 

’HES?3

UNIVERSITY LIBRARI

II IIII IIIIIII

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

IIIIIIIIIIIIIIIIIIIIIIIIIII

3 1293 0169

This is to certify that the

thesis entitled

ATTITUDES 0F EDUCATED AFRICANS TOWARDS
REGIONAL VARIETIES OF ENGLISH
presented by

Mungai Mutonya
has been. accepted towards fulﬁllment

of the requirements for

M.A. degree in Linguistics

 

 

 

Major professor

Date Sept. 18, 1997

 

0-7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

.r "I'V‘

 

LIBRARY
Mlchlgan State
Unlverslty

 

 

 

PLACE IN RETURN Box
to remove this checkout from your record.
To AVOID FINES return on or before date due.

MTE DUE MTE DUE MTE DUE
(1 El”: 9 9 1
v “ r 5 0",?

 

 

 

OCT 199

‘ II ©518%%

iii-63163638

 

 

Novcrnrzjgg .
315% 10802002

‘IJINOO Moot
'2 06 (I?

 

 

 

 

 

JAN QRNUZ’: 5
‘13:)
("X

 

“11 g {OZ-gut

 

 

 

 

 

 

use COMM“

LANGUAGE A'ITITUDES OF EDUCATED AFRICAN S TOWARDS
VARIETIES OF AFRICAN ENGLISH

By

Mungai Mutonya

A THESIS

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

MASTER OF ARTS
Department of Linguistics and Languages

1997

LANGUAGE ATTITUDES OF EDUCATED AFRICAN S TOWARDS VARIETIES
OF AFRICAN ENGLISH

By

Mungai Mutonya

The main objective of this research is to determine the attitudes of
educated Africans towards three major varieties of African English spoken in the
densely multilingual sub-Saharan Africa.

The study has adapted the use of stimulus voices to elicit responses, a
technique utilized in language attitude studies in other settings, but which, to the
best of our knowledge, one not utilized before in the study of language attitudes
in Africa. Therefore, this research has heavily focused on this methodology and
its application in eliciting language attitudes from an African sample.

Effort has been made to maximize the advantages of the traditional tools of
attitudinal research to suit our sample, especially in the preparation of the ratings
scales used by the respondents.

The ﬁndings show a diversity of attitudes by speakers of these varieties
towards their own and other varieties, based on regional afﬁliation and

accentedness.

DEDICATION

Wira uyu ndautegera:
Wakwa wa ngoroini- mwendwa Laura
Imwana ciakwa nyendete ma - Mwariri na Mutonya
na bamiri citu.

iii

ACKNOWLEDGMENTS

The completion of this thesis could not have been possible without the support
and guidance that I received from my committee. I am particularly indebted to my
academic advisor and Thesis Committee chair, Dr. Dennis R. Preston, for his most
valuable intellectual insights, encouragement and inspiration. He nurtured my intuitions
and provided me tutelage in quantitative analysis in sociolinguistic research that is the
backbone of this research. I am also indebted to the other members of my committee: Dr.
David Dwyer for his patience, willingness to listen, and his most valuable insights on the
status of language in Africa; and Dr. Grover Hudson for his thoroughness and stimulating
discussions and suggestions. Dr. Carolyn Harford was a committee member at the initial
stage of my research, and I am equally indebted to her input and support.

I am grateful to the Chairman, teachers, students, and staff of the Department of
Linguistics, Germanic, Slavic, Asian, and African Languages, at Michigan State
University for their friendship and comradeship.

Special thanks to my dear wife Laurah Wambui and our sons Mwariri and
Mutonya, for their love, understanding, endurance, and support. You are so special to me.
I owe every academic accomplishment to my dear parents Mutonya Njuguna and Njoki
Mutonya, who always took great pride in the academic achievement of their children,
and who always motivated us and encouraged us to try and reach the highest peak of any
mountain that we choose to climb. I owe alot to my brothers and sisters who have also

dared to conquer this montain to the delectation of Papa and Mama. All that motivation,

iv

and the support of my in-laws, makes the completion of this academic task even more
meaningful and enjoyable.

A very special word of gratitude to Dr. Mzalendo Kibunjia for his magnanimity -
Ndugu wema wako sitousahau daima.

TABLE OF CONTENTS

List of Tables viii
List of Figures x
Chapter One: Introduction

1.1. Statement of the Problem 1
1.2. Background to the Study 4
1.3 Hypotheses 6
1.4 Signiﬁcance of the Study 6
1.5. Scope and Limitations 7
Chapter 2: Background to the Study

2.1. Introduction 8
2.2. African English? 11
2.3. Regional Varieties of African English 14
2.3.1. West Aﬁ'ican English? (WAE) 14
2.3.2. East African English (EAE) 16
2.3.3. South Aﬁ'ican English (Black SAE) 18
2.4. Languages Attitudes 18
2.5. Language Attitudes in Africa 20

Chapter 3: Methodology

vi

3.1. Introduction: Methods of Elicitation
3.2.0. Elicitation of Attitudes

3.2.1. The Matched Guise Technique

3.3.0. Data Collection Method

3.3.1. Data Elicitation

3.3.2. The Sample

3.4.0. Data Analysis

3.4. 1. Mean Score

3.4.3. Factor Analysis

Chapter 4: Data Presentation and Discussion
4.1.1 Introduction

4.1.2. Mean Score rating and Ranking Analysis
4.1.3. A Correlation Coefﬁcient Analysis
4.1.4 Factor Analysis

4.1.4.1. East African Raters

4.1.4..2. South African Raters

4.1.4.3. West African Raters

4.2.0. Identiﬁcation of Regional Varieties
5.1. Conclusions

6.1. Recommendations

Appendix A : Questionnaire used in the Elicitation of Attitudes

Appendix B: Questionnaire used to Elicit Evaluative Adjectives
List of References

vii

21
22
22
24
25
27
28
28
29

31
31
38
46
47
52
57
61
63
67
68

71
73

LIST OF TABLES

Table 1 - Mean Ratings and Rankings of Mean Ratings of Each Regional
Variety by Eight East African Students

Table 2 - Mean Ratings and Rankings of Mean Ratings of Each Regional
Variety by Eight South African

Table 3 - Mean Ratings and Rankings of Mean Ratings of Each Regional
Variety by Eight West African students

Table 4 - Ranking of the HA Varieties by all Raters
Table 5 - Ranking of WA Varieties by all Raters

Table 6 - Correlation Coefﬁcients of the 12 Semantic Differential
Scales for East Africans

Table 7 - Correlation Coefﬁcients of the 12 Semantic Differential
Scales for South Africans

Table 8 - Correlation Coefﬁcients of the 12 Semantic Differential
Scales for West Africans

Table 9 - Factor Loadin gs for EA Raters

Table 10 - Interpretation of the Factor Loading in Table 9
Table 11 - Factor Loadings for South African Raters

Table 12 - Interpretation of the Factor Loadings in Table 11

Table 13 - Factor Loadings for West African Raters

viii

33

34

36
37
37

40

42

47

48

53

53

57

Table 14 - Interpretation of the Factor Loadings in Table 13

ix

57

LIST OF FIGURES

Figure 1- A comparison of WAE, RP, and EAE vowel systems 17
Figure 2 - An Example of a 7- Point Scale 23
Figure 3 - East African Raters Attitudes Towards the Regional Varieties 50
Figure 4 - East Aﬁ’ican Raters Attitudes Towards Accented Varieties 50
Figure 5 - East African Raters Attitudes Towards Individual Varieties 51
Figure 6 - South African Raters Attitudes Towards the Regional Varieties 54
Figure 7 - South African Raters Attitudes Towards Accented Varieties 55
Figure 8 - South African Raters Attitudes Towards Individual Varieties 56
Figure 9 -West African Raters Attitudes Towards RegionalVarieties 58
Figure 10 - West African Raters Attitudes Towards Accented Varieties 59

Figure 11 - West African Raters Attitudes Towards Individual Varieties 60
Figure 12 - Raters’ Identiﬁcation of the Anonymous Stimulus Voices 61

Chapter One: Introduction

1. 1 Statement of the problem
In spite of the linguistic complexities arising from the language contact

situation in densely multilingual sub-Saharan Africa, very little research has been
undertaken to determine the pattern of language attitudes in the region. According
to Myers (1996), the last comprehensive study on language was published ﬁfteen
years ago. Moreover, Myers hastens to add, these publications did not necessarily
focus on language attitudes.

Adegbija (1994), in his overview study of language attitude research in sub-
Saharan Africa, notes that there have been only ﬁve publications in a period of
about ten years that have focused on language attitude research. However, none of
these particular studies, or any other that we are aware of, have focused on the

attitudes towards regional varieties of English. Adegbija (1994:6) notes:

Also worthy of research interest, are attitudes towards the nativisation or indigenisation
of implanted varieties [Empean languages]...What are the attitudes of those who use the
nativised varieties of implanted languages towards the varieties they use?... much
sociolinguistic work speciﬁcally relating to attitudes of nativisation still has to be done.
This is particularly so when we take cognizance of the fact that these new varieties play
crucial roles in the educational systems of the countries concerned.

The increasing recognition and usage of African languages in the pursuit of
national goals in the political, educational, and social ﬁelds has had a strong impact
on the shift of attitudes towards the English variety spoken by Africans as a second
language. Historically, the RP variety (the ‘Queen’s English’) has been strongly
recommended for use in governmental and private institutions in the former British
colonies and as the model for intellectual achievement and acculturation. The RP

variety was associated with prestige, status, and elitism. The English varieties that

were inﬂuenced by the mother tongue of the non-native speakers continued to be
despised even in the post independence era.

However, the use of English in schools, universities, and broadcast media, as
well as in publications in sub-Saharan Africa, indicate a steady change in attitudes
towards English, as well as towards the various English varieties spoken by citizens
of these countries. The shifting trends in attitudes to English have not been
documented (as far as we know), and this research hopes to make a contribution in
that area.

The change in attitude towards sub-Saharan varieties of English may be
viewed from Moag’s (1982) perspective of the life cycles of non-native varieties in
the South Paciﬁc. Schmied (1991) adopts Moag’s model and identiﬁes ﬁve phases in
the life cycle of English in Africa: contact, institutionalization, expansion, recognition
or repression, and adoption. English in Africa, in Schmied’s view, ﬁnds itself in the
precarious fourth or ﬁfth stage, given the growing tide of cultural revival and social
reassessment in these nations. The two later stages usually witness the steady
institutionalization of the local languages and, implicitly, the change of attitudes
towards the English varieties.

It will be necessary to determine the status of both the RP variety and the
Africanized varieties of English.

A major limitation of the existing research on language attitudes in Africa is
the non-empirical nature of the methods used in data collection and analysis.

Adegbija (1994:53) observes:

Deﬁciency in methodology appears to be the principal weakness of most language attitude

studies carried out hitherto in sub-Saharan Africa. Some of such studies...appear to be
largely impressionistic in approach and sampling techniques and instruments, when
indicated, are often weak.

The solution to this problem, Adegbija proposes, is the utilization of ‘more

sophisticated, ...more thorough...sampling and analytical techniques.’(55) Baker

3

Baker (1992) recommends the use of a combination of approaches in language
attitude measurements as a solution to the problem.

This research is partially a response to the above challenges. It seeks to
ﬁll the research void that exists by attempting to measure and determine the
attitudes of educated Africans, living in the United States of America, towards
the heavily accented (hereafter referred to as HA) and weakly accented
(hereafter referred to as WA) major regional varieties of African English.
Measurement of attitudes is done by utilizing some empirical techniques of
sampling, data collection and analysis that have been used in language attitude
research in other speech communities, but, as far as we know, never with an
African sample.

In a much broader sense, this study seeks to combine, albeit
superﬁcially, the research techniques commonly used in two standard language
attitude research models: the strictly linguistically oriented Labovian model,
which focuses on the variation of the linguistic variable, on the one hand, and
the social psychological approach of Giles, on the other.

William Labov’s (1966, 1972) investigation of New York City linguistic
change and variation revealed a correlation between linguistic variation and the
speaker’s social class, age and speech styles. The application of the Labovian
model in this study is limited to the sampling process, in which distinct features
of the African English varieties (as outlined by Hancock and Angogo: 1982) are
used to determine the level of accentedness and regional afﬁliations of our
respondents. The techniques of indirect language attitude elicitation were
derived from the social psychological approach. The methods of statistical
analysis used in this study are commonly used in both approaches, though more
often associated with the latter.

One limitation to this study is the small size of the sample. The limited

number of respondent did not enable the gathering of data from as large a

4

sample as would have been most appropriate for some statistical models and for
stronger and more binding results. It is appropriate to reiterate, however, that
this is a preliminary study, with a heavy focus on methodology. The study not
only seeks to ﬁll the existing research void, but also aims at laying a
groundwork for future empirical research in language attitude studies in Africa.

1. 2 Backgron to the Study

English has always been the ofﬁcial medium of communication in most
countries in sub-Saharan Africa. The language policies of these African countries
have assigned English a dominant role as the ofﬁcial language of government, the
medium of instruction in schools, and, tacitly, the acknowledged ladder for social
and economic mobility. Consequently, the relation of English to other local
languages in these multilingual societies remains a diglossic relationship, with
English enjoying the prestigious status.

In a majority of cases, the learning of English in schools as a third language
is preceded by the acquisition of the mother tongue, and the national language or a
regionally dominant language. In such cases of multilingualism, mutual linguistic
inﬂuence is bound to take place. The English proﬁciency and the motivation of the
speakers to learn it is affected by social factors such as level of education, degree
of exposure to English, and regional afﬁliation of the speaker.

A general consensus among researchers in this ﬁeld of study is that African
English is not homogenous but that there exists broad regional variation. The
deﬁning linguistic characteristics of these varieties are inﬂuenced by the
indigenous languages spoken by the multilingual speakers of English.

Language contact situations cause the transfer of the features of one
language into the other. For instance, non-native speakers of English in Africa will

unconsciously transfer features of their ﬁrst language into English, resulting in the

5

distinct variation commonly referred to as African English (AB). Angogo and
Hancock (1982) identify four types of African English, which we believe applies
for all Anglophone regions:

Type 1: Native English of African-bom whites and expatriates.

Type 2: Native English of locally born Africans.

Type 3: Non-native English spoken ﬂuently as a second language.

Type 4: Non-native English spoken imperfectly as a foreign language.

Speakers of type 3 are our population for this study since it is:

perhaps the standard for AVE; it is spoken comfortably by people who have learnt an

African language ﬁrst in life, but who. because of their education and profession, have
grown up hearing and using English daily, and who speak it as well as, or maybe even
better than their ancestral language ( Angogo and Hancock 1982:72).

On the basis of broad linguistic commonalties of the many ﬁrst languages
spoken in regions where English is the ofﬁcial language, scholars (as discussed
in Chapter Two) have identiﬁed three regional varieties of African English:
West African English (WAE), East African English (EAE) and South African
English (SAE). The categorization is based on the alleged linguistic
homogeneity of the dominant indigenous languages of the N iger-Congo family
group. Although this study acknowledges the general homogeneity of the Bantu
substratum that may inﬂuence the English speakers in the Southern and Eastern
Africa regions, the Niger-Congo languages in West Africa are relatively
heterogeneous, contrary to Todd’s (1982) assertion (see section 2.3.1.). It is with
this realization in mind that we use the term West African English (WAE)
cautiously and treat the claim (of the existence of West African English) as a
secondary hypothesis in this preliminary study that attempts an empirical
investigation into the attitudes toward varieties of African English.

Insights into the attitudinal trends towards these regional varieties might

provide valuable input into language attitude research in general, as well as in

6

understanding the place of English as the language of a minority but powerful
elite group in Africa.

1.3 Hypotheses

1. Respondents from each region will have a more favorable attitude towards their
own English variety than the other varieties. They will rank their variety higher
than the other varieties

2. Respondents from all the regions will rank the weak accented

(WA) varieties as more prestigious than the heavy accented (HA) varieties.

3. Can respondents correctly identify the regional varieties of English in Africa?

1.4 Significance of the study

This sociolinguistic study aims at enhancing empirical research in language
attitude studies in Africa by testing methods and theories used in similar
research in monolingual and bilingual areas. Monolingual situations do not
provide the linguistic complexity that language-contact oriented research may
be faced with in a multilingual situation like sub-Saharan Africa.

Analysis of language attitudes based on observation or intuition is
undoubtedly less reliable than the measured and quantiﬁable results derived
from empirical research. This research is an attempt at a departure from the
previous attitude studies in Africa, which have largely provided non-empirical
evidence. Moreover, the theoretical approach adopted in this research addresses
to some degree a major argument concerning language attitude studies in
Africa, as stated by Schmied (1991:202):

With regard to the theoretical perspectives which underlie ﬁeldwork, many

sociolinguists (e.g. Jibril 1982) have asked themselves whether the Labovian

paradigm is invalid for empirical ﬁeldwork in Africa, or whether it merely has to be
adapted. Others have suggested a more comprehensive approach. such as the

7
ethnography of speaking suggested by Dell Hymes, or the network compromise

suggested by the Milroys. Up to now the data are far too limited to give any general
answer to such questions.

The study contributes to that list of possibilities and adaptations.

Since English is an integral part of the life of many African nations, it is
an indicator of status and relationships. The continued growth of the nativized
English varieties poses serious challenges, not only to language policy
formulators and educational planners, but also to sociolinguists. The latter have
the task of analyzing and reporting the changing status of languages, dialects
and varieties in such densely multilingual communities. Language attitude

research provides some vital information.

1. 5 Scope and Limitations
Our study focuses on students from English-speaking African countries
who are currently attending colleges and universities in the United States of
America. Our sample is limited to:
(a) those who speak English as a second language
(b) those whose mother tongues belong to the Niger-Congo language family
group
(c) these students are the source of the adjectival pairs used in the attitude
study which form the major part of this study, and they are the
respondents to the male voice samples which were used as stimuli in this
study.
The study is strictly limited to the analysis of language attitudes, and
language use is not analyzed. It does not delve deeply into Moag’s life cycle
of non-native languages’ model, nor does it assess its applicability or

appropriateness to the varieties studied.

Chapter Two: Background to the study

2.1 Introduction

Many studies have been carried out reﬂecting the status of English
language variation. The earliest and most detailed studies were regionally oriented.
These studies, conducted in Britain and the United States, sought to identify the
obsolescent words and pronunciations of the rural, less well-educated and elderly
folk. The language variety of this social group was regarded to be far less
corrupted by the innovations of modernity and other languages than the varieties
of the urban dwellers, the better educated, and of the younger speakers. The main
objective of these researchers was to locate and record those older forms of
English before they disappeared forever (Preston 1989).

Although the early research succeeded in recording those older forms of
English, it failed to reﬂect the diversity and dynamism of English usage in
language contact situations and among non-native speakers.

Modern studies in English dialectology and sociolinguistics of English have
acknowledged some of the factors that account for the diversity of English all over

the world: age, gender, ethnicity, and region. Bailey & Gorlach (1982:l) observe:

scholars now recognize that they must acknowledge the complexity of the diverse language
communities that make up the English-speaking world.

Various research methods have been used to show the correlation between
social and linguistic variables in the variations in English use. The spread of
English to many parts of the world has given new impetus to the study of such
variations, in view of the diverse linguistic environment that the language faces in

the different parts of the world.

F ishman, Cooper, and Conrad (1977) estimate the number of native
speakers of English in the world to be 300 million. The rapidly increasing number
of speakers (both native and non-native) is largely due to the role the language has
played, and continues to play, in trade and in the export of European religion,
education and culture to North America, Africa, the Caribbean, Asia and
elsewhere. The effects of such a spread to regions with diverse linguistic and
social backgrounds is the rise of English varieties which have distinct linguistic
features that are easily associated with those regions. Such English varieties
include at least American, Canadian, Australian, South African, East African and
others.

Language policy declarations and language attitudes of particular speech
communities towards English have had a considerable impact in determining the
institutionalization of either the national or regional varieties of English, as well as
the assigning of value and labeling of particular varieties as standard or
nonstandard, prestigious or stigmatized.

Bailey and Gorlach (1982) observe that the status and function of English
differ in various regions. In Britain, Ireland and the USA, it is the dominant
language in what is predominantly a monolingual situation; in Canada it coexists
with other languages; whereas in the densely multilingual situations in Africa and
India, it is the language of a substantial minority; the elite. It is essential therefore
to consider the correlation of language attitudes to the role and status of a
language variety in a speech community. This is the one of the objectives of
language attitude studies and a main objective of this study.

In multilingual situations, a nonstandard variety of English may appear
(with distinct linguistic characteristics) to be an emerging national or regional
standard among speakers even where the ofﬁcial policy still proclaims the RP

variety as the standard variety. In such a situation, the prestigious, standard variety

10

of that community, or what the community regards to be its standard form, is a
variety whose pronunciation is inﬂuenced by the mother tongue or the ﬁrst
language of the speakers. Although the prosodic as well as the segmental features
of the variety may indicate varying degrees of accentedness that distinguishes it
from other varieties, that speech community nevertheless regards it as a standard
variety. The beliefs and attitudes of the speakers towards these nativized varieties
may be borne consciously and unconsciously, but they play a pivotal role in
determining the role that the language assumes in the speech community. The
knowledge of attitudes towards language varieties, especially in a multilingual
society, is pertinent not only for pedagogical concerns, but for the understanding
of the linguistic and sociological development of a speech community.

The attitude of the native and non-native speakers of English towards such

non-native Englishes varies considerably. Kachru (1982: 43) observes:

The native speakers' attitude towards the development and the nativization of institutionalized
varieties has traditionally not been one of acceptance or ontological recognition. Because of the
linguistic manifestation of the nativization, these varieties have been considered deﬁcient models of
language acquisition.

On the other hand, the speakers of the nativized varieties may be more
receptive to the nativized variety than to the RP variety. Kachru (1982:45) ﬁrrther
claims:

educated Ghanaian English is acceptable; but as Sey (1973: 1) warns us, it does not entail
competence in speaking RP since in Ghana the type that strives too obviously to approximate to RP is
frowned upon as distasteful and pedantic. In Nigeria the situation is not different from Ghana or

India... Many Nigerians will consider as affected or even snobbish any Nigerian who speaks like a
native speaker of English.

This research uses quantitative methods used in sociolinguistic research in

an attempt to identify the attitudes of educated African towards these nativized

11

African varieties of English. Our focus is the East, West, and South African
Englishes.

2. 2 African English?

The term 'African English' has been used as a theoretical abstraction by the
few scholars who have done research on the forms and functions of English in
Anglophone Aﬁica. In these countries, although English is the ofﬁcial language, it
is spoken as a second language by only a minority of the population.

A few scholars have addressed this issue and some of the most notable
publications in this ﬁeld are: Todd (1982), Ladham (1982), Angogo and Hancock
(1982), Pride (1982), Kachru (1986), Cheshire (1991), Schmied (1991), Gorlach
(1991), and Adegbija (1994).

Is it presumptuous to conceive of such an idea as ‘African English,’ taking
into account that English-speaking sub-Saharan African countries are densely
multilingual, and although English is the declared ofﬁcial language, it is mostly
spoken by an educated minority as a second language? Some scholars have
attempted to answer this question as discussed below.

Gdrlach (1991:123) carried out tests among native English speakers of
British and American origins. He found that these students had no great difﬁculty
in identifying a speaker from Nigeria as African. Although this observation does
not provide much evidence regarding the existence of an African English, it
vaguely points to the fact there are some distinct linguistic features that mark the
Africanness of the English utterances of the speakers.

Schnried (1991 :2) deﬁnes African English as :

forms of English spoken by African speakers;this does not imply that there is an acknowledged

variety or that there are several distinct varieties of the language, nor that these forms are already
standardized and codiﬁed in any way.

12

Schmied’s deﬁnition uses the geographical criterion to identify the English
variety. It also groups together all forms of English, from the English-based
Creoles and Pidgins in West Africa, to other Africanized and non-Africanized
English varieties spoken in most English speaking regions. In view of the
linguistic diversity of Africa, this is a misleading deﬁnition.

Bokamba (1982:78) reiterates Gdrlach’s argument that native speakers of
English are able to positively identify an African English utterance. He further
expounds his argument about the existence of such a variety by providing

linguistic evidence that indicates the characteristics of the variety:

these Englishes share certain properties that can be identiﬁed as Africanisms, in that they reﬂect
structural characteristics of Aﬁican languages. Speciﬁcally, these properties can be discovered at all
linguistic levels: phonological, morphological, semantic and syntactic.

Bokamba’s study focuses on the syntactic and semantic properties of
sentences produced by educated Africans. Our area of concern in this study is
phonology.

The Africanisms deﬁned by Bokamba are in Lanham’s (1965) words
‘deviations from the authentic English’ which arise as a result of the transfer of

features and strategies previously known from the mother tongue to the target
language.

In every major linguistic area of the world where English is learnt as a second or foreign language,
there is a characteristic set of deviations from their authentic English, much of which is a point of
easy transfer from the mother tongue to English. With the passage of time these deviations become
institutionalized and give speciﬁc stamp to Indian English, African English in its various forms,
Spanish English, and so on. (Lanham I965 : 198)

The more compelling evidence that points to the existence of African
English is the phonological, syntactic, lexical and morphological evidence that has

been provided by some scholars to instantiate the difference between the African

13

variety and RP. This study focuses on the phonological evidence and especially on
pronunciation. Why focus on pronunciation?

It is not only the ‘most reliable test for localizing a speaker’ (Gdrlach
1991 :24), but it is also:

the most ﬂexible element, which can be used (subconsciously) to express subtle sociolinguistic
messages of speaker identity and of distance from or solidarity with the listener. (Schmied 1991:57)

Schmied continues:

the pronunciation of English in Africa is of particular importance because (non,- standard)
pronunciation features seem to be the most persistent in African varieties i.e., they are retained even
in the speech of the most educated speakers.

Similarities and variations in vowel and diphthong realization are the focus
in this study. In this study, variation in vowel realization is a yardstick used in the
sampling of prototypical speakers of a particular regional variety.

Gimson (1989:3180 notes that the British phoneme system may pose
problems to the foreign learner of the language:

...irr particular the opposition of the close vowels /i:l-/i/, lu:/-/u/, the existence of a long central vowel

/ 9:/ and the delicately differentiated ﬁont vowel set of /i:/- /i/- /e/- /E/ together with the signiﬁcant or
conditioned variations of vowel length, will pose problems to many foreign learners.

Such are the difﬁculties confronting the respondents in this study, whose
ﬁrst languages belong to the Niger-Congo family group. These language family
groups have 5-9 vowels.

Schmied makes three contrasting generalizations about the vowels of RP
and the African variety.

(i) There is a length difference in vowels. The African English short vowels are

longer and more peripheral than in RP, especially the closed /I/, /U/ or /A/.

14

(ii) Central vowel /A/, and /e/ as in but, bird and /a/ are avoided and tend towards
open positions [o,a].
(iii) Diphthongs tend to be monophtlrongized.

Schmied's study is the only one we are aware of that attempts such a broad
generalization about Aﬁican English. Other studies have attempted a descriptive
study of English in individual countries or an analysis of distinctive features in
each given region (cf. section 1.1.1). Such generalizations are not yet corrobated
by any compelling empirical evidence and (for lack of alternative studies in the
ﬁeld), can only be used cautiously as a ﬂame of reference in the sampling stage of
our research, to back up our intuition in African English, in selecting as a very
representative voice sample as possible. Our interest, of course, is on three broad
regional varieties of Aﬁican English, and our focus is on the distinct phonological

characteristics (especially the vowel system) of each region.

2. 3 Regional Varieties of African English

2. 3. 1 West African English (WAE)?

Given the linguistic complexity of the West African region, especially the
phonological, syntactic, and even typological distinctiveness among the cognates
of the Niger-Congo family, (see John Bendor-Samuel: 1989), the use of the term
West African English along with East African (EAE) and South African English
(SAE) as variants of Aﬁican English may be disputed. Whereas EAE and SAE
share a relatively homogeneous Bantu (sub-branch of the Niger-Congo)
substratum, WAE does not. We shall not delve deep into the details of this debate.
As stated earlier, the term WAE is used loosely and cautiously in this preliminary

study, for the purposes of undertaking an empirical study of language attitudes in

15

Africa. However, we present, without prejudice, the perspective of one scholar
who argues for the existence of such a variety.

English is the ofﬁcial language in six countries of the region. Todd (1982)
identiﬁes four types of WAE:
(i) Pidginized and Creoliud variety
(ii) Second language English that is acquired in schools and strongly
inﬂuenced by the mother tongue(s) of speakers
(iii) Standard WAE, which "... with the exception of Liberia " is
equated with British (RP) norms
(iv) Francophone WAE.
Our interest is in types (ii) and (iii).
Todd argues that there are distinct characteristics of WAE:

in spite of regional and educational differences, certain generalimtions can be

made about pronunciations of WAE largely because West African languages are
fairly similar in structure.

Similarity of the structure of the West Aﬁican languages, as claimed by Todd
above, is a highly contentious issue. Even among the Ni ger-Congo languages there
are structural differences; some have SOV word order whereas others have SVO,
moreover, vowel systems are not uniform in this family group (Williamson 1989).
Here we summarize Todd's description of the WAE vowel system:

(a) WAE has fewer vowel contrasts than RP. It utilizes 7 vowels and 3 or 4
diphthongs.

(b) Central vowels and centralized diphthongs are virtually nonexistent. Words
which end in [a] in RP have [a] in WAE, e.g. smoother [smuda], beer [bia], hair
[heal-

(c) Long central vowel fronting; bird [be :d]

(d) [A ] replaced by [a] : [bat] - [bot], [kAla] - [kola]

16

(e) [86] - [a]

Without delving into the WAE discussion, we cautiously use the concept in
this study and seek to determine whether our sample of educated Africans perceive
it as a regional variety and how the respondents judge and identify it relative to the

EAE and SAE varieties.

2. 3. 2 East African English (EAE)

While West Africa had little exposure to native speakers of English, East
Africa had a considerable number of native speakers who had a greater inﬂuence
in government and teaching during the colonial and the post-independence era.
Due to the continued reinforcement of British standards in the schools, EAE never
strayed far from the prestige dialect of England (Angogo and Hancock 1982). This
historical fact will undoubtedly affect the attitudes of raters towards EAE, and, in
fact, the results of this study show a much more favorable rating for the variety.

Hancock and Angogo identify four types of English in the region:
(i) Native English (NE) of whitesgand expatriates.
(ii) NE of Africans of racially or linguistically mixed marriages.
(iii) Non native English (NNE) spoken ﬂuently as a second language. The speech
variety of the educated, mid and higher status Africans, ‘...it is the speech of this
group which may be taken as the norm of the varieties of EAE.’ (308)
(iv) NNE imperfectly used as the foreign language by speakers who have little
knowledge of English.

The samples in this study were drawn from type (iii) speakers who are
educated and have a higher rate of contact with native speakers of English in
America and elsewhere.

The distinct vowel characteristics of EAE are:

(a) Vowel raising : [re n]- [ran]

17

(b) Contrast of [ a ] and [A] is collapsed to /a/
(c) Tensed and lax vowels are not distinguished.

Angogo (1980) illustrates the comparison between the WAE and EAE in
Figure 1.

Example RP WAE EAE
bead i: i 4
bade er e
bed a e e

bird
bad

 

 

bud

 

 

 

Figure l - A Comparison of RP, WAE and EAE Vowel Systems.

As stated above, there is no empirical evidence to support the above
analysis, but it is indicative of the perception and intution that such distinctive
variations exists and need to be investigated, analyzed, and reported in a empirical

manner.

18

2. 3. 3 South African English (Black SAE)

The absence of the SAE vowels in Figure l, is due to the fact that the
available literature analyzes the characteristics of only the native speakers of
English in the region. Literature on Black non-native speakers (Black English) is
scarce and sketchy.

Angogo and Hancock (1982) argue that there is a phonological similarity
between Black SAE and EAE varieties since both varieties are inﬂuenced by
indigenous languages which share a common Bantu substratum.

Such is the summary of evidence for the existence of the regional varieties
of English in Africa. This study seeks to illuminate the nature of attitudes of
educated Africans towards heavily accented (HA) and weakly accented (WA)

varieties of these three major regional varieties of African English.

2. 4 Language Attitudes

The study of attitudes of speakers towards a language or a language variety
has drawn considerable interest among social psychologists and sociolinguists.
The premise for language attitude study is that ‘hearers of language respond to the
personal, ethnic, national gender, class, role, age, and other identities of speakers'
(Preston 1989:50).

Most language attitude work by social psychologists is based on either a
mentalist or a behaviorist view of attitude (Fasold 1984:148). Agheyisi and
F ishman (1970) provide deﬁnitions and a survey of methodological approaches
that have dominated the ﬁeld for many years. Ryan and Giles (I982) outline
various approaches to the ﬁeld. The following summary of various language

attitude studies will place our current study in perspective.

19

Modern language attitude research has its origin in a bilingual setting.
Lambert, Hodgson, Gardner, Fillenbaum (1960) used the technique of a semantic
differential rating and the matched guise presentation of voices (see Chapter 3) to
determine the attitudes of French and English-speaking Canadian college students
to the two languages. Preston (1963) used a similar sample to determine the role of
gender in attitudinal responses to the two languages.

Lambert, Anisfeld and Yeni-Komshian’s (1965) research focused on
ethnicity as a factor in the attitudes of adolescents (from Arab Israel and Jewish
Israel) towards Arabic and Hebrew. Wolck’s (1972) study in Peru provided
evidence that, in bi- or multi-lingual situations, speakers will unconsciously assign
different roles to languages, as the case was for Quecha and Spanish.

Attitudes towards variation in dialects have also been studied. Speakers
with a standard dialect were rated favorably for competence, while speakers with a
nonstandard accent were rated less favorably for competence (e. g. intelligence,
industriousness) but favorably for integrity (e. g. helpfulness, trustworthiness) and
social attractiveness (friendliness and sense of humor) (cf. Ryan and Giles
1982:23). Such, studies were conducted by Harms (1961), Tucker and Lambert
(1969), Fraser (1972), Buck (1968) in America, Strongman and Woosley (1967),
Cheyne (1970), Giles (1971, 1973), Edwards (1977), Milroy and McClenaghan
(1977) in Britain and Ireland; Baker (1966), Mitchell and Delbridge (1965), and
Beechree and Ball (1979) in Australia. Other studies have been conducted in
Canada and Israel, yielding similar results. Language attitude studies have focused
on areas other than the ones we have mentioned above, but this review provides

sufﬁcient examples which contextualize our research.

20

2. 5 Language Attitude Studies in Africa

Adegbija (1994:57) observes that there are only very few studies of
language attitudes in Sub-Saharan Africa: Schmied 1985; Saab 1986; Sure 1991;
Webb 1992 and Adegbija 1992. The primary objective of these studies was to
identify speakers' attitudes towards English and indigenous languages in schools
and work places. The observations from the studies conﬁrmed that positive and
favorable attitudes were given to English in the ofﬁcial functions.

However, Adegbija (1994:53) laments a major shortcoming in these
studies:

Deﬁciency in methodology appears to be the principal weakness of most language attitude studies
carried out in sub-Saharan Africa. Some of such studies (e.g., Saab, 1986) appear to be largely
impressionistic in approach and sampling techniques and instruments, when indicated, are often weak.

The researchers not only used direct methods to elicit language attitudes,
but also relied on observation in making attitudinal judgments of their

respondents. Adegbija further observes that:

...the instruments they used and the statistical techniques employed [were not] powerful enough to
reveal ﬁne nuances and complexities of language attitudes.

He therefore proposes that:

...future research on language attitudes in sub-Saharan Africa would need to be less impressionistic,
more sophisticated in research design and more thorough in sampling and analytical techniques. They
would also need to use more detailed statistical procedures and methods. (55)

This study contributes to research in the ﬁeld of language attitudes in
Africa by taking the assumption that English in Africa is not a homogeneous
entity. There exists distinct interregional as well as intraregional variants. We also
contend that the indirect method is a more effective means of eliciting language
attitudes than the direct method, and that method is explained in detail in Chapter

Three.

Chapter Three: Methodology

3.1 Introduction: Methods of Elicitation

The two most commonly used methods of scientiﬁc investigation in
language attitude studies are the direct and the indirect methods. The direct
method encompasses two techniques: self-reporting and observation.

Self-reporting, as the term suggests, means relying on the respondents’
self evaluation and explicit statements about their attitudes towards a language.
It is a direct elicitation of language attitudes. Through interviews and
questionnaires, the researcher asks the respondents to state categorically their
attitudes towards a particular language or a language variety. (Ryan and Giles:
1982)

Proponents of this technique argue that it is effective in view of the fact
that the researcher’s questions are direct and speciﬁc. The respondent is clearly
aware of the researcher’s intentions. On the other hand, opponents of the
technique argue that respondents are likely to offer responses which only suit
the researcher’s perceived interests instead of the respondents’ ‘real’ attitudes.
Opponents also argue that the observer’s paradox, plus other social and
psychological factors, may be barriers in the search for the revelation of genuine

attitudes. The observer's paradox is the ﬁeld researcher’s realization that:

the more aware the respondents are that speech is being observed,
the less natural their performances will be. (Preston l989:7)

The observation method, also referred to as the content analysis of
societal treatment (Giles and Bourhis: 1976), gives the researcher the sole
prerogative to determine the societal attitudes towards a language by merely

observing the functions and status a language is assigned in society.

21

22

The observation method involves the identiﬁcation of language choices in
both formal and informal settings and language usage in speciﬁc contexts of the
speakers' interaction. The researcher determines how, why, and under what
circumstances the choices are made. Opponents of such a descriptive study
argue that mere observation does not reveal subjective attitudes towards a
language.

The indirect method has been considered by many researchers in recent
years to be more effective in eliciting language attitudes. It makes the
respondents reveal their beliefs toward a language variety without being overtly
conscious of participating in a task of language attitude revelation. Indirect
techniques of elicitation such as the matched- guise used in conjunction with
semantic-differential scale ratios are commonly used in the indirect method.

This research has utilized the techniques of the indirect method in the
elicitation of evaluative judgments of Aﬁican English speakers towards varieties
of Aﬁican English.

3. 2. 0 Elicitation of Language Attitudes
3. 2. 1 The Matched Guise Technique

The original matched-guise procedure was developed by Wallace Lambert
and his associates (F asold 1984). Through time the procedure has had numerous
modiﬁcations in its application. It involves recording the voices of respondents
as they read a word-list or a reading passage. These anonymous voices are
presented to other respondents who are drawn from a broad linguistic and social
spectrum (depending on the scope and objectives of the study) in terms of level

of education, social status, gender, age and regional or ethnic afﬁliation.

23

The subjects are asked to make subjective judgments about each voice on
a 7-point semantic differential scale, which is the main tool of attitude
measurement in this ﬁeld. Typically, a semantic differential scale involves the
evaluation of a concept by rating it on scales comprised of adjectival opposites
(Williams 1983). For instance, a scale testing whether a language variety evokes
in its listeners a sense of social closeness or distance may contain a scale with
adjectives such as ‘ﬁiendly,’ ‘humble,’ ‘trustworthy’ and so on. Such a scale is
presented to respondents as shown in Figure 2.

Voice 1.

Friendly ------- Unﬁiendly
Humble ------- Not humble
Trustworthy ------- Untrustworthy

Rm 2 - An example of a 7-point scale

The evaluators check a space in each item to indicate their impression of
each voice. A check closest to each positively marked feature on the scale
implies a very strong value in that regard; for example, a check closest to
‘ﬁiendly’ means the voice evokes an impression of the speaker being a very
friendly person, while a check closest to ‘unfriendly’ means that the particular
voice sounds to the respondent as that of a very unfriendly person. Each point
on the scale is given a numerical value (1-7). The values are assigned in a
descending order from the point closest to the positively valued adjective, such
as ‘ﬁiendly,’ which is assigned a 7, to the furthest one which is assigned a 1.

We chose eight evaluators from each of the three regions under study.

Each evaluator was asked to rate eight voices (described below) on a semantic

24

differential scale of twelve features (adjectives) for each voice. These adjectives,
as discussed in more detail later in this chapter, were elicited from an African
sample.

In the following sections we discuss the methods of data collection and

analysis used in this study.

3. 3. 0 Data Collection Methods

First, we identiﬁed speakers of the regional varieties of African English.
The identiﬁcation process was based on the knowledge of the speaker’s regional
afﬁliation, but of greater signiﬁcance was our judgment of the speaker’s level of
accentedness as a representative of West Aﬁican, East Aﬁican, or Black South
Aﬁican Englishes. Hancock and Angogo’s (1982) vowel distinctions of the
regions (see p.18 above) and this researcher’s intuitive awareness of those
differences as reﬂected in the speakers’ pronunciation were heavily relied on for
this task.

A short reading passage was created, with deliberate attempts being made
to highlight the vowel contrasts that mark the distinction between WAE and
EAE. Our contention was that a passage provides an opportunity for the readers
to make continous utterances that expose the subtle characteristics of their
speech.

The passage which the respondents were asked to read was as follows:

The hot African sun baked his bare head. Warm salty beads of sweat aggravated his thirst and

he yearned for a cold brew. He longed to lay on his lion- skin bed and drink from his male
buffalo horn. Who could have imagined that at this time of the year he could be this far
away. There were a few birds singing, imaginary pools of water kept disappearing any time he
approached them. He thought he had occasionally heard voices. Voices? Voices from where?
Well, he could not tell, he was not sure; all he knew for sure was that the mysterious voices
were calling his sweet sounding name, loud and clear.

The passage was made up by the researcher for this particular research.

25

3.3.1 Data Elicitation

On the stylistic continuum, the reading passage was preferred over a word
list, in cognizance of the fact that distinctive features of an individual’s speech
are more discemable in the relatively less formal style provided for by the reading
passage than in the more careful word list style.

The ﬁrst step in the elicitation procedure involved determining how
Africans evaluate a wide spectrum of English varieties from different regions of
the world. As discussed in Chapter Two, there is ample documentation from
language attitudinal surveys from numerous regions, but, as far as we know, no
previous language attitude study on Africa has attempted to elicit evaluative
adjectives from Aﬁican respondents. These adjectives are used as the
parameters for language attitude measurements on the semantic differential scale.
Researchers have used the adjectives used to elicit attitudes from non-African
respondents in total disregard of any cross-cultural inhibitions that may occur in
the interpretation of labels.

Our earlier attempt at borrowing adjectives used successfully in similar
studies in other regions was met with strong objections, sharp criticism and
negative responses ﬁorn our African respondents. Adjectives used in previous
research studies include: good upbringing, poor upbringing, faith in God, quick,
slow, brilliant, dull, warm, cold, intelligent, ignorant, diligent, lazy, sharp,
affectionate, smart, happy, and .many more. Our ﬁrst task was to elicit appropriate
adjectives, in this particular regard, for our groundbreaking task in language
attitude measurement in Aﬁ'ica.

We had two stages of data collection and two sets of samples. The

preliminary stage involved the elicitation of evaluative adjectives from an

26

uncontrolled sample of 80 Africans. The core study involved the elicitation of
language attitudes ﬁ'om a controlled sample of 24 Aﬁicans.

For the preliminary investigation, eight carefully selected voices were
recorded. Six from the three Africans regions under study and two from non-
African respondents who also speak English as a second language. The rationale
for adding the two non-African voices was, as discussed above, to elicit
attitudes towards English varieties in general, without focusing on a speciﬁc
variety.

Subsequently, the eight recorded voices were presented to Africans from
English speaking countries. Our sample in this taskwas not restricted by such
social factors as gender, age, status or level of education. Respondents were
asked to listen to each voice and to write down any impressions of the speaker
that struck them when they listened to the anonymous voice (see appendix). To
avoid subjective evaluations based on the identity of the voice, we avoided the
opinion of Michigan State University students who were likely to identify the
stimulus voices. Questionnaires and audio tapes containing the 8 voices were
dispatched to research assistants in three locations: University of Iowa,
University of California at San Diego, and Boston University. A total of 80
questionnaires were ﬁlled out and returned. (see Appendix for questionnaire
sample).

The adjectives which were most frequently used by the different
respondents in evaluating the voices were selected for the subsequent rating
task. The number in parentheses represents the total number of respondents who
identiﬁed that particular adjective for any of the 8 voices. Well-educated (77),
conﬁdent (71), responsible (67), gentle (72), ﬁiendly (78), polite (66), leader-like
(56), patient (62), proud (76), high status (67 ), trustworthy (53), social (73). These

27

adjectives, paired with their negative partners, were subsequently used on the
semantic differential scale.

The adjectives shown above are different ﬁ'om those used in earlier studies
as exempliﬁed in the non-African and African-oriented attitude studies discussed
below. Tucker and Lambert (1972:178-84) used the following adjectives to
determine the attitudes of Whites and Black American students towards various
American-English dialects: Upbringing, intelligent, disposition, speech, faith in
God, talented, character, personality, considerate, among others. Woods’ (1994)
study of attitudes towards French, national languages, and mother tongues in
the Republic of Congo (Africa) elicited attitudes of his Congolese respondents
by using such adjectives as beautiful, intelligent, trustworthy, and friendly.
Woods’ method is different from the one used in this study in two ways. First,
Wood uses the direct method of elicitation, while we use the indirect method.
Secondly, Wood asked respondents to evaluate the language, while we asked
respondents to evaluate anonymous voices of speakers of language varieties

under study.

3. 3. 2 The Sample
The stimulus voice samples were selected from Michigan State
University’s African male student population. Students from English speaking
African countries who learnt English as a second or third language in their early
years of formal education were the focus of this study. All the respondents had
lived in English speaking communities for a period of 1-6 years. The age range of
the respondents was 23-48 years. The sample was comprised of:
(a) Two respondents from West Africa. The respondents were chosen ﬁ'om

speakers whose ﬁrst language is of the Niger-Congo language family group

28

with English as their second or third language. A 39 years old Sierra Leonian, a
speaker of Mende (mother tongue) and Temne languages, who had lived in
Engish-speaking communities for three years, provided the heavy accented
variety. The weak accented voice was that of a Ghanaian, 29 years old, speaker
of Asante (mother tongue) and Akan Fante languages, and who had lived in
Europe and America for 11 years.
(b) Two East Aﬁicans. The heavy accented voice belonged to a Kenyan, 39
years old, and speaker of Ekegusii (mother tongue), Kiswahili and Gikuyu
languages. He had lived in the United States for two years. The weak accented
speaker was a Chichewa and Chinyanga speaking Malawian who had lived in
Britain and United States for a combined eleven years. He was 28 years old.
(c) Two respondents from Southern Africa . Like the other respondents,
English is their third language.The heavy accented speaker was a 22 year old,
Zulu (mother tongue) and Xhosa speaker. His duration of stay in the United
States was 5 years. The Setswana (mother tongue) and Shona speaking, male
from Botswana provided the weak accented voice. He had lived in the United
States of America for a year. He was 24 years old.
Eight respondents from each of the three regions under study were asked
to evaluate each of the six stimulus voices on a semantic differential scale. (see

Appendix A for questionnaire)

3. 4. 0 Data Analysis
3. 4. 1 Mean Score
After the twenty-four evaluators had rated the voices, analysis was done

to determine the raters attitudes towards the various regional varieties of English

29

in Aﬁica. Their rating of the voices on a semantic differential scale is considered
to be a reﬂection of their attitudes towards the variety.

An SPSS (Statistical Program for Social Sciences) computer program was
used to calculate the mean score of the scale for each region and individual voice
on each of the twelve adjectives. Tucker and Lambert's (1972) method of ranking
each trait was used to determine how judges evaluated each voice for each trait.
The mean score is calculated by dividing the total value by the total number of
respondents. The mean score for all other measured features was calculated and
a factor analysis done.

Two methods of analysis were used to get a quick and fast estimate of the
evaluators’ attitudes to the language varieties. First, as stated above, Tucker and
Larnbert's (1972) method of rating and ranking of mean scores was done. It
indicates the evaluators’ preference of the voice and how they associate the
voices with the different traits. Secondly, a correlation coefﬁcient matrix for each
region was calculated to determine the degree of correlation of the traits based
on their level of signiﬁcance. Correlation coeffrcient may not be as efﬁcient as the
factor analysis method, but taking into account the exploratory nature of this
research, we attempted to exploit the relevant techniques that can provide it s
with a thorough and exhaustive analysis of the attitudes towards the varieties of

African English.

3. 4. 2 Factor Analysis

A factor analysis technique is utilized to analyze the mean scores and to
determine the underlying factor groups of the semantic differential scale. Hatch
and Farhady (1982) note that this analytical technique is based on the

assumption that in every test there is probably one or more underlying traits that

30

are being assessed. The technique helps us to condense and cluster together
variables that we tested in terms of correlation of the factors and the observed
score. The scores indicate which items have a common variance and can be
identiﬁed as a factor under a new label, for instance a hypothetical factor called
‘competence' could comprise of variables such as ‘responsible,’ ‘well educated,’
and ‘leader-like,’ if the scores indicate that they correlate as an underlying factor
group.

Factors are rotated using the Varirrrax rotation technique in order to
maximize the correlation between a variable and the factor to which it belongs,
whereas the correlation between a variable and the factor to which the variable
does not belong is minimized. A variable is considered to be positively loaded in

a given factor group if its loading is 0.30 or greater.

Chapter 4. Data Presentation and Discussion
4.1.1 Introduction

In this chapter, the methods of data analysis outlined in chapter 3 are
applied in order to determine the attitudes of the educated African raters towards
the target African English varieties. The methods are rating and ranking,
correlation coefﬁcient, and factor analysis.

As stated earlier, we hope the main contribution of this analysis to
language attitude research in Africa will be in the realm of methodology.

The mean score rating and ranking method provides very preliminary
results of the raters’ attitudes towards the language varieties in a very speciﬁc
trait by trait analysis, as well as an overall comparative ranking of the speakers of
each variety.

The correlation coefficient method seeks to determine how the raters
unconsciously correlate the twelve semantic differential scale adjectives they use
in evaluating the six stimulus voices. The method combines correlated adjectives
into a cluster that gives a more deﬁnite identity to all those members of the set.

The factor analysis is the main analysis, which we have exploited
extensively for the purposes of obtaining the attitudes of the raters towards the

varieties in a more principled reduction of the individual traits into clusters.

4.1. 2 Mean Score Rating and Ranking Analysis

A mean score of each voice on each of the twelve adjectives was calculated
for the three regions. This is the method used by Tuckeranﬂd Lambert (1972), and
the results are shown in Tables 1, 2, and 3. This method provides preliminary

results of the semantic differential scale. The brackets in each column refer to the

31

32

rank of each variety as regards that feature. The overall rank (for each of the six
stimulus voices) is shown in the "sum" column. This ﬁgure is derived from the

addition of all the rank scores for each voice; the ranks are based on high

 

(positive) values, but the sum score is lowest for the most positively ranked
voice since it is based on rank rather than value.

Table 1 shows the ratings and rankings of the 6 stimulus voices by East
African raters. An overview of the rankings indicates a generally more favorable
ranking of the weak accented varieties (hereafter referred to as WA). Note that
two heavy accented varieties (hereafter referred to as HA) take the last two
positions of the ranking. (We shall continue to use the abbreviations of the
varieties to refer to the speakers of that particular variety). Furthermore, there is a
striking numerical break between the sums of the fourth ranked WAE WA (40)
and the ﬁfth ranked EAE HA (54) voice. However, WAE HA is ranked second
(32).

7k EAE WA and EAE HA are especially contrastive. EAE WA is ranked high
(i.e., 3 or above) for all traits except for patience, politeness, and pride. EAE HA is
ranked high only for gentleness, politeness, patience, and social.

There is a close correlation in the rankings of EAE and SAE voices but the
WAE HA voice tends to distort the seemingly orderly and predictable pattern of
ranking of the categories. I

In Table 2, the ratings given by Southern Africa raters, the WA speech
varieties are identiﬁed with the following traits: well educated, conﬁdent,
responsible, gentle, polite, patient, high status, not friendly, proud, not social,
and not leader-like. The HA varieties on the other hand are identiﬁed with the
following traits: gentle, ﬁiendly, polite, patient, leader-like, social, not proud, not

well educated, not conﬁdent, not responsible, and low status.

33

TABLE 1 - Mean Ratings and Rank of Mean Ratings of Each Regional
Variety by Eight East African Students

 

 

 

 

 

 

 

 

Edl con rgp ent frie poli lead pati pro stat trus soci sum
EAE 4.8 4.0 4.6 5.0 5.0 4.8 3.8 3.4 4.2 4.2 5.2 4.8
WA2 [1] [2L [1] 11 i2] [4] [3] 16] [2] I3] [1] II] 27
WAE 4.0 3.0 4.2 4.6 5.0 5.2 5.0 4.4 3.4 4.4 4.8 4.4
HA3 [3] [51 L2] [21 [2] [2] 11] I31 [4] [21 [2] [4] 32
SAE 4.6 4.2 4.0 4.2 4.6 4.4 4.8 4.0 4.4 4.8 2.6 4.6
WA [2] [ll [3] [5] . I3] [51 12] [41 ll] [11 [6] I21 35
WAE 3.8 3.6 3.4 4.4 5.2 5.2 3.4 5.0 3.8 3.6 4.6 4.4
WA 16 J3] [51 [3] [11 [2] [41 [1] l3] [5] [31 14] 40
EAE 3.8 2.2 3.2 4.6 4.0 5.6 3.0 4.6 2.8 3.4 3.6 4.6
HA [6] [6] i6] [2] [5] ll] [6] [2] I 6] I 6] [6] 12] 54
SAE 4.0 3.4 4.0 4.0 3.8 4.2 3.4 4.0 3.0 4.0 3.8 4.2
HA [3] [4] I 31 l6] [6] [61 [41 I4] [5] [4] [4] 16] 55

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1 Ed=educated, con=conﬁdent, resp=responsible, gen=gentle, frie=friendly, poli=polite,
lead= leader-like, pati= patient, pro=proud, stat=status, trus=trustworthy, soci=social

2 WA = weakly accented

3 HA = heavily accented

Among Southern Aﬁica raters (Table 2), the ranking of EAE voices does
not conform with the seemingly well deﬁned split of HA and WA. The EAE HA
and WA varieties fall in the opposite ends of the ranking scale. EAE HA gets the
highest overall making while EAE WA gets the lowest.

34

TABLE 2 - Mean Ratings and Mean Rankings of Each Regional Variety by
Fight Southern Aﬁica Students

 

Edu conf resp gent flien polit lead patie prou stat trust soci sum

 

EAE 5.6 5.6 5.2 4.8 5.2 4.8 4.4 4.8 4.2 4.6 5.2 5.4 23
HA [2] [1] [21 I3] [2] [I] [3] [1] [2] I31 [1] [21

 

SAE 5.2 5.0 5.0 4.8 4.6 4.6 3.2 4.6 4.2 4.8 4.8 5.4
WA [41 12] [3] [31 [51 [21 [5] [2] [21 [1] [31 [2] 34

 

WAE 5.6 5.0 4.4 5.2 4.4 4.6 4.6 4.0 3.8 4.8 3.8 5.2
WA [2] [2] [5] [1] [6] [2] [1] [3] [4] [1] [6] [4] 37

 

WAE 5.2 3.8 4.6 4.8 5.2 4.4 4.6 3.8 4.4 3.8 4.6 4.8
HA [4] 61 I41 [3] [2] I5] [1] [51 Ill [5] [41,151 45

 

SAE 3.8 4.4 4.2 5.0 6.0 4.6 3.0 4.0 3.2 3.8 5.2 5.6
HA [6] [4] [6] [2] [ll [2] [6] [3] [6] I5] [1] I1] 47

 

 

EAE 5.8 4.0 5.6 4.2 4.8 4.4 3.6 3.8 3.8 4.4 4.6 4.6
WA [11 151 [ll [6] [41 [5] I41 [5] [4] I4] [41 I6] 49

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Note the striking numerical break between the sums of the ﬁrst ranked EAE HA
(23) and the second ranked SAE WA (34). The South African raters’ rank their
local variety (SAE WA) second unlike the East Africans (Table 1) who rank their
own (EAE WA) variety ﬁrst.

Finally, Table 3 shows that the West Africans rate the WA variety

speakers higher than the HA variety speakers, and there are no exceptions. The

 

35

WA varieties are associated with the following traits: well educated, conﬁdent,
responsible, gentle, polite, patient, proud, high status, social, not friendly, not
leader like, and not trustworthy. The HA varieties are identiﬁed with the traits
educated, not conﬁdent, responsible, not gentle, ﬁiendly, not polite, leader like,
not patient, not proud, low status, trustworthy, and social.

The West Africans rank their WAE WA variety lower (36) than the SAE
WA (32) and EAE WA (32) varieties. However, the WAE HA is ranked higher
(41) than the EAE HA (47) and SAE HA (49). In Table 2, the South Aﬁicans also
give their HA variety the lowest ranking.

A summary of ranking of the regional varieties is shown in Tables 4 and 5.
The columns represent the regional varieties of English and the ﬁrst three rows

(u

represent the raters from each region. The numbers are drawn from the ranking of

 

 

each category in Tables 1, 2, and 3. The last row in both tables indicates the sum
of the rankings of each variety. For example, Table 4 shows that the East Aﬁican
raters rank WAE HA speakers highest; EAE HA speakers rank second while
SAE HA speakers rank the lowest in that category. The overall ranking of the HA
variety by all raters shows a very favorable attitude towards WAE HA (10),
followed by EAE HA (I 1), while SAE HA (17) gets the least favorable rating.

Table 5 shows that the highest ranked WA variety is the SAE (6), EAE is
second (9), while WAE is last in the ranking (10). It is interesting to note that
whereas the SAE HA gets a very poor score from all the raters including the
South Africans, the SAE WA almost gets the best score from all raters except the
East Africans who give the highest ranking to their own variety.

36

Table 3 - Mean Ratings and Ranks of Mean Ratings of Each Regional Variety by
Eight West Aﬁican Students

 

Edu Co Resp Gent Frie Polit lead patie Prou Stat Trus Soci
sum

 

 

SAE 4.4 5.2 4.4 5.4 4.4 4.6 4.6 5.2 4.4 4.6 5.0 4.3
“M [5] [3] [3] [2] [5] [4] [1] [2] [1] [1] [2] [2] 32
EAE 6.0 5.6 5.2 4.2 4.6 4.8 4.0 5.6 4.2 4.2 4.2 4.8
WA [1] [1] [1] [6] [4] [3] [4] [1] [2] [2] [5] [2] 32
WE 6.0 5.4 4.4 5.0 5.2 5.0 4.0 3.6 3.4 3.8 4.6 4.2
I: [1] [2] [3] [3] [1] [1] [4] [6] [4] [3] [4] [4] 36
WAE 5.0 3.6 4.2 5.6 5.0 5.0 4.4 4.4 2.8 3.0 4.8 4.4
f:- [31 [6] [6] [1] .[2] [1] [2] [3] [6] [5] [3] [3] 41
EAE 3.8 4.8 4.4 4.6 4.8 4.2 3.8 3.8 3.2 3.8 5.2 5.2
: [6] [4] [3] [4] [3] [6] [6] [5] [5] [3] [1] [1] 47

 

 

 

 

 

SAE 5.0 4.0 4.8 4.6 4.6 4.4 4.4 4.2 3.8 2.4 4.2 3.4
"A [3] [5] [2] [41 [4] [5] [2] [4] [3] [6] [5] [6] 49

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

There is little difference in the HA and WA version of WAE, but the
striking reversal of expected scores in the SA ratings of the HA and WA
versions of EA is difficult to account for.

The comparative voice rankings as shown as in the ‘sum’ columns of
Tables 1, 2, and 3 are summarized in Tables 4 and 5 below. The tables show how
each group of raters ﬁ'om each region ranked each individual voice on the basis
of its level of accentedness. For example, in Table 4, the EA raters ranked their
heavy accented variety (EAh) ﬁfth overall, SAh last, but WAh was ranked
second (see Table 1). Overall, the ranking of HA varieties by all raters is not as

 

37

high as the WA varieties (T able 5). Most HA varieties are ranked fourth, ﬁfth,
and sixth, while most WA varieties are ranked in the upper tier.

Table 4 - Ranking of HA varieties by all Raters

 

 

 

 

 

 

EAh SAh WAh
EA 5 6 2
_sz 1 5 4
WA 5 6 4
Sum 1 1 l 7 10

 

 

 

 

 

Table 5 - Ranking of WA Varieties by all Raters

 

 

 

 

 

 

 

 

 

 

EAw SAw WAw
EA 1 3 4
SA 6 2 3
WA 2 l 3
sum 9 6 l 0

 

It is evident from the rating and ranking analysis that the raters ﬁorn all
regions make a distinction between speakers with a HA or WA variety of African
English. The WA varieties get more favorable rankings in general than the HA
varieties, and the HA varieties are associated with the more positive Semantic
Differential Scale traits while the WA are associated with the more negative ones.
The attitudes towards the various regional varieties varies considerable among

the raters.

38

At this level of our analysis the ﬁnding conﬁrms our hypothesis that
attitudes towards speakers of the regional varieties is diverse. However, we need

further analysis to evaluate these and other distinctions more carefully.

4.1.3 A Correlation Coefficients Analysis

The next step involved analyzing the correlation coefﬁcients among the 12
Semantic Differential Scale traits for each region.
_ Correlation coefﬁcient is a measure of linear association between any two
variables, in this case, two semantic differential adjectives. A high correlation

coefﬁcient score, such as .62, indicates a relatively strong linear relationship

'..——e\

between the column variable and the row variable. Conversely, a smaller number
(closer to zero) indicates a weak relationship between two semantic differential
adjectives.

In the following correlation coefﬁcient matrices (Tables 6, 7, and 8), each
entry represents the correlation coefﬁcient between the row variables and the
column variables. Since the correlation matrix is symmetric, only the lower
triangle of the matrix needs to be presented. Since the correlation between the
same variables (such as the column variable 1 and the row variable 1) is always
equal to 1.00, that diagonal of the matrix is omitted.

The matrices can help us determine which of the 12 semantic differential
adjectives used in the study can be clustered together. The criteria used to
determine the strength of the correlation of the variables is their signiﬁcance
level.

In the three correlation coefﬁcient matrix tables below, * indicates a
signiﬁcance level of .05, while ** indicates that the correlation coefﬁcient is

signiﬁcantly different from zero at a signiﬁcance level of .01.

39

Correlation gives a measure of how one characteristic varies from another. Ifthe
two vary together perfectly in a positive direction, then we would obtain a
coefﬁcient of correlation of +1.00. If they varied in the exact opposite direction
from each other, the correlation coefﬁcient would be -1 .00. If they were not related
to each other at all, the coefﬁcient of correlation would be 0.00. Thus, coefﬁcients
of correlation of 0.83, ft! example, would represent a very strong positive
correlation. A value of -0.53 would be a moderate negative correlation. (Fasold
1984: 103) “

Although a subsequent factor analysis will show these relationships
from a more general point of view, the correlations allow the determination of
trends in the categories, item by item. I will not detail all those possibilities but
pick out some of the strongest correlation sets.

/ For example, beginning with ‘education’ for EA raters (Table 6), we can
see that it correlates signiﬁcantly with ‘conﬁdence,’ ‘gentle,’ ‘leader-like,’
‘patience,’ ‘proud,’ ‘status,’ ‘trustworthy,’ and ‘socialness,’ which is a large set.
The following are the other clusters:

‘Conﬁdencez’ correlates with ‘educated,’ ‘responsibility,’ ‘gentle,’

‘leader—like,’ ‘proud,’ ‘status,’ ‘trustworthy,’ ‘social,’ a set very like the

ﬁrst.

‘Responsiblez’ ‘Educated,’ ‘conﬁdent,’ ‘gentle,’ ‘status,’ ‘trustworthy,’ and

‘social.’

‘Gentle:’ ‘Educated,’ ‘conﬁdent,’ ‘responsible,’ ‘friendly,’ ‘patient,’

‘trustworthy,’ and ‘social.’

‘Friendlyz’ ‘gentle,’ ‘polite,’ and ‘patient.’
‘Politez’ ‘gentle,’ ‘friendly,’ ‘patient,’ and ‘social.’
‘Leader-like:’ ‘educated,’ ‘conﬁdent, ‘proud,’ ‘status,’ and
‘social.’
‘Patientz’ ‘educated,’ ‘gentle,’ ‘friendly,’ ‘polite,’ ‘trustworthy,’ and

‘social.’

40

‘Proud:’ ‘educated,’ ‘conﬁdent,’ ‘leader-like,’ and ‘status.’
‘Status:’ ‘educated,’ ‘conﬁdent,’ ‘responsible,’ ‘leader-like,’ ‘proud,’ and
‘social.’
‘Trustworthy :’ ‘educated,’ ‘conﬁdent,’ ‘responsible,’ ‘gentle,’
‘patient,’ and ‘social.’

‘Socialz’ ‘educated,’ ‘conﬁdent,’ ‘rcsponsible,’ ‘gentle,’ ‘polite,’ ‘leader-like.’

Table 6 -.Correlation Coeﬂicients of the 12 Semantic Differential Scale for East

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Aﬁican Raters

edu con resp gent the lit lead pat proud statu trust soci
edu
con .57‘
resp .30 .51‘
_gent .47" .33‘I .43“
liie .21 .17 .30 .62‘
polit .21 .16 .08 .61“ .46‘
lead .49" .34‘ .29 .23 .19 -.l l
pat .38’ .22 .10 .68‘ .49‘ .64‘ .ll
proud .36‘ .52‘ .21 -.I4 .02 -.I9 .44‘ -.07
stat .55“ .57‘ .43“ .15 .01 -.01 .55‘ .18 .50‘
trus .33’ .32‘ .39‘ .33‘ .25 .22 .23 .36’ .01 .29
soci .61" .46" .45" .53" .30 .46" .34' .40‘ .22 .53" .25
*= signif. LE .05 **= signifLE .01 (2-tailed)

An overview of the correlation of the traits as perceived by the East

Aﬁican raters indicate a pattern in the cluster combinations as follows:

41

‘Educated’ correlates signiﬁcantly with all other traits except ‘responsible’
and ‘ﬁiendly.’ ‘Conﬁdence’ and ‘responsible’ have a combination of traits very
similar to ‘educated.’ ‘Conﬁdence’ correlates with all other traits but ‘ﬁiendly,’
‘polite,’ and ‘patient,’ and ‘responsible’ also does not correlate with these three
in addition to ‘leader-like.’

‘Gentle’ correlates signiﬁcantly with ‘educated,’ ‘conﬁdent,’ ‘responsible,’
and ‘ﬁiendly.’

‘Friendly’ and ‘polite’ have a somewhat similar combination to ‘gentle.’
Both clusters correlate signiﬁcantly with ‘gentle’, ‘polite’, and with each other. In
addition to the above-mentioned traits ‘polite’ correlates with ‘social. ’

‘Leader-like,’ ‘proud,’ and ‘status’ are the only clusters that do not
signiﬁcantly correlate with any of the following traits: ‘Gentle,’ ‘ﬁiendly,’ ‘polite,’
and ‘patient.’ Moreover, these three clusters plus ‘educated’ and ‘conﬁdent’ are
the only ones that correlate with ‘proud.’ However, the correlating traits in each
of the three clusters is different as shown above.

‘Patient’ correlates with ‘educated,’ ‘gentle,’ ‘friendly,’ ‘polite,’
‘trustworthy,’ and ‘social.’

‘Trustworthy’ and ‘social’ have the traits ‘educated,’ ‘conﬁdent,’
‘responsible,’ ‘gentle’ as members of the sets. Other members of the
‘trustworthy’ cluster are ‘patient’ and ‘social,’ while ‘polite’ and ‘leader-like’ are

additions to the ‘social’ set.

‘Educated :’ ‘conﬁdent,’ ‘leader-like,’ ‘status.’
‘Conﬁdent :’ ‘educated,’ ‘leader-like,’ ‘status.’
‘Responsible :’ ‘polite,’ and ‘patient.’

‘Gentle :’ ‘ﬁiendly,’ ‘patient,’ and ‘patient.’

‘Friendly :’ ‘gentle,’ ‘polite,’ ‘trustworthy,’ and ‘social.’

42

‘Polite :’ ‘responsible,’ ‘ﬁiendly,’ ‘trustworthy,’ and ‘social.’
‘Leader-like :’ ‘educated,’ ‘conﬁdent,’ ‘proud,’ and ‘status.’
‘Patient :’ ‘responsible,’ ‘gentle,’ ‘polite,’ and ‘trustworthy.’
‘Proud :’ ‘leader-like.’

‘Status :’ ‘educated,’ ‘conﬁdent,’ and ‘leader-like.’
‘Trustworthy :’ ‘ﬁiendly,’ ‘polite,’ and ‘patient.’

‘Social :’ ‘gentle,’ ‘friendly’ and ‘social.’

Table 7 - Correlation Coefﬁcients of 12 Semantic Differential Scales for
Southern Africa Raters

 

edu can rent 4gent ﬁi: jpoht had red rnoud snuu bust soci

edu
con .50"
resp .22 .18

gent -.14 .or .16

ﬁie -.23 -.12 .30 .36"

polit -.18 .01 .37. .15 .33.

lead 37* .46" .22 -.13 -.0s .04

pat -.09 .16 .353 .31‘ .20 .53s -.07

prou .16 .28 .28 .l0 .09 -.02 .34. -.16

stat .65‘ .63’ .02 -.20 -.48“ -.08 .36‘ .21 -.01

trus -.43 -.13 .24 .10 .41" .48" -.32s .49" .03 --35‘

soci -.36 -.05 -.01 .41" .55" .25 -.30"' .10 .16 -.50 .41”

 

. = Signif. LE .05 u = Signif. LE .0l (2-tailed)

43

Among the South African raters, the sets ‘educated’ and ‘conﬁdent’ are a
mirror image of one another. They correlate with each other plus ‘leader-like’ and
‘status’.

An overview of all the clusters indicate two emerging patterns: a group of
clusters that contain any, some, or all of these four traits: ‘gentle,’ ‘ﬁiendly,’
‘polite,’ and ‘patient.’ The second group is a complement of the ﬁrst one: it is a
group of clusters that do not contain any of the four traits.

In the ﬁrst group we have clusters such as ‘responsible’ which
signiﬁcantly correlates with ‘polite’ and ‘patient’. ‘Gentle’ has ‘friendly,’
‘patient,’ and ‘social’ as its members. The clusters ‘ﬁiendly’ and ‘polite’ correlate
with ‘trustworthy’ and ‘social’. In addition, ‘friendly’ correlates with ‘gentle’ and
‘polite’ while ‘polite’ adds ‘responsible,’ and ‘friendly’ to its list of combinations.
The cluster ‘patient’ is made up of the traits ‘responsible,’ ‘gentle,’ ‘polite,’ and
‘trustworthy’ whereas ‘trustworthy’ contains all the core traits of this group (as
stated above) except the trait ‘gentle.’ ‘Social’ correlates with ‘gentle,’ ‘ﬁiendly,’
and ‘social.’

In the second group of clusters, the core features are ‘leader-like,’ ‘proud,’
‘status,’ ‘educated,’ and ‘conﬁdent.’ Each of these traits signiﬁcantly correlates
with one another and the other members of the group except for the trait ‘proud’
which correlates only with ‘leader-like.’ It is worth noting that ‘proud’ is the only

cluster that consists of only one member - ‘leader-like. ’

‘Educated :’ ‘educatcd,’ ‘responsible,’ ‘gentle,’ ‘friendly,’ ‘polite,’ ‘leader- like,’

‘social.’

44

Table 8 - Correlation Coefficients for 12 Semantic Differential Scales for West

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Aﬁican Raters
edu con resp gent ﬁie polit lead pat pro statu trus soci

edu
con .55‘
resp .47" .73"
_gent .45’ .53 " .61‘
ﬁie .51‘ .64‘ .73‘I .73‘
poli .35‘ .53‘ .58" .65‘ .70‘
lead .39‘ .60” .66“ .61” .74“ .71"
pet .31 .56‘ .54‘ .52‘ .49’ .56" .51‘
prou .Ol .13 .14 .01 -.02 .02 .06 -.08
stat .ll .66‘ .60‘ .38‘ .44" .51‘ .45‘ .41‘ .36"I
stat .26 .56" .69“ .63” .69" .58" .53" .39‘ .00 .57"
soci .32‘ .44" .44“ .49" .53" .39" .34‘ .48" -.13 .42" .53”

" = Sigrrif. LE .05 “ Signif. LE .01 (2- tailed)

‘Conﬁdent :’ ‘educated,’ ‘responsible,’ ‘gentle,’ ‘friendly,’ ‘polite), ‘leader-
like,’

‘patient,’ ‘status,’ ‘trustworthy,’ ‘social.’

‘Responsiblez’ ‘educated, ‘conﬁdent,’ ‘gentle, ‘friendly,’ ‘polite,’ ‘leader-Iike,’

‘patient,’ ‘status,’ ‘trustworthy,’ ‘social.’

‘Gentlez’ ‘educated,’ ‘conﬁdent,’ ‘responsible,’ ‘friendly,’ ‘polite,’ ‘leader-like,’

‘patient,’ ‘status,’ ‘trustworthy,’ ‘social.’

‘Friendlyz’ ‘educated,’ ‘conﬁdent,’ ‘responsible,’ ‘gentle,’ ‘polite,’ ‘leader-like,’

‘patient,’ ‘status,’ ‘trustworthy,’ ‘social.’

‘Polite,’ ‘educated,’ ‘conﬁdent,’ ‘responsible,’ ‘gentle,’ ‘friendly,’ ‘leader-like,’

‘patient,’ ‘status,’ ‘trustworthy,’ ‘social.’

 

45

‘Leader-like:’ ‘educated,’ ‘conﬁdent,’ ‘responsible,’ ‘gentle,’

‘friendly,’ ‘polite,’ ‘patient,’ ‘status,’ ‘ trustworthy,’ ‘social.’

‘Patient:’ ‘conﬁdent,’ ‘responsible,’ ‘gentle,’ ‘ﬁiendly,’ ‘polite,’ ‘leader-like,’
‘status,’ ‘trustworthy,’ ‘social.’

‘Proudz’ ‘status.’

‘Statusz’ ‘conﬁdent, ‘responsible,’ ‘gentle,’ ‘friendly,’ ‘polite,’ ‘leader-like,’

‘proud,’ ‘trustworthy,’ ‘social.’

‘Trustworthyz’ ‘conﬁdent,’ ‘responsible,’ ‘gentle,’ ‘ﬁiendly,’ ‘polite,’ ‘leader-
like,’ ‘patient,’ ‘social.’
‘Social:’ ‘educated,’ ‘conﬁdent,’ ‘responsible,’ ‘gentle,’ ‘friendly,’ ‘polite,’

‘leader-like,’ ‘patient,’ ‘status,’ ‘trustworthy.’

The correlation matrix of the West African raters indicate a large group of
clusters that share a uniform membership. The commonality of the group draws
from the correlation of all the traits with each other, and most importantly their
lack of correlation with the trait ‘proud’. The seven traits are; ‘conﬁdent,’
‘responsible,’ ‘gentle,’ ‘friendly,’ ‘polite,’ ‘leader-like,’ and ‘social’.

Three other clusters do not correlate with ‘proud’ as well as other traits,
these are ﬁrst, ‘educated’ which also does not correlate with ‘patient,’ ‘status,’
and ‘trustworthy,’ secondly, ‘patient’ which also does not correlate with
‘educated,’ and ﬁnally, ‘trustworthy,’ which does not correlate with ‘educated’
and ‘status’.

‘Proud’ correlates only with ‘status’. ‘Status’ does not correlate with
‘educated’ and ‘patient’.

The most noteworthy result of Table 8 is row 9 (proud) which deviates

from the correlation pattern. The only variable that correlates with ‘proud’ is

‘status’ which has a weak but signiﬁcant correlation score (36*) .

46

While the tables of rating and ranking presented a general overview of the
rank of the varieties in respect to each evaluative adjective and in relation to one
another, the correlation coefﬁcient matrices on the other hand, show the pattern
of the broader perceptions of the raters towards the voices they evaluated.

The matrices have also indicated that the raters from the three regions
have different scales for evaluating language varieties. The different linguistic

and social realities in the respective regions may account for the differences.

4.1.4 Factor Analysis

A factor analysis is an even more reliable method for such a test, for it
maximally ensures that each member of a factor group truly belongs to that
group. In the correlation coefﬁcient matrices, the clustering of variables was
characterized by overlapping and combinations of some features which had
strong and weak correlations. The Varirnax rotation seeks to rectify this anomaly
by ensuring that there are no overlaps and that the correlation between a variable
and a factor that it belongs to is maximized, and minimized in a factor in which it
does not belong. The outcome of a Varirnax rotation is a reﬁned set of factor
groups.

In this analysis, the scores of each region are computed anew using the
factor analysis technique with principal component extraction and Varirmx
rotation.

One interesting observation in this study is that our African respondents
have different factor groups for each region. Furthermore, these factor groups are
unique in the sense that they are different from the traditional ones identiﬁed in
previous language attitude studies.The difference is especially notable in the

selection of the elements of the factors.

47

A variable belongs to a factor if its loading is .30 or hi gher.The following
tables show the factor groups of each of the regions.

4.1.4. 1 EAST AFRICAN RATERS
Table 9 - Factor loading for EA raters

Trait Factor 1 Factor 2 Factor 3 Factor 4 Factor 5

patient *-.95

responsible *.80 .42 36
educated *.71 .49 .48

polite ‘-.71 -.36 .49 .33
proud , *.87 .32

conﬁdent .45 *.86

social *.95

gentle *.79 .59
leader-er *.97

status .48 .34 *.78

friendly .60 .34 *.62
trustworthy *.98

Table 10 interprets the scores into features and an umbrella label for the
factor group is sought. An attempt is made to ﬁnd the best representative label
for the factor group, and as such, the highest scores in each set (indicated with‘)

are given a greater consideration in the task.

48

Table 10 - Interpretation of the Factor Loading in Table 9

FACTOR l FACTOR 2 FACTOR 3 FACTOR 4
‘Not patient * proud *social *leader-like
*responsible *conﬁdent * gentle *high status
*educated educated educated responsible
*not polite not polite polite proud
conﬁdent high status friendly
high status ﬁiendly

 

 

 

EGOCENTRIC PATRONIZING CHARMING PROUD

FACTOR 5
*trustworthy
* ﬁiendly

"' gentle
polite

responsible

 

DEPENDABLE

The mean scores of the features that have a signiﬁcant loading on each
factor are calculated using the initial mean scores as shown in Tables 1, 2, and 3.
For example, the Factor 1 mean score is calculated by calculating mean of the
mean scores for ‘patient,’ ‘responsible,’ ‘educated,’ ‘polite,’ ‘conﬁdent,’ and
‘status’ as presented in Table 1 (East African raters). The new mean of means

constitute the factor 1 mean. The same computation is done for the other factor

49

groups with Special attention being paid to the analysis of the speech variety
under study. For instance, to determine factor 1 (egocentric) for the weak and
heavy accented varieties, we calculated the mean of the mean score of each of
the six traits that loaded signiﬁcantly into factor 1 as mentioned above, for the
weak’a'c'cented varieties ﬁrst and then for the heavy ones. The same procedure is
followed in calculating the means for the broad regional varieties such as EAE,
SAE, and WAE, as well as for the individual variety. The results are represented
in form of charts that clearly explain the attitudes of the particular raters towards
the varieties of African English.

As Figure 5 shows, three, perhaps four, out of ﬁve features are important
in deﬁning the East African raters' attitude towards the speakers of regional
varieties of English. The odd factor, egocentrism, is considered by the East
African raters to be relatively even among the speakers of the three regional
varieties.

The EAE speakers are perceived to be least patronizing, least arrogant,
most charming , and very dependable: a very positive evaluative. The SAE
speakers are a little more patronizing, arrogant, least charming, and least
dependable of all the speakers. The WAE speakers are judged as most
patronizing, most arrogant, but, ironically, the most dependable members of the
AE speech community.

The WAE speakers are also thought to be rather charming. The EAE
speaker gets the most positive evaluation, SAE gets the worst, while the WAE

speakers gets a mixed evaluation.

50

Figure 3 - East African Raters' Attitudes Towards the Regional Varieties

 

 

 

 

 

 

 

 

5.5

5
4.5 l I m
4 ll SAE
El WAE

3.5

 

3
Egocentric Patronizing Charming Proud Dependable

The attitude of the educated East Africans towards the weak and heavy
accented of African English (Figure 3) indicate a clear distinction in the

perception of the varieties.

Figure 4 - East Aﬁican Raters’ Attitudes Towards Accented Varieties

 

5.5-

 

5.I

 

 

 

I Weak accent
Heavy Accent

 

 

 

Egocentric Patronizing Charming Proud Dependable

WA: egocentric, very patronizing, charming, very proud, and dependable.
HA: less egocentric, far less patronizing, charming, far less proud, and

dependable.

51

The distinction between the two varieties is most evident in considering
the mean score differences of factor group 2 (patronizing) and factor group 4
(arrogance). The speakers of the weak accented varieties are considered to be
much more patronizing and proud than the HA speakers. The mean score
difference for the more positive features, charming and dependable do not have
as big a differences as the two negative traits.

These results are very similar to the East Aﬁ'ican raters rankings of the
varieties as shown in Tables 4 and 5. As shown in Figure 4, the raters give a

relatively higher ranking to the weak accented voices.

Figure 5 - East Aﬁican Raters Attitudes Towards Individual Varieties

 

 

 

        

     
 

 

 

    

 

 

 

 

 

 

 

 

5.5-
5
. \\’/ 1' \ V.
.4 \ 2. \ IEAh
4’54 \v’ 8 V |_ ., " R
§ 8 " V .. 8“ §¢ ”SAW
. . 4i E \ E \
4 a; = § 2 lg = is ,l sg can.
\/ : 4: : \ "z, \/
v — \, — V — \ —. \/ EWAw
V — h — V — \ — v
3.5- \/ Z \’ I \/ Z \l :‘ \/
$2 a i? a; I§él 2 |§g 2., I|§§ IJIWAh
: \ : : . :'
3,, $é : Né : NZ : $6 :" gé

            

      

      

          

Egocentric Patronizing CInrming Proud Dependable

Most egocentric : EAw, SAW, WAh
Most Patronizing : EAw SAw

Least patronizing : EAh, SAh.
Charming : EAw, Eah, , WAh ,WAw.
Most proud: EAw, SAw, WAh.

Far less proud: EAh, SAh

52

Most dependable: EAw WAh, and WAw.

The East African raters (as shown in Figure 5) have an overall negative
attitude towards the SAw variety speakers who score relative high for the
negative traits (egocentric, patronizing, proud) but low for the positive traits
(charming and dependable). Such a trend is also reﬂected in the evaluation of
WAh variety speakers. Inversely, EAh and SAh have relatively lower mean
scores for ‘proud’ and ‘patronizing.’ Based on this apparent parallelism, the
inference that East Aﬁican raters have a more positive attitudes towards
individual heavy accented varieties than the weak accented ones can be made.

A sense of loyalty and identity to the local varieties is clearly evident in
view of the fact that these East African raters perceive their heavy and weak
accented varieties as the most charming, followed by the West African varieties.
The EAh speaker is regarded to be least proud and least patronizing. The most
dependable speakers are the ones who speak EAw followed by the speakers of
the West African varieties.

The factor analysis results for the East African raters correlate with the
earlier rankings as summarized in Tables 4 and 5 in which the East African raters
rank the WA more favorable than the HA varieties. In the Table of making 1, as
the case is in Figure 5, EAw ranks high for all the features, WAh second, and
SAh ranks last.

4. l. 4. 2 SOUTH AFRICAN RATERS
Table 11 - Factor Loadings For South Aﬁican Raters

Traits Factor 1 Factor 2 Factor 3 Factor 4

conﬁdent *.95

53

responsible *.95

gentle .33 *-.93

ﬁiendly *-.92

polite *.95

leader-like "'.88
patient *.96

proud .31 *.88
status 56 *.79

trust .39 *-.83 32

social ‘.75 -.48 -.36
educated *.65 .48 .56

Table 12 - Interpretation of the Factor Loadings in Table 11

FACTOR l FACTOR 2 FACTOR 3 FACTOR 4
*conﬁdent *educated *responsible *leader-like
*polite *not ﬁiendly "' not gentle ’proud
*patient *high status educated educated
*social *not trustworthy proud not social
status trustworthy

trustworthy not social

gentle

DEPENDABLE EGOCENTRIC RESERVED PROUD

54

The results of Table 7 indicate that there are two broad categories of
clusters: the cluster that represent the positive values contains components such
as ‘gentle,’ ‘fliendly,’ ‘polite,’ and ‘patient,’ - the group that loads as Factor 1
(dependable). The other broad cluster represents the negative values and does
not contain any of the four mentioned virtues. In the factor analysis, the other

three groups represent that broad combination of perceived vices.

Figure 6 - South African Raters Attitudes Towards the Regional Varieties

 

 

 

 

 

, .

~v
'—
—-
p—n'
-—
.—
—.
—.
—.
..__'
h..—
_
.—
u—
—-—I
‘—

Dependable Egocentric Reserved Proud

Dependable ( Factor I): SAE, EAE, WAE.
Egocentric (Factor 2) : EAE, SAE, WAE.
Reserved (Factor 3): EAE, WAE, SAE.
Proud (Factor 4) : WAE, EAE. Far less arrogant: SAE
The South African raters regard the speakers of their regional variety
(SAE) to be the most dependable, least proud and least reserved of the speakers

of Aﬁican English varieties.

55

EAE speakers are perceived to be the most egocentric and most reserved. The
WAE speakers are seen as the least dependable as well as most arrogant of the

lot. However, WAE speakers are regarded to be the least egocentric.

Figure 7 - South African Raters’ Attitudes Towards the Accented Varieties

 

 

, a

 

I Weak Accent
El Hea Accent

 

 

 

Dependable (factor 1) : HA, WA.
Egocentric (factor 2) : HA, WA
Reserved (factor 3) : HA, WA.
Proud (factor 4) : HA, WA.

The SA raters do not make a strong perceptual distinction between the
heavy and weak accented varieties. Although the HA gets the highest score in
all categories, there is not a very big range in the assignment of degrees of
reservedness, pride and dependability, but in egocentrism, HA speakers are
perceived to be higher. The mean score range for both varieties are not very
sharp except for the factor ‘ egocentrism’. The results of ranking as shown in
Tables 4 and 5 somewhat corresponds to this perception of minimal differences

of the two varieties, except for egocentrism.

56

Figure 8 - South African Raters’ Attitudes Towards Individual

Varieties

 

 

5.

!.

illlllllllllllllllllllg ‘

AN ,

 

Z/Jﬁﬁ

f/.-' ' Iy,’ ./ I x/ .I .I //,- I z' .I.- ,-
.' .r r . .- . - .- .- ,- I
’c- tic-’1 //.-’1 l ,7, / /.r.- /,- {.-

 

’/ / / /.-' I / 1

_ - J
fﬁﬂ¢féf¢¢ffé€cc

 

 

\\\\\\\\\\\\\\\\\\\\\\\\\‘\\
:\\\\\\\\\\\\\\\\\\\\\\\\\\‘;

 

Illlllﬂlllllllllllllh
-s-__
lllllllllllllllllllllll

llllllllllllllllllllil

WY/iZ/Zfr';

Q-
Q
\:
\j
. .
.
r,
3;
s
.\.
n“.
.5.

'll/ia

\

Dependable Egocentric Reserved Proud

The EAh is the dominating variety in all the characterizations by the SA
raters, which also reﬂects their perceptions towards the HA and WA varieties as
discussed above. The EAh speaker is thought of as the embodiment of both
negative and positive values: most egocentric, most reserved, most proud, and
surprisingly the most dependable speaker, an accolade which we hypothesized
would be accorded to the raters' local varieties, as the case was with the East
Aﬁ’ican raters. The only other notable result in the chart is that SAw is judged to
be the least proud and least reserved. SAh speaker is the least egocentric
whereas the WAh and EAw speakers are the least dependable.

The overwhelming dominance of the EAh in all the categories mirrors the
results in Tables 2 and 4 in which the SA raters rank EAh speakers higher than all
the other speakers. SAw speakers are not remarkably different from the perceived

EAh ‘model’ in both analyses.

4. l. 4. 3 West African Raters
Table 13 - Factor Loadings for West African Raters

Traits
Friendly
proud
patient
gentle
responsible
leader-like
trustworthy
status
social
conﬁdent
polite

educated

57

Factor] Factor 2 Factor 3

*-.89
*.87
‘286

.49
.61

.32

*.98
*-.82
*.64
"258

-.41
32
-.34

Table 14 - Interpretation of the Factor Loadings in Table 13

Factor 1

* not friendly
‘proud

patient
leader-like
responsible
not trustworthy
high status

EGOCENTRIC

Factor 2

* gentle

‘not responsible

*leader-like

*trustworthy
not conﬁdent

not educated

polite

CHARMING

Factor 3

* status
*social
*conﬁdent
trustworthy

not leader-like

HUMBLE

Factor 4
.43
-.41
.51 -.55
*.93
*.89
*.77
.92
*.93
Factor 4
‘educated

*polite
not trustworthy
friendly

OUTGOING

58

The main correspondence with the correlation coefﬁcient analysis is that
the results of Table 8 make a distinction between the large group of clusters that
do not correlate with ‘proud’ and the very few that do, just as the fact that it is
only in Factor 1 that ‘proud’ loads signiﬁcantly and strongly (.87). The other
‘non-proud bearing’ factor groups represent positive values as perceived by the

raters.

Figure 9 - West African Raters’ attitudes Towards Regional Varieties

 

 

 

 

 

 

 

 

Egocentric Charming Humble Outgoing

Egocentric: SAE, EAE, WAE (least egocentric).
Charming: WAE, SAE, EAE.
Humble: EAE, SAE, WAE.
Out-going: WAE (very), EAE, SAE.

The West African raters regard the WAE speakers as the most outgoing,
most charming, least egocentric and least modest. Speakers of SAE are seen as
charming, modest, somewhat more egocentric and least outgoing. The FAB

speakers are the most humble.

59

Figure 10 - West African Raters’ Attitudes Towards Accented Varieties

 

 

 

 

5
4.8
4.6 - III
4 4 I Weak Accent
4.2 555 5;; Han accent

Egocentric Charming Humble Outgoing

Egocentric: WA (most), HA (least).
Charming : WA, HA.

Humble : WA (most), HA (least)

Outgoing : WA (most), HA (least).

These factor analysis results for West African raters reﬂect the rating and
ranking results in Tables 4 and 5 in which these raters have a seemingly more
favorable rating of WA varieties than the HA ones. However, a closer scrutiny
reveals a rather mixed evaluation of the two varieties. Considering that
‘egocentn'c’ is the only factor that characterizes a negative value, speakers of the
WA varieties are considered to be far more egocentric than the HA speakers,
which is a favorable evaluation of the HA speakers. Also we may argue that the
West African raters consider ‘humble’ to be a negative feature. Consider the
evaluation of the regional varieties in Figure 9 above. Whereas it has been a
consistent trend for the raters to evaluate their regional varieties high for the
positive features, WAE is least ‘humble’ and least ‘egocentric’ but most
‘charming’ and most ‘outgoing.’ The sharp difference in the evaluation of WA

as most ‘humble’ and HA as least ‘humble’ is consistent with that analysis.

60

Figure 11 - West Aﬁican Raters’ Attitudes Towards Individual Varieties

 

 

 

 

 

    

 

 

 

 

Egocentric Charming Humble Outgoing

Egocentric: SAw, EAw.

Charming : WAw, SAw, WAh.
Humble: SAW, EAw & EAh, WAw.
Outgoing: WAw, EAw, WAh.

The speakers of the WA varieties dominate in the positive and negative
factors which reﬂects the evaluation of the WA and HA varieties as discussed
above. The very favorable evaluation of the WAw speakers in the factors
‘charming’ and ‘outgoing’ indicate a very strong sense of loyalty to that local
variety. SAw speakers are considered to be the most ‘egocentric’ and most
‘humble,’ a feature considered to be negative as discussed above. Note that the
WAw and WAh are evaluated less favorably for the feature ‘humble’ as the case

for ‘egocentric.’

4.2.0 Identiﬁcation of Regional Varieties

61

Figure 12 - Raters’ identiﬁcation of the Anonymous Stimulus Voices

 

 

 

 

 

 

 

 

 

EA (5) SA (5) WA (4) Total
__E_Aw 3 3 4 10 (70%)
SAW 2 5 3 10 (70%)
WAW 3 4 2 9 (66%)
_§_4_h 2 5 2 9 (66%)
WAh 4 3 2 9 (66%)
EAh 4 2 l 7 (50%)
18 (60%) 22 (73%) I4 (58%)

 

 

 

 

 

15 out of the 24 raters who evaluated the six stimulus voices on the
semantic differential scale also took the voice and region matching test (section
B of the questionnnaire. See Appendix). The above calculations were based on
the evaluations of 4 West Africans and 5 raters from East and Southern Aﬁica
respectively. The results show a generally balanced recognition of all the
regional varieties of English under study.

1. The most recognized varieties are the EAw and SAw each with a 66%
recognition by raters from all the regions. EAh is the least recognizable while
the West Aﬁ'ican varieties, which are of special concern to this study rank
favorably in this regard. A

2. The raters with a highest percentage in the correct identiﬁcation of the
stimulus voices are the South Africans (73%). The East Africans were correct

60% of the times while the West Aﬁ'icans had a 58% identiﬁcation rate.

 

62

One of the objectives of this study was to determine whether the raters
would positively identify the regional varieties, especially the WAE in view of
the region’s linguistic heterogeneity. Based on the data we obtained, it is evident
that raters ﬁom all regions recognize the three regional varieties of African
English.

Although this is a preliminary study, limited by the lack of diversity of its’
small sample and the consequent non binding nature of its results, the
pioneering nature of the study, and the absence of results from such a study,

should, however, warrant the drawing of preliminary conclusions.

63

5. 1 CONCLUSION

The methodology we have utilized in this study has proven to be
empirically reliable and more efﬁcient in the elicitation and analysis of language
attitudes in African than the observation techniques that have preoccupied
researchers in the region for a long time. The methods have revealed new and
interesting information regarding language attitude study in Africa and the
perception of English among educated Africans.

The attitudes of the educated Africans towards the varieties of African
English are diverse when considered from a region to region basis as indicated in
the Tables of Rating and Ranking as well as the factor analysis results. However,
considered from a broad viewpoint there are glaring similarities in the evaluative
criteria used by the all raters. In the factor analysis, ‘egocentric’ is identiﬁed by
all the raters, ‘arrogant,’ ‘charming,’ ‘dependable,’ are identiﬁed by raters from
two regions, while ‘patronizing,’ ‘modest,’ ‘humble,’ and ‘outgoing,’ are
evaluative features used by raters from one region.

These attitudinal features can be categorized as negative or positive on
the basis of the combination of traits of a factor loading and the trends evident in
the evaluation of the speech varieties.

One trait that strikingly stands out as the embodiment of negative values,
and one which the raters utilize extensively in their judgment of the stimulus
voices is the trait ‘proud.’ The correlation coefﬁcient analysis, the factor
analysis, and even the rating and ranking method support this ﬁnding.

The positive features are ‘dependable,’ ‘charming,’ and ‘outgoing.’

‘Egocentric,’ ‘proud,’ ‘patronizing,’ and ‘humbie’ are considered to be the

64

negative values. The commonality of these features is not surprising. We believe
that the judgment of the raters is, as Mbiti (1969) argues, shaped by a common
experience of the African culture in which the individual will is largely considered
to be subservient to the social will, hence egocentrism and pride are the deviant
characteristics of a tacitly held code of social behavior. These values are aptly
reﬂected in language attitude analysis.

A region to region analysis gives us the following results:

1. East Aﬁ'ican raters.

These raters consider speakers of EAE as the most charming, least
arrogant, and least patronizing. They not only consider speakers of WA varieties
as most egocentric, most patronizing, most arrogant, but also charming and
dependable almost as much. as the HA speakers are. The evaluation of the
individual speech varieties reveals a strong sense of loyalty to the East African
varieties. EAh speaker are regarded as the least arrogant and least patronizing
.EAw speakers are regarded as the most charming and (most dependable speakers
of African English.

2. South African Raters.

The SAE speakers are rated most favorably. The HA speakers are
considered to be most egocentric, reserved, and arrogant. The speakers of the
EAh are the standard bearer for these raters in all the factors. This is an issue
i that requires further investigating to determine the overwhelming high evaluation
of the EAh speakers as evident in the table of ranking and the factor analysis
results.

3. West African Raters.

A strong sense of loyalty is also evident in evaluation of WAE varieties.

Speakers of the variety are regarded as the most charming, most outgoing, as

well as the least egocentric and least humble of the African English speakers.

65

Surprisingly, humble is considered to be a negative trait by these raters. A more
thorough research is required including sociological information to verify this
seemingly odd fact. The HA speakers are considered to reﬂect the positive traits
while the WA reﬂect the negative ones.

In summary, we can conclude that the educated Africans can positively
identify the regional, accented as well as individual localized varieties of Aﬁ-ican
English. They judge speakers of a language based on the values drawn ﬁ'om their
common cultural heritage and experience. It is therefore necessary for future
language attitude researchers in the region to strongly consider eliciting the
appropriate evaluative adjectives to be used in the semantic differential scales
from African respondents, as the evidence of this study shows.

Raters show a strong sense of loyalty to their regional varieties. Speakers
of the HA varieties are judged favorably by East and West African raters, which
indicates solidarity with the distinct local varieties. The general dislike for WA
by these educated raters may reﬂect the association of RP variety with colonial

institutions and experience. Kachru (1982 :45) writes:

...Educated Ghanaian English is acceptable, but as Sey (1973:1) warns us, it does
not entail competence in speaking RP since in Ghana, “ the type [of speakers]
that strives too obbviously to approximate RP is frowned upon as distasteful and
pedantic.” In Nigeria the situation is not different from Ghana or India Many
Nigerians will consider as affected or even snobbish any Nigerian whoo speaks

like a native speaker of English.

Our ﬁndings concur with Kachru’s observation and further shows that the
attitudes towards RP (or WA) is shared among speakers of the regions we
investigated.

The factor analysis results of the individual regional varieties does not

provide very explicit correlation patterns in terms of how the raters evaluate their

66

preference of the varieties. These facts may be due to the main shortcoming of
our very exploratory research, which as stated numerous times elsewhere in this
work, is our limited sample.

Further research is required to verify these preliminary results derived

from such a limited sample.

67

6. l RECOMMENDATIONS

A much larger and more diverse sample should be used. This study
focused on educated male Africans who are based in the United States of
America, and whose ﬁrst language belongs to the larger Niger-Congo family
group. It should be interesting to analyze the language attitudes based on
gender, age, status, education of Africans in Aﬁ'ica as well as those living out of
the continent.

Moreover, the empirical analysis used in this study can be used in the
language attitude analysis of African languages, dialects and lingua francae for
pedagogical as well as language planning purposes.

APPENDICES

APPENDIX A

Appendix A: Questionnaire Used in the Elicitation of Attitudes.

Michigan state university,
Dept. of Linguistics, Germanic, Slavic, Asiatic and African Languages.

Sociolinguistic survey.

Please listen to the eight recorded voices and rate each voice on the scale
provided. The adjectives provided at each end of the scale are meant to guide you
in making your independent evaluation of the speaker. Indicate your rating of the
speaker by checking ( \l ) on the seven point scale your impression of the speaker
‘ on each characteristic provided.

For example to indicate your impression of voice # 1's ﬁ'iendliness, you check
closest to the feature ﬁ'iendly if you feel the speaker sounds to be a very ﬁiendly

person or closest to unﬁiendly if you feel so or any other point depending on your

evaluation.

ﬁiendly ------- unﬁiendly = absolutely ﬁ'iendly
ﬁiendly ------- wrﬁiendly = absolutely unfriendly
ﬁiendly ------- unﬁiendly = very friendly
ﬁiendly ------- unﬁiendly = just friendly etc.

After your rating of the voices, please try to identify each speaker's place of
origin. ( section B)
Finally, ﬁll your biographical sketch. (section C)

68

69

Voice #
well educated ------- not well educated
conﬁdent ------- not conﬁdent
responsible ------- not responsible
gentle ------- not gentle
friendly ------- not friendly
polite ------- not polite
leader-like ------- not leader-like
patient ------- not patient
proud ------- not proud
highstatus - - - - - - - lowstatus.
trustworthy ------- not trustworthy
social ------- not social
Voice #
Well educated ------- not well educated
conﬁdent ------- not conﬁdent
responsible ------- not responsible
gentle ------- not gentle
friendly ------- not friendly
polite ------ - not friendly
leader-like ------- not leader-like
patient ------- not patient
arrogant ------ - not arrogant
highstatus ---.---- lowstatus
trustworthy ------- not trustworthy

social ------ - not social

70

Section B
Match each voice with its place of region. Try to be speciﬁc e.g. West Africa,
Middle East or Nigeria, Palestine etc.

Voice # 1
Voice # 2
Voice # 3
Voice # 4
Voice # 5
Voice # 6
Voice # 7
Voice # 8

 

 

 

 

 

 

 

 

Section C

Please provide the following information about yourself (biographical inf.).
1. Gender : ( a) Male (b) Female
2. Place of origin: (a) East Aﬁica (b) West Africa (c) Southern Africa
3. Age: (a) under 20 (b) 203 (c) 305 (d) 40s (e) over 50

4. a) Number of years you've lived out of Africa

 

b) Places you've lived while out of Africa

 

5. Which languages do you commonly use?

 

Thank you for participating in this research.

APPENDIX B

Appendix B. Questionnaire Used to Elicit Evaluative Adjectives.

Michigan State University,
. Dept. of Linguistics and Languages.

A Langgge Survey.

Instructions.

Please listen carefully to the eight recorded voices. Write in the space provided
for each voice the impressions that each voice evokes in you. Feel free to write
especially the very ﬁrst impressions of the speaker you get after listening to the
voice for the very ﬁrst time. Listen to the voice again and write all the
impressions that you can get. Anything and everything you note will be greatly
appreciated. Thank you.

71

LIST OF REFERENCES

REFERENCES

Adegbija, BE. 1994. Language Attitudes in Sub-Saharan Africa. A Sociolinguistic
Overview. Clevedon: Multilingual Matters Ltd.

Agheyisi, R. and Fishman, J .A. 1970. Language Attitude Studies: A Brief Survey
of Methodological Approaches. Anthropological Linguistics
12, 131-57.

Angogo, R. and Hancock, 1. 1980. English in Africa: Emerging Standards or Diverging
Regionalisms? English World- Wide: A Journal of Varieties
of English 1(1): 67-96.

Bailey, L, and Gorlach, M. (eds) 1982. English as a World LanguagirAnn Arbor:
University of Michigan Press.

Baker, S. 1966. The Australian Language. Sydney: Currawong.

Bokamba, E. 1981. The Africanization of English. In Kachru, B.B.(ed.).The Other
Tongue. English Across Cultures. Urban: Illinois University Press, 77-98.

Buck, J. 1968. The effects of negro and white dialect variations upon attitudes of college
students. Speech monographs 35: 181- 86.

Chesire, J. (ed) 1991. English Around the World: Sociolinguistic Perspectives.
Cambridge. Cambridge University Press.

Cheyne, W. 1970. Stereotyped reaction to speakers with Scottish and English regional
accents. British Journal of Social and
Clinical Psychology 9: 77-9.

Edwards, J .R. 1977. Students’ reaction to Irish regional accents.Language and speech

72

73
20: 280-6.

Fasold, R 1984. The Sociolinguistics of Society. Oxford: Blackwell.
Fairclough, N. 1989. Language and Power. London: Longman.

Fraser, B. 1973. Some ‘unexpected’ reactions to various American English dialects. In
Shuy, R. and F asold, R (eds), Language attitudes: CurrentTrends and Prospects.
Washington: Georgetown University Press, 28-35.

Giles, H. 1971. Patterns of evaluation in reactions to RP, South Welsh and Somerset
accented speech. British Journal of Social and Clinical Psychology 10: 280-1.

Gimson, A.C. 1980. Introduction to the Pronunciation of English.London:
Longman.

Girlach, M. 1987. English as a world language- the State of the art. English World- Wide 9: 1-32.

1991. Englishes. Studies in Varieties of English 1984-88. Amsterdam: John
Benjamins.

Hancock, 1. F. and Angogo, R. 1982. English in East Africa. In Bailey, R. and Garlach,
M. (eds.), English as a World Language. Ann Arbor: University of Michigan
Press, 306-23.

Harms, LS. 1961. Listener’s judgements of status cues in speech. Quarterly Journal of
Speech, 47,164-68.

Hatch, E. and Farhady, H. 1982. Research Design and Statistics for Applied Linguistics.
Rowley: Newbury House Publishers.

Herbert, R.K.(ed) 1992. Language and Society in Africa. The Theory and Practice of
Sociolinguistics. Cape Town. Witwatersrand University Press.

Kachru, B.B.(ed) 1982.'Ihe Other Tongue: English Across Cultures. Urbana: University
of Illinois Press.

74

Labov, Williams.1966. The Social Stratiﬁcation of English in New York City.
Washington D.C: Center for Applied Linguistics.

1972. Sociolinguistic Patterns. Philadeiphia: University of Pennyslavania

 

Press.

1984. Field methods of the project on linguistic change and variation. In
John Bauch and Joel sherzer (eds), Language in Use: Readings in
Sociolinguistics. New Jersey: Prentice-Hall.

 

Lambert, WE. et a1 Evaluational reaction of Jewish and Arab adolescents to dialect and
language variations. Journal of Personality and Social Psychology, 1965, 2, 84 -
90.

1960. Evaluational reaction to spoken languageJournal of Abnormal Social
Psychology, 60:44-51.

Lanham, L.W. 1982. English in South Africa. In Bailey, R. and Gorlach, M.
(eds). English as a World Language. Ann Arbor: University of
Michigan Press, 353-83.

Mbiti, J. S. 1969. Aﬁ'ican Religious and Philosophies. New York. Doubleday.

Milroy, L. and McClenaghan, P. 1977. Stereotype reactions to four educated accents in
Ulster. Belfast Working Papers in Language and Lingustics, 2,(4). V /) 3, .2

Moag, R. 1982. The life cycle of non-native Englishes: A case study. In Kachru, Braj B.
(ed.), The Other Tongue: English Across Cultures. 270-88.

Mutonya, M. 1991. Maudhui Katika Ushairi wa Hassan Mbega. Unpublished masters
thesis. University of Nairobi. Kenya.

Preston, D. R(ed) 1993. American Dialect Research. Philadelphia: John Benjamins
publishing company.

75

1989. Sociolinguistics and Second Language Acquisition. New York:
Blackwell.

Pride, J. (ed.). 1982. New Englishes. Rowley, Mass: Newbury House.

Ryan, EB. and Giles, H.(eds).1982. Attitudes Towards Language Variation. Social and
applied contexts.London: Edward Arnold.

Saah,K.1986.Language use and language attitudes in GhanaAnthropological Linguistics
28 (3): 367-77.

Samarin, William J. 1996. Language Attitudes in Sub-Saharan Africa: A Sociolinguistic
Overview by Efurosibina Adegbija A Book Review in Anthropological
Linguistics Vol. 38 no. 2 July 1996.

Schmied, J. 1991. English in Africa: An introduction. London: Longman.

Scotton, C. Myers. 1996. Language Attitudes in sub-Saharan Africa: A Sociolinguistic
Overview by Efurosibina Adegbija. A Book Review in Language and Society 25:
2 (1996) pp 311-313.

Shuy, R. and Fasold, R.W.(eds). 1973. Language Attitudes: Current Trends and
Prospects.Washington, D.C: Georgetown University Press.

Sridhar, Kama] K. 1985.Sociolinguistic theory and non-native varieties of English,
Lingua 68, 39-58.

Strevens, Peter. 1971. World English and the world's Englishes - or whose language is it
anyway? Journal of the Royal Society of Art 130: 418-31.

Sure, K. 1991 . Language functions and Language attitudes in Kenya.English World- Wide
12 (2), 245-60.

Tucker, GR. and Lambert, W.E. 1972. White and Negro Listeners' Reactions to Various
American Dialects. In Fishman, J .A. (ed.) Advances in the Sociology of
Language 11. The Hague: Mouton & Co., Printers.

76

Williams, J. 1987. Non-native institutionalized varieties of English. English World- Wide
8, 161-99.

Williams, R, et al.1969. Exploration of the Linguistic Attitudes of Teachers.
RowleyzNewbury.

Williamson, Kay. 1989. Niger-Congo overview. In Samuel, J .B.(ed) 1989.
The Niger-Congo Languages. Lanham. University Press of America.

Woods, David R 1995. Attitudes towards French, national languages, and mother
tongues across age and sex in Congo. In Akinbiyi Akinlabi (ed.)Theoretical
Approaches to African Linguistics.Trenton: Africa World Press. 403-15.

Vogt, Paul W. 1993. Dictionary of Statistics and Methodology. A Non-Technical Guide
for Social Sciences. London. Sage Publications.

Zahn, CJ. and Hopper, R. 1985. Measuring language attitudes: The speech evaluation
instrument. Journal of Language and Social Psychology, Vol.4, 2.