of the Moroccan speakers in this study.

6.2

Methods for Measuring Bilingual Ability

Since there are no standard tests to operationalize bilingual skills, each investigator has developed his or her own method. In Poplack's study "(...) the informants responded to a language-attitude questionnaire to tap various aspects of language skill by self-report as well as community attitudes towards Puerto Rican language and ethnicity." (Poplack, 1980:593). Poplack not only wanted to show that the ability to code switch intrasententially depends on bilingual skill but also that it is used as a measure of bilingual competence. To avoid circularity, it is necessary then to determine bilingual proficiency independent of intrasentential code switching. In a comprehensive study of English loanword usage in Canada, Poplack, SankofF and Miller (1988) also determined bilingual proficiency on the basis of self-reported data: All speakers were coded as belonging to (...) one of four levels of an index of English proficiency calculated from speakers' self-reports of their English competence and usage in various situations and domains. (1988:99) Berk-Seligson (1986) remains rather vague about how she measured bilingual proficiency, although her speakers were divided into three groups according to their degree of bilingualism. All she mentions is that "Each informant was classified as either (1) fluent bilingual, (2) Spanish dominant, or (3) Hebrew dominant." (1986:321). Sociolinguistic factors were hypothesized to be relevant to bilingual ability, including place of birth, and —if not native to Israel—, age of arrival, native language of the speakers' mother. She does use observed bilingual ability as a criterion, but does not explain in detail how these observations were made. A category that is missing both in Poplack's and Berk-Seligson's work is that of speakers who perform poorly in both languages. This omission may be a consequence of the way bilingual proficiency was determined. There were only three possibilities: speakers were proficient in LI, in L2 or equally proficient in both. Psychological studies often advocate that variables should be operationalized according to the so called 'multi-trait-multi-method design', in order for the results to be reliable (e.g., Cook and Campbell, 1979). Applying this research design to the present study implies that the construct (or trait) I wanted to measure, i.e. bilingual

6.2. METHODS

FOR MEASURING

BILINGUAL

ABILITY

97

proficiency, should yield the same results for each different way it is measured (multimethod). In this study I have used four different methods to measure bilingual proficiency: 1. Self-report by speakers who were asked what language they spoke best and used preferably. 2. Actual use of monolingual Dutch, Moroccan Arabic and a Dutch-Moroccan Arabic mixed variety calculated for each speaker. 3. Judgement of monolingual proficiency in either language by native speakers of Dutch and Moroccan Arabic, respectively. 4. Error analysis of deviant linguistic features in the Dutch (and Moroccan Arabic) of the speakers. In table 6.1, Khadija, Abdelhaq, Abderrahman and Wali are marked with an asterisk, because their speech will not be included in the analyses later in this chapter for reasons explained in the following sections. In these analyses, rank orders will be used, based on the data from tables 6.1 through 6.7. Eleven speakers will be included. Consequently, eleven positions have to be filled in the rank orders. The ordering is always from Dutch (1) (most, preferred or best) via both Dutch and Moroccan Arabic to Moroccan Arabic (11). In table 6.1 the rank order is given in the last column: Ahmed, Hassan, Mohammed and Youssef share the first four positions, which gives them an average of 2.5 each. Kaltum takes the only position in the middle, because she was the only one preferring both Dutch and Moroccan Arabic. The remaining six speakers share the positions 6 through 11, which gives each of them an average rank order of 8.5.

6.2.1

Self-report

In the conversations the speakers expressed which was their best language, the preferred language for everyday communication. Their preferences are summarized in table 6.1. None of the speakers was explicitly asked to express his or her preference for either language. Ahmed, however, who told the interviewer on a different occasion that he preferred Dutch, gave an indication of his problems with Moroccan Arabic in Morocco during one of the conversations 1 : "...it may go on for the first weeks, well, then it goes very badly, you know, and then all of a sudden you come along quite a bit (..). You need 1 ...dat duurt misschien de eerste weken, nou ja, dan gaat't heel slecht, weet je, en dan daama dan schiet je ineens een heel stuk naar voren (...). Je moet die beginstukken hebben. Als je dat hebt, dan kom je al heel ver, weet je, en door middel van gebaren, andere talen erbij, lukt 't wel. (ahma215-243)

CHAPTER 6. BILINGUAL LANGUAGE

98

Name Ahmed Hassan Khadija* Mohammed Youssef Kaltum Abdelhaq* Abderrahman* Driss Kadi Karima Morgan Nadia Souraya Wali*

Dutch

Moroccan Arabic

other

X X X X X X

X X X X X X X X X

Berber

PROFICIENCY

rank order 2.5 2.5 n.a. 2.5 2.5 5 n.a. n.a. 8.5 8.5 8.5 8.5 8.5 8.5 n.a.

Table 6.1: Languages preferred by the speakers to have the starting points. Once you have that, you come a long way, you know, and with the help of signs and other languages you succeed." Mohammed also spoke about his poor command of Moroccan Arabic when he was in Morocco for the first time after a long period in the Netherlands. In the other conversations, the speakers did not refer specifically to their preference for one of the languages. As can be seen in table 6.1, Kaltum was the only speaker who reported to have no particular preference for either language. Most of the other speakers preferred Moroccan Arabic (8), while Wali reported Berber to be his best language (see also 5.1.3).

6.2.2

Actual use of Dutch, Moroccan Arabic and a mixed variety

For all speakers the time they spent uttering monolingual stretches in either language as well as in a Dutch-Moroccan Arabic mixed variety has been recorded and the percentages have been calculated for each variety. A monolingual stretch is defined here as at least one utterance by one speaker in one language. When the utterance is followed or preceded by an utterance by the same speaker in the other language, it is interpreted as the use of a Dutch-Moroccan Arabic mixed variety. Inter-speaker shifts were not interpreted as mixing. The use of more than one language within a

6.2. METHODS

FOR MEASURING

BILINGUAL

ABILITY

99

Monolingual:

[Sentence]^eaker

A

+ [Sentence]sp e a k e r

A

Monolingual:

T2 [Sentence] S p e a k e r

B

Li + [Sentence] S p e a k e r

A

Mixed:

[Sentence] L1

Mixed:

[Sentence]^eaker

[Sentence]^eaker

A

311(1 1 2

A

+

Figure 6.1: Formal representation of possible language use sentence is always scored as an instance of mixing, even when it only concerns one item. The different possibilities of language use are represented formally in figure 6.1.

In addition to Dutch and Moroccan Arabic, Abderrahman, Morgan, Nadia and Kaxima used some French in their speech; this was not included in the analysis. I will return to the use of French in chapter 9. Results are given in table 6.2. In general, Lkoundi stressed the use of Moroccan Arabic in the conversations thus creating a potential bias towards Moroccan Arabic. However, this bias is consistent for all speakers. As was done in table 6.1, the speakers in table 6.2 are also rank ordered from use of Dutch (1) via both Dutch and Moroccan Arabic, to Moroccan Arabic (11). In the speech of Ahmed, Hassan and Youssef, the highest percentage is Dutch. Both Kaxima and Mohammed used a mixed Dutch-Moroccan Arabic variety in over half of their utterances, while the utterances produced by the remaining six speakers, show the highest percentage of Moroccan Arabic. The rank ordering is given in the last column. It appears from table 6.2 that four speakers (of course, not including Khadija, whose speech will not be further analyzed) use more Dutch than Moroccan Arabic: Ahmed, Hassan, Mohammed and Youssef. They are the same speakers who reportedly preferred Dutch over Moroccan Arabic (table 6.1). Some speakers show an extremely low use of monolingual Dutch (Nadia, Karima, Driss, Kaltum, Souraya), or Moroccan Arabic (Mohammed) but they do not seem to avoid Dutch, because it does appear in the mixed variety they frequently use (Abdelhaq and Abderrahman are not considered here). Only in the case of Karima and Mohammed is the mixed variety used more often than either two languages. All other speakers use one language more often than the mixed variety and the other language less often. Kaltum is quite exceptional in severed respects. In the first place she is the only one who reported to be a balanced bilingual with no particular preference for Dutch

100

CHAPTER

Name Abdelhaq* Abderrahman* Ahmed Hassan Youssef Karima Mohammed Morgan Souraya Kadi Kaltum Driss Nadia Khadija* Wali*

Dutch 0 0 55.2 52.9 49.0 2.6 41.0 22.6 8.7 17.0 4.4 3.3 1.7 100 18.5

6. BILINGUAL

MA 100 100 17.2 16.3 19.6 44.3 8.0 41.5 53.2 56.8 62.6 63.3 66.7 0 55.5

mixed 0 0 27.6 30.8 31.4 53.1 51.0 35.8 38.1 26.2 33.0 33.3 31.6 0 26.0

LANGUAGE

French

PROFICIENCY

rank order

x

x X

X

1 2 3 4.5 4.5 6 7 8 9 10 11

Table 6.2: Actual language use in the conversations, in percentages or Moroccan Arabic. Secondly, as was already mentioned in 5.2, the conversation in which she was involved had a rather formal character. This might explain why she spoke Dutch only 4.4% of the time. It should be added here that the other conversations with speakers who scored low on percentage of monolingual Dutch were not particularly formed. Kaltum seemed to be the only speaker who was noticeably sensitive to the potentially formal character of a conversation. This was also clear from the fact that she tried to use the highest possible variety of Arabic (i.e., the one closest to Classical Arabic) and tried to avoid Dutch. The data from table 6.2 are visualized in figure 6.2. It is clear that the percentages of monolingual Moroccan Arabic (between 8% and 66.7%, average 42.1%) are higher than those of monolingual Dutch (between 1.7% and 55.2%, average 23.1%). In either case, the standard deviation is rather high, because there is a lot of variation between individual speakers: 21.2 for Moroccan Arabic and 20.9 for Dutch. Data on the use of the mixed Dutch-Moroccan Arabic variety show a different picture: scores vary between 26% and 53.1%, the average is 34.8% and the standard deviation is only 8.8. These results imply that the speakers rather consistently spend between a quarter and half of their time producing mixed utterances, while there is considerable variation between individual speakers in how the remaining percentage of their speaking time is spent. As in the other analyses, only real speech time for each speaker was measured. Pauses and speech turns of the interviewer or of other people present have not been measured. Table 6.3 indicates how long each of the 12 informants spoke.

METHODS FOR MEASURING BILINGUAL ABILITY

Nadia Karima Driss Kaltum Souraya Kadi Morgan Mohammed Youssef Hassan Ahmed

Figure 6.2: Actual language use ordered by percentage of Dutch

Length of conversation 42 42 28 30 30 60 31

Speech time

Other people present

14.5 12.3 26.5 8.8 12.2 45.5 17.3

Driss Youssef Nadia Kaxima

31 31 62 26

15.0 5.1 36.0 19.2

Wali

31

10.8

Hassan, Lkoundi Ahmed, Lkoundi Lkoundi Mohammed, Lkoundi Kadi, Lkoundi Lkoundi Karima (5 min.), Khadija, Lkoundi Youssef, Lkoundi Driss, Lkoundi Lkoundi Souraya, Khadija (5 min.), Lkoundi 3 friends (15 min.), Lkoundi

Name Ahmed Hassan Morgan Kadi Mohammed Kaltum Souraya

Table 6.3: Actual speech time of the informants (in minutes)

102

CHAPTER

6. BILINGUAL

LANGUAGE

PROFICIENCY

Concluding this section it is fair to state that Moroccan Arabic is used most and Dutch least, while the use of the mixed variety lies somewhere between the two. Considering the fact that Lkoundi predominantly used Moroccan Arabic, it is not surprising that speakers use more Moroccan Arabic than Dutch. The fact that they use Dutch at all is an indication of the speakers' high level of confidence in their Dutch proficiency.

6.2.3

Judgement by native speakers

In 5.1.2 I briefly mentioned that native speakers were involved in judging the speakers' proficiency in Dutch and Moroccan Arabic. Here these judgements will be discussed in detail. To enable native speakers of Dutch and of Moroccan Arabic to judge the language proficiency of the speakers in the study, I compiled two audio-cassettes, one with Dutch and one with Moroccan Arabic monolingual fragments extracted from the conversations. On the cassettes, randomly selected sentences or larger units were grouped together by speaker. The sentences had to be uttered in one language, and could not be interrupted by other speakers or by background noise. For some speakers this left little material left to choose from. Because their privacy had to be guaranteed, fragments containing specific names or places, or compromising statements could not be involved. The resulting marerial consisted of 82 Dutch fragments from 13 speakers and 67 Moroccan Arabic fragments from 14 speakers. When there are many fragments from one speaker, they are usually short. When only a few fragments are included, they usually consist of longer stretches. The number of speakers involved was not the same on both cassettes: Khadija did not speak Moroccan Arabic at all, so only her Dutch could be judged. Abderrahman and Abdelhaq mainly spoke Moroccan Arabic (cf. table 6.2). During the conversation in which they were involved more people were present who all liked to talk and did talk a lot, unfortunately. After 15 minutes, Abdelhaq and Abderrahman left, leaving Wali and Lkoundi behind. For this reason uninterrupted stretches of speech from Abderrahman and Abdelhaq were hard to find. The 15 minutes of conversation between Wali and Lkoundi did provide useful material in which code switching between Dutch and Moroccan Arabic occurred. For the judgements native speakers were asked to listen carefully to the fragments and give their opinion on a five point scale. Not only were they asked about their general impression, they were also asked for their opinion about pronunciation and grammatical competence. Vocabulary (the choice of right or wrong words) was also judged. In addition, they were invited to give comments and examples whenever possible. The Moroccan native speakers were also asked to identify a possible regional accent. The tasks described here have been carried out for every speaker involved in the study. Eventually, there were 13 speakers whose Dutch was judged and 14 speakers who were judged for their Moroccan Arabic proficiency. 17 native speakers of Dutch were the judges of the informants' Dutch proficiency.

6.2. METHODS

FOR MEASURING

BILINGUAL

ABILITY

103

It was a problem to find Moroccan Arabic native speaker judges. Since I wanted them to recognize interference from Dutch, I believed it would be better if they were not familiar with the Dutch or Moroccan Arabic spoken by people who may have been in contact with Dutch for years. Therefore, it was necessary to find native speakers of Moroccan Arabic in Morocco. Despite these strict criteria, I was able to find four Moroccan Arabic speakers who were willing to act as judges. Professor Moha Ennaji from Fes, Morocco, was kind enough to arrange for four Moroccan Arabic native speakers who judged the taped Moroccan Arabic fragments. Of course, I would have preferred to include more Moroccan Arabic native speakers as judges. Given the difficulties in finding them at all I must manage with the few I could get. In table 6.4 the judgements by the Dutch and Moroccan Arabic judges axe presented. The judges scored between one and five points on three variables: general impression, pronunciation, and grammaticality. For each speaker the average score on the three variables was calculated. Since there was never more than a 0.2 point difference between the scores on the three variables, only the average is given in table 6.4 (in the columns "Total Dutch" and "Total Mor(occan) Ar(abic)"). In the next column, the difference between the average Dutch and Moroccan Arabic scores are given for each speaker. The difference reflects how close the degree of proficiency for Dutch and Moroccan Arabic was for each speaker. The Dutch native speakers tended to be very rigid in their judgements compared to the Moroccan judges. This is not a problem as long as they were rigid consistently, which seems to be the case. What counts is the rank order for the Dutch and the Moroccan Arabic scores, which would not change even if one point were added to all Dutch scores. The standard deviations of the 17 scores per speaker ranged from 0 to .83, which implies that the opinions of the native speakers were rather homogeneous. In the last column of table 6.4 the rank order of the speakers is given. The speaker with the smelliest difference between total Dutch and total Moroccan Arabic scores supposedly has the greatest proficiency in Dutch compared to proficiency in Moroccan Arabic. According to the native speaker judges, none of the speakers had vocabulary problems. An occasional unusual word was observed by the native speakers, but this did not detract from the overall positive impression of lexical proficiency in either language. It should be noted that on the five point scale Dutch native speaker judges only sporadically used 1 (very bad) and 5 (excellent), while the Moroccan native speaker judges seemed less afraid to do so. Moroccan Arabic proficiency was generally rated higher than Dutch proficiency. Of course, this may reflect reality, but it must be borne in mind that other factors might have exerted their influence on this difference. In general the Dutch listeners seemed to apply higher standards than the Moroccans. Kahadija, for example, scored an average for Dutch of 3.9. However, the first time I heard her speak, I thought she was a native Dutch speaker with a clearly recognizable regional accent. The native speaker judges did remark in their comments that she sounded "native-like", "accentless" (i.e. without a foreign accent) and "fluent" but,

104

CHAPTER

6. BILINGUAL

LANGUAGE

Name of Average Average Difference between Dutch Mor Ar Dutch and Mor Ar speaker 4.4 0.4 Ahmed 4.0 0.6 3.8 4.4 Mohammed 0.7 3.0 3.7 Youssef 4.7 0.9 3.8 Kaltum 4.2 1.2 Hassan 3.0 1.4 Nadia 3.4 4.8 2.0 3.0 5.0 Morgan 4.7 2.1 Driss 2.6 4.7 2.1 Kadi 2.6 2.7 4.9 2.2 Souraya 2.4 Karima 2.6 5.0 n.a. 4.8 n.a. Abdelhaq n.a. 4.8 n.a. Abderrahman n.a. Khadija 3.9 n.a. 2.4 2.9 0.5 Wali n.a.: not applicable The scale ranges from 1 (very bad) to 5 (excellent)

PROFICIENCY

Rank order 1 2 3 4 5 6 7 8.5 8.5 10 11 n.a. n.a. n.a. n.a.

Table 6.4: Average judgements by native speaker judges on a five point scale. nevertheless, only four out of seventeen judges considered her to have an excellent command of Dutch. Another four of them even found her Dutch not worth more than 3, only halfway on the five point scale, while considering her to be the most proficient speaker of Dutch.

6.2.4

Error analysis of spoken D u t c h in t h e corpus

As a native speaker of Dutch, I did not feel competent to analyze the speakers' Moroccan Arabic with respect to interference or other unusual constructions. As a consequence, this was only done for the Dutch part of the corpus. For the Moroccan Arabic material only specific comments from the native speaker judges in Morocco were considered. Although it is not possible, then, to make a real comparison of errors in the two languages, the analysis of the available information will shed some light on the consequences of intensive language contact. First it must be made clear that I consider errors to have an exceptional status, since they appeared only sporadically. I have defined an error as the appearance of a lexical item or a construction that a native speaker would never use. Table 6.5 shows the number of errors the speakers made when they were using Dutch. The rank order in the last column goes from lowest (1) to highest (11) error

6.2. METHODS FOR MEASURING BILINGUAL

ABILITY

105

index. For the error analysis the use of Dutch throughout the conversations has been investigated, not only in the monolingual Dutch stretches, but also in the mixed parts. In what follows errors that may be due to interference from Moroccan Arabic will be discussed first, followed by a look at the other constructions and lexical items that would not be used by native speakers of Dutch. The value of these results should not be overestimated because they axe based on a small number of occurrences, but they illustrate that there is quite some variance in the number of non-native-like constructions and lexical items in the informants' Dutch. A closer analysis of the results shows that only a small proportion of the unusual occurrences resemble words or constituents in Moroccan Arabic: only 17 out of 122 errors. This means that possible interference from Moroccan Arabic can explain only 14% of the unusual words and constructions. Among these 17 errors, omission of the obligatory Dutch indefinite article "een" occurs seven times. The resulting constructions were all grammatical according to Moroccan Arabic grammar, but were ungrammatical in Dutch. The following example illustrates this: (133)

We hebben wel *(een) oplossing voor die dingen. (mora023) We do have *(a) solution for those things.

Twice the Dutch definite article "de" or "het" was not realized, which also resulted in constructions that axe probably only acceptable in Moroccan Arabic grammar, but are unacceptable in Dutch: (134)

Ik heb ook *(de) zelfde. (kada332) I have also *(the) same.

Four times the copula was deleted, as is usual in Moroccan Arabic, resulting in sentences that are not acceptable in Dutch: (135)

Islam *(is (de)) beste godsdienst. (soua435) Islam *(is (the)) best religion.

Twice a verb in the present tense was used where past tense was required, possibly in analogy to the different use of tense and aspect in Moroccan Arabic: (136)

Wij wonen daar als eerste. (youa435) We are the first to live there (years ago).

00

00 io n TT es ci co 00 o

errors

o o lO Ci

•aV t •w — Ol -o bp O t-, «e S

t Ci CS o Ci CS CS —

-

- co co CS CS

CS

cs to

•

•

•

•

CO CS

•

-

•

•

•Si X

CO CO

co t-

-

co o •V T

cs CO CS t•

-

•

1—1 cs •

CS

CS

1 CS

-

TT

-

•

- -

•

•

cs

I Mohammed 1 Kaltum Ahmed Youssef Hassan Nadia Morgan Kadi Driss Souraya K ari m a Wali* 1 Total

total

-o o

] wrong tense cop deleted 1 het deleted 1 een deleted Name

-»S-> o

r— i * oo

•

•

•

-

>ß Iß iß

- -

c O CS CS cs CS

1

CS

>> •

-

CS CO

iß co Iß

•

m i— o

•

•

CS f—t

iß iß O

1

co -

•

•

uaV vM