Upload
others
View
11
Download
0
Embed Size (px)
Citation preview
1
CHAPTER ONE: INTRODUCTION
1.0 INTRODUCTION & GENERAL OVERVIEW
This research explores phonetic and phonological variation between a native speaker
English accent and a non-native speaker English accent. With special emphasis on how
language convergence and divergence impinge on a speaker‟s intelligibility to a listener,
the study focuses on geographical/regional variation, a sub-field of sociolinguistics; while
its main thrust is variation as a determinant of accent intelligibility.
The study measures the level of understanding between native speakers and non-native
speakers of English. More specifically, it measures the level of understanding between
speakers of the Standard British English accent and the speakers of the Nigerian English
Accent. Both accents – the Standard British English accent (also known as Received
Pronunciation or simply RP) and the Nigerian English Accent – have certain distinctive
speech patterns which demonstrate marked accent variation. Correlation of RP distinctive
speech features and intelligibility to Nigerian listeners is the main aim of the research.
As no accent is homogeneous or without variation, proper contextualization of the accents
which constitute the focus of the study is necessary. Several types of the RP accent have
been identified but the RP accent of interest in this study is the RP accent variety which
constitutes a clustering of features identified by Wells as well as Cruttenden as
mainstream RP (279) and General RP (78) respectively. Other varieties of RP identified
by Wells are U-RP or upper-crust RP, adoptive RP and Near RP. These accents – including
2
mainstream RP – share one important social characteristic which is that they are usually
spoken by members of the upper class socio-economic group. However, they differ in
terms of phonetic characteristics, with each accent representing an area within an accent
continuum. In the same way, the Nigerian accent of English is also not homogeneous as
several varieties of it have been identified. Attempts at describing the Nigerian accent of
English include those made by Adetugbo, Brosnahan, Bamgbose, Jibril, Udofot and Banjo.
Of these attempts, Banjo‟s classification is the most often cited and perhaps represents the
most realistic description of the Nigerian accent of English. This is because it is close to
present-day realities of language use in the country. Banjo‟s description identifies Variety
One, Variety Two, Variety Three and Variety Four respectively. However, Variety Three is
the variety on which this study focuses because this is the accent variety used by the group
of Nigerians in which we are interested. These are educated Nigerians at the undergraduate
level who speak a brand of English which Banjo refers to as “the variety of Spoken
Nigerian English which is internationally intelligible and acceptable….the most appropriate
endonormative model” (26).
Generally, intelligibility may be studied across the linguistic levels of grammar, semantics,
lexis and phonetics/phonology. However, speech intelligibility is studied at the linguistic
level of phonetics/phonology where intelligibility is seen as a construct of speech and “the
hearers‟ response is perceived as appropriate only if the linguistic forms which constitute
the speakers‟ utterances are selected appropriately” (Catford 2). The utterances of the RP
speakers represent the linguistic forms, while the hearers‟ perceptions of RP speech forms
constitute the listener‟s response in this study. In line with the intelligibility testing
3
framework, both the RP speakers‟ utterances and the Nigerian listeners‟ written responses
are subjected to critical phonological analysis in order to determine instances of
intelligibility failure, identify the RP speech features responsible for instances of
intelligibility failure and most importantly, offer phonological explanations as to why
intelligibility failure occurs. The phonological analysis also enables us calculate the
intelligibility index of the RP speakers‟ accent to Nigerian listeners with the percentage
value being an indication of the number of items in the RP speakers‟ utterances which were
correctly received by the Nigerian listeners. This percentage value therefore represents the
level of intelligibility of RP to Nigerians.
Phonological considerations generally involve segmental and suprasegmental sounds.
Consonants and vowels constitute the segmentals while the suprasegmental sounds include
the features of stress, rhythm and intonation. This research explores the realizational
qualities of the segmental sounds of RP through the examination of RP monophthongs,
diphthongs, triphthongs and consonants. Beyond the segments, the impact of the
superimposed features of stress, rhythm and intonation are also investigated. These aspects
of RP accent phonology are examined in the study with a view to determining their effect
on intelligibility. They are similar to dependent variables, manipulated to discover their
influence, if any, on the independent variable which in this case is intelligibility. Some
other features of RP speech such as excessive reduction of unstressed syllables and certain
phonological processes which are known to affect intelligibility are also tested for their
effect on the Nigerian listeners‟ level of understanding. Speaker‟s gender and listener
4
familiarity with accent are two extra-linguistic factors known to affect intelligibility. These
variables are also explored in order to determine their influence on speech intelligibility.
The instrument designed for the study comprises four tests representing aspects of RP
segmental and suprasegmental phonology. The administration of these tests produces both
quantitative and qualitative data which are subsequently analyzed using suitable statistical
procedures. Qualitative analyses are done through correlations of the RP speakers‟
recordings and the Nigerian listeners‟ responses. On the basis of the analysis,
interpretations are made and general conclusions drawn on the intelligibility level of the RP
accent to Nigerians. Specific statements are also made on peculiar characteristics of RP
which constitute intelligibility problems for Nigerians.
1.2 BACKGROUND TO THE STUDY
1.2.1 Language Variation
Nearly everyone who uses and listens to other people use language has experienced
variability in language across several dimensions. However, the consciousness of this
experience seems to be most noticeable among phoneticians and linguists generally who
have identified that language variation occurs across the dimensions of time, style, social
space and geographical space, while variations in language are studied across the structural
levels of grammar, lexis/vocabulary and phonology. The dimensions and structural levels
of language variation are considered in the section below.
5
Linguists have generally acknowledged that the dimension of time is an ideal platform for
the observation of language variation. Studies based on this dimension involve the adoption
of time as an independent variable whose effects on linguistic and extra-linguistic variables
are then investigated. Two main approaches of language variation in time are the
diachronic (real time) and the synchronic (apparent time) dimensions (Fasold 56). While
diachronic studies make observations of language use among similar populations over
intervals of many years, synchronic studies observe language use by different age groups at
the same time. Examples of language varieties on the basis of the time dimension include
Old English, Middle English and Shakespearean English.
Another level of the observation of language variation is the dimension of style.
Considerable discrimination is usually exhibited in language use depending on the degree
of familiarity between the participants in a conversation. Observations of stylistic variation
are usually dependent on the manipulation of style as an independent variable in which
linguistic variants correlate with a range of speech styles spanning formal to casual style.
Established stylistic varieties that are routinely employed are word list style (the most
formal), reading passage style, interview style and casual style which is the most informal
(SP 80).However, the formal and informal stylistic varieties are the observable forms in
natural human interactions. An illustration is the use of man rather than guy where the
former is indicative of formal speech and the latter of informal speech. High fluency and
proper articulation also characterize formal speech situations while utterances are more
rapid and filled with more phonological assimilations and coalescences in informal
situations such as those involving interacting with family and close friends. Another
6
linguistic cue of highly formal speech is the pronunciation of the indefinite article „a‟ as
with being more common in casual speech.
Two types of space may be identified in linguistics: social space and geographical space.
The two form an important element of spatiality and cannot exist independently of one
another (Britain 620). This is because geographic space is the environment occupied by the
human settlements that appropriate the space. The natural consequences of this
appropriation are observable on the occupied geographical space in the form of language
and other characteristics of human interaction. Inequalities in the distribution of wealth,
priviledge and opportunity is a feature of human communities which results in the creation
of social classes. Although social class structure is not always clearly delineated in all
societies, the most common social class divisions are upper class, middle class and working
class, with the divisions usually based on the indices of occupation, income, education and
education. Apart from social class structure, other social attributes of human communities
include sex, age and ethnicity. Thus, in any linguistic analysis which employs any of these
socially significant factors as the independent variable, varieties or sociolects may be
identified. Some sociolects relate to sex/gender while others relate to the age of the
speaker. An example of a class-related speech feature is the use of rather than for
-ing words among native speakers of English. For instance also, a creaky or tremulous
voice quality and lower voice pitch are usually indicative of an aged man or woman.
Differences in vowel patterns among different generations of speakers of the same accent
can also be observed.
7
The inter-relatedness of the dimensions of social and geographical space can be observed
through the differentiation of language occasioned by social and physical factors. Other
social processes including migration, urbanization, labour indentation, colonization
etcetera, also influence the innovation and spread of linguistic features across Euclidean or
geometric space (Britain 615).
One contingent effect of the aforementioned social factors and physical processes on
human communities is variation in language. Thus, politically-motivated boundary
constructions pervade all modern human communities and explaining the linguistic
significance behind such reformations of the geographical landscape is the goal of geo-
linguists. The identification and description of spatial or geographical varieties constitute
the main findings of this sub-discipline of linguistics. Examples of spatial or geographical
varieties which may be equated to national and regional associations include British
English, American English, Caribbean English, and Nigerian English. Each of these
national varieties has features which differentiate them, one from the other. For example,
the use of the voiced inter-vocalic /t/ (in better, pity, bottom) usually identifies a speaker as
an American while syllable-timed speech rhythm may be associated with African speech.
In order to make adequate statements about language, the linguist often makes reference to
a set of approaches referred to as levels of analysis. Based on different types of features of
language, the approaches are phonology and grammar, following the main activities of
language which are speaking and writing. Grammar is concerned with the formal patterning
and arrangement of the written form of language while phonology examines the patterning
8
and arrangement of the phonetic features of language. However, the study of language
variation includes other fields as it also considers the levels of lexis/vocabulary and
semantics. These levels correspond to the lexical aspect of language and meaning
respectively.
Variation in language across the dimensions of time, style, social space and geographical
space as well as across the linguistic levels of grammar, lexis/vocabulary, semantics and
phonology gives rise to dialects. Dialect comprises the patterns of grammar, vocabulary,
semantics and sounds of a language while the sounds alone constitute the accent. Based on
a geographical classification, dialects of English include Scottish dialect, Irish dialect,
Northern England dialect, Indian dialect, Malaysian dialect etcetera. These in turn have
sub-dialects that are representative of particular speech communities which have common
speech features. While on one hand, dialect comprises the patterns of grammar, vocabulary
and semantics of a language, on the other hand, accent is the totality of the sound system of
a language or language variety, comprising the phonemic contrasts and the tone group. In
other words, while dialect refers to varieties distinguished from each other by differences
of grammar and vocabulary etcetera, accent refers to variations in pronunciation alone.
Examples of accents of English include Indian accent, West African accent, French accent
and Black American accent. In a similar vein, these accents also have sub-accents that are
representative of particular speech communities with common norms of pronunciation.
Therefore, the phonemic contrasts and tone group features of an accent which constitute the
pronunciation serve as the platform on which one may observe the phenomenon of accent
9
variation. This may lead one to ask the question: why does accent variation occur? In an
attempt at providing an explanation for accent variation within English, Wells argued that:
[T]he phenomenon of language change is the fundamental reason
responsible for accent variation. English pronunciation changes as
time passes; and the developments which have arisen and become
established in different places and among different social groups have
not been identical (94).
Wells‟ argument may be said to provide an adequate explanation of the nature of accent
variation within English as it not only identifies the relationship between the major
dimensions of time, place and accent variation but it also highlights how the interplay of all
these factors leads to differences in accents. The awareness of the fact that English
pronunciation has changed over the years as earlier pronunciation patterns have been
modified and new ones adopted, provides evidence that the factors of time, social and
geographical factors do lead to changes in pronunciation. The causal effect of language
change and accent variation is further established by two postulations which explain how
innovations in pronunciation not only arise but also spread. They are the theory of least
effort and the naturalness of segments. Lindblom (367) establishes this link through the
theory of least effort or the H&H theory. The theory provides an explanation for “new
pronunciation patterns which occur as a result of the tendency to pronounce words and
sentences in a way which involves the minimum of articulatory effort consistent with the
need to maintain adequate communication” (415). This results in what is known as hypo or
10
under articulation which is the deletion of consonants and vowels & reduced articulatory
gestures (co-articulation of consonants and vowels). An example of this tendency is
consonant deletion resulting in the loss of historical /r/ in certain environments in some
accents: e.g. to pronounce start as without /r/ is simpler than to pronounce it as
.
The nature of the segments in an accent also affects the extent of variation which the
accent may have. Wells points out that differences occur in accents as languages constantly
work towards naturalness (95). This implies that less natural segments have a tendency to
be modified towards more natural pronunciation. Segment types which are more natural
than others include those learnt earlier by children and those found more widely in the
languages of the world. For example, a pronunciation change involving modification
towards a more natural segment is the dark /l/ which occurs in certain positions. In many
areas, this variant of /l/ is now pronounced with lip rounding which results in a more
natural segment type. In this way, the interplay of several factors contributes to the
development of changes in pronunciation and hence accent variation.
1.2.2 Varieties of English in the British Isles
The heterogeneous nature of accent variation may be illustrated using the array of spoken
English varieties that exist in the United Kingdom of Great Britain/British Isles. A first-
time visitor to this group of islands is immediately aware that the varieties of English there
are as widely varied as the English, Welsh, Scottish and Irish inhabitants of the islands.
Linguistic reconstruction has helped to identify the language family from which these
11
accents of English are derived as the Indo-European family and Germanic sub-family of
languages. Thus, this dialect of English has its original home on the group of islands known
as the British Isles. Located off the North West coast of Europe, the islands include Great
Britain, (the largest), Ireland and over one thousand smaller islands. The island of Great
Britain is divided into Scotland in the North, England to the South and Wales to the West.
The island of Ireland is divided into the Republic of Ireland in the south and the six
counties of Northern Ireland in the north. England, with a population of about 51.2 million
has the lion share of an estimated 61.7 million for the entire United Kingdom comprising
England, Scotland, Wales and Northern Ireland (Crystal 68). With a highly industrialized
economy, it is the largest economy of the four countries of the United Kingdom. London as
the political, legal, administrative and commercial centre of the UK also occupies a crucial
position in the linguistic landscape of England because it is home to a major spoken
English variety.
Based on a linguistic rather than geographical classification, two major language varieties
have been identified in England namely: North of England and South of England varieties.
This classification serves to further emphasize the North-South dichotomy which has
pervaded the consciousness of Britons from the eighteenth century till present and which
also formed the discourse of a number of nineteenth-century novels which are regarded as
social commentaries regarding this period. Described by Mugglestone as “prime signifier
of the social divide” (96), these novels include Gaskell‟s North and South, Dickens‟ Hard
Times and Benjamin Disraeli‟s Sybil. Commenting on the latter for instance, Disraeli
12
describes: “two nations… who are formed by a different breeding, are fed by a different
food, and are ordered by different manners” (Mugglestone 84)
Besides the two major North/South classifications, minor varieties which are a sub-set of
the North and South varieties have been identified in Hughes, Trudgill and Watt as follows:
North-east, Central north, Central Lancashire, Merseyside, Humberside, North-west
Midlands, East Midlands, West Midlands, South Midlands, East south-west, West south-
west, South-east and East Anglia (70). From these areas arise regional varieties such as
London English, Norwich English, Bristol English, Liverpool English etcetera. However,
in line with the characteristic nature of language variation in general, the varieties listed
above are not distinct spoken varieties as such but exist on a continuum with each accent
consisting of inter-related pronunciation features.
Aside from a regional classification such as that presented above, socially stratified
varieties of English originating from the social class divisions of the 17th century can also
be identified within England as follows: Upper class, Upper middle class, Middle class,
Lower middle class, Upper working class, working class, Lower working class and
Underclass (Hughes, Trudgill and Watt op cit.). Factors such as home, attitudes, clothing,
mannerisms, occupation, position, education and speech determined this assessment. In
England, at the uppermost rung of the social ladder are the title holders of nobility and their
relatives, some with substantial inherited wealth, while individuals reliant on state benefits
for income are those at the lowest rung, forming the underclass. The characteristic fluidity
within the regional varieties is also present in the social class varieties as it can be observed
13
that many individuals do not stay within particular social classes throughout their lives.
Movements in and out of social classes occur due to changes in occupation, financial
circumstances and social standing and these also tend to affect speech characteristics. The
aspirations of members of the lower classes for upward mobility in terms of economic
circumstance also occurs in language use as the stigmatized features of the speech of the
lower classes is often abandoned in favour of the prestige speech of the members of the
upper classes. However, despite these aspirations, it is not always possible to erase lower
class speech features from the speech of upper-class speakers. To the discerning therefore,
when an Englishman speaks, it is usually possible to identify his/her geographical origin
and social status.
However, at the apex of these regional and social class varieties of English within England
is a regionless spoken English variety with origins in the south-west of England named
Received Pronunciation (RP).Thus, the earlier reference to the Standard British English
accent, refers to the totality of the phonemic contrasts and tone group associated with this
accent (Received Pronunciation). This accent is spoken in England and is also variously
referred to as BBC English (British Broadcasting Corporation English), Standard English,
Southern British Standard, and Public School Pronunciation (PSP). It is more commonly
referred to “talking proper or “talking posh” (Mugglestone 209).
Received Pronunciation (abbreviated RP) is the name with which linguists generally refer
to this accent. Received in this name refers to the nineteenth-century sense of the accent as
“accepted in the most polite circles of society” (Ramsaran 47) . RP is said to be regionless
14
because even though it is historically linked with the speech of London and surrounding
areas, it is not the accent of any particular region. This is due to its propagation by the UK
public schools which are in practically every region of England. Thus, RP is linked with
public school education, being the accent of those educated at these elite private schools
which are beyond the means of many parents. The elitist nature of RP is probably why it is
the model taught to foreign learners of British English. It is also the most widely used
accent among the British Broadcasting Network (BBC) newsreaders, the most widely
understood accent within England, the most thoroughly described English accent and the
accent of the British colonial officers of the sixteenth century.
1.2.2 The Development of the Nigerian English Accent (NEA)
The phenomenon of language variation which explains the existence of varieties such as
RP mentioned above is a linguistic situation which is sometimes further complicated by
language spread. Language varieties which have been created by such complex interplay of
factors include national and regional varieties of English such as American English,
Caribbean English, Singaporean English, South African English and Nigerian English. Of
particular interest in this study is Nigerian English whose origins can be traced to the
British colonial explorations of the sixteenth century.
The spread of English beyond its native shores to Nigeria and other former colonies of
Britain has been well documented by several scholars (cf. in South Asia by Kachru; in
Africa by Spencer; and in the West Indies by Craig). The spread of the language in these
regions has led to the development of dialects of the English language referred to as non-
15
native dialects. These are dialects of English which are used by non-native speakers of
English. The English language is not just spreading fast in former colonies of Britain but it
is also spreading fast in China, the former Soviet Republic and even the Arab world where
it is used for specific purposes such as education and international communication. Kachru
represents the global profile and overwhelming presence of English across the world in a
model referred to as the Concentric circles of English (47). The model comprises three
concentric circles which reflect the nature and roles of the English language in countries
around the world. The model provides examples of Inner Circle countries where English
is the mother tongue ( i.e. USA, UK, Canada, Australia and New Zealand); Outer Circle
countries are those where English has been transplanted as a result of colonialism and the
language functions as a second and official language (i.e. Nigeria, Ghana, Malaysia,
Zimbabwe, Singapore, Bangladesh) and the Expanding Circle countries where English is
strictly learnt as a foreign language ( i.e. China, Japan, Saudi Arabia, Israel, Caribbean
countries). „Expanding‟ refers to the increasing number of countries adopting the language
for the purpose of international communication.
In a detailed description of the nature of English in these countries, Kachru refers to the
pluricentricity of these Englishes to reflect the gradual development of new local centres
for authentication of the models and norms of acquisition, teaching, and creativity of non-
native Englishes (31). Further analysis classifies the users of English into two distinct types
– the norm providing and the norm depending. Two levels of norm-providers are
recognized as L1 and L2 norm-providers for Inner Circle and Outer Circle norms
respectively while the Expanding Circle countries are the norm dependents. In the Outer
16
and Expanding Circle countries where there was already a language or even several in
existence before the introduction of the English language, accent variation is even more
complex due to bilingualism and multilingualism. The sociolinguistic implications of such
contact between languages include the nativization and acculturation of the English
language in the new environments and this has led to the emergence of distinct varieties of
the language with different national and regional names such as Indian English,
Singaporean English, Ghanaian English, Chinese English, and Nigerian English.
The process of the emergence of the Nigerian variety of English is not unlike what takes
place in other language contact situations where numerous indigenous languages co-exist
with the English language. In Nigeria, interference is another consequence of language
contact which plays a crucial role in shaping the characteristics of this second language
variety of English. These interference features occur at every level of language description,
but are perhaps most noticeable at the phonological level where these peculiar features
characterize the accent. Mother tongue interference features are further complicated by the
improper teaching methods adopted by a majority of second language teachers which
either: (a) assume that there is no need to teach listening comprehension to students, or (b)
teach the four basic principles for acquisition of language proficiency in reverse order. The
second method involves the teaching of writing skills or spelling of words first, followed
by reading passages/poems and then the student is encouraged to speak. In this way, the
typical order of listening, speaking, reading and writing (LSRW), which is the order
unconsciously adopted when learning the mother tongue, is reversed. This technique has
17
also often produced individuals who are very proficient in the writing of English but lack
similar levels of oracy.
This peculiar ordering and interference features are some of the factors responsible for the
evolution of this accent leading to characteristics which make it unique and different from
its British counterpart. However, similar to RP, the totality of the phonemic contrasts and
tone group associated with the educated variety of English spoken in Nigeria constitutes the
Nigerian English Accent (NEA). It is against this backdrop of language variation that we
intend to investigate the topic: Intelligibility of the Standard British English accent to
Nigerians.
1.3 RESEARCH PROBLEM
In recorded history, the present global spread and use of English is unparalled. An estimate
provided by Crystal of the number of English speakers is about two billion (8). Year 2001
estimates provided by Crystal puts Inner Circle/native speaker population at 380 million
while the Outer circle and Expanding circle populations are put at 500 million and 1,000
million respectively (61).
This global spread of the English language in terms of use and characteristics has resulted
in concerns over the issue of maintaining mutual intelligibility between diverse speakers of
the language (Christophersen, Kachru, Munro, van der Walt, and Rajadurai). The most
recent of these concerns have made calls for rigorous comparisons and contrasting of native
and non-native accents with a view to assisting interlocutors achieve communication goals.
18
These concerns have been voiced in relation to all the dimensions of language study but the
achievement of intelligibility is perhaps most crucial at the level of phonetics/phonology
where variation is most noticeable.
The imperative of achieving speech intelligibility in communication can only be grasped in
light of the current critical situation among speakers of different accents of English.
Sometimes, variation between accents is so marked that speakers of different accents of
English have even imagined their interlocutor to be speaking another language entirely and
not merely another accent of English. Numerous instances of misunderstanding have been
observed and while some of the incidents are hilarious (Smith and Christopher 92-3,
Shockey 119); others are serious, sometimes leading to disastrous consequences (Brown 1,
Atechi 60, Rocha 98). An example of a funny incident arising out of intelligibility failure is
that narrated by Smith and Christopher (op cit.) in which Salmon Rushdie‟s wife, an
English woman met an Indian intellectual on the train. They were both conversing in
English but each thought the other was speaking another language. The traumatic
experience of a 19-year-old Dutch undergraduate has also been narrated. After studying
English for 6 years at a grammar school and 1 year at a Dutch university, this student went
to England for the first time and had to ask for directions to the bus station from a porter.
However, he could not understand a single word the porter said to him. Brown also
reported that in England, “many overseas students are unable to understand English as
spoken by university and college lecturers, sometimes to such an extent that they give up
their course of studies” (1). One tragic consequence of lack of intelligibility and what is
possibly the most deadly accident in aviation history occurred in the year 1977. In this
19
incident which occurred on the Spanish island of Tenerife, 583 people lost their lives in an
air crash because the airport traffic controller and a German pilot did not understand each
other even though they were both speaking English. This incident led to major
communication and language training reforms in the airline industry.
Other scholars have even gone beyond merely expressing concerns over the issue of
maintaining mutual intelligibility between speakers of different varieties to suggesting that
English is undergoing a process of radical change which will eventually “lead to
fragmentation of the language into a family of languages” (McAuthur 39). This is
suggestive of the devolvement of the Romance languages (i.e. French, Italian, Spanish,
Portuguese and Romanian) from popular Latin.
While the same process of radical diversification may not yet have occurred between RP
and the Nigerian accent of English (NEA), there are variations between the two accents
which may affect intelligibility. These differences have been highlighted in relation to the
unique features of the NEA. Areas of divergence of the NEA from RP have also been well
documented. These areas of variation are well documented by scholars such as Adetugbo,
Awonusi, Udofot, Jowitt, Banjo and Bamgbose. Unique features of NEA which
differentiate it from RP have been identified by these scholars and are listed as follows: a
reduced vowel system, a reduced intonation system, non-differentiation in some aspects of
length, absence of glottalization in some contexts, the voicing of non-voiced consonant
endings, the insertion of vowels in syllabic consonants, the insertion of epenthetic vowels
in some consonant clusters, substitution of alveolar fricatives for interdental fricatives
20
etcetera. These are features which characterize the Nigerian accent of English and lead to
intelligibility problems between speakers of the NEA and their interlocutors who may not
be familiar with the accent. Tiffen‟s 1974 study of the intelligibility of Nigerian English
identified some of these features as constituting intelligibility problems between Nigerian
speakers and British listeners. In other countries, intelligibility studies have measured the
level of intelligibility of non-native speakers to native speakers of the English language.
Bansal on Indian English, Elalani on Jordanian English, and van der Walt on South African
English are a few examples. The main aim of these studies is usually to make
recommendations for the second language learner in order to improve the level of
intelligibility to the native speaker. The outcome of such studies is usually a description of
second language speech as “annoying and burdensome” to the native speaker, thereby
focusing on the native speaker as the listener-judge and ascribing to him a higher rank in a
subtle hierarchical ordering of the communication process (Elalani 84). Another example
of condemnation of non-native English speech is a commentary on the English
pronunciation of a black Namibian newsreader that: “[H]e mispronounced almost every
word, messed up almost every sentence, and sent my blood pressure roaring with every un-
understandable news item” (Harlech-Jones 276).
As true as the positions above are that lack of proficiency often hampers the speech of the
second language speaker, anecdotal evidence does suggest that non-native speakers have as
many difficulties understanding native speakers as the native speakers have being
understood by the non-native speakers. However, many studies on intelligibility have not
considered a reversal of the communication process in which the native speaker is the
21
speaker and the educated non-native speaker, the listener and judge of intelligibility. Even
more specifically, none that we know of has considered the level of intelligibility of any
native speaker accent to Nigerians. There is often no consideration of the difficulties which
the Nigerian L2 speaker faces in understanding native speaker RP speech. This study
attempts to examine intelligibility from a different (non-native speaker/Nigerian)
perspective based on the following assumptions:
a. That variation exists between the Standard British English accent and the Nigerian
English accent.
b. That as accent variation leads to problems of intelligibility in face to face
communication involving Nigerian speakers and British listeners, it will also lead to
intelligibility problems in face to face communication involving British speakers
and Nigerian listeners.
c. Empirical research has revealed that non-native speakers (not Nigerians) do not
always find native speakers highly intelligible showing that native speaker
phonology is not inherently intelligible. (Smith and Rafiqzad, Bent and Bradlow,
Munro, Deterding & Kirkpatrick, Gupta ). This may also be true with Nigerians.
There is a need for this study in view of the crucial role of the RP accent as the
pronunciation model in Nigeria. It is hoped that this study will explore this area as there is
presently a dearth of research in this field.
22
1.4 RESEARCH OBJECTIVES
In exploring how speakers of the Standard British English accent and Nigerians negotiate
intelligibility in the face of accent variation, this study is guided by the following
objectives:
1. To measure the general intelligibility of the Received Pronunciation accent to
Nigerians and identify the features which determine intelligibility.
2. To establish the extent to which intelligibility of RP to Nigerians is affected by the
segmental features of the Standard British English accent (RP).
3. To ascertain the influence of the nuclear/emphatic stress features employed by RP
speakers on intelligibility of the accent to Nigerians.
4. To investigate how intelligibility is influenced by the intonation contours employed
by the speakers of RP at the suprasegmental level of phonological analysis.
5. To explore the impact of accent familiarity on intelligibility negotiation between
speakers of the Standard British English accent (RP) and Nigerians.
1.5 RESEARCH QUESTIONS
The following questions have been formulated to guide the exploration of the
sociolinguistic and intelligibility issues central to this study.
1. What is the intelligibility level of the Standard British English accent (RP) to
Nigerians and what features determine it?
2. To what extent do the segmental features of the Standard British English accent
(RP) determine its speakers‟ degree of intelligibility to Nigerians?
23
3. Do the nuclear stress patterns of the Standard British English accent (RP) affect the
degree of its speakers‟ intelligibility to Nigerians?
4. To what degree do the intonation features employed by the speakers of the Standard
British English accent (RP) affect its speakers‟ intelligibility to Nigerians? (at the
suprasegmental level of phonological analysis).
5. What is the impact of the Nigerian listeners‟ extent of familiarity with the Standard
British English accent (RP) on the degree of intelligibility?
1.6 SCOPE & LIMITATIONS OF THE STUDY
The range of disciplines covered by this study includes intelligibility, sociolinguistics and
phonetics/phonology. Catford provides a broad view of intelligibility as “level of
understanding” (3). However, as understanding may take place at various levels, further
delimitation is necessary to enable us specify the precise issues covered by the study.
Three categories of understanding which may take place in communication are identified
by Smith and Nelson (429) as follows:
(a) word/utterance recognition;
(b) word/utterance meaning; and
(c) meaning behind word/utterance.
This study is concerned with the first category of word/utterance recognition which on the
continuum of intelligibility corresponds to phonological intelligibility. The other categories
involve other levels of language study which roughly correspond to semantics and
pragmatics.
24
Phonological intelligibility is the area of intelligibility testing which involves the use of
particular vowel or consonant sounds and particular rhythmic, intonational and other
prosodic features. It also includes the structural and systemic interrelationships between
them and the phonological representations of the rules governing the relations between all
these features of an individual or a group‟s pronunciation. These pronunciation features
constitute accent which in this study involves two accents of English: the Standard British
English accent (RP) and the accent of English used by educated Nigerians (NEA).
Extra-linguistic variables such as age, sex, socio-economic class, educational background
and educational attainment are some of the factors which are relevant to the study and
therefore necessitated the introduction of certain themes in the realm of sociolinguistic
theory for adequate explication. These themes are explored by focusing on a specific social
and linguistic group in a university in England and another in Nigeria, the two locations of
the research study. The undergraduate level of education was selected as the educational
level of interest in order to achieve a parallel comparison of both societies. Besides the
appeal to qualitative sociolinguistics, the quantitative nature of intelligibility studies is also
explored as the objective framework of intelligibility testing rather than the subjective is
used for data generation, analysis and interpretation. The objective framework involves
assigning numerical values to various test materials on the segmental and suprasegmental
features of RP speech which were designed for the study. The combination of both the
quantitative and qualitative modes of analysis and interpretation not only emphasizes the
25
interdisciplinary nature of the study, it also reveals previously unexplored areas on
intelligibility of this native speaker accent of English to Nigerians.
The findings of this study may have certain limitations with regards to the veracity of
claims of the subjects concerning their parents‟ educational level as there was no way of
verifying the truthfulness or falsity of the information supplied in the questionnaires.
However, it is assumed that the rigorous pre-selection tests which successful informants
were required to undergo is adequate enough to overcome such limitations. Another
measure that was taken to overcome this observer‟s paradox is that the subjects were
reassured of their anonymity as they were instructed not to supply their individual
identities.
1.7 SIGNIFICANCE OF THE STUDY
This study differentiates itself from most studies in the field of intelligibility research
because it represents a change in perspective from the previous view of intelligibility as a
concept which may be examined only from a native speaker viewpoint. The traditional
view of intelligibility has a subtle suggestion of the linguistic superiority of native speaker
English accents over non-native English accents in the hierarchical order of the
communication process. By attempting to judge the level of intelligibility of a native
speaker English accent such as RP through the examination of the responses of Nigerians, a
unique attempt is made to examine the process of communication from a Nigerian
perspective. This study is necessary in view of the insights offered by this substitution of
roles which serves to reveal the nature of intelligibility negotiation between speakers of RP
26
and speakers of the Nigerian English accent, with the educated Nigerian non-native speaker
of English playing the role of listener and hence judge of intelligibility.
The study further distinguishes itself from other studies due to its establishment of gender
as a viable variable in native/non-native speaker communication. The study goes beyond
the correlation of socio-economic, linguistic, educational background and educational
attainment alone with intelligibility. It attempts to establish the link between gender and
intelligibility in relation to the peculiarities of RP and the Nigerian English accent.
In addition to this, the study gives due cognizance to the present unprecedented migration
of Nigerians to the United Kingdom and seeks to examine the effect this may have on
intelligibility. Therefore, several Nigerians living in the United Kingdom were selected to
represent Nigerian Diasporic dwellers in the UK who form a significant population. This
enables us examine the effect of accent familiarity on intelligibility. This correlation of
listener factors with intelligibility from a Nigerian perspective reveals the effects of listener
factors (which are external to the speaker) on intelligibility.
27
CHAPTER TWO: LITERATURE REVIEW
2.0 INTRODUCTION
Having established the basis for this study in the introductory chapter, a review of relevant
literature and authourities on intelligibility will now be attempted. Relevant works relating
to the concept as well as aspects of previous studies on the intelligibility are examined in
this chapter. Factors affecting intelligibility are also discussed while the chapter is
concluded with a consideration of descriptions of the Nigerian accent of English.
2.1 SPEECH INTELLIGIBILITY
Intelligibility, like any other concept in language is controversial in nature. This seems to
be due to a lack of consensus among scholars on appropriate terminologies for the
explication of the concept. Jenkins for example, acknowledges that: “there is as yet no
broad agreement on a definition of the term intelligibility: it can mean different things to
different people” (149).
However, the ambiguity involved in the use of the terminologies may be simplified by
reconciling the various definitions with different levels of linguistic analysis. On that basis,
it becomes obvious that the different terminologies such as „effectiveness‟,
„comprehension‟, „word/utterance recognition‟, „interpretability‟ etcetera, seem to have
been employed to refer to basically the same notion. In the following section, major works
that employ various terminologies and definitions are reviewed in relation to the
28
dimensions of linguistic analysis to enable us determine the specific level to which they
belong.
Catford identifies the steps towards the achievement of intelligibility between a native
speaker of a language and a non-native speaker as follows:
Step 1: Selection of appropriate words, morphological and syntactic systems and the
appropriate sounds by a speaker.
Step 2: Execution of the selected forms through the production of an utterance.
Step 3: The transmission of sound through the physical medium.
Step 4: Discrimination between perceived sounds and the sounds of the listener‟s
own language. A failure in identification occurs if the hearer fails to
distinguish between two sounds that are auditorily similar but affect
meaning.
Step 5: Correct interpretation of the message by the association of the linguistic
forms with the appropriate elements in the situation. Correct interpretation is
demonstrated by making a correct response. If the listener‟s response to the
utterance is not in accordance with the expected response, then a failure in
interpretation has occurred.
Catford also argues for the crucial role of linguistic and situational context in the
attainment of intelligibility. Linguistic context is described as: „…the given form of a word
which consists of the words or other linguistic forms with which it is associated or
surrounded‟ while situational context is “everything else in the situation which is relevant
29
to the speech act…includes, therefore, not only the speaker and the hearer themselves, and
their relative positions and actions at the moment of utterance… but, the hearer‟s linguistic
background and experience, his educational and cultural background, etc”.(13)
Furthermore, Catford distinguishes between a „lower threshold‟, „a medium threshold‟ and
a „higher threshold‟ of intelligibility. A hearer‟s threshold of intelligibility is defined as the
point on a scale at which a speaker‟s utterance in a given context becomes completely
intelligible for that hearer.
Regarding the use of terminology, Catford uses the terms „intelligibility‟ and
„effectiveness‟ to describe the process of communication. However, „intelligibility‟ is
involved from the first stage up to the fourth stage out of the five stages discussed above,
i.e. the stages of selection of linguistic forms, execution of linguistic forms, transmission
from speaker to hearer and hearer‟s identification of linguistic forms. The utterance is said
to be „effective‟ only if the last stage of the process which involves the hearer‟s
interpretation of the linguistic forms is achieved.
Although Catford distinguished between intelligibility and effectiveness: „the effectiveness
of an utterance …is not identical with what is usually meant by the term intelligibility‟, it is
not clear why it is included in the process of communication given the following reasons.
First, the speaker may be quite intelligible to the hearer, but the hearer may be unable to
make the appropriate response due to certain limitations. Secondly, the hearer may
willfully make inappropriate responses even though the speaker was perfectly intelligible.
30
That „intelligibility‟ may occur without effectiveness being achieved shows that
effectiveness is outside the realm of intelligibility. It belongs, perhaps in pragmatics
because it was stated that it depends on extra-linguistic factors for its attainment.
A more practical viewpoint of intelligibility is provided by Smith and Rafiqzad which is
rather different from Catford‟s perspective. In what is possibly the largest empirical study
of intelligibility, Smith and Rafiqzad investigated the validity of the proposition that the
educated native speaker of English is more likely to be more intelligible than the educated
non-native speaker. The research spanned eleven countries and involved 1,386 people from
Hong Kong, India, Japan, Korea, Malaysia, Nepal, The Phillipines, Sri Lanka, USA,
Bangladesh, China, Hong Kong, Indonesia and Thailand. Collaborators were employed
from these countries to choose educated speakers of English between the ages of 20 and 40.
Other criteria used for the selection is that none of the participants had: (1) spent more than
four consecutive months in any English speaking country; (2) been formally educated in
schools directed by native speakers of English; or (3) ever lived with English speaking
families or groups.
Each speaker was taped while giving a ten minute speech to a group of 10 educated fellow
countrymen. A cloze procedure test of the speech was then prepared and every sixth word
in the passage was deleted. This passage was then presented to the listeners along with a
questionnaire to test listening comprehension. The conclusion was that the native speaker
was not always the most easily understood person.
31
Due to the empirical nature of this work, the notion of intelligibility which is proposed
seems to be practical in orientation. It is closely related to the research design of the study.
It is stated that “the operational definition for intelligibility is capacity for understanding a
word or words when spoken/read in the context of a sentence being spoken/read at natural
speed” (371). Therefore, the more words the listeners were accurately able to fill in on the
cloze test, the higher the speaker‟s intelligibility.
The three concepts of „intelligibility‟ and „comprehension‟ and „understanding‟ are used
interchangeably in the study as it is initially argued that “we realise that the greater the
comprehension of the context material, the more likely intelligibility will occur” and also
that “…the difficulty rank ordering correlated highly with the rank ordering of both
intelligibility and understanding” (375). Also, the questionnaire presented to the listeners
along with the cloze test was to test how the results from this questionnaire would compare
with the intelligibility results of the cloze procedure test. This further shows that to them,
intelligibility and comprehension are closely related. However, a somewhat contrary
opinion is expressed thus: “we feel comprehension involves a great deal more than
intelligibility” (367). This statement appears to be contradictory as no definition or
description of comprehension is offered anywhere in the study. The lack of adequate
description makes it difficult to ascertain the level of linguistic analysis each of the notions
refers to. But it seems that it may be safely assumed that both „intelligibility‟ and
„understanding‟ correspond to phonological and grammatical intelligibility while
„comprehension‟ belongs with semantics which involves another level of analysis.
32
Another empirical study which provides a definition of intelligibility which is related to the
methodology is that conducted by Munro and Derwing. They employed three terms,
„accentedness‟, „comprehensibility‟ and „intelligibility‟ and attempted to measure the
interrelationships among them through an examination of second language speech. The
study involved 18 native speakers of English and 10 Mandarin speakers of English. The
Mandarin speakers were presented with a page of cartoon and asked to describe the events
depicted within a sound-treated room. Short excerpts of their speech was then extracted and
presented to the 18 native speakers. All the speakers had some linguistic or phonetics
training. This was done deliberately. The listeners performed two different tasks. The first
was an intelligibility assessment task which involved writing down what they heard in
standard English orthography. The second task was to measure comprehensibility on a
scale of 1 to 9 where „1‟ is equal to extremely easy to understand and „9‟ is equal to
impossible to understand. A third test was done a few days later to test the degree of
foreign (non-English) accent of the Mandarin speakers. This was done to avoid skewing of
the results due to exposure to the accent. Again, a 9-point scale was used, where 1 is equal
to no foreign accent and 9 equal to very strong foreign accent. Three sets of scores were
generated based on the assessments of intelligibilility, comprehensibility and accentedness
of the speech samples.
The three notions were measured using different tests and they were also defined separately
as follows: intelligibility as „the extent to which a speaker‟s message is actually understood
by a listener‟ (76); accentedness as „variables that caused the speech samples to sound
deviant‟ (91); while comprehensibility was simply „ease of interpretation‟. Munro and
33
Derwing assert that while there is a relationship between the three concepts, they
„correspond to related but partially related dimensions‟ (90). However, the correspondence
between the notions and the dimensions is never provided as no correlation was found
between the scores obtained in the three tests: “the distributions of perceived
comprehensibility and accent scores were noticeably different…the strength of the
correlation among any of the three possible pairings of dimensions tended to be in the
moderate range for most listeners; there was not one perfect correlation” (90). Another
finding that supports the assertion that there is no correlation between the tests is the
evidence that “foreign accent scores did not predict intelligibility very well” (91).
These findings seem to provide credence to the view that the three notions belong to
different dimensions though they are never implicitly or explicitly linked to any dimension.
However, it may be concluded that because the intelligibility scores were based on
speakers‟ phonemic and phonetic errors, intelligibility relates to the dimension of
phonological recognition. Comprehensibility was measured through listeners‟
impressionistic judgements on how easy or difficult it was to understand the accent and this
also relates it to understanding of phonetic and phonological content of the utterances. It is
generally acknowledged that accent involves prosodic elements of language varieties and
so accentedness may be said to also belong to the phonetic/phonological dimension.
The next propositions of intelligibility provide brief theoretical viewpoints. They are
similar to Catford‟s essay in terms of being theoretical in nature. The first consideration is
Kenworthy‟s proposal. Two terms “intelligibility” and “understandability” are employed
34
interchangeably to refer to the dimension of understanding of the speaker‟s words which is
defined as „the more words a listener is able to identify accurately when said by a particular
speaker, the more intelligible that speaker is‟ (13). However, the third term seems to be
unequivocal as it is identified as: „communication‟ and described as, „involves reading the
other‟s intention‟ (13). Communication clearly refers to the more complex level of
understanding of the speaker‟s intention. It is also asserted that:
[C]ommunication is much more complex than intelligibility and
understandability, as intentions only exist in the other person‟s mind, and
listeners need to make use of all the information available to them in order
to guess the speakers‟ intentions. The information comprises the knowledge
listeners have about the situation, the knowledge speakers and listeners
share, and so on (16).
Kenworthy‟s definitions of intelligibility, understanding, and communication seem to refer
to two dimensions. The first two refer to the phonological level while the third notion is
related to pragmatics.
Dalton and Seidlhofer propose the terms “accessibility” and “acceptability”. They argue
that a speaker makes his/her utterance accessible if “he/she succeeds in making an
interlocutor understand it” while acceptability involves the listener factor of attitude which
is defined as “the value interlocutors ascribe to the speaker‟s accent (19)”. They relate the
two concepts to pronunciation and to „social and psychological factors‟ (9). Accessibility
is determined not only by the interlocutor‟s expectations and feelings, such as „experience
35
with and tolerance of low prestige or foreign accent‟ (10). Acceptability on the contrary,
depends on the value the interlocutors ascribe to the speaker‟s accent. Thus, for them, an
utterance may be phonetically accurate but not acceptable. Dalton and Seidlhofer‟s
accessibility clearly refers to the dimension of phonological intelligibility while
acceptability cannot be readily linked to any linguistic dimension as it relates to
psychological value-judgements that are non-linguistic in nature.
A one-term proposition is suggested by Bamgbose to cover the three dimensions of
understanding the speaker‟s words, understanding the speaker‟s meaning and also grasping
the intentions. The term is intelligibility which is defined as “a complex of factors
comprising recognizing an expression, knowing its meaning and knowing what that
meaning signifies in the sociocultural context” (10). It is described as a communicative act
which „involves a speaker and addresse, both participants contribute to the speech act and
its interpretation, and part of this contribution is making an allowance for the accent and
peculiarities of the other person‟s speech‟ (11). Thus, Bamgbose‟s single-term definition
can be linked to the three levels of phonology, semantics and pragmatics.
In a similar manner, James proposes one cover term “comprehensibility” to refer to the
three dimensions of understanding the speaker‟s words, understanding the speaker‟s
meaning and also grasping the intention. His definition of comprehensibility as “ a cover
term to refer to all aspects of the accessibility of the content - opposed to the form - of
utterances” (56) clearly relates the single term to three different dimensions.
36
Jenkins defines the concept of intelligibility as “the production and recognition of the
formal properties of words and utterances and in particular, the ability to produce and
receive phonological form” (10). This definition suggests that her concept is linked to only
the first level of language use which is phonological understanding.
Field proposes two terms: „intelligibility‟ and „comprehensibility‟ which were defined as
“the extent to which the content of the message is recognizable” and “the extent to which a
speaker‟s message is understandable, thanks to a combination of appropriate vocabulary,
correct (or appropriate) syntax, sensitive pragmatics and mastery of basic features of
pronunciation” respectively. Field‟s definitions, thus, correspond to two levels:
understanding of the speaker‟s words and understanding of the speaker‟s intentions which
involve the phonological and pragmatic levels respectively.
Smith and Nelson propose three terms – “intelligibility” which they define as
“word/utterance recognition”; “comprehensibility”- which they define as “the meaning of a
word or an utterance”; and “interpretability” which they define as “the meaning behind the
word or utterance” (334). To enable us link these definitions to the appropriate levels, we
may have to refer to the context of the empirical research in which they offered the
definitions. The definitions were based on a cloze procedure test in which the listeners
filled in the appropriate word(s). To Smith and Nelson therefore, the intelligibility of a
passage is high if the reader is able to copy the passage or fill in the blanks of the missing
words without much difficulty; a passage read is comprehensible if the reader is able to
make sense of the sentences or paraphrase them and interpretability is high if the reader is
37
able to identify the author‟s intentions. The first two terminologies seem to correspond to
Catford‟s “intelligibility” which is linked to understanding of the speaker‟s words while
“interpretability” is clearly linked to the higher dimension of understanding the speaker‟s
intentions.
To conclude this discussion on the inter-relationship between terminologies and levels of
language use, mention must be made of Smith and Nelson‟s assessment of intelligibility in
which they concluded that the terms “comprehensibility” and “interpretability” relate to the
discourse notions of “locutionary force” and “illocutionary force” respectively while
intelligibility relates to word/utterance recognition, corresponding to phonological
intelligibility. Intelligibility is further described as „recognising words and other sentence-
level elements of utterances‟ (430). Comprehensibility is described as „the category of
understanding meaning or the „speech act of request for a particular activity‟ (334).
Interpretability is also described as „the apprehension of intent, purpose, or meaning behind
an utterance‟. To illustrate the inter-relationship between the dimensions, these three
components are presented in an interactional scenario which is narrated below: An
Australian woman was having a conversation in English with a taxi driver in Istanbul.
Things were going well until she asked the driver to turn off the interior light as the driver
refused sharply. Since her request seemed innocuous to the passenger, and since a mutual
compatibility in English had been established by the preceding conversation, she thought
there had been a simple failure of intelligibility or comprehensibility – that the driver had
misheard or misunderstood some part of her utterance - so she repeated it, only to receive a
38
near-hostile negative response and marked silence until the end of the trip, which
terminated in the driver almost snatching the fare from her and driving away rapidly.
The story illustrates the fact that cultural differences (relating to pragmatics) cannot be
divorced from intelligibility. Turning off the light while a man and a woman are alone
together is almost taboo in Istanbul while it does not have the same cultural significance in
English culture. The driver heard the woman but did not successfully grasp the meaning
behind it because her reasons for wanting the light turned off were misinterpreted by the
driver. These discourse notions refer to the situational, social and cultural interpretations of
utterances which are extralinguistic as they depend on factors outside the linguistic form
for their interpretation. On this basis, it may be concluded that „intelligibility‟ (the level of
understanding the speaker‟s words/utterances), is limited to the field of
phonetics/phonology and does not extend to either „comprehensibility‟ or „interpretability‟
which correspond to higher levels of word/utterance meaning. These two are major
preoccupations in the field of discourse analysis/pragmatics.
These clarifications were necessary to enable us establish the scope of this investigation. It
has assisted in defining the interests of the present study as being limited to the recognition
of words/utterances, thereby linking it to phonetic/phonological intelligibility. Thus, this
area of study serves as the background for our investigation of accent intelligibility.
Another area which requires some discussion is factors affecting intelligibility. Background
noise, speaker‟s voice quality, speech rate, and familiarity are some of the variables that
39
have been linked to intelligibility in terms of either aiding it or hindering it. The
relationship between noise and intelligibility was suggested by Flege, Einstein & Verdi,
Clopper & Bradlow and Rogers & Dalby. Any unwanted or disturbing sound during a
communication process constitutes noise. Examples of disturbances which may impair
intelligibility include the ringing of telephones, traffic noises and extraneous conversations
etcetera. According to Rogers and Dalby, noise can occur at different levels of low,
medium and high depending on the volume. An investigation of different noise levels on
intelligibility was conducted which involved experimentation with three levels of low,
medium and high noise to determine effect on dialect recognition. Their conclusion was
that:
listeners are sensitive to dialect variation in noise---although
dialect classification is quite difficult under all listening conditions,
comparisons between the results…revealed that overall accuracy
and interpretability…were better in the more favorable listening
conditions than the degraded conditions (31).
However, noise can be avoided by ensuring that the recordings for intelligibility testing are
done in quiet places, free from unwanted distractions. Sound attenuated booths may also be
used during recordings. These are specially designed sound-proofed rooms which help to
ensure noise-free recordings. High quality recording microphones and digital recorders
could also be utilized to avoid electronic interference which also constitutes noise.
40
Apart from background noise, the voice quality of the speakers in an intelligibility test also
affects intelligibility. Voice qualities such as breathy, creaky, nasal, tense, hoarse, harsh
and whispery usually serve two purposes. First, they perform communicative functions of
speech prosody such as assigning stress and conveying linguistic distinctions and emotional
tone. Secondly, they convey information about the physical, psychological and social
characteristics of a speaker. But apart from providing linguistic and extra-linguistic
information about the speaker, Ramig has reported that voice characteristics may also
adversely affect speech intelligibility. Based on the examination of various acoustic and
perceptual reports of voice characteristics, he argues that “voice quality may have
detrimental effects on speech intelligibility” (124).
Speech rate is defined by Deterding as “the number of syllables spoken per second” (225).
The speed of an utterance may be calculated as the total duration divided by the number of
syllables. In the calculation of speech rate, the final syllable is usually excluded while the
value is stated in syllables/s. Previous research on speech rate has yielded a common
finding that a fast speaking rate results in reduced speaker intelligibility because a lot of
reduction or under-articulation of segments takes place in fast speech. Native speaker
speech is often described as being faster than non-native speech, so we can therefore
conclude that speed of delivery may contribute to intelligibility (Anderson-Hsieh &
Koehler, Munro & Derwing).
Two types of familiarity often affect intelligibility. They are accent familiarity and topic
familiarity. However, while it has been sufficiently established that topic familiarity
41
influences intelligibility, the same conclusion has not been adequately proven for accent
familiarity. For topic familiarity, there is some evidence that the use of technical words and
specific register reduces intelligibility. For instance, Smith and Rafiqzad observed that
intelligibility level differed according to the difficulty level of the passages used in their
experiment. However, there is no such consensus for the conclusions on accent familiarity
and intelligibility. On one hand, Gass & Varonis, Gupta and Major all agree that
intelligibility is affected by accent. Bross also supports this position by claiming that “the
key to intelligibility is the concept of calibration” (37). Calibration is the listener‟s ability
to make rapid correlations between the different sounds of an unfamiliar accent with the
sounds of the listener‟s own accent. It is argued that once the listener has been able to
calibrate the message, then intelligibility will naturally take place. On the other hand,
Eisenstein & Verdi, Strevens, Eisenstein and Hopper all disagree with the proposition that a
high degree of accent familiarity guarantees intelligibility. They base their conclusion on
the results of empirical studies which seem to indicate that accent familiarity is not a
significant determinant of intelligibility.
2.2 PREVIOUS STUDIES ON INTELLIGIBILITY
Intelligibility studies involving the comprehension of non-native accents of English to
speakers of Received Pronunciation will first be considered and thereafter, Tiffen‟s 1974
study which is the major investigation on the intelligibility of Nigerian English will be
examined. Other relevant studies on the intelligibility of various varieties of English will
also be reviewed in this section.
42
Bansal‟s study of the intelligibility of Indian English is probably the first major work on
intelligibility. It was a pioneering effort which set the trend for several others in terms of
perspective. These studies are informed by a viewpoint that non-native language varieties
are in some way deficient compared with native varieties. Bansal, for example stated the
need for the study as due to:
the emergence of a wide variety of foreign dialects of English ... These
foreign dialects do not diverge to any appreciable extent from native
English, as far as morphology, syntax and vocabulary are concerned, but in
pronunciation they seem to differ so much from native English that
sometimes they do not sound English at all and are hardly intelligible (6).
In order to achieve this aim, an objective approach of intelligibility testing was adopted. It
involved twenty four educated Indian speakers of English from different mother tongue
backgrounds and 178 listeners of varied nationalities including the UK, the US, Nigeria,
and Germany. Various types of test materials were used for data gathering. They were:
connected speech in which the speaker was asked to speak about himself, work, interests
and hobbies for a few minutes, several reading passages, sets of sentence lists and word
lists. The listeners made both oral and written responses to the speakers‟ utterances. In
other words, the listeners gave oral repetitions of what the speakers had said. In certain
cases when the listeners‟ oral responses were not clear, they were then asked to write down
what they had heard. In addition to this, an opinion assessment scale of effort, the Richards
43
and Swaffield test was presented to the listeners. The listeners were to indicate on a scale of
1 to 9 the amount of effort that was required to understand the Indian speaker. Both
quantitative and qualitative methods of assessment of intelligibility were employed. This
involved attaching a numerical value to the Indian speakers‟ utterances correctly identified
by the listeners which was then followed by phonetic analysis of the instances where the
responses were incorrect.
The main conclusion was that the average intelligibility level of Indian English to speakers
of Received Pronunciation is 70% with intelligibility scores for individual speakers ranging
from 53% to 95%. From the analysis of the speech of the least intelligible Indian speakers,
the major causes of intelligibility failure were identified as follows:
a large number of divergencies in the distribution of vowels and consonants,
and in word stress, sentence stress and intonation. What makes these
speakers less intelligible is the frequency of the divergences from the normal
patterns of the distribution of vowels and consonants and the patterns of
stress and intonation in RP (147).
Besides this, the use of wrong stress patterns by the Indians was also identified as
being largely responsible for lack of intelligibility while phoneme substitutions
affected intelligibility only to a lesser extent. The most common substitutions
involved the following RP consonants and vowels:
and . Other causes of
44
unintelligibility were wrong sentence stress and rhythm, lack of clear articulation,
unfamiliar proper names, incorrect vowel length, and lack of aspiration in
voiceless plosives. To avoid these problems of intelligibility, the recommendation
for the Indian speakers was that “the correct distribution of English consonants
and vowels should be learnt by the constant use of a pronouncing dictionary”
(173).
Comparison of the performances of the British, German and American listeners
revealed that:
The German listeners understand Indian English less well than
British listeners. This is particularly marked in the connected
speech and reading passage tests. The conclusion is that it is
features of connected speech that present special difficulties to
German listeners…there is no significant difference in the scores
obtained by the American listeners (118).
The averages obtained by the 16 Nigerian listeners were 52% (connected speech), 34%
(reading passage), 54% (sentence list) and 66% (word lists). The findings regarding the
Nigerian listeners are as follows:
[T]he scores with the Nigerian listeners are significantly lower
throughout. The fact that Nigerian listeners understand Indian
English much less well than British listeners is understandable.
45
Nigerian English and Indian English have diverged from British
English in different directions and they are now farther apart
from each other than either is from the parent language, British
English (115).
Bansal‟s study is significant in terms of the fact that it is the first major study of the
intelligibility of second language speech. The relatively large sample of listeners also
makes the conclusions quite reliable and the adoption of a dual analytic mode also
enhanced the reliability of the intelligibility scores obtained by the speakers. However, the
selection of the speakers was without any identifiable set of criteria. Of major concern is
the fact that the educational and English proficiency levels of both the speakers and the
listeners varied widely as “thirteen of them (Indian speakers) were students but their
subjects of study covered a very wide range…one speaker was employed in the tea trade as
a tea taster” (39). The sample also included university lecturers of English and postgraduate
students of English phonetics. The same trend was observed with the listeners as some of
them were undergraduate and postgraduate students of English. Apart from the widely
differing proficiency levels, 12 out of the 24 Indian speakers were resident in England and
some of the American listeners also lived in England. It is rather doubtful that the
performances of these Indian and American speakers and listeners would have remained
unaffected by their stay in the native speaker community. Their interaction with native
speakers of English may have resulted in significant modifications to their speech. In terms
of institutional viewpoint, Bansal‟s orientation seems to support the view of the inherent
superiority of the native speaker. This view is supported by the recommendation that
46
approximation of native speaker speech norms should be the target of the second language
learner. Thus, he maintains that:
[A]mong the Indian speakers, the best speakers of English are those who
consciously tried to imitate the BBC accent, and those who tried to change
…pronunciation by attending a practical course in phonetics and spoken
English….the less efficient speakers were (those) who have not made any
effort to conform to any minimum standard of English pronunciation
based on British RP (168).
Following the same ideological and methodological frameworks employed by Bansal were
studies such as: Elalani on Jordanian English, Strevens and thereafter Brown on Ghanaian
English and Tiffen on Nigerian English.
With 15 Jordanian teachers and university students in addition to 48 British listeners as
speakers and listeners respectively, Elalani studied the intelligibility of Jordanian English
with the aim of the investigation being “to determine the linguistic variables causing
interference in the Jordanian use of English” (9). The recordings were made in Jordanian
classrooms, with both teachers and children joining in, and in addition, group conversations
were also taped. From these recordings, extracts were made for the listeners to listen to.
The study employed both the objective and subjective approach to intelligibility testing,
with the listeners being required to write down utterances as well as rate them according to
a 3-point scale as follows:
47
Intelligible, i.e. readily comprehensible
Just intelligible, i.e. difficult to understand, but of which sense could be made
Unintelligible, i.e. impossible of comprehension.
The main finding was that “defective pronunciation and inadequate language selection”
(45), were the major causes of unintelligibility. Problem areas for the Jordanian speakers
which constituted problems for the British listeners were the consonants and vowels, lack
of weak forms and differing stress and rhythmic patterns.
However, the specific segmental or suprasegmental features which led to lack of
intelligibility were not stated, neither was a hierarchy of features leading to unintelligibility
done. This makes it impossible to conclude on causes of intelligibility failure. Besides this,
the study seemed to have two methodological defects: firstly, the recordings were done in
Jordanian classrooms which implies that the environment was noisy; secondly, extracts of a
recording was used for the test and the listeners may have found this material to be rather
disjointed.
Two studies have been made on the intelligibility of Ghanaian English. Strevens conducted
“a quantitative assessment of the intelligibility which a West African pronunciation
possessed for speakers of …two accents” (119). With one RP speaker and about a hundred
Ghanaians from the Southern part of Ghana, the conclusion was that:
48
the average score of Received Pronunciation-speaking subjects was
84%, while the average score of Type 2-speaking subjects was 62%.
When it came to the recordings made by the speakers of Type 2
pronunciation, the averages were 27% by Received Pronunciation–
speaking subjects, and 35% by Type 2-speaking subjects. Even allowing
for imperfections in the design and conduct of the tests, it seems an
inescapable conclusion, on the basis of these figures that Type 2
pronunciation is a less efficient means of communication (120).
This conclusion was rather surprising because it claims that the speakers of RP were better
understood by Ghanaian listeners than other Ghanaian speakers of English as the general
observation in intelligibility testing is that speakers of the same accent understand one
another better than speakers of other accents. However, Brown‟s conclusions were
conversely different as he found that Ghanaians understood each other better than they did
the Britons. With an RP speaker and 45 Ghanaians from different mother tongue
backgrounds, the average intelligibility of the two varieties is as shown below:
Twi listeners: Twi reader 83%
RP reader 72%
Ewe reader 72%
Ewe listeners: Ewe reader 78%
Twi reader 73%
RP reader 70%
49
The scores indicate that all the Ghanaian speakers found other Ghanaians considerably
more intelligible than the RP speakers while Ghanaians with the same mother tongue
backgrounds were more intelligible to one another than Ghanaians from different mother
tongue backgrounds.
Tiffen conducted a major study on the intelligibility of Nigerian English to speakers of RP.
It remains the most comprehensive research on the intelligibility of Nigerian English till
date. The participants in the study were 24 Nigerians and 240 British listeners. While the
24 Nigerians were all undergraduates from Northern Nigeria and Southwest Nigeria, the
British speakers had widely varied levels of education. Some of them possessed
educational qualifications as low as primary education and others had higher levels such as
postgraduate degrees. The motivation for the study was stated as the need to ensure that
Nigerian speakers of English should try to approximate to the norm, which of course is RP.
It is clearly stated in the opening paragraphs:
For this reason, it is important that the type of English used
should keep within certain norms, if speakers are to be mutually
intelligible. If English should become so distorted - and there is
some evidence that this is already happening – as to become
incomprehensible both within and without Africa, one of the
main purposes for which it has been learnt will have been
thwarted (13).
50
Five types of tests were employed in the study namely: connected speech, reading passage,
phonemes, stress and attitude and intonation. A dual method of assessment was adopted as
the British listeners had to not only listen and write down what they heard but they also had
to fill in an opinion assessment scale of how easy/difficult it was to understand the Nigerian
speaker. The results of the opinion assessment scale were found to correlate with the
intelligibility scores as the most intelligible Nigerian were found to be the most easily
understood while the least intelligible Nigerians were the most difficult to understand. The
conclusion was stated as follows: “the listeners‟ subjective opinions of the speakers‟
intelligibility correlated closely with the scores obtained on Test 1. Averaged over 10
listeners, subjective grading of speakers is a satisfactory form of assessment” (187). As for
the intelligibility scores of the objective method of assessment, the average intelligibility of
the Nigerian undergraduates to British listeners is found to be 64.4% with a range of 29.9%
and 92.7%. A summary of the intelligibility scores for the various tests is presented below.
Mean scores (%)
Connected speech 64.4
Reading 65.5
Phonemes 76.4
Stress 40.4
Intonation 60.1
RP speaker 99.4%
Most intelligible Nigerian speaker 92.7% (Hausa speaker)
Mean of Nigerian speakers 64.4%
Least intelligible Nigerian speaker 29.9% (Yoruba speaker)
51
From the intelligibility scores, it was concluded that stress is the most crucial speech
component affecting intelligibility between Nigerian speakers and British listeners as stated
below: “faulty stress and rhythm was the most frequent cause of intelligibility failure”
(187). This conclusion was also supported by the scores of the connected speech test which
was the test assumed to be the most fundamental for intelligibility testing. Other features of
the Nigerian English accent were arranged in order of the impact they had on intelligibility.
This is presented as follows: incorrect word stress, incorrect phrasing, mispronunciation of
vowels, mispronunciation of diphthongs, mispronunciation of consonants,
mispronunciation of consonant clusters, incorrect elision, incorrect assimilation,
metathesis, incorrect and unfamiliar lexis and unusual syntax. Other conclusions reached in
the study include the finding that the Hausa speakers were more intelligible to the Britons
as the scores revealed that „10 of the 12 Hausa speakers scored above the mean, while 9 of
the 12 Yoruba speakers scored below the mean‟ (187). The final conclusion of this study
echoes Bansal‟s conclusions on Indian speakers of English that education correlates highly
with proficiency in English as Tiffen infers that:
[T]he above average speakers had better qualification at Credit
and Distinction level in Oral English than the below average
speakers. More of the above average speakers also had
qualifications in English at HSC level…the above average
speakers began English later and used it as a medium later than
the below average speakers at primary school, but at secondary
52
school they had more and earlier contact with native English
teachers (187).
This implies that the Nigerian speakers whose speech was closest to that of the native
speaker were the most intelligible. It is note worthy that the same conclusion had also being
reached by Bansal, thus establishing that education correlates very highly with
intelligibility.
This study is significant because it makes useful contributions in several ways including:
providing detailed characterization of the Nigerian accent of English, identification of the
features that are most crucial for the achievement of intelligibility, assisting to provide a
better framework for the teaching of spoken English. However, certain weaknesses put
limitations on the generalizations of the findings: perhaps the most important being that
Igbo speakers (from Southeastern Nigeria were excluded from the study. The Igbos of
Eastern Nigeria represent a significant population, being the third major ethnic/language
group in Nigeria apart from the Yorubas and the Hausas. Their exclusion is significant as it
may have influenced the findings of the study in terms of the conclusions reached on
positions as regards the most intelligible language group. Besides this, while the Nigerian
listeners were a homogeneous group in terms of educational level, there was no
corresponding homogeneity of educational level among the British listeners. The levels
varied from between primary education to post graduate training. This may have influenced
the listeners‟ listening proficiency. Another source of concern raised by the study is the
insistence on native speaker norms as the standard. Descriptions of the areas in which the
53
Nigerians deviated from native speaker speech norms were labelled as „errors‟,
„mispronunciations‟, „misinterpretations‟ and „incorrect‟. This seems to imply an
expectation of rigid compliance with RP standards, which is not only unattainable but also
undesirable given the existence of endonormative standards which are nationally
intelligible. Strict compliance with native speaker speech norms is even more undesirable
given the evidence from various empirical sources that native speaker accents are not
always highly intelligible.
A number of significant studies have reached the conclusion that native speaker accents are
not always among the most intelligible. Among such studies include those conducted by
Smith and Rafiqzad on speakers of Standard American English, Major et al on Standard
American English, Smith on both RP and Standard American English, Munro and Derwing
on RP, Deterding and Gupta on RP and lastly Atechi on both Standard American English
and RP. These studies are examined below in relation to their relevance as to the provision
of empirical evidence towards the establishment of the fact that native English phonology
is not inherently intelligible.
The study conducted by Smith and Rafiqzad is perhaps the foremost pioneering effort on
the testing of native English accents to non-native speakers. The opening statements
demonstrated that the focus of the study was clearly different from previous studies on
intelligibility which are informed by the ideology of the inherent intelligibility of native
speaker phonology. The aim of the study was therefore to discover “whether or not these
propositions are well founded” (371). To achieve this aim, 1386 Asians and Americans
54
were involved in the study which spanned 11 countries and involved nationals of Hong
Kong, India, Japan, Korea, Malaysia, Nepal, The Phillipines, Sri Lanka, USA, Bangladesh,
China, Hong Kong, Indonesia and Thailand. Collaborators were employed from these
countries to choose educated speakers of English between the ages of 20 and 40. Other
criteria used for the selection are that none: (1) had spent more than four consecutive
months in any English speaking country; (2) had been formally educated in schools
directed by native speakers of English; or (3) had ever lived with English speaking families
or groups. Each speaker was taped while giving a ten minute speech to a group of 10
educated fellow countrymen. A cloze procedure test of the speech was then prepared and
every sixth word in the passage was deleted. This passage was then presented to the
listeners along with a questionnaire to test listening comprehension.
Based on the combination of the objective approach which consisted of a cloze-procedure
test and the subjective approach which consisted of a listening comprehension
questionnaire, the results are as follows:
[B]ased on the intelligibility averages in eleven countries, the ordering
would be as follows: Sri Lanka 79%, India 78%, Japan 75%, Malaysia 73%,
Nepal 72%, Korea 68%, Phillippines 61%, United States 55%, Hong Kong
44%. (376):
55
The comprehension questionnaire also yielded the following results:
Speakers Average
Hong Kong: 57%
India 76%
Japan 75%
Korea 61%
Malaysia 76%
Nepal 67%
The Philippines 69%
Sri Lanka 76%
U.S.A 64%
The conclusions of the study seemed to disprove the proposition that native speakers are
always easily understood as the evidence shows that the native speaker (American) was
always among the least understood. This conclusion is further reinforced by the rank order
correlation of the scores which is presented below:
Difficulty Intelligibility Understanding
(Least to most) (Most to least) (Most to least)
India Sri Lanka India
Japan India Sri Lanka
Malaysia Japan Malaysia
Sri Lanka Malaysia Japan
Philippines Nepal Philippines
56
Nepal Korea Nepal
United States Philippines United States
Korea United States Korea
Hong Kong Hong Kong Hong Kong
(378)
Even considering a prominent methodological shortcoming which was the differing
difficulty levels of the cloze procedure passages used in this research which as admitted by
the researchers “influenced the degree of intelligibility and understanding of the speakers”
(376), it may still be concluded that the native speaker is not always found to be highly
intelligible. This is clearly stated in the final paragraphs: “since native speaker phonology
doesn‟t appear to be more intelligible than non-native phonology, there seems
to be no reason to insist that the performance target in the English classroom be a native
speaker” (380).
Major, Fitzmaurice, Bunta and Balasubramanian also attempted to measure the
intelligibility of a native speaker accent to non-native listeners. According to them, the
research was necessary because „it is widely believed that listeners understand some
dialects more easily than others, although there is very little research that has rigorously
measured the effects‟ (37). The research participants consisted of 180 potential takers of
TOEFL (Test of English as a Foreign Language) who were all enrolled at either pre-degree
courses or were undergraduates at various American colleges and universities. They
comprised Chinese, Japanese and Spanish second language speakers of English as listeners
and 60 undergraduate and postgraduate native speakers of Standard American English from
57
U.S universities as speakers. The objective system of assessment was employed through a
test based on recordings of 12 short lectures. The participants were questioned based on the
recordings. For adequate comparison, other speakers were included in the test. They
include speakers of Southern American English, African American English (AAVE),
Australian English and Indian English. After the recordings were made, they were edited
and test tapes were constructed from the initial studio recordings. Multiple-choice answers
followed the test items. Statistical analysis showed that „both native and non-native
listeners are affected by a speaker‟s dialect--- and for non-native listeners, there were no
significant differences between Standard American and South American speakers‟ (58).
The conclusions further buttressed the point that the speakers of Standard American
English were not significantly more intelligible than speakers of other varieties because it is
maintained that:
[T]he results do not…support…that native-English speaking
listeners and ESL listeners would perform better on listening
comprehension tests in English based on lectures delivered by native
speakers of Standard American English (58).
Despite the fact that the focus of this study was listening comprehension rather than
phonological intelligibility, the results are relevant to intelligibility testing as
comprehension is related to understanding. However, the study suffered from a limitation
which is common to almost all listening comprehension test in that there was no control for
background knowledge. Some listeners may have had undue advantage over other listeners
if they had previous background knowledge of the topic.
58
Smith examined the intelligibility of both RP and Standard American English to second
language speakers of English. Among the other English varieties which were tested were
educated accents from China, India, Indonesia, Japan, Papua New Guinea, the Phillipines
and Taiwan. The major aim of the study was to help determine: „what differences, if any,
there are in the intelligibility…of selected taped material of nine national varieties. There
were 29 undergraduate participants and they were all balanced for age and sex. The test
material consisted of a recoding on forms of address used in each speaker‟s country to
address outsiders in English. The difficulty level of the test materials was controlled and
judged to be about the same level through the following means:
[B]oth speaker and respondent were fully proficient in English and
believed themselves to be educated speakers of their national variety
of English, each person spoke clearly and the number of embedded
sentences and the speed of delivery were approximately the same.
Thereafter, a cloze procedure of the passage was constructed with every 7 th word deleted
from the passage, leaving ten blanks which the listeners had to fill in. The passage was then
presented to the listeners to test intelligibility (word/utterance recognition). In addition, a
subjective questionnaire was also presented to the listeners. The listeners were to indicate
how easy or difficult it was to understand the speaker.
59
The important findings of this study include the following results: (a) the native speakers
were not as able to correctly identify their fellow native speakers as speaking standard
varieties of English an the non-native speakers were; (b) the non-native speakers was more
critical and seemed to have stricter criterion for identification of standard English, many
thought that the speaker of Standard American English was speaking non-standard English;
and lastly, (c) the results easily support the interpretation that it is possible for Standard
English to be spoken with many different accents. But what was perhaps the most striking
results of the investigation is the conclusion that „native speakers (from Britain and the
United States) were not found to be the most easily understood---thus, being a native
speaker does not seem to be as important as being fluent in English‟ (441).
A likely limitation of this study is the reliability of the findings of a study based on just ten
items. A ten-item test is perhaps not sufficient to measure the intelligibility level of an
accent. Perhaps a more extensive test would have been more adequate to achieve the aims
of the study. Nevertheless, the findings do seem to strikingly support the view that native
speaker accents is not inherently intelligible.
A similar study which is similar in orientation was that which was conducted by Atechi on
the intelligibility of two native speaker varieties to Cameroonians. This study is significant
as it is the perhaps the first to consider the intelligibility of a native speaker accent to
African non-native speakers. Thus, besides attempting to disprove the notion that native
speaker accents are inherently intelligible, the researcher also set out to:
60
[D]ismiss the trepidation nursed by scholars that the emergence of non-
native varieties around the world would cause English to disintegrate into
mutually intelligible varieties…and to contribute to the debate on the level
of phonological analysis that is suggested to pose the greatest threat to
intelligibility between native and non-native speakers (10).
The selected Cameroonians were all undergraduates while the native speakers of English
included both students and non-students. The age range of the Cameroonians was between
18 and 28 years while that of the native speakers extended to 60 year olds. Five types of
tests were designed to measure intelligibility in the study. They are as follows: connected
speech, passage reading, sentences with embedded phonemic contrasts, nucleus placement
in words and nucleus placement in sentences. The objective mode of assessment which
involves attaching a score to the items correctly identified by the listeners was adopted.
Based on this, the result for the American speakers was: „there is a wide range in the
intelligibility of native speakers of English…intelligibility scores in this column vary from
86.4% to 26.6%, with a mean score of 56.3%‟ (125); and for the Britons „the range (of
intelligibility) is fairly wide as the scores vary from 86.4% to 38.4%, with average of
58.7%‟ (127).
Atechi‟s study was a two-way investigation in the sense that it also considered the
intelligibility of speakers of Cameroonian English to Americans and Britons. A comparison
of the intelligibility scores obtained by all three groups of Cameroonians, Americans and
Britons is presented below:
61
Speakers and Listeners Intelligibility scores in %
Cameroonian speakers to British listeners 62.9
Cameroonian speakers to American listeners 59.7
British speakers and Cameroonian listeners 58.7
American speakers and Cameroonian listeners 53.9
An important inference can be drawn from these figures: the non-native accent was more
intelligible than the native speaker accents. However, perhaps these results should be
interpreted with a little caution. Several limitations are observable in the study as regards
the following: First, the wide margin between the speakers‟ ages and the listeners‟ ages (18
to 28 for the Cameroonians and 18 to 60 for the Americans and Britons); second, the
number of listener participants in the study is rather low – 40 – as it has been suggested in
the literature that a relatively large number of listeners is required for any useful
conclusions; third is the possibility that the Americans and Britons may have had their
speech influenced as they were resident in Germany; and lastly, there was no clarification
provided as to the specific British or American accent used by the native speakers. These
factors make it necessary to approach these conclusions with some caution. Nevertheless,
the conclusion reached is that: „this clearly shows that the native speaker of English is not
necessarily the most intelligible English speaker nor is… the native variety the most
intelligible‟ (130).
62
Another significant achievement of the research is the identification of the features of
American and British native speaker accents which affected intelligibility to the
Cameroonians. The study revealed that „the suprasegmental level was the most problematic
area of native English speech for the Cameroonian listeners as it led to the highest number
of intelligibility failures (46.5%) while the segmental level (40.2%), phonotactic
differences (11.8%) and lexical differences (01.5%) were less problematic‟. These
conclusions are in line with previous findings on the intelligibility of native speakers of
English, however it does seem that the number of listeners involved in the study was rather
low compared to the usual number of listeners in intelligibility studies as it has been
suggested that a relatively high number of listeners is necessary to provide any meaningful
conclusions.
To conclude this section of our review, it is pertinent to state that the current focus of
accent intelligibility studies seems to be the intelligibility of native speakers of English
(usually speakers of either RP or Standard General American) to non-native speakers. This
current re-orientation is not only crucial but timely because as argued by Rajadurai, “there
is need for a reconsideration… as empirical research has raised doubt about the
intelligibility of Inner Circle speech worldwide… the Inner Circle speaker was always
found to be among the least intelligible, showing that L1 phonology is not always
inherently most intelligible” (95). In the light of these intelligibility assessments, one may
consider the nature of RP and how Nigerians who are L2 speakers of English perceive the
accent in terms of intelligibility. These considerations as well as other related issues are the
problems which we shall attempt to resolve through this research. In order to resolve these
63
problems, a comprehensive description of both RP and NEA is necessary in order to fully
appreciate the degree of variation between the NEA and RP. Therefore, a discussion of the
NEA is attempted in the following section.
2.3 THE NIGERIAN ACCENT OF ENGLISH (NEA)
It has been established earlier that human language is subject to change. However, several
other factors apart from the natural processes of language change have contributed to the
evolvement of the Nigerian accent of English. Notable scholars such as Awonusi have
argued that these factors include:
mother-tongue interference, the history of the introduction of
English, the tradition of teaching English in Nigeria, the influence
of orthography or spelling, articulatory settings, social-cultural
values and incipient foreign values (207).
Interference is a phenomenon which characterizes multilingual societies such as Nigeria.
An estimated 505 languages are spoken in Nigeria (Grimes and Grimes 452). Of the
indigenous Nigerian languages, Igbo (spoken in the South-East), Yoruba (spoken in the
South-West) and Hausa (spoken in the North) are the major languages. Many Nigerians
are bilingual or multilingual with a command of several Nigerian languages and non-
indigenous languages such as English, Arabic and French. In view of this, Awonusi has
identified that interference which is characterized by the adaptation of the system of a
second language to that of a mother tongue is therefore observable in the speech of many
64
Nigerians, with an illustration of the dental and interdental fricatives providing an
illustration of this phenomenon. The absence of these sounds in the phonological systems
of the three major Nigerian examples is usually adjudged to be responsible for the inability
of some Nigerians to articulate these sounds accurately. British trading missions of the
sixteenth century which later developed into colonial rule also influenced the evolution of
the Nigerian accent of English. The majority of the British colonial officers in Nigeria
came from the upper or middle classes of British society, speaking RP. The initial contact
with RP developed into greater dominance of the accent with the establishment of
missionary schools where English was taught by native speakers of RP. Since then, the
teaching standard has been RP. However, other influences such as a language attitude
which seems to favour a „local‟ accent rather than a native-like accent has also contributed
to the evolution of this accent.
As a result of these influences, there is a diversity of spoken English in Nigeria. A further
complication is the lack of rigorous attempts at codifying these varieties. As of now, no
uniform and universally accepted description of a standard Nigerian spoken English exists
(Gut 818). What exists are some experimental and instrumental studies and a large body of
impressionistic studies. Due to the lack of quantitative data and the great variability of
Nigerian spoken English, most descriptions of spoken English in Nigeria are based on the
phonological tendencies of three major varieties. These major varieties correspond to the
three major Nigerian languages of Hausa, Igbo and Yoruba, thus leading to spoken English
varieties such as the Hausa accent of English, Igbo accent of English, and the Yoruba
accent of English.
65
The descriptions of NEA that have been undertaken have employed the criterion of
educational background for the description and classification of English accents in Nigeria.
Many studies which involve the correlation of speaker competence and educational
background include Brosnahan, Adetugbo, Jibril, Banjo and Udofot. A discussion of the
classifications provided by Banjo and Udofot is presented below as the two views seem to
represent the linguistic realities in Nigeria. Banjo proposed four varieties with distinct
linguistic features:
- Variety I is used by those Nigerians who picked up English as a result of the requirements
of their occupation. They are possibly semi-literate people with only elementary school
education. It is characterized by a high transfer-rate of phonological features from the
mother tongue and is unacceptable even nationally.
- Variety II speakers are likely to have had at least primary school education. It features
some transfer from the mother tongue and does not make „vital phonemic distinctions‟.
This variety of English is accepted and understood both nationally and internationally.
- Variety III is associated with university education and is recommended and is
recommended as the model for Nigerian Standard English. It is most widely accepted in
Nigeria.
-Variety IV is equal to British English and is less accepted in Nigeria than Variety III as it
sounds affected.
Udofot claims that Banjo‟s Variety IV is not a variety of Nigerian English and that spoken
Nigerian English can be divided into at least three sub-varieties. These sub-varieties
66
collectively show phonological differences from British Standard English in both
segmental and prosodic terms, and in many cases, the speaker‟s education is correlated
with the degree of proficiency. Three varieties are proposed by Udofot. They are Non-
Standard Variety, Standard Variety and Sophisticated variety. They seem to correspond to
Banjo‟s Variety I, II and III respectively.
In order to present a comprehensive description of the varieties of spoken English in
Nigeria, our review attempts a global view of the NEA through reference to the general
English accent used in West Africa. Certain trends have been observed in both the NEA
and the West African accent of English. These trends include the high correlation of
variables such as education and competence for the classification and description of both
accents of English. More importantly, many areas of convergence have also been identified
between all West African varieties of English including the NEA. Areas of convergence
have been identified by researchers such as Bobda, Gorlach and Kachru. Convergence has
been linked to one common influence which is colonization. In recognition of this,
Omoniyi has argued that „no doubt…(this) may be a consequence of the same general
political experience, i.e. colonization‟ (172). Therefore, our review includes references to
areas of convergence between the three main accents of English in Nigeria and the broader
West African accent of English. This is done in order to provide a more globalized
description of these Nigerian accents of English. Notions often associated with issues of
accent convergence which are employed here include: common core or nucleus and
periphery. Common-core or nucleus has been described by Nelson as „the set of features
and characteristics which all varieties have in common‟ (738). „Periphery‟, is a related
67
concept used to refer to „those features which are unique to individual varieties, and are not
shared with any variety‟ (738). Therefore, this aspect of the review will consider certain
features which form the common core features of West African Englishes and also include
the periphery features – that is, features which are specific to the NEA. Another
consideration which informs this review is that as Omoniyi rightly observes, the literature
on second language English varieties is mostly written from two main viewpoints. The first
is those written from a micro-analytical framework which “seeks to establish and describe
the nature of deviation or difference from default native–speaker Englishes”, while the
second involves a macro-analytical framework that “explores contact situations for the
sociopolitical relationships they promote and the impact of these on English, indigenous
languages, and societies as language users” (175). It is added that a major consequence of
writing from these orientations is that “while the former retains the trinity of ENL, ESL and
EFL, the latter dissolves it”. It is pertinent that we state this distinction because it divides
the literature in this section along these lines. While some of the descriptions present the
phonology of West African Englishes as mere deviations from native speaker norms (RP),
others examine them as legitimate varieties without reference to native speaker varieties.
The intention is to present some sort of pronunciation atlas for Nigeria and the section
draws mainly from the work of Gut, Awonusi, Udofot and Mesthrie. These studies are
some of the most current descriptions in this area and they also provide what we consider
to be adequate descriptions of the NEA. It must also be stated that the entire description
presented below is guided by Schneider‟s general comment about the difficulty of
cataloguing pronunciations of English:
68
[T]he amount of detail of sound realizations - idiosyncratic, phonologically
conditioned or not, socially or regionally motivated – is extremely difficult
to grasp and categorize...difference in levels of details...ranges from minute
phonetic analyses with lots of diacritics to essentially broad phonemic
categorizations (1111).
Besides this, other principles which guide the review are as follows:
(a) a certain degree of intra-speaker, inter-speaker and stylistic variation characterizes
spoken language and the observations here are no exception. Therefore, it may not
be right to make generalizations about all the features at all times.
(b) Well‟s standard lexical sets are used for easier differentiation between phonemes.
When different accents are being compared, this system is a better alternative to
using minimal pairs. The key word for each standard lexical set appears
conventionally in capital letters, and is shorthand for a whole list of other words
sharing the same vowel.
(c) Reference words are underlined, and;
(d) iNEA, yNEA and hNEA stand for Igbo accent of English, Yoruba accent of
English and Hausa accent of English respectively.
2.3.1 Short Vowels
West-African varieties of English have either a 5- or 6-vowel system for the short
monophthongs. Two other features which characterize the vowel system is (a) schwa is
marginal and (b) length distinction between vowels is not a general feature. The educated
69
accent of Nigerian English (henceforth ENEA) has a 6-vowel system comprising the KIT,
DRESS, TRAP, LOT, STRUT and FOOT vowels.
KIT
This is the front, high, spread vowel. Generally in West Africa, the realization is , a
tense, long variety. The same realization occurs among Yoruba speakers of English in
Nigeria (henceforth yNEA), and Igbo speakers of English in Nigeria (henceforth iNEA).
This lengthening leads to a lack of distinction between word pairs such as sit and seat.
Some Igbo speakers also realize this vowel as a pharyngealized , while the realization
among the Hausa speakers is a close approximation of the RP realization .
DRESS
The front, mid, half-open vowel has two variants within RP. While is found among
younger RP speakers, is more common with older speakers. The same realization is
the variant commonly found in West Africa while free variation between and occurs
in yNEA and iNEA. The Hausa speakers of English in Nigeria (henceforth hNEA)
have a different realization of this vowel, oscillating between the schwa, and .
TRAP
This is the central/back, open, low vowel. Two variants of this vowel occur in RP. in
Modern RP and in Traditional RP. There is uniform realization throughout the varieties
of spoken Nigerian English with being the realization of the vowel. However, there is
usually a raising to .
70
LOT
The back, low vowel is realized as in both traditional and modern RP. In West Africa,
this vowel is variously realized as , or but it is realized as in both yNEA
and iNEA. In hNEA, it is .
STRUT
This vowel is fairly central, neutral and half-open. There is usually a lowering or backing
of the vowel in West Africa. This leads to realizations such as and . is the
variant found in yNEA and iNEA, while , an allophone close to RP or sometimes
also is found in hNEA.
FOOT
This is the high back, half-close vowel. Variants in West Africa include a weakly rounded
or well- rounded or a short . In yNEA and iNEA, the distinction between full and
fool is neutralized due to the realization of this vowel as . In addition to this, a
pharyngealised variant of , which is is sometimes also found in iNEA. A very
similar realization to RP is the variant in hNEA.
2.3.2: Long Vowels
A 5-vowel system is usual for West African Englishes for long vowels (Mesthrie 2004).
This system includes,, or the lexical distribution sets of FLEECE, GOOSE,
THOUGHT, NURSE and BATH. It excludes the RP lexical sets of CLOTH and PALM.
71
Besides this, the set of BATH is most likely the most distinguishable feature of traditional
and modern RP. While traditional RP consistently has , modern RP realizations are
and . With being more prevalent.
BATH ~
In West Africa, this vowel has a front realization which is . Other less common variants
are or . In both yNEA and iNEA, the realization is such that word
pairs such as march and match are homophones. For hNEA, the realization is .
CLOTH
The realization for this vowel in modern RP is while for traditional RP, it is between
~. The variant for yNEA, iNEA and hNEA is while hNEA also includes the
variant .
NURSE
There is immense variation in the realization of this vowel with six variants in West Africa
alone. The vowel is in Ghanaian English, in Northern Nigerian English, in
Southern Nigerian English and also ,, and . The variant for yNEA is depending
on the spelling, in work, in girl, in dirty, in perch, in sir. The realizations
are the same for iNEA but hNEA has only the vowel for the set of NURSE.
72
FLEECE
As with most long vowels, shortening occurs in West African varieties. The realizations in
West Africa are or . For both yNEA and iNEA, the realization is , while it is
for hNEA.
PALM
A front variety of this vowel occurs in West Africa. This variant is also the realization
in yNEA, iNEA, and hNEA. The implication is that word pairs such as march and match
are homophones.
THOUGHT
Short forms of this vowel characterize West African varieties. They are variants such as
, and unlengthened or . In yNEA, the realization is so that caught and
cut become homophones. The same realization occurs in iNEA but in hNEA, the
realization is , a closed variant of the RP vowel.
2.3.3: Diphthongs
FACE
The variants of this vowel found in West Africa include or . is the variant
most commonly found in yNEA, in iNEA and so that the distinction between let and
late is neutralized.
73
PRICE
Realizations such as are common in West Africa but is also an alternative form.
The realization in yNEA and iNEA is but in hNEA, the first element may be
centralized. When this occurs, the onset may be raised to or .
CHOICE
This vowel shows relatively little variation. Realizations in West Africa include and
. is the variant found in yNEA, iNEA and hNEA.
MOUTH
This vowel also shows slight variation as it is throughout West Africa. In yNEA and
iNEA, it is also but the realization differs slightly in hNEA as the first element may be
centralized to .
GOAT
This vowel is usually monophthongized into throughout West Africa. The realizations
in yNEA and iNEA vary from to while it is in hNEA.
SQUARE
Immense variation occurs with this accent with realizations including half-open and half-
close qualities such as ,,,, in West Africa. The general realization in the
74
Nigerian accent of English is a diphthongal realization with some speakers realizing it
as .
NEAR
Diphthongal realizations occur throughout West Africa as and . In yNEA,
iNEA and hNEA, the realizations are or with an epenthetic as .
START
In yNEA, iNEA and hNEA the realization for this vowel is .
NORTH
In the accents of English in Nigeria, this vowel is usually realized as .
FORCE ~
The realization in yNEA, iNEA is , while in hNEA, the vowel is realized as ,
or .
CURE
There is a great deal of variation found in West African accents of English with this vowel.
Variants include , , and the monophthongal realization .
75
2.3.4: Other Vowels
happY
This vowel is realized as and . Generally in Nigerian accents of English the vowel
is realized as . It is also realized sometimes as and sometimes as in words such as
greeted because of the influence of spelling pronunciation
lettER
The West African variants for this vowel are and . It is realized as in yNEA,
iNEA and hNEA.
horsES
Variants in Nigerian English are and .
commA
The realizations of this vowel in West Africa are and . , one of the West African
variants is also the realization in yNEA, iNEA and hNEA.
2.3.5: Triphthongs
Awonusi points out that “NEA has virtually no triphthongs” (220). Instead of triphthongs,
glide formation processes apply in the accents of Nigerian English. This process changes
the middle vowels into the corresponding semi-vowels as in:
76
fire
lion
power
our
2.3.6: Consonants in NEA
The description of NEA consonants will be brief as consonants are not usually crucial in
issues of accent variation. General commentaries about consonant sounds in NEA are as
follows:
NEA is non-rhotic. Both yNEA and hNEA speakers use the tap in all positions while some
iNEA speakers use the roll. Besides this, while all NEA accents use the linking , the
sandhi or intrusive is “practically non-existent.” (Awonusi 244). With regards to the
nasals, only two of them , are phonologically significant for NEA. exists in the
speech of very few speakers. The same occurs for the postalveolar fricative .
Another general feature of the NEA is the spelling-induced pronunciation of words. This
occurs generally as in debt and bomb pronounced with a terminal , loan words such as
elite , plateau , and the pattern which Awonusi (op.cit) described as h-
restoration which is the use of in words which are less in RP e.g. heir and hour.
Another spelling-induced feature is the voicing of certain non-voiced consonants such as
increased - because the orthography suggests it. Voicing of in maximum
and laxity is also reported by Gut (821).
77
Several consonant cluster simplification strategies occur in NEA but the most significant
for the educated speakers is the insertion of epenthetic vowel or between word- final
syllabic consonants and the preceding consonant as e.g. in bottle t.
Other phonological processes in NEA include metathesis, e.g. the pronunciation for
ask.
2.3.7: Sentence Stress
Many scholars have made commentaries on the nature of Nigerian English prosody. They
include Jubril, Jowitt , Udofot, Eka , and Simo Bobda. A summary of their observations
will be attempted here as there are no systematic studies available.
Five features which are general to sentence stress in the Nigerian accent of English are as
follows:
(a) Sentence stress is rarely used for emphasis or contrast. Instead, extra words are
introduced for emphasis. For instance, instead of saying “Mary did it” Nigerians
tend to say “It was Mary who did it”.(Gut 826)
(b) Given information is hardly de-accented.
(c) An overall preference for “end-stress”, i.e. the placement of the nucleus, on the
most prominent accent.
(d) A large number of extra accented syllables occur in which virtually all verbs,
adjectives and nouns are accented.
(e) Accent, which is defined as “the phonetic realization of stress in speech” (Gut
890), is primarily realized by tone with lexical words receiving high tone on the
78
stressed syllable and non-lexical words receiving low tone. The stressed syllable
of lexical words is produced with a high tone, which then spreads to the end of
the word. Any unstressed syllables preceding the stressed syllable are produced
with a low tone.
2.3.8: Rhythm
The rhythm of the Nigerian English accent has been described as “syllable-timed”
(Adetugbo 131), “inelastic” (Eka 9) and “full-vowel timed” (Udofot 20). These labels are
pointers to the roughly equal intervals between syllables in Nigerian English. This has also
been confirmed by Akinjobi and also by Gut. Suggestions that Nigerian English is syllable-
timed follows the traditional division of the languages of the world as either syllable-timed
or stress-timed. Speech rhythm, according to Gut is the “periodic recurrence of events such
as syllables in .... syllable-timed languages and the feet in…..stress-timed languages. (cf
Gut 818). Therefore, the presence of full and strong vowels rather than reduced vowels in
unstressed syllables gives the Nigerian accent its characteristic timing which differs from
the stress-timed rhythm of Standard British English.
An alternative to the traditional view of speech rhythm as discrete units has been proposed
by Dauer . This view is based on the assumption that the rhythm of any language is a result
of phonological, phonetic, lexical and syntactic facts about the language. In Dauer‟s view, a
composite of these features will locate any language along a rhythm continuum instead of
in a distinct class of syllable-timed vs. stress-timed. This approach provides a better
explanation because it offers a fluid classification for all world languages. The approach
79
also applies to the NEA as it supports the views expressed by Eka that NEA should not be
categorized as strictly syllable-timed, but should be described as a language variety with “a
tendency towards syllable-timing” (10).
Speech rhythm is central to this study because it is one of the features in which marked
accent variation occurs between the two accents in focus in the study and is therefore a
crucial determinant of speech intelligibility between Nigerians and the speakers of RP.
2.3.9: Intonation
This is the final segment in our consideration of the NEA. A striking difference exists
between NEA intonation tunes and RP tunes. These differences are noticeable in terms of
the patterns and range of the tunes. Adetugbo, Tiffen, Eka, Jowitt and Udofot and Ufomata
concur that the mastery of RP intonation patterns is a problematic area for Nigerians. To
overcome these limitations, Nigerians use “more words, longer utterances and gestures in
place of using tonic placement to agree, disagree or express doubt” (Ufomata7).
Nevertheless, several empirical studies have been undertaken to attempt a description of the
intonation patterns employed by Nigerians and two of them (Udofot 2002 & Jowitt 2000)
which seem to be the most comprehensive have similar conclusions. The prevalence of
falling tones over rising tones and unidirectional tones over more complex tones was
revealed in Udofot; while Jowitt considers Nigerian English intonation system to be
„simplified‟. The principal elements of Nigerian English intonation were listed as follows:
1. falls predominate in statements, wh-questions, and commands;
80
2. rises predominate in yes-no and tag questions;
3. complex tomes are seldom used, although there is some use of the fall-rise
in non-final subordinates and of the rise-fall for strong emphasis;
4. “end-stress” is a general rule for intonation groups;
5. where the “core” pattern includes a falling nucleus, it is a low-falling nucleus.
The picture that has been drawn here is that the NEA presents marked differences in
comparison with RP. However, the degree of variation between the two accents cannot be
fully appreciated until the description of RP is also presented. The following chapter
presents some of the major commentaries on the phonetic/phonological features of RP,
particularly areas which may impinge on intelligibility.
CHAPTER THREE: ASPECTS OF RECEIVED PRONUNCIATION
3.0 Introduction
81
In this chapter, an attempt will be made to establish the phonetic/phonological description
of our RP accent variety with a view to further delimiting the focus of the present study. In
our consideration of RP, the names with which this accent has been variously referred is
first examined. Several sociolinguistic themes which run through the major names of RP
are also discussed. This is subsequently followed by an assessment of sub-varieties of the
accent, while phonological features which constitute a “mainstream” accent of RP are
thereafter discussed.
3.1 Received Pronunciation
Standard British English accent is one of the names given to a particular accent of British
English. But what is probably the most common name and also the name which linguists
use to refer to this accent is Received Pronunciation. The earliest recorded use of this
term is by Ellis where it is described as follows:
in the present day we may…recognize a received pronunciation
all over the country, not widely differing in any locality, and
admitting a certain degree of variety. It may be considered as the
educated pronunciation of the metropolis, of the court, the pulpit,
and the bar (23).
„Received‟ is used by Ellis in the 19th century archaic sense of the word when it meant
„accepted in the most polite circles of society‟ (25). In the first half of the 20th century, the
term Received Pronunciation was again adopted by Daniel Jones, the first Professor of
Phonetics at University College, London. Although he initially used the term Public
82
School Pronunciation in 1916, in subsequent works he used the term „Received
Pronunciation‟ to refer to an accent which is “the everyday speech of the families of
Southern English persons whose men-folks have been educated at the great public boarding
schools”(45). Jones later expanded the users of this accent to include non-Southern English
people “but who have been educated at these (public) schools” (45). These descriptions
focus on the educational background of the speakers of RP of which Jones himself is a part
having attended a public school. Jones also based the description of RP in the editions of
his pronouncing dictionary (English Pronouncing Dictionary) on his own pronunciation.
Thus, it may be concluded that a public school education is a criteria for the identification
of the speakers of RP.
The association of this accent with public school education probably provides an
explanation as to why the accent is also referred to as Public School Pronunciation (PSP).
This term was first used by Daniel Jones in the first edition of English Pronouncing
Dictionary (EPD) in 1917. Windsor Lewis uses the term General British, while Wells and
Colson use Southern British Standard to also refer to the same accent. The term BBC
English is used by Roach and Hartman to refer to this accent probably because of its
familiarity as the accent used by announcers and newsreaders in the BBC (The British
Broadcasting Corporation). This company was founded in 1922 with a policy of employing
“men who in the presentation of programme items, the reading of news bulletins and so on,
can be relied upon to employ the correct pronunciation of the English tongue”
(Mugglestone 273). However, accent leveling has forced the BBC to reassess its policies
and practices and this has led to lesser linguistic hegemony, leading to comments by BBC
83
authourities in the early 1970s that “a far greater variety of presenters and attendant
linguistic patterns being available on-screen, with a stronger sense of different voices”.
(Mugglestone 280). This re-visioning makes has made the term BBC English somewhat
misleading. However, the most widely known term for this accent remains Received
Pronunciation, as Wells also notes that:
the accent which enjoys the highest overt prestige in England is
known to phoneticians as Received Pronunciation (for short, RP).
This name is less than happy, relying as it does on an outmoded
meaning of received („generally accepted‟)…It is what English
people mean when they say that someone „hasn‟t got an accent‟
(though to Americans it is a typical British accent). (117)
A careful look at the naming practices associated with RP reveal certain recurrent themes.
The principal ones are non-localizability of the accent, upper-class society and
association with public school education. These themes are examined below:
3.1.2 RP and Non-localizability
A philologist, who was dubbed „father of modern phonetics‟ by the Philological Society,
Alexander J. Ellis through this description of RP as: „a received pronunciation all over the
country, not widely differing in any locality, and admitting a certain degree of
variety…may be considered as the educated pronunciation of the metropolis, of the court,
the pulpit, and the bar‟ (23: emphasis added), established the geographical neutrality of this
accent. It has often been said that RP is not the accent of any particular region because it
84
may be encountered as the native accent of people who come from all over Britain.
However, its origin is traceable to the south-east of England, most especially London and
its adjacent counties. This is an area that has for long been the bastion of wealth in
England. Nevertheless, its origin is merely traceable to this region as RP cannot be said to
be „synchronically southern‟ (Ramsaran 15). Certain key phonological characteristics of the
south-eastern accent such as t-glottalling in prevocalic environments and the use of [
rather than in the happY lexical set readily differentiate it from RP (Wells 303). Other
markers also distinguish RP from all other regional accents in England, thereby making it
geographically neutral. Hughes, Trudgill and Watt summarize the key phonological
characteristics of regional accents of English in the British Isles in a table showing the
salient features of each accent. The table is reproduced below to demonstrate the unique
differences between regional accents within the British Isles and RP.
in in in in in in in in in in mud path palm hazy bar pull harm sing few gate
Scotland & + - - - + - + - + -
N. Ireland
85
S. Ireland + + + + + + + - + -
Northeast - - + + - + + - + -
Central north - - + - - + - - + -
Central Lancs. - - + - + + - + + -
Merseyside - - + + - + - + + +
Humberside - - + + - + - - + -
NW. Midlands - - + - - + - + + +
E. Midlands - - + - - + - - - +
W. Midlands - - + + - + - + + +
S. Midlands + + + + - + - - - +
E. South-west + - - + + + - - + +
W. South-west + - - + + + - - + -
South-east + + + + - + - - + +
East Anglia + + + + - + + - - +
Wales + - + + - + - - + -
(Hughes, Trudgill & Watt 71)
Though RP has distinct features which differentiate it from all other regional accents in
Britain, care must be taken not to create the impression that each of these accents
represents discrete varieties. Rather than this, the varieties form a continuum, a gradual
changing of pronunciation which occurs from Scotland to the Southwest of England to
Wales. RP occupies one extreme end of the accent continuum due to its total lack of
regional features. At the other extreme are the broadest forms of regional accents which
86
contain localized features. Besides the continuum of regional variation, the association of
RP with the upper classes also implies that accents in Britain may be identified on a
continuum of social variation with RP again occupying one extreme while accents used by
the lower classes occupy the other extreme. The relationship between RP, regional
variation and social variation is often modelled as having the form of an equilateral triangle
or a pyramid as in the figure below:
social
variation
regional variation in pronunciation
(After Hughes, Trudgill & Watt 10)
The base of the triangle is broad, implying considerable amounts of phonological variation
between the different regional accents spoken by the lower classes. Going upwards from
the base, the increasing narrowness of the triangle implies decreasing regional variation
between the accents of speakers higher up on the social scale. Similarly, the point at the top
of the triangle indicates the total lack of regional variation which we have already noted is
characteristic of the RP accent, spoken as it is by people at the top of the social scale. There
is no doubt that this model is an effective one because it explicates a well known fact which
87
is that it is impossible to tell where an RP speaker comes from. Between the top of the
triangle and the base are speakers with different degrees of regionally marked speech; and
the higher the person is on the social scale, the less regionally marked will be his or her
accent. Therefore, it is usually possible to tell from which broad origin of the British Isles,
middle-class speakers come from. Working-class speakers can usually be pin-pointed even
more accurately to their geographical origins. At the apex of the triangle is RP whose
speakers are impossible to identify as coming from any particular region, hence the
description of the accent as non-localizable. The relationship between RP and regional
accents is described rather succinctly by Wells:
[T]he more localizable (and hence non-upper-class) characteristics an
accent has, the „broader‟ we say it is. A maximally broad accent reflects
(i) regionally, the highest degree of local distinctiveness, (ii) socially, the
lowest social class, and (iii) linguistically, the maximal degree of
difference from RP. An important and defining characteristic of RP is thus
its non-localizability within England. (14)
3.1.3 RP and the Upper-class society.
The class-specificity of RP is often linked to the high social status, substantial wealth and
political power of its speakers. It is not co-incidental that most RP stereotypes have upper-
class connotations. References to RP stereotypical speakers include: „dowager duchesses,
88
certain army officers, Noel Coward-type sophisticates, and popular images of elderly
Oxbridge dons‟ (Wells 133); and also „retired admirals and brigadiers; dukes and their
families‟ (Ramsaran 39). This stereotype is even more finely ingrained on the
consciousness when it is realized that all the twentieth century English Prime Ministers and
of course the Queen of England, Queen Elizabeth are all speakers of RP.
Therefore, the use or non-use of RP may be said to polarize the British society because the
accent acts as a segregating influence, dividing the population into those who speak RP and
those who do not. This division has been likened to a linguistic shibboleth (Judges 12: 5-
6), and an accent bar (Abercrombie 226). The story is told in the bible of how language was
used as a means of group identification. In the book of Judges, Jephthah had mobilized the
Gileadites army to attack the Ephraimites. However, many of the Ephraimites escaped but
the Gileadites waited for them at the Jordan River. As the Ephraimites attempted to cross
the river, they were asked if they were Ephraimites. In order to confirm their responses,
they were asked to pronounce the word shibboleth because the Gileadite guards knew that
the Ephraimites „could not frame to pronounce it right‟ (Judges 12:6). Instead, they
articulated sibboleth. In this way, the Gileadites were able to identify the Ephraimites and
they were subsequently executed.
This story has become rather symbolic of the way language acts as a marker of group
identity just as RP has become the linguistic symbol of upper class British society. But
instead of language being a marker of ethnic identity as it is in this biblical account, it is a
marker of class identity in England. An observation of the social significance of RP within
89
England informs this conclusion. The significance of RP within England seems to go
beyond a mere means of communication, hence, Waller‟s observation of RP serving as „an
instrument of communication and also of ex-communication‟ (16). This was stated in
reference to the manner in which RP speakers are accorded certain priviledges in the
society which non-RP speakers are not accorded. There is a lot of anecdotal evidence of
how RP has been used as an instrument for the ex-communication of non-RP speakers
through exemption from certain employment priviledges. However, the evidence remains
anecdotal because discrimination is not permitted in most modern societies including
England, so there is scant documentary evidence of this. But Trudgill has confirmed this
observation regarding RP that the accent „has been the necessary passport to certain kinds
of jobs such as BBC newscaster and presenter positions and telephone sales jobs‟ (267).
RP has also been likened to an „accent bar‟ by Abercrombie. This again refers to the way in
which language affects access to priviledges in the English society. This term was coined
on the model of the „colour bar‟ which exists in some societies. It is an analogy used to
describe the accent prejudice in England which favours speakers of RP who are „on the
right side of the bar‟ (49). One of the consequences of the accent bar was this story of a
young English girl reported in the press which was provided by Abercrombie. As the
consequence of a divorce, this girl had gone to live with her father, while her brother and
sister lived with the mother. The father sent the girl to an expensive school, the result being
that she became an RP speaker. Her brother and sisters continued to speak with their local
accents. A judge had to take a decision on the girl‟s future and the judge‟s decision was
that she should stay with her father because she now spoke RP and had moved into a
90
different social class „so as not to have to mix with speakers of a socially inferior accent‟
(Abercrombie 50).
Sociolinguistic interviews conducted by Fabricius also stress that accents other than RP are
often judged to be deficient and consequently put their speakers at a disadvantage. The
interviewee, a male undergraduate RP speaker claims that:
I think it would be …you‟d be hard pushed to say at the moment
that your accent doesn‟t make a difference to your chances of
getting into Cambridge for example, and I don‟t think at least I
would like to think that none of it‟s deliberate but I think if you
have you know if you have a very strong North Welsh accent for
example, it‟s or strong Scottish accents or something which seems
quite alien I don‟t think I think if you‟re having to concentrate that
much more on listening to it, I can well believe that you‟re not
consciously saying “dear oh dear they‟re not as good” but
somehow you feel as if you‟re not getting on with them as well
because you‟re having to put so much more effort in just talking to
them um and I don‟t I hope none of this is done deliberately but I
can see that if you‟re… as inevitably happens at your interview,
hundreds of people, you‟re relying to an extent on a gut feeling.
(40)
3.1.3: RP and Public School Education
91
The association of the RP accent with public school education started in the nineteenth
century with the establishment of the first public boarding schools in England for boys.
These schools catered for boys aged between eleven to eighteen years. While at these
schools, the boys were usually isolated from their surroundings and only met pupils of
other public schools when playing against them at various sports. This system of Victorian
education for boys has been thoroughly described by Mugglestone as „officially established
through the 1860 appointment of the Royal Commission as an endowed place of education
of old standing, to which sons of gentlemen resort in considerable numbers‟ (209). The first
set of schools included Eton, Westminster, Winchester , Harrow, Rugby, Charterhouse and
Shrewsbury. Of these, Winchester dates back to 1382, Eton to 1440, Westminster and
Shrewsbury to 1560, and Harrow to 1571. Although they were originally founded to absorb
intakes from the less-privileged members of the society, „the poor and needy of the local
area‟, this changed later on in the century. Initially, the masters were, in addition to private
(fee-paying) pupils also allowed to take a limited number the poorer children on
scholarship so that the patterns of intake included sons of earls, bishops, vicars, stewards,
blacksmiths and cooks. However, a gradual change began to take place as the rich and
affluent began to displace the „poor and needy‟ at a number of these schools (210). For
example, as noted by Mugglestone, out of 3000 pupils who entered Eton between 1755 and
1790, only thirty eight were the sons of tradesmen and by 1930, this had dropped to zero.
Even though some of the headmasters of these schools attempted to put a stop to the
discrimination in the admissions policies, this apartheid system of schooling became so
pervasive that the School Inquiry Commission set up in 1868 to conduct an investigation
92
into the problem. Mugglestone noted that one of the headmasters that were questioned
responded that:
[H]ad very numerous applications from „persons of good standing in
the world and good fortune‟, and that he had invariably been asked,
„What is the character, station and position of the home boarders?‟
When he answered that they were „of all classes down to the sons of
blacksmiths and washerwomen‟, the application had immediately
been withdrawn. Of these „home boarders‟‟‟, he stated that he had
succeeded in gaining them perfect fair play in school‟, but that he
had had to separate them out of school and that „mainly for the sake
of the village boys‟. He felt that if he allowed them to associate, „he
should have a constant fear of their being ill-treated. (223)
The reason for this behaviour was pointed out by the headmaster: “It is not the fault of the
boys, it is the fault of society…I never yet saw a man who would send his boy to a school
in order to associate with those lower than himself” (223). Some of the advantages of these
prestigious schools which were usually monopolized by the elite included the inculcation
of religious and moral principles, gentlemanly conduct and intellectual ability. Thus, the
products of these schools tended to be a blend of social elegance, refinement, wealth, good
manners, and perfect gentlemanly conduct.
However, education at such public schools was not only for the purpose of producing a
gentleman but was also for the purpose of remedying „linguistic provincialities…a means
93
of integration into the linguistic as well as social properties of an elite… to protect its
pupils against those habits of faulty pronunciation, together with those vulgar and offensive
tones in reading and speaking‟ (Mugglestone 229). It appeared that this aim was generally
well attained because in referring specifically to Eton, one of the first such schools, Wells
comments on the homogenization influence of these schools noting that “an Old Etonian
sounds much the same whether he grew up in Cornwall or Northumberland” (119).
Abercrombie identifies the linguistic influence in these schools as RP: „Just as RP is a
unique accent, a unique institution provides its basis. This institution is the English public
school‟ (49). The sustenance of RP in the public school system was no doubt ensured by
peer pressure because even though RP was never explicitly taught in these schools, the
students seemed to effortlessly acquire it from fellow pupils who came from RP-speaking
homes. The social and linguistic hegemony of these schools continued for much of the 19th
century as Mugglestone states: „though in previous decades it had been acceptable an even
advisable for a gentleman to be educated at home…by the end of the nineteenth century,
the dominance of the public school tradition was such that this was the archetypical
education for a gentleman, or for those who wished to be numbered amongst such‟ (217).
This is also reinforced by the conclusion that at the close of the nineteenth century:
it was… the homogenization of social environment amongst the pupils
at such schools which in itself prepared the ground for the
homogenization of accent later associated with them. They reinforced
notions of „proper‟ English through a sense of collective identity in
which this too operated as a sign of membership and
integration…Conceptions such as these tended to function as a self-
94
perpetuating paradigm, confirming the associations of the „best‟ accent
and the „best‟ education in schools of this order by means of the
emphasis given to the social composition (and social sensibilities) of
their intake (282).
In present day England, the public school institutions are now officially referred to as
independent schools and the focus remains high academic achievement as expressed in
Mugglestone:
UK independent schools achieve the very highest academic standards. Of
the 500 schools listed by The Times as achieving the highest GCSE re-
sults in 1999, about 380 were independent schools. About 80% of pupils at
independent schools (including special schools) gain five or more GCSE
passes at grades A-C (compared with a national average of 43%). Eighty
per cent of independent school A-level candidates gain three or more
passes, compared with a national average of 61%. Nine out of ten-A-level
leavers from independent schools go on to higher education. At the
primary level, most prep schools taking part in national curriculum testing
report attainment levels well above the national average. It is sometimes
claimed that this academic success is due to selective admissions policies.
Some independent schools do admit only children of the highest academic
ability; many, however, admit a much wider range of ability. Evidence
from the Department for Education and Employment (DfEE Statistical
95
Bulletin no 4/95) suggests that pupils at all levels of ability do better in
independent schools. Comparing A-level results from schools of all kinds,
it concluded that, at every level of ability, there was a clear tendency for
candidates in independent schools to achieve higher A/AS51 level scores
than those in maintained schools (251).
However, it appears that the same social and linguistic homogenization tendencies of the
public schools of the preceding centuries continue to characterize the independent schools.
The schools remain elitist charging fees well beyond the reach of the average parent. The
least expensive day schools tend to charge around £1000 per term, while the the most
exclusive boarding schools charge over £5000 per term, for full boarding. Fabricius
provides details of the linguistic influences of present day public schools through a series
of interviews held with former independent school students in 1997 and 1998. By then, the
interviewees were university undergraduates and graduates. Their ages ranged from 18 to
31 years. The interview protocol comprised questions which were primarily aimed at
gathering information on the social and linguistic influences that they experienced while
attending the independent school. For example, when asked if there were changes in the
way they spoke after they were admitted into an independent school, a male interviewee
responded that:
I think it happens entirely involuntarily um, so many times I‟ve rung
home and my sister had picked up on something I‟ve said and maybe
phrase I‟ve used or even just the way that I‟ve formed you know just
said the word and pointed out quite how you know public school it
96
sounds but when you‟re living amongst you know boys who are all
speaking the same effectively then then it doesn‟t it doesn‟t you don‟t
doesn‟t occur to you but it‟s sad because I mean in the same way as all
dialects in this country are all sort of gradually you know meandering
towards a form of London English, the boys at school who came from
Yorkshire (or) whatever were… people did pick up on their accents and
joke about them you know they all ganged up on them you know
because so many of us all came from um the south and all spoke the
same way it was you know kind of group mentality I mean not in a nasty
way of course but it was noticed and to notice it is as much to say that‟s
not the norm (54).
When also asked if they changed the way they spoke after having left the independent
school, a female speaker responded that:
[N]ot, after I left school yes, um because when I came out of
[public school] I I played it down basically I my… both my brother
and my sister um, it‟s something you notice more in boys or at
least I notice more in boys that they tend to have a more plummy
accent or that they‟re not, less aware of having it… when I came
out of school, I had a very very very plummy accent and I I still do
but not as much as I used to I‟ve played it down a lot. (55)
3.2: Sub-varieties of RP
97
Having considered the naming of RP, we turn to a consideration of sub-varieties within RP.
These include U-RP, Refined RP, Regional RP, Adoptive RP etcetera. There is the sub-
variety used by the royal family, the aristocrats, upper-class families, professions which
have traditionally recruited from upper-class families (i.e. the Navy) which Wells (279)
identifies as U-RP (meaning upper-crust RP) and Cruttenden (78) refers to as Refined RP.
Wells distinguishes this from other forms of RP, describing the stereotypes associated with
this variety as:
…special voice quality and manner of delivery. …the speech…of an
elderly Oxbridge don; and to that of the popular image of a jolly-
hockey-sticks schoolmistress at an expensive private girls‟ school
(280).
Similarly, Cruttenden‟s classification describes Refined RP as follows:
that type which is commonly considered to be upper-class…a
speaker of Refined RP has become a figure of fun and the type of
speech itself is often regarded as affected…Refined RP reflects a
class distinction (78).
The descriptions of this variety given by both Wells and Cruttenden coincide in terms of
being strictly upper-class and inciting a negative attitude.
Regional RP is a somewhat contradictory term which Cruttenden adopts to represent
varieties which unlike Refined RP and U-RP do not reflect class distinction but regional
differences. It is described as “the type of speech which is basically RP except for the
presence of a few regional characteristics which go unnoticed even by other speakers of
98
RP” (78). This variety supports the idea of an RP accent with regional features such as
Northern England RP and Cockney RP. The contradictory nature of such descriptions is
also noted by Cruttenden that “part of the definition of RP is that it should not tell you
where someone comes from” (78).
RP varieties which reflect other considerations rather than class or region are adoptive RP
and near-RP. Speakers of adoptive RP are “adults who did not speak RP as children”
(Wells 281). Such adults are motivated to learn RP as a result of social pressure from RP
speaking circle of friends or office colleagues. They may be regarded as „social climbers‟
who recognize the advantage that an RP accent may offer in terms of better job
opportunities etc. However, there is a distinction between such people and native speakers
of RP. This is that they “lack control over the informal and allegro characteristics of RP”
(284). Characteristic features of native RP informal speech include reduction phenomena
such as elision, assimilation, and smoothing. However, speakers of adoptive RP tend to
avoid such characteristics, consciously or unconsciously regarding such variants as „lazy or
slipshod‟ (284).Without this distinction, it would otherwise be difficult to draw a line
between adoptive RP and native RP.
Near-RP is another variety which is neither based on class or regional considerations.
Wells (297) describes this variety as:
any accent which, while not falling within the definition of RP,
nevertheless includes very little in the way of regionalisms which would
99
enable the provenance of the speaker to be localized within England (or
even as Australian, New Zealand or South African).
It seems that near-RP coincides with Cruttenden‟s Regional RP as Wells supports the
argument of the existence of not one but “several” varieties of near RP similar to
Cruttenden‟s view of the existence of “regional RPs in the plural” (78). Besides this,
several usages which are representative of varieties of near RP but also differ significantly
from RP based on the criteria of phonemic, phonotactic, phonetic and lexical distinctions
are provided by Wells. They include the absence of any of the phonemic distinctions found
in RP particularly between and ; and , rhoticity in non-prevocalic
environments, yod dropping, diphthong shifting, and the use of rather than in -ed and
-es suffixes.
However, it is note-worthy that despite the identification of these features, the
difficulties involved in treating accent varieties as monolithic entities are
recognized by Cruttenden:
General RP, Refined RP and Regional RP are not accents with
precise enumerable list of features but rather represent
clusterings of features…thus there are not categorial boundaries
between the three types of RP or between RP and regional
pronunciations; a speaker may, for example, generally be an RP
speaker but have one noticeable feature of Refined RP (79).
100
The same justification regarding the impossibility of making clear distinctions
between RP accent varieties is stated by Wells:
[I]t is open to question whether it is desirable or possible to draw
firm lines of demarcation between RP, Near RP, and non-RP.
The justification for my proposing the term „Near-RP‟ is that it
conveniently refers to a group of accent types which are clearly
„educated‟ and situated away from the lower end of the socio-
economic scale, while differing to some noticeable degree from
what we recognize as RP (301).
The difficulty involved in drawing lines of demarcation between accent varieties is also
revealed in the controversy involving Estuary English. While some writers have identified
it as a distinct accent variety, others simply describe it as RP with influences from Cockney
(Rosewarne,Coogle, & Haenni). Cruttenden also states that “Regional RP will vary
according to the region involved” and identifies a London influenced form of RP as
„Estuary English‟ (79). This name was coined by David Rosewarne and was first published
in his article in the Times Educational Supplement in 1984. This accent is described as a
“middle ground” between Cockney and RP which is characterized by a “mixture of non-
regional and local south-eastern pronunciation and intonation”. The name Estuary English
was used for this accent because of its link to the geographical spread of the accent as the
features associated with the accent are features which spread outwards from London along
the Thames Estuary into Essex and North Kent in England.
101
Initially used only among journalists, this name eventually became popular among
phoneticians who described it as a modification of RP towards Cockney. It is also similar
to RP in the sense that it is an upper/middle class accent which is adopted by “those
wishing to avoid the stigma of RP as „posh‟ and by upwardly mobile speakers of local
dialects” (79)who usually speak RP. As Mugglestone concurs,
this …(accent) is the mode of speech which could be adopted by
Diana as the „People‟s Princess…and similarly deployed by
Tony Blair as a representative of Labour as the „People‟s Party‟,
an „everyman‟ who speaks to all (286).
Furthermore, Estuary English has been proclaimed by David Rosewarne in his newspaper
articles of the mid 1980s as „the new standard English‟, „tomorrow‟s RP‟ and the new
mode of „talking proper‟. Evidences of the declining prestige of RP are observable in
various dimensions. First is the negative value judgements of arrogance and snobbishness
which was observed by sociolinguistic studies of the 1990s. Such negativity may be
perceived in the following commentary on the present status of RP:
whereas once, people aspired to be posh; it was the voice of the
people in power – in the law, in the City, in the Establishment. Now
there are plenty of people who would be ashamed to speak like that.
A posh voice is seen as naff and unfashionable. (Mugglestone 281):
This negative trend is also reinforced in the media industry as RP‟s role as the accent of
choice in modern films exhibits role reversals whereby heroes and heroines are no longer
102
depicted using RP accents as in many nineteenth century novels. Instead, it has regularly
been deployed for those roundly depicted as villains as Fabricius notes:
[T]he humour of, for example, Monty Python’s Flying Circus, tended to
exploit such affected accent stereotypes, while Disney films for children
seem to reflect the more negative, slightly sinister (because of being
exclusive and arrogant) aspects of U-RP speech: most of the villains of
these films have been U-RP or conservative RP speakers (31).
Trudgill confirms that advertising in Britain has regularly moved away from RP as the
most effective endorsement of a product‟s claim because „messages couched in RP …
proved to be less persuasive than the same messages in local accents‟ (176).
Some writers confirm the decline of RP which Coogle has noted that „many younger
priviledged people make an effort not to sound too “posh”, as they know that this makes
them unacceptable in their peer group‟ (retrieved on the 14th August 2011 from
www.phon.ucl.ac.uk/home/estuary/estufaqs.htm.) While the phonetician Susan Ramsaran
comments that „a real or assumed regional accent has come to have „a greater (and less
committed) prestige for younger speakers‟ (86).
Other writers have sought to determine the nature of Estuary English on the basis of its
reported linguistic features. Haenni, reported in Fabricius discusses selected accent features
(such as t-glottalling, l-vocalisation, and certain vowel phenomena) used to determine
whether they can define a rigid boundary between Received Pronunciation, Estuary English
103
and Cockney. The survey fails to find any single clear-cut „marker of Estuary English‟ On
the basis of this examination of academic and journalistic comments on the linguistic
features of Estuary English, as well as of people‟s awareness of the term, it is concluded
that:
It is thus very difficult to uphold the notion of Estuary
English as a distinct variety in its own right … it appears
rather daring to assign it a place among the stereo-typed (or,
at least, the most „dialectally prominent‟) accents of Britain.
(34).
Estuary English has continued to generate commentaries but the general consensus as to its
existence as a distinct variety of spoken English seems to be skepticism.
3.3 The Phonological Core of a Mainstream RP Accent.
We will now attempt to identify the phonological features which may be found in a
mainstream accent of RP. By this, we mean a clustering of features which form a central
tendency of all RP accent varieties. The description includes a list of both phonemic
distinctions and phonetic features of the accent. For the purposes of this study, we define a
mainstream RP accent as an amalgam of the features listed in Collins and Mees, Wells and
also Cruttenden. The entire section presents a summary of views expressed in notable works
on the phonology of RP such as Wells, Upton, Cruttenden, Brown and Collins & Mees. Our
list of RP phonemic distinctions is an uncontroversial list of 20 contrastive vowels and
diphthongs as stated below:
, , , , , , , , , , , , , , , , , , ,
104
These vowels and diphthongs are all recognized RP phonemes except which has a very
limited occurrence and the variable pronunciation as . Wells points out that “there are
plenty of RP speakers who pronounce some or all of poor, moor, your and sure with ,
and they are on the increase” (287). Besides this, lexical variability is also an important
consideration of the phonemic status of RP vowels and diphthongs. Pronouncing dictionaries
sometimes offer as many as five alternative pronunciations for one word, all within
mainstream RP.
The consonant sounds of our mainstream RP accent are less controversial and are stated
below:
, , , , , , , , , , , , , , , , , , , , , , , .
The phonemic status of these 24 consonant sounds are not in doubt but the voiceless labio-
fricative is not included here even though it is a possible phoneme on the basis of
minimal pairs such as which and witch. According to Wells, “the use of /hw/ in where, wheel
etc is restricted to the speech conscious,” (285) meaning that it has a low frequency of
occurrence.
Statistics for consonant frequencies in RP is provided by Cruttenden and the data reveals the
same 24-consonant phoneme inventory presented above ranging from - 1.97% to -
0.10%. and also show relatively high percentage frequencies (3.56% and 2.81%
respectively). It was noted that structural conditioning occasioned by the frequent
occurrence of determiner the and pronouns in <th-, wh-> probably accounts for the
comparatively high scores of these particular phonemes. Another observation is that “the
105
alveolar phonemes emerge as those which occur most frequently in English, this being a
generalization which appears to be applicable to many languages” (196).
Setting aside the segmental phonemes of RP for now, a consideration of the supra-segmental
features of the accent also reveals certain phonetic/phonological features which define
mainstream RP.
3.3.1 Stress & Rhythm
Stress and rhythm are inter-related isochronic features of speech. For the purpose of this
study however, consideration of stress is limited to sentence stress while rhythm is fully
explored. Native English speech is usually classified as „stress-timed‟ in direct opposition to
second language English speech varieties which work on the principle of „syllable-timing‟
(Gut 828).
In mainstream RP as well, stressed syllables tend to occur at roughly equal intervals of time
as the unstressed syllables in between give the impression of being compressed if there are
many and expanded if there are few (Collins & Mees 108). The shortening of vowels from
their full forms to reduced forms or weak vowels such as , , in addition to various
phonetic/phonological processes of connected speech give mainstream RP speech its
characteristic rhythm. These features may be classified into three broad areas namely:
1. Allophonic variation: Realisational variations of phonemes due to adjustment of
tongue positions to facilitate co-articulation with adjacent or near-adjacent
segments. (Wells 42; Collins & Mees 116).
106
2. Assimilation: The replacement of a phoneme by a second under the influence of
a third phoneme (Collins & Mees 118).
3. Elision: Deletion of a phoneme which necessitates a change from the ideal form
in connected speech (Collins & Mees 118).
These three areas are crucial to the present study because these patterns of allophonic
variation, assimilation and elision are distinctive to mainstream RP speech. These processes
are not only limited in the Nigerian English accent but when they do occur, they follow a
different patterning (Gut 825). Variations such as these may constitute intelligibility
problems which form the central point of the present study. The views are generally in
agreement and cover the areas of allophonic variation, assimilation and elision particularly
as they relate to recent changes within mainstream RP.
3.3.2 Allophonic Variation
The height and degree of centralization of the vowel varies, but is largely relatively open and
central. It is also usually symbolized by unstressed syllables.
The same central and open qualities of are associated with this vowel and it is also
symbolized by .
107
A fronted and more open realization of which is symbolized by usually occur with
this vowel.
This vowel usually involves a glide from to . This makes it diphthongal in all
contexts except before fortis consonants. Another notable phonological characteristic of this
vowel is what is termed „happY tensing‟. This feature affects the final vowels in words such
as happy, lucky and coffee. It involves the realization of the final vowel with a closer
quality (). This quality is similar to a short variety of .
There is considerable variation in the realization of this vowel with qualities ranging from
close mid to open mid. This results in realizations of for unaccented syllables.
It is diphthongal , fronted and becoming gradually less rounded particularly following
palatal and palato alveolar and .
Relatively central, lacking lip-rounding. This often results in a realization which is similar to
. It is also noted that some speakers take this process even farther so that the realization
of the vowel is close to .
108
This diphthong has changed its traditional starting-point to a more fronted position while the
second element is unrounded and also fronted making it similar to .
This either has the same starting point with or is backed to a position similar to
.
The starting point of the diphthong is central and the realization is very similar to or
. The diphthongal movement is very little and there is also minimal lip rounding.
An important phonological process which often characterizes the realization of this phoneme
is the process of smoothing. It is a process of monophthongization of this vowel through the
omission of the glide. This smoothing process also applies to other diphthongs and is
discussed more extensively below under elision.
Because of the monophthonging process described above, a long monophthong is the
usual realization of this diphthong.
109
The most usual realization of this vowel is
The staring point for this diphthong varies between close-mid and open-mid. It is also
usually monophthongized into .
Monophthongization of this vowel also occurs in some cases.
RP triphthongs
All the RP triphthongs namely , , and have their individual vowel
elements maintained in careful speech. But the second element is usually omitted in faster
speech.
Setting aside RP vowels, and turning to RP consonants, only the alveolar plosive is
noted to present any considerable allophonic variation. Glottal reinforcement (also called
pre-glottalisation) and t-glottalling (also known as glottal replacement) are the phonological
features which are involved in variable pronunciation. Although both features were
considered to be restricted to lower class accents such as Cockney, they are certainly
observable in certain environments within the RP accent. This observation is made by Wells
(201):
Glottalling is the switch from an alveolar to a glottal articulation of ,
whereby in a range of syllable–final environments. This is by now
110
very firmly established in casual RP before obstruents … and is increasingly
heard before other consonants (, , ). Among younger RP speakers it can
even be heard finally before vowels… or in absolute final position.
Fabricius (145) also supports this observation in a commentary of the current
sociolinguistic status of t-glottalling in modern RP:
1. T- glottalling in modern RP is stable in pre-consonantal environment in both
speech styles and is accepted by these speakers in formal and non-formal
speech.
2. It has entered modern RP as a vernacular change (spreading out from
London), but its vernacular status is obscured by other factors.
3. It has to some extent lost its stigma, but not yet acquired prestige in word-
final pre-pausal and pre-vocalic environments.
3.3.3 Assimilation
This term refers to the effect created as a result of phonetic conditioning, when one phoneme
is replaced by a second phoneme. Assimilation is one of the features which create the special
allegro characteristics of mainstream RP speech, thereby differentiating it from the speech of
the typical Nigerian. Some of the commentaries on this speech feature which are examined
below show that the patterns of assimilation in RP speech are distinct and rather more
extensive than that of the Nigerian English accent. For instance, Collins and Mees basing
their typology of mainstream RP assimilation on direction and type of assimilation recognize:
leading, lagging, reciprocal, place, energy and manner assimilations . Leading
111
assimilation is described as a situation whereby as a result of phonetic conditioning, features
of an articulation may lead into those of a following segment. On the other hand, lagging
assimilation involves the holding of the articulation features of a preceding segment over
another segment. Cases involving a two-way exchange of articulation features were termed
reciprocal assimilation. Place, energy and manner articulations were terms used to refer to
„types of influences in assimilation‟. Three types of influences were described: assimilations
involving a change in place of articulation (place assimilation), assimilation involving a
reduction of the fortis/lenis contrast (energy assimilation) and assimilation involving a
change in the manner of articulation (manner articulation). Likewise, Cruttenden (301-303)
provides a detailed explanation of this phenomenon by identifying regressive or
anticipatory, coalescence and progressive assimilation. Regressive/anticipatory
assimilation and coalescence assimilation correspond to Collins and Mees‟ leading and
manner assimilation respectively. Progressive/perseverative assimilation is described as an
uncommon process which may occur when a plosive is followed by a syllabic nasal and the
nasal undergoes assimilation to the same place of articulation as the preceding plosive.
Collins and Mees also add that greater complexity be may introduced into the process of
assimilation in two ways: (a) when two types of assimilation affect one phoneme; and (b)
when more than one phoneme is affected by an assimilation.
3.3.4 Elision
A historical view of elision is presented by Collins and Mees (118) in their examination of
the status of this phenomenon in both older varieties of English and contemporary English.
„Historical assimilation‟ is the term that is suggested to cover cases where the deletion of a
phoneme in a word has become fixed and the original ideal forms have become extinct. The
112
„silent letters‟ of English illustrate cases of historical elision (e.g. talk, comb, know, iron).
However, Cruttenden, Wells, Brown and Gimson provide accounts of the processes of elision
in present-day English. Elision is generally described as a process whereby a phoneme is
deleted or elided, unlike assimilation which involves the replacement of a phoneme by a
second one. Cruttenden (303) provides two major environments in which elision may occur
in mainstream RP. They are word-internal position and word-final or word-boundary
positions. Elision in the former usually involves the loss of a vowel, consonant or entire
syllable involving weak accents. The alveolar plosives and and consonant
clusters are also identified as the commonly elided segments in word-boundary positions.
Another type of elision which involves mainstream RP diphthongs and triphthongs is an
innovatory phonological process known as smoothing. Described by Wells (238) as a
„monophthonging process‟ in which the quality of the resultant monophthong is that of the
starting-point of the underlying diphthong, it involves vowels with front and back; mid to
close quality. Commenting on the same process, Cruttenden admits that it results in “the
reduction of the phonetic sequences , to , … new homophones are
produced in this way, e.g. tyre, tower; shire, shower; sire, sour (145). Of all the RP
phonetic/phonological connected speech processes discussed in this section, „smoothing‟ is
perhaps the most crucial for intelligibility and Gimson may well be right in his observation
that „this monophthongization of and and their coalescence with is
likely to be one of the most striking sound changes affecting southern British English” (140)
3.3.5 Intonation
113
The final supra-segmental feature to be considered in this review is intonation.
Differences exist in the intonation of RP and NEA and this affects intelligibility as
agreed by Cruttenden that:
while the variation in intonation between languages (and between
dialects of English) is not as great as that involved in segments, it is
nonetheless sufficient to cause a strong foreign accent and in some
cases lead to misunderstanding (270).
Intonation is a speech feature which is usually acoustically manifested by pitch with three
principal functions. However, for the purpose of the present study, only the discoursal and
attitudinal functions of intonation are relevant. The traditional approach which employs the
nuclear tone to describe the shape of tunes is adopted in the discussion of RP intonation.
Generally, there is agreement among scholars on RP intonation tunes and their meanings.
Roach and Cruttenden reflect this agreement as regards the following:
(i) Declaratives: falling tones
(ii) Yes/no-interrogatives: low-rising tones
(iii) Tag-interrogatives: a falling tone (high fall or low fall ) or a rising
tone (usually low rise).
(iv) Polite imperatives: Rising tone (low rise or fall rise
(v) Exclamatives: Falling tone.
114
Cruttenden also includes a description of tones for certain formulaic expressions such as
„thank you‟, „good morning‟. In such situations, it is claimed that falling tones and high fall
tones generally show sincerity in the first and second expressions respectively.
In conclusion, the examination of aspects of segmental and suprasegmental variation
between NEA and RP has revealed striking differences. There is little doubt that these
differences are bound to affect intelligibility. What we do not know are the specific features
which are crucial for intelligibility and the extent to which intelligibility is affected by
these phonetic and phonological differences. These are the questions which we will attempt
to answer in the subsequent chapters of this study.
CHAPTER FOUR: THEORETICAL & METHODOLOGICAL FRAMEWORK
4.0 Introduction
This study attempts to examine the issue of accent variation and how it impinges on
intelligibility. It is based on the assumption that language is heterogenous rather than
115
homogenous and therefore requires a framework which recognizes the heterogeneity of
language for adequate investigation and explanation. The sociolinguistic theory which
represents a heterogeneous view of language provides a general view of linguistic variation
and is discussed in this chapter. In addition, sub-fields of this theory which are more
directly related to the explication of the issues of accent variation involved in the present
study are also subsequently identified and explored in the present chapter. The framework
of intelligibility testing is an appropriate procedure for the achievement of the main
objective of the study which is to measure the level of intelligibility of the Standard British
English accent to Nigerians. In due recognition of the suitability of this framework, an
attempt is made to establish its appropriateness as a tool in this chapter. The section is
concluded with a discussion of the procedures which were adopted for the data gathering.
4.1 The Sociolinguistic Theory
Emerging in the 1960s, sociolinguistics represents a departure from the previous view of
language structure as a homogenous unit, by recognizing the heterogeneity of language. It
is a theory which seeks to account for the full variety of the individual‟s linguistic
behaviour in three major ways. The three approaches are as follows: (a) language variation
in different social circumstances (b) variations existing between different
individuals/groups within the same speech community, and (c) variation along geographic
divisions. Sociolinguistics can therefore be described as that branch of linguistics which
seeks to describe language in terms of variation along several dimensions.
116
There have been various attempts to describe sociolinguistics as a branch of the social
sciences (Downes; Fasold , Eckert and Rickford, Chambers, Trudgill & Schilling-Estees
etc.) In all of these attempts, the themes tend to include language and social structure,
language use as a social phenomenon, how language contributes to making community
possible, social aspects of language, linguistic aspects of society, language variation etc.
Language and society are therefore two central themes in sociolinguistics and it serves as a
meeting ground for linguists on one hand and social scientists on the other hand. Coulmas
recognizes two „centers of gravity‟(8) within sociolinguistics which correlate with the twin
themes of language and society. They are macro and micro sociolinguistics or
sociolinguistics in the narrow sense and sociology of language respectively. These views
represent different research agendas. While macro-sociolinguistics involves investigations
of language attitudes and language shift among other things, micro-sociolinguistics is
engaged with language centered issues such as how language varieties correlate with
various attributes. In this distinction, sociolinguistics is concerned with investigating the
relationship between language and the society with the goal being a better understanding of
the structure of language and how languages function in communication; the equivalent
goal in the sociology of language is trying to discover how social structure can be better
understood through the study of language. Hudson has described the difference as follows:
sociolinguistics is „the study of language in relation to society‟, whereas the sociology of
language is „the study of society in relation to language‟ (4). In other words, in
sociolinguistics, language and society is studied in order to find out as much as we can
about what kind of thing language is and in the sociology of language, the direction of
interest is reversed. Using the alternative terms above, Coulmas points out that:
117
micro-linguistics investigates how social structure influences the
way people talk and how language varieties and patterns of use
correlate with social attributes such as class, age and sex. Macro-
linguistics on the other hand, studies what societies do with their
languages, that is, attitudes and attachments that account for the
functional distribution of speech forms in society, language shift,
maintenance, and replacement, the delimitation and interaction of
speech communities. (2)
An examination of these descriptions shows clearly that the issues in this study center
around micro-sociolinguistics. This particular area is the focus of this study because of its
concern with the issue of language variation. Investigations of language variation
undertaken in the area of micro-sociolinguistics are often referred to as variation studies
while the theory usually employed for the explication of language variation issues is the
variation theory.
4.2: Variation Theory
The variationist approach is a language-centered viewpoint within the realm of
sociolinguistics. Its concern is the internal variation which exists in languages and the way
in which speakers make use of the many different possibilities that language offers to them.
The major task of variationists is to try to specify the norms of linguistic behaviour that
exist in particular groups and then try to account for individual behaviour in terms of these
norms. In the accomplishment of this task, several issues have been identified by
118
Wardhaugh as being of great concern to variationists. They are: identity, group
membership, power and socialization (6). These themes are described in relation to
language variation as follows: Identity is perceived in terms of individual and group
identity and variationists try to explicate language behaviour in terms of „people seeking to
negotiate, realize, or even reject identities through the use of language‟. Group membership
is seen to be a major determinant of group behaviour as a group‟s language behaviour is
influenced by the desire of its members „to achieve a sense of solidarity‟. Power also plays
a significant role in language variation as its use produces linguistic effects which result in
certain consequences such as the following: the creation of standard languages, standard
accents, gendered styles of speaking and specific types of discourses. The themes identified
above correlate with the major divisions of variation studies which are geographical and
social variation. Differences in pronunciation, choices and forms of words and in syntax
over a wide region constitute geographical or regional variation.
4.2.1 Geographical/ Regional Variation
The study of geographical or regional variation is informed by a fundamental assumption
of the geographical sensitivity of linguistic forms. It recognizes that geographical varieties
arise over time and as a result of historical processes. One of the earliest explanations
which informs this theory is the classical wave theory put forward by Schmidt. In this view,
linguistic changes are visualized as waves or meteorological fronts which spread at
different rates and in different directions, thereby leading to different innovative features.
This implies that geographical variation is similar to the act of dropping a stone in a pond.
This produces ripples that move farther and farther away from the center, producing a
119
wave-like pattern. In linguistic terms, as forms move farther and farther away from the
center, more divergent forms or dialects are created in a language. Schmidt‟s view seems to
explain the emergence of the Nigerian dialect of English as a form which developed from
the English used in England, thereby linking the two varieties. A similar view is expressed
by Dittmar who argues for the equality of language varieties. It is claimed that all regional
dialects and social dialects are functionally equivalent in their expressive possibilities and
their capacity for logical analysis since they have a common source:
[W]hat we find is not well-formed and regulated activity as
opposed to ignorant or careless, but rather a conflict of definite,
fixed locutions, one of which for some reason is “good” while the
others are “bad”...The scientific view…will bring us farther. It has
the disadvantage of not been based on a more extensive survey of
various languages and of their history than any one person could
make; also it has the advantage of methodical approach. This last
means that we shall not operate with the terms “good” or “bad”
language or their equivalents, since it is precisely these that we are
trying to define ….. In this way, wherever there are lines across
which communication is hampered- water, mountains, deserts,
political boundaries and the like – we find differences in speech
(113).
From these arguments, it may be inferred that because all varieties of a language may be
traced to a common origin, none should be considered as intrinsically better than the other.
This seems to suggest that considerations of the Standard British English accent as being
superior to the Nigerian English accent may not be based on any empirical evidence as
such but on mere conjecture.
120
Two main types of processes have been identified in terms of the creation of regional
varieties. They are: relocation diffusion and expansion diffusion (Britain 622). The
process of the creation of varieties through the migration of individuals or groups from one
location to another is termed relocation diffusion while expansion diffusion involves the
process of the creation of a variety through day to day passing on of new variants from
those who have acquired them. The process which led to the formation of the Nigerian
accent of English may be likened to relocation diffusion described above. The migration of
British traders and colonial government officials to Nigeria at around the sixteenth century
is very similar to this process.
Another theorist whose views have a bearing on this study is C.J Bailey. Bailey proposes a
theory of speech variation in the frame of the competence theory developed by Noam
Chomsky; but stresses that it can only succeed when the Chomskyan view of the
homogeneity of the language community is abandoned in favour of an assumption of
heterogeneity. Employing the Chomskyan notions of deep and surface structure, an attempt
is made to establish the proposition that two or more varieties of a language can be said to
be dialects of a language if they have the same deep structure identity. Arguing that
regional varieties of a language can be described by co-existent grammars, this model does
away with the traditional linguistic distinction between synchronic and diachronic, in
favour of a polylectal model.
This model represents the adult listener‟s internally constructed unified grammar which
subsumes dialect differences by enabling him/her to summarize variety differences in a
121
form which enhances understanding. The same ability is perhaps employed by the non-
native speaker in understanding, interpreting and perhaps even predicting the speech of
native speakers. This concept attempts to explain the process of negotiating intelligibility
between interlocutors who speak different accents. The conclusion which may be drawn
from these arguments is that interlocutors speaking different varieties are able to achieve
mutual intelligibility due to a shared underlying form of both accents.
Many other scholars have contributed to the development of this theory in terms of
empirical, theoretical and practically oriented works. Theoretical works on languages in
contact by Weinreich and Ferguson have strongly influenced variation research as some of
the concepts are also used in the study of language variation. This includes the notion of
interference (phonological, lexical, and grammatical), and the notion of transfer. These
concepts are useful in studies involving the comparison of accents. Other useful concepts
include contagion or spatial diffusion concept which were introduced by Bailey. The
concept represents an attempt to build in historical as well as social variation into
geographical variation and explain how variants spread from central focal areas. These
concepts are useful in an accent variation study such as this because they provide
explanations for this phenomenon. A relatively new but related area of study which focuses
on the notions of convergence and divergence within ethnolinguistic diversification is geo-
linguistics. Identified by Britain, it is a sociolinguistic concept within the realm of
spatial/geographical variation which attempts to show how language varieties converge and
diverge at the same time. Focusing on the interrelatedness of accent variation and
geographical variation, it emphasizes the important role of topography in accent
122
delimitation. Britain argues that wherever natural impediments exist (such as rivers and
mountain) creating boundaries between communities, one is likely to find both areas of
similarities and differences in the speech of the communities.
Most of the views that have been examined above have explored geographical and social
variation within the field of accent phonology. The following subsection deals with this
concept and how it creates social divisions in human societies.
4.3 ACCENT PHONOLOGY
Accent phonology aims at identifying the structure behind linguistic systems in spoken
language. Based on the sub-systems of segmental and supra segmental phonology, it
attempts to show the systematic nature of the consonants, vowels, tonal patterns, stress
patterns, syllable shape etc., which exist in individual languages or in languages generally.
It basically seeks to provide a linguistic explanation for phonological structure through the
establishment of permissible and non-permissible phonological patterns which may exist in
phonological space. In describing the two-pronged approach which accent phonology
entails, Wells employed the terms „bottom up phonology‟ and „top down phonology‟ (40).
On one hand, he argues that bottom-up phonology „involves the segmentation of the mass
of speech sound and the identification within it of relevant features of sound patterning…so
that the phonetically essential is separated out from the phonetically redundant or
irrelevant‟. On the other hand, top-down phonology involves the prediction of
pronunciations in a way that illuminates the differences between geographical and social
varieties.
123
These explanations for phonological structure are usually made in relation to the notions of
standard vs. nonstandard which are used in descriptions of accent variation. Hence, accent
varieties are commonly described in terms of variation from or approximation to the sounds
of a standard variety of accent.
The notion of a standard accent is central to accent variation studies. Standard accents are
usually associated with extended education and the use of prestige variants rather than
stigmatized variants. Stigmatized variants are usually associated with non-standard accents.
Both standard and non-standard accents can be represented on the accent continuum –
developed by Stewart and Bickerton - and used to classify linguistic features as spread
along a line, showing clusterings around the acrolect (standard variety), the mesolect (the
speech of the majority of the population) and the basilect (broadest possible accent forms).
The notions of hyperlect and paralect, introduced by Honey refer to clusterings between the
acrolect and the mesolect.
The description of accent variation is central to both phonological and sociolinguistic
theory, but at the intersection of sociolinguistics and phonology is sociophonology.
4.4 Sociophonology
This field is an aspect of sociolinguistics that focuses on socially significant pronunciation
differences rather than all aspects of language variation (Honey 92). The social variables
124
which are identified as being significant in sociophonology are socioeconomic class, region
age, sex (gender), ethnic group, social network and contextual style (Labov, Chambers and
Trudgill, & Milroy). Central to Sociophonological studies such as those listed above, are
distinctions between variants in terms of their relationship to social significance. One of the
leading scholars in sociophonological studies, Labov, identified the following core
notions of the field:
(a) indicators: variants to which little or no social significance is attached, and may
indeed be perceived by observers with linguistic training;
(b) markers: which are readily perceived variants with social significance; and
(c) stereotypes: which are popular and conscious but imprecise general
characterizations of the speech forms of particular social groups.
The association of these notions with linguistic and social variables has been the focus of
several theoretical and empirical studies in the area of sociophonology. (Labov Trudgill,
Petyt, Milroy and Milroy; Chambers and Trudgill) However, in ESL situations such as that
which exists in Nigeria, the social variable which is most often correlated with
pronunciation is education. Major studies involving such correlation which have been
undertaken in Nigeria include: Jubril , Banjo and Jowitt and they all seem to agree that
there is a close correlation between spoken English and educational attainment.
Another important contribution to sociophonology which is relevant to this study is Le
Page and Tabouret-Keller‟s concept of acts of identity. This concept offers an explanation
of “how and why speakers adjust, in any given situation to what seems to them as the
125
appropriate speech norms” (106). They claim that interlocutors are strongly motivated to
adjust their speech behaviour in line with the speech norms of their interlocutor.
Assessments of this claim in relation to the speakers of Nigerian English have been
attempted by Bamgbose and Awonusi. The major conclusion is that attempts at
approximating to the RP accent or any foreign accent for that matter is usually met with
disdain in Nigeria as such individuals are regarded as “wanna bees” implying the ridiculing
of non-localised speech norms.
These notions are useful in a study of this type which involves the investigation of areas of
divergence and convergence between the Standard British English accent (RP) and the
Nigerian English accent and the implications for intelligibility. While due cognizance must
be given to the importance of sociolinguistics and its sub-fields in the explication of the
phenomena of accent variation, it does have limitations as regards the actual measurement
of intelligibility. Therefore, in order to fully achieve our major objective of measuring the
intelligibility of the Standard British English accent to Nigerians, a more suitable tool
ought to be employed. The framework of intelligibility testing which has been successfully
applied in studies of intelligibility measurements is proposed as an appropriate framework
for adoption in the present study. The following section presents a description of this
framework.
4.5 THE CONCEPT OF INTELLIGIBILITY
The framework of intelligibility testing was first proposed by Catford in a landmark
treatise which represents the first serious attempt to grapple with intelligibility as a
126
construct of speech. It is stated that speech is intelligible “if the hearer understands the
words, i.e., if his response is appropriate to the linguistic forms of the utterance” (3). For
Catford, a speaker achieves complete intelligibility if linguistic forms are selected
appropriately, that is, if the sounds are appropriate. He further distinguishes between
intelligibility and another inter-related term, „effectiveness‟ which goes beyond
intelligibility to include the hearer‟s grasping of the speaker‟s intention. According to
Catford, “it is normally the speaker‟s intention that the hearer should respond to his
utterance in a manner which is appropriate to his purpose in speaking” (4). He goes further
by demonstrating that intelligibility depends for its realization on at least four out of five
aspects:
1. Selection (of words/utterances by speaker)
2. Execution (of words/utterances by speaker)
3. Transmission (of words/utterances from speaker to hearer)
4. Identification or recognition (at the word/utterance level)
5. Interpretation (of meaning behind utterance as intended message)
It is also argued that intelligibility level can be measured spanning a continuum between
„low intelligibility‟ and „high intelligibility‟ (7), demonstrating that it is a scalar quantity
rather than a binary one. Decades after Catford‟s landmark treatise, Smith and Nelson
advocated the notion of „global intelligibility‟(“World” 441) based on the abstract
parameters of co-operation, appropriateness, effectiveness and threshold of intelligibility.
Proposing three terms for the measurement of intelligibility, they identified a higher level
of understanding which they termed as “comprehensibility” referring to “the meaning of a
127
word or utterance” (“World” 438). This is a level of perception higher than mere
recognition of words. The introduction of interlocutor‟s attitude as a variable affecting
intelligibility was done by Dalton and Seidlhofer. The inclusion of this variable emphasizes
the subjective nature of intelligibility judgements as a complex of both speaker and listener
factors. Jenkins presents another approach to intelligibility which is defined by her focus
on the functions of English as a lingua franca between speakers of English from different
L1 backgrounds. Perceiving intelligibility as “Interlanguage Talk” ILT (199). It is proposed
that the recognition of the phonological form, the most basic level of intelligibility as a
precondition for effective communication success among speakers of English in ILT. The
interdisciplinary nature of speech intelligibility was presented by Smith and Nelson based
on the application of concepts used in the Speech Acts and Pragmatics theory to
intelligibility. They divided intelligibility into three categories of intelligibility,
comprehensibility and interpretability in which the last two categories correspond to
locutionary and illocutionary force respectively (441). They argue that the pragmatic
effects of utterances, observable only at the advanced end of the intelligibility continuum
cannot be successfully interpreted without situational, social and cultural awareness.
The conclusions we can draw is that intelligibility can be measured at several levels,
spanning phonological intelligibility to pragmatic intelligibility, but the level which is
relevant to this study is phonological intelligibility or what is otherwise known as
“word/utterance” recognition (Catford 5). The second level of intelligibility, identified by
Catford as effectiveness, will not be investigated. This is because the speaker and listener
will both be in an experimental situation and the listener will not be required to make any
128
responses but merely understand what was said. The usefulness of intelligibility testing in
determining the level of understanding in oral communication is well established and we
hope to further demonstrate this.
4.6 RESEARCH METHODOLOGY
4.6.1 The Test Material
Four types of tests were designed for this investigation. The objective was to elicit not only
the entire range of the phonemic inventory of the RP accent but to also ensure that higher
levels of language analysis such as stress and certain connected speech processes which
researchers have identified as distinctly RP were included. The test material consisted of
the following: Test1 - Connected speech, Test 2 - Phonemes, Test 3-Stress and Test IV-
Intonation.
Test 1 -Connected Speech
Of the various types of materials which can be used for intelligibility testing, spontaneous
connected speech is the most important because the elicited text is representative of all the
levels of phonological analysis, from the segmental to the suprasegmental. Another
advantage of this type of material is the naturalness of the text as this type of material is the
closest reproduction of real life conversational communication. However, the effects of
accommodation will of necessity be present in the sort of interview setting created here but
these influences were kept to the barest minimum by letting the interviewee do most of the
talking. Since the focus of the investigation is the interviewee‟s speech rather than that of
129
the interviewer, the mutual adjustment which usually takes place in any conversation was
minimized by adopting a monologic style. After the initial introductions and explanations,
the interviewer kept quiet. Questions had been prepared to help interviewees who needed
prompting but the prompt questions were hardly used. The topic presented to the speakers
was a very familiar one and therefore they were all able to speak freely, choosing their own
words and they all spoke for longer than ten minutes which was the desired recording limit.
The chosen topic was their last holiday and as the subjects are Britons who generally enjoy
going on holidays and have had at least one in the past, the conditions were relaxed and
thus elicited naturalistic language which the subjects were likely to use in everyday
conversation. Technical language and difficult words were not used by the speakers as
there was no need for them because of the simplicity of the topic.
Although this is an experiment, all efforts were made to make it as close to real life
situation as possible. This is the rationale behind including the connected speech test in
spite of certain objections towards its use on the grounds that it provides contextual clues
unlike word lists tests. People do not use isolated words alone in conversation. They use
longer stretches with similar structure to the structure of connected speech. Therefore, the
connected speech test is closer to real life situation. Besides, initial exposure to connected
speech with its contextual clues and redundancies instead of isolated words allows the
listeners some time to become familiar with the speakers‟ accent for sufficient „calibration‟.
All speakers experience the phenomenon of calibration which is an attempt to
automatically correlate the sounds of other accents with the sounds of their own accent to
facilitate understanding.
130
Test 2
Apart from connected speech tests, phoneme tests are also crucial in intelligibility testing
because articulation tests enable us test specific phonemic contrasts and certain
phonetic/phonological processes in the accents. This list was drawn up based on the
researcher‟s own experience during interactions with British RP speakers and from the
experience of other researchers in the field.
Three other important considerations were involved in the compilation of the word list.
They are as follows: the exclusion of polysyllabic words, the exclusion of certain form
words and the issue of whether or not the words should be embedded in the context of a
meaningful sentence.
Polysyllabic words are generally easier to recognize than monosyllables and it was thought
that the inclusion of polysyllabic words will make the test too easy for the listeners as it
will then mean that the words will be easily recognized. This is the reason why it was
decided that the list should have mostly monosyllabic words, and disyllabic words should
be brought in only where it was necessary for testing a particular phoneme or process. Rare
and unfamiliar words were avoided.
It was considered desirable to omit certain form words such as articles, conjunctions and
monosyllabic prepositions which are usually regarded as weak forms in connected speech
but would have to be given strong forms when said in isolation.
131
The issue of whether or not the words should be embedded in sentences to provide some
context was a serious consideration. The advantage of placing the words to be tested in a
context is that this represents a more realistic language situation. In addition, this elicits
more natural speech because the speaker is usually unaware of what is being tested.
However, the disadvantages include the fact that a lower number of items can be tested and
this may make the list less reliable.
Testing the phonemic contrasts in isolation has several advantages. Chief among them is
that it enables both speaker and listener to concentrate on the phoneme/process being tested
and this makes for easier analysis. In addition, more items can be tested as more time is
saved. Time was of essence here because all the speakers and listeners were volunteers who
would not like to be delayed unnecessarily. However, there are two overriding reasons for
testing the words in isolation:
(a) it enables the speaker to concentrate solely on the phoneme being tested; and
(b) isolated words are easier to analyze as extraneous influences are absent.
Therefore, it was decided that the words will be tested in isolation.
Test 3 - Placement of nucleus in sentences (sentence stress)
Test items on sentence stress were designed to elicit certain nuclear stress patterns in
groups of three similar sentences. For example in the sentence No, Tom had fried chicken
132
for dinner, the nucleus can be placed in at least three places, depending on the emphasis the
speaker wishes to give. Thus, we have:
No, `Tom had fried chicken for dinner (i.e. not James)
No, Tom had `fried chicken for dinner (i.e. not roasted)
No, Tom had fried chicken for `dinner (i.e. not lunch)
The researcher (called the interviewer in the tests overleaf) asked the RP speakers to
contradict as emphatically as possible the statements put to him/her in the first line of each
dialogue. The statements were read to the speaker with a high falling tone on the key word
the speaker was supposed to contradict, as follows:
Did James have fried chicken for dinner?
Did Tom have roasted chicken for dinner?
Did Tom have fried chicken for lunch?
The 16 speakers produced 380 recorded examples of nucleus placement in sentences.
Test 4- Attitude and intonation
Items were designed to elicit sentences conveying different attitudes by means of
intonation. The test was designed to judge how well Nigerian listeners would understand
different attitudes conveyed by means of varied intonation patterns. To elicit the desired
intonation tunes, short situational dialogues were constructed in which both researcher and
speaker took part. Some of the items contained short prompts for the speakers, in the form
of instructions as to the tone of voice or attitude expected, viz. surprise, doubt, agreement.
133
Before the actual recordings were done, the researcher and speaker had a practice run-
through of all the items, in order to familiarize the speakers with the material.
Item 49 aimed at eliciting from the speaker a falling statement pattern, in contrast to item
50 where a questioning, rising nucleus commonly associated with information-type
questions is expected. Other intonation tunes include the following: falling nucleus
commonly associated with information-type questions, rising or fall-rise contour used in
surprised questions, final low falling nucleus and final low rising nucleus to differentiate
between a set of complete and incomplete alternative choices, the falling tone and fall-rise
tone used to convey positive and doubtful attitudes, a low-rising nucleus and a high-rising
nucleus conveying a normal enquiry and a surprised enquiry respectively. Finally, the last
items concern the intonation patterning of question tags.
Test involving a Nigerian Speaker and Nigerian Listerners
In order to provide some measure of validity to the tests as a whole and to test each
listener‟s capacity to listen accurately, a male Nigerian undergraduate was recorded in
Lagos. He performed all the tasks required of the RP speaker.
4.7 POPULATION SAMPLE
134
Both speakers and listeners were selected using the stratified random sampling procedure.
The principles which informed the stratification and the recoding details are discussed
below.
4.7.1 The speakers and recording procedures .
There were several important considerations involved in the selection of the speakers for
this study. The first consideration was this: which type of RP accent is the study focused
on? This question is important because no accent is a homogeneous invariant monolith -
including RP - as a lot of variability is found within all accents. For the sake of proper
contextualization and clarity, mainstream RP which contains features which form a
central tendency for all RP accents was selected as our RP variety. Therefore, 16 Britons
who use mainstream RP accent were selected. 8 male speakers and 8 female speakers were
selected. 16 was considered an adequate number to generate enough data representative of
the RP accent. A smaller number of speakers might not generate enough data while a larger
number would have resulted in an over-proliferation of data to be analyzed. The second
consideration was the level of education of the speakers. As the Britons in consideration are
mother tongue speakers of English, level of education was really of no consequence, but in
the interest of homogeneity of sample population, university undergraduates were selected
so that both speakers and listeners would have a comparable level of education.
The selection of speakers was done within certain principles. The principles were set down
to ensure a certain level of homogeneity amongst the informants. The principles are stated
below:
135
(i) born and bred in Britain, more specifically in England
(ii) parents must be educated (at least up to university level)
(iii) never have been outside England for a considerable length of time
(iv) a university undergraduate
(v) attended a public school.
4.7.2 Method of selection
The speakers were carefully selected from an initial pool of 70 informants. They were
individually engaged in a chat on their personal biography and information on these issues
was sought. Based on their responses, 16 speakers were finally selected. The principles
which informed the selection will now be examined.
The investigation is concerned with only English people, born and bred in England and
therefore excludes other British citizens such as the Scottish and the Irish. Scotland and
Ireland have their own regional accents as well as educated standard accents which is not
RP. This is why England was specifically stated. In addition, all the RP speakers should be
brought up in mono-lingual situations, English being the only language they speak.
The next principle concerns the parents of the informants. Several sociolinguistic studies of
RP have stressed the prestige connotations of native-speaker RP. Chief among the non-
linguistic factors with which RP is usually highly correlated is education. The accent used
in the home will most likely be that adopted by a child. This is why it was felt that parents
136
who had at least a university degree will most likely speak RP and most importantly, their
children will also speak RP.
The third principle is important because of issues concerning „accommodation‟ during
interaction. Individuals who are English by birth but have spent considerable time outside
England are likely to have had their speech influenced by other accents that they have come
in contact with. It is to avoid such undesirable influences that individuals who stated that
they had spent a considerable length of time outside England (more that 6 months at a
stretch) were identified and excluded from the study.
The fourth criterion was included simply to ensure that both speakers and listeners have
comparable levels of education. The study is concerned with educated Nigerians,
specifically, undergraduates and since it is our intention to select undergraduate listeners,
selecting undergraduate speakers will make for homogeneity of both groups of informants.
It should be added here that all the undergraduate RP speakers were students of the
University of Leeds. However, students from the School of English, University of Leeds
and other language departments were excluded for the obvious reason that they would have
had specialized training which might affect the results of the study. But it was possible to
select individuals from widely scattered geographical areas within England due to the large
and widely varied student population of that university.
Lastly, the status of the secondary school attended by the informant was considered. In
England, there exists a distinction between public/private/independent schools and state-
137
owned schools. Exclusive boarding/day schools which are privately owned are referred to
as public/private/independent schools. These schools are elitist, charging fees ranging from
5000 pounds sterling per term for boarders and 2000 pounds sterling per term for day
students. RP is the accent norm in these schools and in fact, one of the earliest references to
the RP accent by Daniel Jones referred to the prestige accent as Public School
Pronunciation (PSP). All 16 informants involved in the study were ex-public school
students. Letters were sent out to the students selected for interview. The purpose of the
interview was to collect some background details of the speakers, to have them sign the
consent form and to ensure that none of the speakers suffered from speech defects such as
stuttering or other serious idiosyncratic speech habits.
4.7.3 Recording Details
The speakers were recorded in various locations on the campus of the University of Leeds.
Care was taken to choose quiet locations such as Postgraduate rooms and Computer rooms
in various departments on the campus. It was not possible to record all the speakers at one
location because of time and convenience constraints. Locations close to each speaker‟s
department were therefore chosen for the recordings. However, it was still possible to
obtain clear and noise-free recordings because of the highly sophisticated recording
equipment used. The samples of speech were recorded using a Sony ICD-UX60 IC
Recorder. The IC recorder offered MP3 stereo recording quality and a USB Direct PC link
which was used to transfer the recorded data directly into a Dell Inspiron laptop computer.
The data was then copied into audio compact discs. An Ahuja CTP-10DX body contact
microphone was attached to the IC recorder and clipped as close to the speakers‟ mouths as
138
possible to enhance high quality recording. The recording speed was 192kbps. Each
recording session lasted approximately thirty minutes.
4.7.4 Recording Procedures
4.7.4.1 Recording Test 1
Each speaker was asked to talk on the topic: My last holiday. The speaker was told the
topic and asked to think about the subject for 2 or 3 minutes with the aid of the prompt
questions written below. These questions were meant to be a stimulus to enable the speaker
to speak fluently and not be at a loss for words. The recording was started when the speaker
signaled readiness. However, the prompt questions were hardly used as virtually all the
speakers were able to speak for five minutes on the chosen topic.
4.7.4.2 Recording Test 2
The word list illustrating various phonemic representations and various phonological
processes, described earlier were presented to each speaker. They had the opportunity of
reading them silently before recording. The speakers were to pause slightly between each
word and to read at normal speed.
4.7.4.3 Recording Test 3
139
As already explained earlier, this section required the speaker to place the nuclear accent on
two groups of three similar sentences with the desired word presented with strong visual
clues e.g. TOM, DINNER. It was explained to the speaker that he/she should contradict the
questions put to him as strongly as possible. Before the recording began, the researcher and
speaker had a practice run in order to let the speaker familiarize himself/herself with the
text.
4.7.4.4 Recording Test 4
As with Test 3, all the sentences were read through once by the researcher and speaker
before the recording. It was explained that the sentences not preceded by a prompt (e.g. the
instructions „doubtful‟, „surprised‟) were to be read, as far as possible, in a normal tone of
voice. As with the earlier tests, there was a run-through session.
Tests 1, 2, 3 and 4 were administered to all the 16 speakers, which resulted in about 6 hours
of recorded speech before the editing. The recordings were made between the 20th of
September and the 25th of October, 2008.
4.8 The Listeners and Listening Procedures
In order to achieve a truly representative sample of educated Nigerian listeners, a slightly
different set of criterion had to be adopted for the selection of listeners. However, to ensure
a relatively comparable set of speakers and listeners, the criterion of School Certificate as
the minimum qualification which was used for the selection of speakers was also used for
the selection of listeners. Another consideration was the competence of the listeners.
140
Competence was determined by the ability of the listener to obtain a high score with the
recorded Nigerian speaker. Those that failed to obtain a minimum score were rejected on
account of being „incompetent‟ and other listeners were substituted. Also, listeners who
had considerable exposure to foreign English accents were not acceptable and neither were
listeners with any special training in phonetics or spoken English.
However, a group of listeners made up of Nigerians living in the United Kingdom was
included in the sample for comparison purposes. They represented the Nigerian diasporic
dwellers in the UK. This group was incorporated into the study because it was considered
necessary to investigate the effect of familiarity on intelligibility.
4.8.1 Criteria for Selection
Education is the common criterion which was used in the selection of both speakers and
listeners. For the purposes of this research „education‟ meant a minimum educational
qualification of School Certificate. Other criteria used in the selection of listeners are as
follows: Firstly, the listener must be a Nigerian, born and bred in Nigeria of Nigerian
parents from any of the three major ethnic groups. Secondly, he/she must speak a Nigerian
indigenous language with some fluency. Thirdly, listeners must be second language
speakers of English.
16 Nigerians who had spent a minimum of 10 years in the UK and all living in either
London or Leeds were selected to represent „Nigerians in diaspora‟. 16 was considered an
adequate number as the ratio of Nigerian listeners living in the UK to other Nigerian
141
listeners involved in the study would then be 1:10. For this group of Nigerians, their ethnic
orientation was not an important consideration. The criteria used in their selection were as
follows: Firstly, the listeners must be born in Nigeria and to Nigerian parents. Secondly,
they must speak a Nigerian language with moderate fluency. Lastly, the minimum
educational qualification of School Certificate was a necessity for this group of listeners
just as with other listeners and speakers.
Based on these criteria, 160 Nigerian listeners were finally selected. 12 others were rejected
on the grounds of their scores when listening to the Nigerian speaker. Acceptable listeners
were given a number 1-160, preceded by L, to distinguish them from the speakers‟
numbers. Details of each listener were recorded on a form and they were all required to
sign a consent note.
Of the 160 listeners, 108 were female and 52 were males. The preponderance of females
was because females appeared more willing to volunteer than males.
4.8.2 Listening Procedures
Due to the large number of listeners required, it was not possible to have them all in one
location for the listening sessions. Different locations were used for the listening sessions.
Some of the sessions were held in listeners‟ hostels, study areas and even religious centres.
The question of whether to hold individual or group listening sessions was considered. The
advantage of group listening sessions is that it saves time as a larger number of responses
142
can be obtained at one sitting. Eventually, both individual and group sessions were held as
the major consideration was the listener‟s convenience.
The use of earphones was considered, but rejected because there were no obvious
advantages they offered. The playback was done using a Dell laptop computer fitted with
external speakers and the sound quality was satisfactory. Care was taken to use rooms free
from distractions and extraneous noises. The listeners were arranged in semi-circles round
the loudspeaker, not more than two rows deep, to ensure that no one listener was too far
away from the speakers to hear properly.
On all occasions, listeners produced written responses. The possibility of having listeners
provide oral responses was rejected on three grounds. Firstly, oral responses were only
possible with individual listening sessions; secondly, there was no guarantee that the
listener would reveal to the researcher all the words he had not understood; thirdly, no
permanent record could be obtained for further study and analysis at a later date. Written
responses had all these advantages.
4.8.1.1 Listening to Test 2
The method of listening adopted for this study is the written response method in which the
listener was to write down what he/she heard. This method produced permanent and easily
verifiable records. The RP speakers‟ connected speech texts were edited to make them all
approximately the same length. The researcher‟s voice was deleted from the recordings.
143
The texts of each speaker were then divided into units of approximately 9 or ten words.
This length was considered adequate for the listener to be able to recollect what was heard
sufficiently for accurate writing. Wherever possible, units were divided into sense groups,
ranging in length from 3 to 16 words with a mean length of 9.5 words.
On a number of occasions, however, units had to be divided not according to the sense
groups determined by grammar, but according to the hesitations or pauses made by the
speaker. Pauses and hesitations did not appear to affect the listener‟s ability to comprehend
the whole sentence.
Before each listening session began listeners were told in advance the subject matter of the
text which was „My last holiday‟. Listeners were also told to ignore hesitation phenomena
and pause fillers such as „ehm‟, „well‟, „you see‟.
A problem was caused by the use of abbreviations, place names and some other proper
nouns that would be unfamiliar to some Nigerian listeners. No doubt this will interfere with
the intelligibility of the listeners because it would mean a reduction in the number of items
scored correctly. To avoid this, it was decided that abbreviations, place names and other
unfamiliar items should be revealed to the listeners in advance. These were dictated
immediately before the unit in which they occurred was played back.
During the listening sessions, the connected speech texts were played unit by unit, the
compact disc being stopped by pressing the pause button on the computer while the
144
listeners wrote down what they had heard. This procedure continued till all the units were
played through.
4.8.2.2 Listening to Test 2
The procedure adopted for listening to this section was essentially the same with the one
used for the first section. The entire word list was played back one word at a time and the
listeners were given time to write after each word.
4.8.2.3 Listening to Test 3
It will be recalled that different stress patterns were elicited by using short, contextualized
dialogues in which both the researcher and the RP speakers took part. Before presenting the
RP speakers‟ utterances to the listeners, the researcher‟s own speech was erased from the
compact disc. After playing back each utterance, the listener would then have to indicate
which sentence was being contradicted out of several possible ones. The test as presented
to the listeners can be found in the appendix.
4.8.2.4 Listening to Test 4
The listening procedure for this section was done in a similar manner as that adopted for
Test 3. It is an attitude and intonation test in which the listeners were to indicate which
attitude they thought the speaker was trying to convey. The researcher‟s speech was erased
from the recorded contextualized dialogues before being presented to the listeners. It was
explained to the listeners that the term „attitude‟ was an elastic one that also included the
notion of completeness and incompleteness, agreement and disagreement as well as the
145
more conventional meaning of that term. The test as presented to the listeners can be found
in the appendix.
4.8.2.4 Order of tests
Tests were presented to all of the 160 acceptable listeners in the following order:
Test 1– Connected Speech
Test 2 – Phonemes
Test 3 – Stress
Test 4 - Intonation
Each listener took part in one listening session only. In this way, the possibility of
„learning‟ through increasing familiarization with the RP speakers was kept to a minimum.
The allocation of listeners to speakers was done on a random basis. Each session lasted
approximately 45 minutes.
One final remark on data: as a second language speaker of English attempting to make
statements on native speaker speech, I am fallible. I therefore had to enlist the help of a
native speaker as a consultant on certain grey areas during the transcription of the native
speaker data.
CHAPTER FIVE: DATAPRESENTATION AND ANALYSIS
5.0 Introduction
146
In this chapter, the intelligibility scores obtained by the 160 Nigerian speakers in the four
tests are presented in the first section in addition to the phonetic/phonological analysis of
the RP speakers‟ utterances. The analysis of the speakers‟ utterances and the written
responses of the Nigerian listeners was done in order to achieve two purposes:
(a) identify the instances of intelligibility breakdown; and
(b) identify the cause of intelligibility breakdown.
The second section presents the analysis of both the quantitative and the qualitative data for
the four tests administered to the participants in the study. However, an explanation of the
scoring procedures adopted for each of the four tests is discussed first.
5.1 DATA PRESENTATION I: INTELLIGIBILITY SCORES
5.1.1 Scoring Procedures and Intelligibility Scores
In Chapter 4, it was stated that the Nigerians produced written responses for all the tests.
This procedure has many advantages; chief of which is that it provides a permanent record
for comparison and analysis. These written responses were compared to the original
recorded text, making it possible to discover where intelligibility had broken down. It was
necessary to devise a scoring system that would measure the intelligibility of the individual
speakers, compare speakers with one another and to make statistical statements not only
about the performance of the speakers but also about the listeners. To this end, different
scoring procedures had to be adopted for the various tests as it was not possible to utilize
147
the same pattern for all of them. The section that follows provides details of the scoring
procedures that were adopted for the study.
5.1.2 Scoring Procedure for the Connected Speech test (Test 1)
As explained in the previous chapter, the speakers‟ texts were presented to the listeners in
units of between nine to ten words, each unit comprising a single sense unit. For the
scoring of this test, each unit was taken as a whole, with no partial or fractional scores such
as half or one third. Thus, a correctly written response had to contain all the key content
and structural words giving meaning to a particular unit and such a correct unit was
awarded one mark. Units were marked either correct or incorrect, with units which had
important elements misinterpreted by the listener marked as incorrect, and units which had
all important elements correctly interpreted by the listener marked otherwise. Hesitation
phenomena and pause fillers such as „I mean‟, „you see‟, „so‟, „erm‟ „aah‟ etc were
discountenanced during the division into units and the subsequent grading. In other words,
it did not matter whether the listener correctly interpreted them or not as grades were not
affected. Also discountenanced were proper nouns such as names of places and people.
This was done because it was observed that a majority of the proper nouns were unfamiliar
to the listeners and this would have negatively affected their scores. Minor verb tense
changes, plural marker changes and substitution of determiners etcetera., were also ignored.
This method may be regarded as favouring the poorer speakers because inability to
correctly interpret a word attracted the same zero score as does the inability to correctly
identify up to seven or eight words. However, this method was considered to be the most
148
effective because if partial scores are awarded, there would be subjective judgements of
whether to award half or one-third. Too many subjective judgements such as this may
affect the reliability of the intelligibility scores. Besides this, it was considered that the
poorer speakers were already at an advantage due to the numerous contextual clues which
connected speech provides and this method would provide some balance to ensure that
neither group is favoured. The system of rating each speaker by ten listeners is an added
precaution that was taken to mitigate against skewing in the intelligibility scores. This
procedure entailed having ten listeners listen to a single speaker. Each of the ten scores was
converted to percentages and the speaker‟s final score was the mean of all ten listeners‟
scores. Thus if a speaker‟s text consisted of 42 units and 23 units were successfully
conveyed to a listener, the speaker‟s score with that particular listener would be 54.8%. In
this way, each speaker was given ten scores and the speaker‟s final score was the mean of
all ten scores.
5.1.3 Scoring Procedure for the Phonemes test (Test I2)
The scoring procedure adopted is similar to that of the connected speech test. The total
score for this test was 40. Each speaker‟s score was calculated based on the number of
correct responses given by the listener. The scores were then converted into a percentage.
Thus, if a speaker correctly conveyed 22 out of the 40 items in this test, the speaker‟s
percentage score would be 55%.
5.1.4 Scoring Procedure for Stress (Test 3)
149
In a similar manner to the previous test, the number of correctly identified items
represented the speaker‟s score which was then converted into percentages.
5.1.5 Scoring Procedure for Intonation (Test 4)
The items on intonation were presented in this test. The same pattern was employed for the
scoring of this test as the speaker‟s score was also converted into percentages.
5.1.6 Scoring the Nigerian speaker
The listeners‟ responses to all the tests were marked following the same procedure adopted
for the responses to the RP speakers.
5.1.7 The Intelligibility scores.
Following the procedures discussed above, the intelligibility scores obtained in the four
tests by all 16 speakers are shown in the tables that follow:
Intelligibility Scores (%)
150
Speaker No. FRP1
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L1 61 29 100 54
L3 66 28 100 61
L5 62 23 100 47
L7 64 24 100 45
L9 63 31 100 49
L10 77 50 100 75
L11 70 20 80 50
L13 58 17 100 46
L15 79 26 100 49
L17 45 28 100 56
Mean Score 64.5 27.6 98.0 53.2
Intelligibility Scores (%)
151
Speaker No. FRP2
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L2 55 27 100 67
L4 64 24 80 57
L6 71 28 100 65
L8 59 31 100 55
L12 52 29 100 58
L14 78 22 100 64
L16 68 26 100 59
L18 51 19 100 61
L20 85 55 100 75
L22 63 16 100 58
Mean Score 64.6 28.7 98.0 61.9
152
Intelligibility Scores (%)
Speaker No. MRP3
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L19 53 31 100 67
L21 52 26 100 53
L23 49 24 100 58
L25 51 14 100 45
L27 39 26 100 64
L30 67 54 100 73
L29 49 34 80 50
L31 43 21 80 62
L33 41 36 100 48
L35 28 32 100 67
Mean Score 47.2 29.8 96.0 58.7
Intelligibility Scores (%)
153
Speaker No. MRP4
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L24 55 34 100 44
L26 52 46 100 55
L28 50 49 100 50
L32 58 32 100 48
L34 63 29 100 58
L36 67 29 100 67
L38 61 35 100 53
L40 71 51 100 70
L42 60 42 100 49
L44 54 46 80 58
Mean Score 45.1 39.3 98.0 55.2
Intelligibility Scores (%)
154
Speaker No. MRP5
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L37 57 48 100 67
L39 42 43 100 53
L41 61 47 80 50
L43 55 43 100 52
L45 45 40 100 43
L47 58 41 100 56
L50 77 57 80 82
L49 56 48 100 51
L51 51 43 100 62
L53 60 39 100 40
Mean Score 56.2 44.9 96.0 55.6
Intelligibility Scores (%)
155
Speaker No. MRP6
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L46 51 34 100 71
L48 44 49 100 65
L52 59 48 100 52
L54 50 37 80 44
L56 45 31 100 56
L58 59 63 100 83
L60 74 63 100 83
L62 41 23 100 59
L64 48 31 100 50
L66 52 28 100 53
Mean Score 50.7 37.8 98.0 58.5
156
Intelligibility Scores (%)
Speaker No. MRP7
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L55 38 26 80 65
L57 67 33 100 75
L59 50 43 100 75
L61 56 38 100 77
L63 54 38 100 69
L65 59 28 100 64
L67 61 41 100 56
L69 31 23 100 59
L70 73 57 100 79
L71 47 31 80 61
Mean Score 53.6 35.8 96.0 68.0
157
Intelligibility Scores (%)
Speaker No. MRP8
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L68 65 34 100 50
L72 67 33 100 50
L74 60 28 100 52
L76 58 43 100 43
L78 67 40 100 43
L80 81 58 100 69
L82 58 34 100 34
L84 58 35 100 41
L86 52 35 100 38
L88 49 23 100 38
Mean Score 61.5 36.3 100 46.1
158
Intelligibility Scores (%)
Speaker No. FRP9
Listener No. Test 1
Connected
Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L73 50 43 80 33
L75 53 38 100 58
L77 73 43 100 75
L79 82 33 100 50
L81 70 43 100 27
L83 62 33 100 33
L85 68 38 100 50
L87 53 28 100 42
L89 68 50 100 38
L90 82 50 100 74
Mean Score 66.1 38.7 98.0 48.0
Intelligibility Scores (%)
159
Speaker No. MRP10
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L92 58 24 100 34
L94 47 24 100 53
L96 63 38 80 45
L98 52 24 100 45
L100 64 52 100 59
L102 43 19 80 42
L104 58 28 100 55
L106 53 28 100 57
L108 48 36 100 47
L112 68 21 100 39
Mean Score 55.4 29.4 96.0 47.6
Intelligibility Scores (%)
160
Speaker No. FRP11
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L91 70 48 100 42
L93 63 48 100 42
L95 70 33 100 51
L97 65 48 100 64
L99 68 19 80 68
L101 78 33 100 53
L103 69 43 100 64
L105 67 28 100 55
L107 75 24 100 43
L110 83 57 100 58
Mean Score 70.8 38.1 98.0 54.0
Intelligibility Scores (%)
161
Speaker No.FRP12
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L114 67 48 100 36
L116 73 48 100 45
L118 65 25 80 45
L120 93 67 100 73
L122 71 45 100 36
L124 69 44 80 43
L126 67 38 80 53
L128 65 33 100 55
L132 68 24 100 50
L134 72 31 100 25
Mean Score 71.0 40.3 94.0 46.1
Intelligibility Scores (%)
162
Speaker No. FRP13
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L109 73 32 100 71
L111 81 28 100 65
L113 79 42 100 69
L115 62 38 100 63
L117 76 43 100 67
L119 67 33 100 61
L121 75 38 80 62
L123 75 32 80 73
L125 77 38 100 61
L130 83 59 100 79
Mean Score 74.8 38.3 96.0 67.1
Intelligibility Scores (%)
163
Speaker No. FRP14
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L136 50 43 100 38
L138 53 33 100 43
L140 82 52 100 64
L142 68 24 80 34
L144 62 43 100 48
L146 63 48 100 45
L148 73 48 100 50
L152 53 43 100 29
L154 72 38 100 38
L156 62 48 80 42
Mean Score 78.6 37.3 96.0 43.1
Intelligibility Scores (%)
164
Speaker No. MRP15
Listener No. Test 1
Connected
Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L127 52 38 100 35
L129 62 26 100 51
L131 66 43 80 42
L133 64 33 100 52
L135 67 34 80 50
L137 72 39 100 58
L139 66 47 80 50
L141 64 38 100 33
L143 56 32 100 58
L150 75 54 100 43
Mean Score 64.4 38.4 94.0 47.2
Intelligibility Scores (%)
165
Speaker No. FRP 16
Listener No. Test 1
Connected Speech
Test 2
Phonemes
Test 3
Stress
Test 4
Intonation
L145 78 35 100 45
L147 65 38 100 35
L149 79 42 80 44
L151 75 47 100 46
L153 65 45 80 42
L155 69 48 100 45
L157 78 46 80 35
L158 72 42 100 40
L159 56 48 100 40
L160 80 54 100 48
Mean Score 71.7 39.6 96.0 42.0
5.2 DATA PRESENTATION II: PHONETIC/PHONOLOGICAL ANALYSIS
166
At the initial stage of the analysis, it was difficult to ascribe the instances of intelligibility
failure to any particular feature of the RP speakers‟ utterances. However, as the analysis
proceeded, a number of clear-cut categories of causes of intelligibility failure emerged. The
phonetic and phonological analysis of the utterances is presented under these categories in
order to present the data in a coherent form. A note on each category follows the list.
5.2.1 Tools for Analysis of instances of intelligibility failure .
The categories finally decided on are as follows:
1. Smoothing
2. Schwa absorption
3. Schwa suppression
4. -reduction
5. Nasal relocation
6. t-alteration
7. r-intrusion
8. t-glottalling & glottal reinforcement
9. Unilateral idiomaticity
10. Large drops in volume and salience
Smoothing
“Smoothing” is the term used to represent instances of the monophthongal realization of
certain sets of diphthongs and triphthongs in a prevocalic environment. It was observed that
whenever smoothing occurred in the speech of the RP speakers, there was no movement of
the articulators from the first to the second element of the diphthong. Instead, there was a
monophthongal realization which is similar but not identical to the starting-point of the
underlying diphthong. For example speaker FRP2 produced poor area as . None
167
of ten speakers wrote the correct response. But a few who attempted to respond wrote
„pouring‟ showing that it was the monophthongal realization of the diphthong which led to
intelligibility failure.
Schwa Suppression
“Schwa Suppression” is the term used to represent a context-free process of
monophthonging which was observed to affect the diphthongs. Schwa Suppression
involved the incorporation of a schwa into a neighbouring vowel of a more peripheral
nature. The schwa was usually assimilated by the neighbouring vowel causing the vowel to
be perceived as totally absent and it led to instances of intelligibility failure. Examples are:
„say about‟ which was produced as by MRP4 and „curry a lot‟ which was
produced as byMRP4. These utterances elicited responses such as „say down‟,
„say ban‟, „carry lot‟ , and „carry not‟.
Schwa Absorption
The term “schwa absorption” is used to describe cases where something else in the vicinity
of a schwa takes on its syllabic property but the resulting realization does not have the
openness of a vowel, i.e. the sound that is left has the articulatory qualities of a consonant
but the syllabic qualities of a vowel. Many instances of this occurred in the speech of the
RP speakers and led to intelligibility breakdown. Examples of instances involving this
category include: „people and‟ produced as by (FRP1) and „general‟ produced as
by FRP2. Responses by the Nigerians include: „people in‟ „then run‟, and many
instances of question marks and omission marks.
168
Nasal Relocation
This term is used to refer to the process of nasality which was observed with the
phonological sequence VNC (vowel + nasal + consonant), especially where the final
consonant was a voiceless stop. The phonetic reflex produced by the RP speakers was a
nasalized vowel + consonant. For example MRP3 produced the phrase “and I think my” as
and all ten listeners failed to write a word in response.
-reduction
This term is used for the process whereby initial [] in words such as „the, this, that‟
became assimilated to a previous alveolar consonant. Several phonetic realizations were
observed among the RP speakers, ranging from moving the dental to alveolar and also
including a variety of voicing characteristics such as voicing assimilation, manner
assimilation and complete assimilation. The phonetic realizations also included a
lengthened alveolar or a double consonant process referred to as „degemination‟. These
variations led to a lot of confusion among the Nigerian listeners. Examples include „from
the‟ produced as by FRP2 and „all this‟ produced as s] by MRP3.
Responses include „from now‟ and „honeys‟ respectively.
Final alteration
Final t-alteration is the term used to refer to several phonetic variants of /t/. It was observed
that word final or syllable final /t/ led to many instances of lack of intelligibility because
this consonant seemed very prone to change in the sense that it was usually realized in a
169
variety of forms or totally elided in the speech of the RP speakers. Examples include
„night- life‟ which was produced with the [t] elided as by FRP2 and „quite
steep‟ produced as by MRP5.
Glottallized /t/ and glottally reinforced /t/.
This term is used to refer to the switching of /t/ from an alveolar to a glottal articulation
This switch was observed to have affected the intelligibility of the RP speakers to
Nigerians. Examples include „beaches‟ which was produced as by MRP13. The
wrong responses included „bitching‟ and „bit cheese‟.
R-liaison
This term is used to refer to linking-r and intrusive-r processes which occurred in the
speech of the RP speakers. Linking-r refers to instances where there is an <r> in the
spelling while intrusive-r or r Sandhi refers to cases in which there was no spelling-induced
<r>. Example include: „sore elbows‟ which was produced as by MRP3 and
„shore and‟ which was produced as by MRP8.
Unusual Vocabulary Items
This refers to lexical items used by the RP speakers which were unfamiliar to the Nigerian
listeners. These are cases of „unilateral idiomaticity‟ where particular lexical items used by
the British speakers became problematic because they were not known to the Nigerians.
However, these items were not unusual in themselves but were simply unfamiliar to the
Nigerian listeners due to the different cultural settings (England/Nigeria) in which the
170
informants live. Examples include veg by MRP6 and interrailing by
MRP10.
Large Drops in Volume and Salience at Utterance End
This phrase was used to describe a progressive drop in volume which was observed in the
utterances of the male RP speakers. It caused parts of the utterances to be almost inaudible.
This speech characteristic caused the Nigerian listeners to find the speech of the male RP
speakers very difficult to understand. Most of the listeners responded with devices such as
space and omission marks.
5.3 Presentation of Data: Speaker by Speaker.
Before the speaker by speaker speech analysis, it is important to state that the conventions
of phonetic representation were strained in order to draw attention to the features of
informal speech presented here. During the analysis, it was frequently observed that it was
not possible to come up with a satisfactory phonetic representation of speech in this mode
as the stretches of obscure acoustic blur often do not permit representation on a segment by
segment basis. In order to avoid any misinterpretation of data occasioned by this and the
limitations of a second language speaker/researcher, the assistance of a native speaker was
employed in certain instances of the transcription process which required such intervention.
FRP1
Intelligibility errors due to smoothing:
there I
171
boy and
tired
train in
player
poor or
Intelligibility errors due to schwa suppression:
there
and
war and
Intelligibility errors due to schwa absorption:
similar
people and
unusual
and they
train and
172
station and
Intelligibility errors due to nasal relocation:
VN
when was it
VN
went with
VN
parents
VN
then we
VN
found our
VN
and I think
VN
went around that
VN
spent all
Intelligibility errors due to -reduction:
ALV CONS
on the
ALV CONS
173
and they
ALV CONS
all the holes
ALV CONS
and then
ALV CONS
from the
Intelligibility errors due to final alteration:
bullet (holes)
got (these)
great (photos)
t
quite (grumpy) t
Intelligibility errors due to glottal reinforcement:
right (into)
Intelligibility errors due to r-liaison:
similar in
174
war and
area and
Intelligibility errors due to unusual vocabulary items:
via
bizarre
lisp
FRP2
Intelligibility errors due to smoothing:
holiday I
year and
go out
ago I
ago it
poor area
now and then Intelligibility errors due to schwa suppression
city and
175
go away
bare
know a lot
lower and n
few of v
the amount
Intelligibility errors due to schwa absorption:
couple
balanced
was
general
and they
Intelligibility errors due to nasal relocation:
5VN
camp
VN
176
springs and
VN
unless
Intelligibility errors due to -reduction:
ALV CONS
on that
ALV CONS
even that
ALV CONS
from the
Intelligibility errors due to final alteration:
night- life
feet were
that land
Intelligibility errors due to glottal reinforcement:
country
177
Intelligibility errors due to r-liaison:
were up
lower and
MRP3
Intelligibility errors due to smoothing:
there are
air and
go and
lower and
buy a house
buy a
way around
way I
day and
Intelligibility errors due to schwa suppression
ago in
178
bizarre and
wander around
sore elbows
really afford
Intelligibility errors due to schwa absorption:
and they
Intelligibility errors due to nasal relocation:
VN
when we
VN
resent you
VN
and I think my
VN
mint so
Intelligibility errors due to -reduction:
ALV CONS
from the
ALV CONS
179
all this s
ALV CONS
all that
ALV CONS
feel that
ALV CONS
as though
Intelligibility errors due to final alteration:
got sore
that side
can‟t wander
Intelligibility errors due to glottal reinforcement:
cooped
Intelligibility errors due to r-liaison:
sore elbows
Intelligibility errors due to unusual vocabulary items:
snowboarding
swanky
180
Intelligibility errors due to large drops in volume and salience at utterance end:
I got sore elbows so it was good…. it‟s quite bizarre, really bizarre… so I spent a couple of weeks….
wander around I won‟t like to be cooped up…
MRP4
Intelligibility errors due to smoothing:
poor areas
go off
may end
they are
there are
Intelligibility errors due to schwa suppression
curry a lot
181
variety and
say about
Intelligibility errors due to schwa absorption:
cows and
fruit and
awful
Intelligibility errors due to nasal relocation:
VN
I think
VN
and I don‟t think
VN
something
Intelligibility errors due to -reduction:
ALV CONS
was there
ALV CONS
and they
182
Intelligibility errors due to final alteration:
white person
Intelligibility errors due to glottal reinforcement:
peeped
Intelligibility errors due to large drops in volume and salience at utterance end:
down the street you‟ll be peeped at…. go wearing their sort of skin revealing stuff…
which is really bizarre… helped release some baby tortoises into the sea… sort of more of traditional dress…
it just seems awful…
MRP5
Intelligibility errors due to smoothing
holiday I
stay in
holiday and
care of r
there again
183
prey on pr
Intelligibility errors due to schwa suppression:
try again
go again
Intelligibility errors due to schwa absorption:
the lake
problem
another
Intelligibility errors due to nasal relocation:
VN
and a
Intelligibility errors due to -reduction:
ALV CONS
and then
ALV CONS
because they ze
184
ALV CONS
from the
Intelligibility errors due to final alteration:
quite steep
Intelligibility errors due to large drops in volume and salience at utterance end:
you can try again… and I‟d like to go again…
out of the city and in another surrounding…. pre-occupied with going out…
to catch up and go out…
MRP6
Intelligibility errors due to smoothing:
everywhere I
try it
guy I
I also
holiday a
185
now and
Intelligibility errors due to schwa suppression:
menu and
try again
Intelligibility errors due to schwa absorption:
gilt and
oyster in
Intelligibility errors due to nasal relocation:
VN
spent a
VN
refreshment
Intelligibility errors due to -reduction:
ALV CONS
man the
ALV CONS
asked the
186
ALV CONS
but then
Intelligibility errors due to glottal reinforcement:
stretching
teach
Intelligibility errors due to unusual vocabulary items:
touristy
veg
Intelligibility errors due to large drops in volume and salience at utterance end:
a little bit of red wine…
was incredibly ornate inside…
diverse and the herbs that they use…
just for a refreshment or two…
a little bit noisy…
seven months voluntary work…
setting questions etcetera…
MRP7
Intelligibility errors due to smoothing:
187
there and
there otherwise
poor areas
try and
how effective
there is
Intelligibility errors due to schwa suppression
there at
Intelligibility errors due to schwa absorption:
you can
brothel
Intelligibility errors due to nasal relocation:
VN
percent
VN
was apparent
188
VN
there and
Intelligibility errors due to -reduction:
ALV CONS
and then
ALV CONS
was this
ALV CONS
because they
ALV CONS
and there
Intelligibility errors due to final alteration:
eight percent
quite hard
Intelligibility errors due to r-liaison:
the east
Intelligibility errors due to large drops in volume and salience at utterance end:
running under such a tight grip…
really just incredible…
too often even in Beijing….
it‟s quite hard to breathe….
189
just to get into Tibet…
earthquake was as well…
which obviously they didn‟t have…
MRP8
Intelligibility errors due to smoothing:
go off f
Intelligibility errors due to schwa suppression:
for about
shore and
nasty and
more of
airy and
tour around
Intelligibility errors due to schwa absorption:
similar (design)
currency
190
Intelligibility errors due to nasal relocation:
VN
I think
VN
brilliant (exhibition)
Intelligibility errors due to -reduction:
ALV CONS
round the
ALV CONS
down there
ALV CONS
and there
Intelligibility errors due to final alteration:
sight-seeing
Intelligibility errors due to r-liaison:
shore and
Intelligibility errors due to large drops in volume and salience at utterance end:
places to go out… from how it is now…
was quite fun…
the place was massive…
191
FRP9
Intelligibility errors due to smoothing:
wear a
their own r
Intelligibility errors due to schwa suppression:
saw amazing
Intelligibility errors due to schwa absorption:
snorkeling
thousand
kilometre
Intelligibility errors due to unusual vocabulary items:
snorkeling
Intelligibility errors due to nasal relocation:
192
VN
camp near
VN
when we went
VN
and she
VN
when we
VN
and then I
Intelligibility errors due to -reduction:
ALV CONS
and they‟re
ALV CONS
and they
ALV CONS
relaxed there
ALV CONS
in there
Intelligibility errors due to final alteration:
boat trips
Intelligibility errors due to glottal reinforcement:
t
tattoo
MRP10
Intelligibility errors due to smoothing:
193
bathing in
blower or
drier in
shower as
scare at r
buy a
Intelligibility errors due to schwa suppression:
go away
two of us
touristy and
camel leather
collapsing Intelligibility errors due to schwa absorption:
people and
pool and
social
194
people and
Intelligibility errors due to nasal relocation:
VN
and then I
VN
and it was
VN
whenever
Intelligibility errors due to -reduction:
ALV CONS
and then
ALV CONS
„cause the
ALV CONS
ended then
Intelligibility errors due to final alteration:
sat by
ate loads d
195
Intelligibility errors due to r-liaison:
weather in
Intelligibility errors due to unusual vocabulary items:
touristy
interrailing
Intelligibility errors due to large drops in volume and salience at utterance end:
the South of Morrocco…
sat by the pool…
which was good…
which was quite fun…
FRP11
Intelligibility errors due to smoothing:
row up
196
there are
say is
Intelligibility errors due to schwa suppression:
twenty at
tour of v
the only
actually afford
wander around
venture out
Intelligibility errors due to schwa absorption:
properly
famous
197
sucker and
travel again
interrailing
to go
bubble and
Intelligibility errors due to nasal relocation:
VN
and my
VN
and went
VN
point and
VN
prudent in
Intelligibility errors due to -reduction:
ALV CONS
when this
ALV CONS
loved that
ALV CONS
because they
198
Intelligibility errors due to final alteration:
upset because
ate fish
FRP12
Intelligibility errors due to smoothing:
there I
wear a
idea about r
Intelligibility errors due to schwa suppression:
actually it
fairly
199
westerner and
money and
to eat
idea about
chilly and
were all
Intelligibility errors due to schwa absorption:
westerner and
westerner I‟m
summer and
Intelligibility errors due to nasal relocation:
VN
long way
VN
I think my
Intelligibility errors due to -reduction:
200
ALV CONS
treat them
ALV CONS
that they
ALV CONS
and they‟re
Intelligibility errors due to final alteration:
FRP13
Intelligibility errors due to smoothing:
holiday and
clear and r
year or
nearby and
now I
area and
stay in
201
lay out
way as
Intelligibility errors due to schwa suppression:
you about
wealthy and
whether it
the other
wore an
way as
Intelligibility errors due to schwa absorption:
recent
to get
celebrity
202
rural and
Intelligibility errors due to nasal relocation:
VN
spent a
VN
recent years
VN
spent a
VN
different in
Intelligibility errors due to -reduction:
ALV CONS
and they
Intelligibility errors due to final alteration:
quite rural
Intelligibility errors due to glottal reinforcement:
beaches
FRP14
203
Intelligibility errors due to smoothing:
stay in
their animals z
there I
here I
my exams
our anniversary
power
Intelligibility errors due to schwa suppression:
two and
extra energy
for about
summer I
my exams
204
Intelligibility errors due to schwa absorption:
asylum
criminal
moral
Intelligibility errors due to nasal relocation:
VN
I think
VN
then went
VN
went in
Intelligibility errors due to -reduction:
ALV CONS
in this
ALV CONS
all this
Intelligibility errors due to final alteration:
t
quite cold
Intelligibility errors due to glottal reinforcement:
205
teach
MRP15
Intelligibility errors due to smoothing:
there and
buy a
idea as r
go on
know everything
violence
Intelligibility errors due to schwa suppression:
more of
more and n
206
community and
clever and
me about
Intelligibility errors due to schwa absorption:
electoral
again and
colonizers
prison and
Intelligibility errors due to nasal relocation:
VN
spent the
Intelligibility errors due to -reduction:
ALV CONS
problems there
ALV CONS
and then
ALV CONS
207
all these
Intelligibility errors due to glottal reinforcement:
beach
Intelligibility errors due to large drops in volume and salience at utterance end:
also did go to the rift valley…
tree all afternoon…
they had no money… they‟ve got nothing to do…
tree all afternoon…
about that…
FRP16
Intelligibility errors due to smoothing:
way of
day and
208
enjoy and
row and
also about
try and
there every
science
Intelligibility errors due to schwa suppression:
nanny and
pasta and
colorfully and
more of
camel
also about
row and
209
Intelligibility errors due to schwa absorption:
along
and they
rational
Intelligibility errors due to nasal relocation:
VN
garment as
VN
to think
Intelligibility errors due to -reduction:
ALV CONS
in their
ALV CONS
and the
Intelligibility errors due to unusual vocabulary items:
veg
5.4 Mean Intelligibility scores for Test 1 (Connected Speech)
The percentage values presented below were arrived at by first converting the individual
scores obtained by each Nigerian speaker into percentages. The percentage scores were
then added and the total was divided by 10 (the Nigerian listeners were in groups of 10) to
arrive at the average score with each RP speaker. In other words, the percentage values
210
below are the average scores of 10 listeners involved for each particular RP speaker. Table
1.1 below shows the intelligibility scores for all the RP speakers. The scores are based on
the connected speech test as spontaneous speech performance is usually taken to be the
most reliable indication of intelligibility. The scores are based on the number of units
correctly identified by the listener.
Table 1.1: Mean Intelligibility scores for Test I (Connected speech)
Speaker ID Scores in %
FRP1 64.5
FRP2 64.6
MRP3 47.2
MRP4 45.1
211
MRP5 56.2
MRP6 50.7
MRP7 53.6
MRP8 61.5
FRP9 66.1
MRP10 55.4
FRP11 70.8
FRP12 71.0
FRP13 74.8
FRP14 78.6
MRP15 64.4
FRP16 71.7
Average 62.2
The results presented in Table 1.1 show that the average intelligibility level of the Received
Pronunciation (RP) accent to Nigerians is 62.2%. The figure is based on the calculation of
the number of items correctly received by the Nigerian assessors in the continuous speech
test. From the figures above, it can be observed that the RP speakers‟ performance vary
considerably in levels of intelligibility (as assessed by Nigerian listeners). The RP
speakers‟ scores vary from 45.1% (MRP4) to 78.6% (FRP14) with an average score of
62.2%. Seven (7) of the RP speakers scored lower than the average score of 62.2%, while
nine (9) of them scored higher than the average score. The scores for this test (Connected
speech) are relatively high, possibly because contextualized speech stimuli has a lot of
clues and redundancy. The percentage value of intelligibility was calculated based on the
connected speech test (Test 1) because the spontaneous speech test was considered to be
the closest to real life interactional communication and was therefore assumed to be the
212
criterion of fundamental importance in assessing the English speakers level of
intelligibility. The implication of this percentage value is that the Nigerian listeners
involved in this study correctly received about 60% (sixty percent) of the British speakers‟
utterances in spontaneous unscripted speech which is highly similar to real life
communication.
Further investigations of the instances of intelligibility failure showed the features of RP
which caused lack of intelligibility. Following the phonetic/phonological analysis of the
instances of intelligibility failures, the ten features of RP which led to lack of intelligibility
were identified. The following table (Table 1.2) presents these features and the figures of
intelligibility failure relating to each feature.
213
Table 1.2
Causes of intelligibility failure
(Figures relate to the number of listener failures in each category)
Speaker ID
RP speech feature
FR
P14
FR
P13
FR
P16
FR
P12
FR
P11
FR
P9
FR
P2
FR
P1
MR
P15
MR
P8
MR
P5
MR
P10
MR
P7
MR
P6
MR
P3
MR
P4
Tota
l
1. Smoothing 2 3 1 6 7 8 5 3 9 6 6 6 6 6 9 6 89 2. Schwa suppression 1 8 6 2 5 7 3 6 6 5 2 2 5 7 5 2 72
3. Schwa absorption 3 3 2 3 3 3 3 7 4 4 2 2 4 5 1 6 55
4. -reduction 5 3 3 3 2 2 2 3 1 3 4 3 3 3 5 5 50
5. Nasal relocation 1 2 2 1 3 2 3 4 4 1 3 2 3 3 4 8 46
6.Large drops in volume and salience at utterance end
4 5 3 5 4 5 4 6 36
7. t-alteration 4 1 1 1 1 1 2 1 2 2 2 3 3 4 28
8. t-glottalling & glottal reinforcement 1 1 1 1 1 1 2 1 1 1 1 12
9. Unilateral idiomaticity 1 1 2 2 2 3 11
10. r-liaison 1 1 1 2 1 3 9
Total number of failures 18 20 17 16 22 23 18 25 30 25 24 26 30 35 35 44 408
214
5.5 Rank Order Presentation of Percentage Scores For Test 1
The following section presents the analysis of the scores obtained in the connected
speech test (Test 1). Based on the ten groupings of our phonetic analysis of the previous
section, the section shows the instances of intelligibility failure in numbers. The
individual speaker‟s errors listed in the previous section are summarized in rank order
according to category. The following table (1.3) shows the causes of intelligibility failure
among the RP speakers. At the foot of this table, it will be seen that there is some
variance between the order based on the speakers‟ percentage scores and the order based
on the number of listener failures. This is because the speakers‟ percentage score, as
explained previously was based on the number of correct units. In some cases, incorrect
units contained more than one error.
Among the RP speakers, the process of smoothing is the largest single cause of
intelligibility failure, accounting for 21.8% of the instances of lack of intelligibility. This
is closely followed by schwa suppression (17.6%), schwa absorption (13.4%), -
reduction (12.2%), nasal relocation (11.2%), large drops in volume and salience at
utterance end (8.8%), t-alteration (6.8%), t-glottalling & glottal reinforcement (2.9%),
unilateral idiomaticity (2.6%) and lastly, r-liaison (2.2).
215
Table 1.3
Summary of speakers showing % errors
FRP14 FRP13 FRP16 FRP12 FRP11 FRP9 FRP2 FRP1 MRP15 MRP8 MRP5 MRP10 MRP7 MRP6 MRP3 MRP4 Sub
total %
1. Smoothing 2 3 1 6 7 8 5 3 9 6 6 6 6 6 9 6 89 21.
8 2. Schwa
suppression
1 8 6 2 5 7 3 6 6 5 2 2 5 7 5 2 72 17.
6 3. Schwa
absorption
3 3 2 3 3 3 3 7 4 4 2 2 4 5 1 6 55 13.
4
4. -reduction 5 3 3 3 2 2 2 3 1 3 4 3 3 3 5 5 50 12.
2 5. Nasal relocation 1 2 2 1 3 2 3 4 4 1 3 2 3 3 4 8 46 11.
2 6.Large drops in
volume and salience at
utterance end
4 5 6 6 7 7 4 4 36 8.8
7. t-alteration 4 1 1 1 1 1 2 1 2 2 2 3 3 4 28 6.8
6 8. t-glottalling &
glottal
reinforcement
1 1 1 1 1 1 2 1 1 1 1 12 2.9
4
9. Unilateral
idiomaticity
1 2 2 2 3 11 2.6
9 10. r-liaison 1 1 1 2 1 3 9 2.2 Total number of
failures
18 20 20 21 22 23 24 25 26 26 28 28 30 30 35 38 408
Speaker’s % score
78.6 74.8 71.7 71.0 70.8 66.1 64.6 64.5 64.4 61.5 56.2 55.4 53.6 50.7 47.2 45.1
216
One general characteristic which was found to contribute most highly to lack of
intelligibility is extensive reduction in unstressed syllables. This generally included
vowel reductions involving additions, changes and loss of characteristics. It was found
that this feature led to many Nigerians being either completely at a loss or writing
incorrect responses during the listening sessions. These specific features are smoothing
and schwa suppression/monophthongization. These features led to almost half of the
total number of instances of intelligibility failure in connected speech. Smoothing is the
monophthongal realization of certain set of diphthongs and triphthongs in a prevocalic
environment. Examples from our analysis include the realization of poor in poor (in) as
which led to intelligibility breakdown as most of the Nigerian speakers
responded with pour, bore, pouring or blanks, showing that they did not understand the
speaker. Other examples include the following: boy (and) ; player
; buy (and) ; wear (a) . Instances
of schwa suppression/monophthongization include: war and ;
few of v, bizarre and sore elbows
. Features involving the segmental features accounted for another one-third
of the instances of intelligibility failure. This include schwa absorption, -reduction and
t-alteration. Examples of schwa absorption include: saw amazing
; two of us ; pool and . Instances of -
reduction are: ALV CONS ; in there ALV CONS ;
Examples of t-alteration include: sat by ; The features of nasal
relocation, glottal reinforcement and r-liaison also had a significant impact on
intelligibility. Examples are: and then, camp near VN beaches
217
, and drawing . Two RP speech features of
unilateral idiomaticity and large drops in volume and salience were responsible for about
a tenth of the instances of intelligibility failure . Examples from the speech of the RP
speakers include: veg, interrailing, touristy and swanky for unusual vocabulary items
which were classified as unilateral idiomaticity. ‘Large drops in volume and salience‟
refers to a peculiar habit of fading off at the end of an utterance which occurred in the
speech of only the male RP speakers and which left many of the Nigerians at a loss. It
was observed that this was not an idiosyncratic speech feature because it was found to
occur in the speech of all the male speakers.
5.6 INTELLIGIBILITY SCORES FOR TESTS 2-4
The scores of the tests involving relatively decontextualized speech are presented below.
5.6.1 Presentation of Scores Obtained in the Phoneme Test (Test 2)
The following tables present the data for the intelligibility of the RP segmental features in
relatively decontextualized speech (word pairs). The presentation of the test results is
shown in three different tables: Table IIa: Vowels and level of intelligibility; Table IIb:
Diphthongs and level of intelligibility; Table IIc: Triphthongs and intelligibility; and
Table IId: Consonant contrasts and level of intelligibility.
218
Table IIa: Vowels and level of intelligibility
Word Pairs
Number of intelligibility
failures
Total number of utterances
Vowel Contrast
1. 294 320
2. 256 320
3. 230 320
4. 224 320
5. 93 320
6. 81 320
Total 1178 1920
These results show the descending order of vowels and number of times in which
intelligibility failure occurred. Out of a total of one thousand, nine hundred and twenty
(1920) utterances/responses, intelligibility failure occurred one thousand, one hundred
and seventy eight times (1178). Certain phonemic contrasts were found to lead to a
higher number of instances of intelligibility failure than others. With the RP vowels, the
central vowels in particular were discovered to have the highest contribution to
unintelligibility, while phonologically absent segments and vowel length were also found
to cause confusion in many instances. For example, the RP speakers produced
for the pair of words: fur and for and many Nigerians responded with devices such as
space and omission marks. Other responses included: for in place of fur and vice versa,
from, fun and fall.
219
Chart Ia: Intelligibility level of RP vowel contrasts to Nigerians in relatively
decontextualised speech (Phoneme Test)
The scores obtained in the test of phonemes (vowels) show that there was a high
intelligibility failure rate in the test involving relatively decontextualised speech. The bar
chart below highlights the intelligibility vs. unintelligibility rate for RP vowels.
294
256
230 224
9381
-30
20
70
120
170
220
270
320
1 2 3 4 5 6
1
2
3
4
5
6
Key: 1
2
3
4
5
6
220
These results show the descending order of vowels and number of times in which
intelligibility failure occurred. Out of a total of one thousand, nine hundred and twenty
(1,920) utterances/responses, intelligibility failure occurred one thousand, one hundred
and seventy eight times (1,178).
221
Chart Ib: Intelligibility level of RP vowel contrasts to Nigerians in relatively
decontextualised speech (phonemes)
The scores obtained in the test of phonemes (vowels) show that there was a high
intelligibility failure rate in the test involving relatively decontextualised speech. The
percentage pie chart below highlights the percentage intelligibility vs. unintelligibility
rate for RP vowels.
RP vowels: %
intelligible
39%
RP vowels: %
unintelligible
61%
RP vowels: % intelligible
RP vowels: % unintelligible
222
Table IIb: Diphthongs and level of intelligibility
Phoneme
(Diphthongs)
Number of
intelligibility failures
Total number of
utterances
Vowel contrast
1. 289 320
2. 266 320
3. 228 320
4. 162 320
5. 76 320
Total 1020 1600
The table above shows the descending order of diphthongs and number of times
intelligibility failure occurred. Of a total possible number of one thousand six hundred
(1,600) times, intelligibility failure with the Nigerian listeners occurred one thousand, and
twenty one (1,021) times. The peculiar RP process of monophthongization of diphthongs
was observed to be the major cause of lack of intelligibility. For example, were
monophthongized into . The items for these vowels were beer and bear. The
responses included bee, be, bin, bill etcetera for beer and beg, bell, bend, bad etcetera for
bear. Incidentally, both diphthongs have the realization // in NEA which is quite far in
phonetic space from the RP realizations of these two diphthongs.
223
Chart IIa: Intelligibility level of RP vowel contrasts (diphthongs) to Nigerians in
relatively decontexualized speech.
The scores obtained in the test involving phonemes show a high rate of intelligibility
failure in the test involving relatively decontextualised stimuli (word pairs involving
diphthongs). The percentage bar chart below highlights the intelligibility vs.
unintelligibility rate for RP diphthongs.
289
266
228
162
76
-30
20
70
120
170
220
270
320
1 2 3 4 5
1
2
3
4
5
Key: 1
2
3
4
5
224
Chart IIb: Intelligibility level of RP vowel contrasts (diphthongs) to Nigerians in
relatively decontexualized speech.
The scores obtained in the tests of minimal pairs show a high rate of intelligibility failure
in the test involving relatively decontextualised stimuli (minimal pairs involving
diphthongs). The percentage pie chart below highlights the percentage intelligibility vs.
unintelligibility rate for RP diphthongs.
RP diphthongs: %
intelligible
36%
RP diphthongs: %
unintelligible
64%
RP diphthongs: % intelligible
RP diphthongs: %
unintelligible
225
Table IIc: Triphthongs and intelligibility
This table shows the descending order of triphthongs and number of times intelligibility
failure occurred. Of a total number of one thousand two hundred and eighty (1280)
utterances, intelligibility failure with the Nigerian listeners occurred one thousand one
hundred and six times (1106) times. The peculiar RP process of smoothing was observed
to be responsible for the high level of lack of intelligibility of the RP triphthongs. For
instance, the triphthongs in items such as fire, tire, layer and payer were realized as /
for the first pair of items and /e for the second pair. The responses included far, farm,
tar, tap, lay, lame, prey and pay which were all wrong.
Phoneme
(Triphthongs)
Instances of intelligibility
failure
Total number of
utterances
Vowel contrast
1. 289 320
2. 278 320
3. 272 320
4. ; 267 320
Total 1106 1280
226
Chart IIIa: Intelligibility of vowels (triphthongs) in relatively decontextualized
speech
The scores presented in Table 11c above show that there was a high rate of intelligibility
failure in the test involving relatively decontextualised stimuli (word pairs involving
triphthongs). The bar chart below highlights the intelligibility vs. unintelligibility rate for
RP triphthongs.
289
278
272
267
240
250
260
270
280
290
300
310
320
1 2 3 4
1
2
3
4
Key: 1
2
3
4 ;
227
Chart IIIb: Intelligibility of vowels (triphthongs) in relatively decontextualized
speech
The scores presented in Table 11c above show that there was a high rate of intelligibility
failure in the test involving relatively decontextualised stimuli (word pairs involving
triphthongs). The percentage pie chart below highlights the percentage intelligibility vs.
unintelligibility rate for RP triphthongs.
RP triphthongs: %
intelligible
14%
RP triphthongs: %
unintelligible
86%
RP triphthongs: % intelligible
RP triphthongs: %
unintelligible
228
Table IId: Consonant contrasts and level of intelligibility
Phoneme (consonants) Instances of
intelligibility
failure
Total number of utterances
Consonant contrasts
1. Final 125 320
2. Medial 119 320
3. Initial ; 79 320
4. Initial 12 320
Total 335 1280
The table shows consonant contrasts, their positions in the word (i.e. word initial
position, word medial position and word final position) and the number of times that
intelligibility failed in the phoneme test (Test 11). Of a total of one thousand two hundred
and eighty (1280) possible times, intelligibility failure occurred three hundred and thirty
five times (335) times. The instances of intelligibility failure for the RP consonants were
few and the commonest errors were concerning the items lose; loose which many of the
Nigerians could not correctly identify and also the items hill; ill which were also not
correctly identified.
Chart IVa: Intelligibility of consonants in relatively decontextualized speech
229
The scores presented in Table 11d above show the rate of intelligibility failure in the test
involving relatively decontextualised stimuli (minimal pairs involving consonant
contrasts). The bar chart below highlights the intelligibility vs. unintelligibility rate for
RP consonants.
125 119
79
12
-30
20
70
120
170
220
270
320
1 2 3 4
1
2
3
4
Key: 1
2
3 ;
4
The results presented below are the listener scores for Test III (Sentence stress test).The
scores were averaged over ten (10) listeners for each speaker and then converted to
percentages.
Chart IVb: Intelligibility of consonants in relatively decontextualized speech
230
The scores presented in Table 11d above show the rate of intelligibility failure in the test
involving relatively decontextualised stimuli (minimal pairs involving consonant
contrasts). The percentage pie chart below highlights the percentage intelligibility vs.
unintelligibility rate for RP consonants.
RP consonants: %
intelligible
74%
RP consonants: %
unintelligible
26% RP consonants: % intelligible
RP consonants: %
unintelligible
The results presented below are the listener scores for Test III (Sentence stress test).The
scores were averaged over ten (10) listeners for each speaker and then converted to
percentages.
231
Table III: Mean intelligibility scores for contrastive/emphatic stress
Speaker ID Scores in (%)
FRP1 98.0
FRP2 98.0
MRP3 96.0
MRP4 98.0
MRP5 96.0
MRP6 98.0
MRP7 96.0
MRP8 100
FRP9 98.0
MRP10 96.0
FRP11 98.0
FRP12 94.0
FRP13 96.0
FRP14 96.0
MRP15 94.0
FRP16 96.0
Average 96.7
The average score for this test is 96.7%. The high score indicates that the Nigerian
listeners were able to correctly identify almost all the emphatic stress positions in the
speech of the RP speakers. With reference to the nuclear stress results presented in Table
III, it was found that in the area of contrastive/emphatic stress positions in sentences, the
average degree of intelligibility was calculated to be 96.7%, implying that the Nigerian
listeners were able to correctly identify the contrasts in practically all the utterances. The
232
British speakers indicated contrast on some utterances, each consisting of a single high
falling nuclear tone on the first, second or third stressed syllable. Contrast was achieved
by placing nuclear stress on the appropriate stressed syllable. The ability to identify the
part of the sentence being contrasted is similar to that required for the identification of
nuclear stress placement positions which distinguish word classes such as nouns/verbs
(e.g. export (noun) and export (verb). Since most educated Nigerians are quite proficient
in distinguishing between such word classes in utterances, it was not very surprising that
very high scores were obtained in this test. Besides this, the nuclear tunes employed by
the British speakers to signify contrasts were relatively simple tunes which Nigerians are
quite familiar with. For instance, the Nigerians correctly identified the emphatic stress
positions (in capitals) in the following statements:
(a) No, the JEALOUS woman slapped her husband‟s lover.
(b) No, the jealous WOMAN slapped her husband‟s lover.
(c) No, the jealous woman SLAPPED her husband‟s lover.
(d) Table IV: Mean intelligibility scores for attitudinal intonation.
(The scores are averaged over the ten listeners involved for each particular speaker.)
Speaker ID Scores in %
233
FRP1 53.2
FRP2 61.9
MRP3 58.7
MRP4 55.2
MRP5 55.6
MRP6 58.5
MRP7 68.0
MRP8 46.1
FRP9 48.0
MRP10 47.6
FRP11 54.0
FRP12 46.1
FRP13 67.1
FRP14 43.1
MRP15 47.2
FRP16 42.0
Average 53.2
Test IV was designed to elicit varying intonation contours based on different attitudes
expressed by various sentences. The Nigerian listeners‟ performance in this test was
average. The mean score was 53.2% and the range was between 42% and 68%. With
reference to Table 4, an average intelligibility rate of 53.2% was observed for the
234
intelligibility level of RP intonation contours to Nigerians. This implies that the Nigerians
were able to correctly identify just about half of the various attitudes conveyed by the
various intonation contours used by the RP speakers. Generally, the Nigerians were able
to identify the intonation patterns conveying a simple question. This utterance involved
the use of a relatively simple, unidirectional intonation pattern. However, utterances
involving more complex bi-directional tones left many of the Nigerians at a loss. The
items which left many Nigerians completely defeated were those involving
complete/incomplete listing of choices, and statements conveying different attitudes
including sincerity and boredom. In addition to other areas that have been pointed out
earlier, the area of intonation contours is another one in which the NEA exhibits
divergence from the RP accent. From a functional perspective, while intonation tunes
seem to serve a lone purpose of provision of information in Nigerian English, they serve
the dual function of provision of information and expression of attitude in the Standard
British English accent. Therefore, compound and complex intonation tunes are hardly
employed by Nigerians, previous analysis of Nigerian speakers‟ utterances have shown a
preponderance of simple tunes. It may be due to lack of familiarity with compound and
complex intonation contours that is responsible for the average score obtained in this test.
Table V below shows the scores of the 16 Nigerian Diasporic dwellers (UK) in the four
tests. These listeners were selected to enable us examine the effect of familiarity on
intelligibility.
Table V: The effect of the listener’s degree of familiarity on intelligibility
Listener ID Test 1 Test 2 Test 3 Test 4
235
Connected Speech Minimal Pairs Sentence Stress Intonation
L10 77 50 100 75
L20 85 55 100 75
L30 67 54 100 73
L40 71 51 100 70
L50 77 57 80 82
L60 74 63 100 83
L70 73 57 100 79
L80 81 58 100 69
L90 82 50 100 74
L100 64 52 100 59
L110 83 57 100 58
L120 93 67 100 73
L130 83 59 100 79
L140 82 52 100 64
L150 75 54 100 43
L160 80 54 100 48
Average 77.9% 55.6% 98.7% 69.0%
Sixteen (16) out of the one hundred and sixty (160) Nigerian listeners involved in this
study were Nigerian undergraduates in the UK. These 16 are representative of Nigerian
diasporic dwellers (UK). The scores indicate that they correctly identified a higher
236
number of units than the other one hundred and forty four (144) Nigerians who are
resident here in Nigeria.
5.7 ANALYSIS OF DATA: 1
In this section, detailed explanation is made of the phonetic causes of intelligibility
failure according to the features that have been identified in the previous section. As
stated earlier, performance in connected speech is considered to be the fundamental basis
on which statements on intelligibility should be made. Therefore, the results of the
connected speech test (Test 1) are discussed first. Thereafter, the results obtained in the
tests involving relatively decontextualised speech are discussed in order to have greater
insight into how intelligibility is negotiated in the face of accent variation.
5.7.1 Causes of intelligibility failure in Connected Speech
Smoothing
“Smoothing” is the term used to represent instances of the monophthongal realization of
certain sets of diphthongs and triphthongs in a prevocalic environment. Smoothing
accounted for 21.8% of the instances of intelligibility failure in connected speech,
translating to about a quarter of all the instances of breakdown in communication. A
diphthong is a complex vowel and the pronunciation has a diminuendo (or falling) effect
in that the prominence decreases as we pass from one element to the second. However, it
was found that when smoothing occurred, there was no movement to the second element
of the diphthong. Instead, there was a monophthongal realization which is similar but not
identical to the starting-point of the underlying diphthong. Smoothing was also found to
237
apply across word boundaries. Furthermore, the following observations were made
concerning smoothing.
First, smoothing was most common with the diphthongs and triphthongs with vowels
having a front mid to close quality and also vowels with a back mid to close quality.
These vowels are ;;;;;. Smoothing was found to apply most readily to these
vowels and the resulting monophthongal realizations had a range of qualities which the
Nigerian listeners found confusing.
Secondly, with the triphthongs, smoothing led to the reduction of the number of syllables
by one because in some cases, it involves the incorporation of a schwa into the
neighbouring vowel. Due to this assimilation of the schwa by a neighbouring vowel,
perceived syllabicity was not preserved and this led to many instances of intelligibility
failure. Besides this, in many of the realizations involving smoothing, the remaining
vowel (monophthong or diphthong) seemed longer than normal.
Apart from this, the diphthongs ; and showed a lot of variability as
not only were the second elements missing, the first elements had a different realization
and this also caused confusion for the Nigerian listeners. Lastly, for some of the
diphthongs, two realizations were observed:
(a) in instances when <r> begins the following word, there was a linking r in the
pronunciation; and
(b) in other instances without a spelling <r>, an intrusive r was often added.
238
These realizations further led to lack of intelligibility. However it was generally observed
that the usual effect of smoothing was of the production of a monophthong with the
phonetic quality of the starting-point of the underlying diphthong. A selection of
instances where smoothing led to intelligibility breakdown is presented below:
Instances of smoothing
tired (FRP1)
there I (FRP1)
poor or (FRP1)
air and (MRP3)
buy a (MRP3)
row up (FRP11)
nearby and (FRP13)
Smoothing with a different vowel quality
there I (FRP1)
poor area (FRP2)
everywhere I (MRP6)
wear a (FRP9)
blower or r (MRP10)
Smoothing with linking/intrusive r
239
care of r (MRP5)
idea as r (MRP15)
scare at r (MRP10)
their own r (FRP9)
clear and r (FRP13)
idea about r (FRP12)
Schwa Suppression
“Schwa Suppression” is the term used to represent a context-free process of
monophthonging which was observed to affect the diphthongs. It commonly affected the
centring diphthongs ,,, as these diphthongs were generally realized as long
monophthongs. This process was responsible for 17.6% of all the instances of
intelligibility failure in the Connected Speech test (Test 1). This is significant as it
represents about a quarter of all instances of breakdown in communication. Schwa
Suppression involved the incorporation of a schwa into a neighbouring vowel of a more
peripheral nature. The schwa was usually assimilated by the neighbouring vowel. It was
also observed that the diphthongs which were subsequently reduced to monophthongs
were sometimes derived from , and that this process occurred very commonly
across word boundaries.
Apart from this general monophthonging process, a duality and multiplicity of processes
involving Schwa Suppression and other processes were also observed to lead to lack of
240
intelligibility. The other processes include elision, assimilation, syllabic consonant
formation and the addition of intrusive r. Several instances are presented below:
Instances of Schwa Suppression
the amount (FRP2)
bare (FRP2)
know a lot (MRP3)
really afford (MRP3)
say about (MRP4)
curry a lot (MRP4)
go again
try again (MRP6)
saw amazing (FRP9)
go away (MRP10)
tour of v (FRP11)
more of (MRP15)
Schwa Suppression with elision and intrusive r
for about (FRP14)
Schwa Suppression with linking r
lower and n (FRP 2)
241
idea about (FRP12)
Schwa Suppression with syllabic consonant formation and assimilation
more and n (MRP15)
Schwa Absorption
The term “schwa absorption” is used to describe cases where something else in the
vicinity of a schwa takes on its syllabic property but the resulting realization does not
have the openness of a vowel, i.e. the sound that is left has the articulatory qualities of a
consonant but the syllabic qualities of a vowel. This process was responsible for 13.4% of
the instances of intelligibility breakdown.
The process was observed to occur in several environments:
(a) In the environment of a lateral
(b) In the environment of a nasal
(c) In the environment of syllabic „r‟ and „w‟
(d) Schwa absorption involving voiceless vowels
In all the environments stated above, the loss of the schwa was observed but the number
of syllables remained constant due to the formation of the syllabic consonant. Instances
of this process are identified below.
Schwa absorption involving laterals
242
It has long been an axiom of English phonology that certain sound can be syllabic under
the right circumstances. In the instances below, it was observed that the „l‟ became
syllabic following the loss of the schwa. However, the loss of the schwa did not affect the
number of syllables in the word/phrase (both within a word and across word boundaries).
It was as if the reduced vowel was simply a syllabic place holder. However, the reduction
and subsequent syllabification of „l‟ led to many instances of lack of intelligibility.
Instances in the speech of the RP speakers are presented below:
people and (FRP1)
unusual (FRP1)
general (FRP2)
awful (MRP3)
the lake (MRP5)
brothel (MRP7)
similar (design) (MRP8)
snorkeling (FRP9)
kilometre (FRP9)
pool and (MRP10)
bubble and (FRP11)
rural and (FRP13)
Schwa absorption involving nasals
and they (FRP1)
243
cows and (MRP3)
fruit and (MRP3)
problem (MRP5)
gilt and (MRP6)
currency (MRP8)
thousand (MRP9)
Schwa absorption involving syllabic „r‟ and „w‟
It was observed that what was produced by the RP speakers as a syllabic„w‟ is similar to
a rounded schwa. Thus the vowel and consonant gestures overlap completely and the
resulting segment sounded like neither a consonant nor a vowel, thereby causing a lot of
confusion for the listeners.
Instances of syllabic „r‟ occurred across word boundaries in the sequences containing []
+ [r]. Examples are produced below:
interrailing (FRP11)
electoral (MRP15)
Schwa absorption involving voiceless vowels
to go FRP11
to get FRP13
again and MRP15
244
Nasal relocation
This term is used to refer to the process of nasality which was observed with the
phonological sequence VNC (vowel + nasal + consonant), especially where the final
consonant is a voiceless stop. The phonetic reflex produced by the RP speakers was a
nasalized vowel + consonant. It is similar to the homorganic nasal process which usually
occurs in final -nt clusters. This process is common among native speakers of English. It
is usually explained as a re-timing of velum lowering and oral closure, thus nasalization
begins earlier than one might expect from the citation shape and articulator contact is
later. The segment deletion which accompanied this process also added to the confusion
experienced by the Nigerians. Instances are provided below:
unless FRP2
when we MRP3
and I think my MRP3
I think MRP4
something MRP4
and a MRP5
percent MRP7
camp near FRP9
and went FRP11
I think my FRP12
245
-reduction
This is the process whereby initial [] in words such as the, this, that became assimilated
to a previous alveolar consonant. Several phonetic realizations were observed among the
RP speakers, ranging from moving the dental to alveolar and also including a variety of
voicing characteristics such as voicing assimilation, manner assimilation and complete
assimilation. The phonetic realizations also included a lengthened alveolar or a double
consonant process referred to as „degemination‟. These variations led to a lot of
confusion among the Nigerian listeners.
Instances are produced below:
on that FRP2
from the FRP2
all this s] MRP3
as though MRP3
and they MRP4
down there MRP8
and there MRP8
and then MPR10
when this FRP11
and they‟re MRP12
in this FRP14
and the FRP16
246
Final alteration
Final t-alteration is the term used to refer to several phonetic variants of /t/. It was
observed that word final or syllable final /t/ led to many instances of lack of intelligibility
because this consonant seemed very prone to change in the sense that it was usually
realized in a variety of forms or totally elided in the speech of the RP speakers. Although
it was hardly elided in syllable initial position (probably due to the fact that most elisions
occur in the least obvious part of the syllable - the final position). It was observed that
when /t/ occurred in this position, it was quite obscure.
Two main variants of /t/ in the speech of the RP speakers which was found to affect
intelligibility to the Nigerian listeners were the non-glottalized and glottalized phonetic
variants. There are three non-glottalized variants of /t/ and they are: /t/ without audible
release, tapped /t/ and elided /t/. /t/ without audible release is the variant which involves
tongue transition to alveolar position, followed by a hold stage, without an aspirated
release. This variant occurred before stops which made the closure for the following
consonant masked by the release of the preceding alveolar stop. Tapped /t/ occurred pre-
vocalically in word-final environment while elided /t/ occurred mainly across word
boundary. Elided /t/ was also observed when the /t/ was preceded and followed by
consonants especially labials.
Glottalised phonetic variants of /t/ were also observed and will be discussed in the
subsequent section. Instances of the alteration of /t/ are presented below:
247
/t/ without audible release
quite (grumpy) t FRP1
night- life FRP2
got sore MRP3
that side MRP3
white person MRP4
quite steep MRP5
eight percent MRP7
sight-seeing MRP8
boat trips FRP9
sat by MRP10
ate loads MRP10
ate fish FRP11
quite cold FRP14
Tapped /t/
bullet (holes) FRP2
feet were FRP2
can‟t wander MRP3
quite hard MRP7
248
Elided /t/
great (photos) FRP1
that land FRP2
quite rural FRP13
Glottallized /t/ and glottally reinforced /t/.
This term is used to refer to the switching of /t/ from an alveolar to a glottal articulation
This switch was observed to have affected the intelligibility of the RP speakers to
Nigerians. Both glottalling and glottal reinforcement were observed to have occurred in
the following range of syllable-final environments: before obstruents, before l, m and w,
when ,,, are in syllable final position, or are preceded by a vowel, a liquid or a
nasal. Instances are presented below:
right (into) FRP1
country FRP2
cooped MRP3
peeped MRP4
quite steep MRP5
teach MRP6
tattoo FRP9
ate loads d MRP10
beaches MRP13
teach MRP14
249
R-liaison
This term is used to refer to linking-r and intrusive-r processes which occurred in the
speech of the RP speakers. Linking-r refers to instances where there is an <r> in the
spelling while intrusive-r or r Sandhi refers to cases in which there was no spelling-
induced <r>. Generally, /r/ was inserted after certain vowels before a following vowel,
optionally across a morpheme or word boundary. The vowel endings to which an /r/ link
was added include /, / and also single/complex vowels containing final such as
, ,,,. R-insertion also occurred word internally (internal intrusive-r) and
this affected intelligibility to the Nigerians. Instances are presented below:
war and FRP1
area and FRP1
were up FRP2
sore elbows MRP3
the east MRP7
shore and MRP8
weather in MRP11
drawing MRP15
Unusual vocabulary items
This refers to lexical items used by the RP speakers which were unfamiliar to the
Nigerian listeners. These are cases of „unilateral idiomaticity‟ where particular lexical
250
items used by the British speakers became problematic because they were not known to
the Nigerians. However, these items were not unusual in themselves but were simply
unfamiliar to the Nigerian listeners due to the different cultural settings
(England/Nigeria) in which the informants live. These items represented salient cultural
references which of course the British used freely in expressing certain cultural ideas but
these items completely defeated the Nigerians. This was signaled by the use of devices
such as omission marks, space and question marks. However, some of the Nigerians did
make orthographic representations which were phonologically sensible, suggesting that
the listeners were not struggling so much with accent as with culture. Cultural factors
appear to be as important as accent as the interviews were close to real life interactional
communication and therefore they were not culturally vacuous. Instances are presented
below:
snowboarding FRP2
swanky FRP2
touristy MRP6
veg MRP6
interrailing MRP10
Large drops in volume and salience at utterance end
This phrase was used to describe a progressive drop in volume which is a feature of the
RP accent used by many male speakers. It has been noticed even among male news
presenters on the BBC (British Broadcasting Service). It appears to be associated with a
presentation of self as unassertive and unthreatening but it causes parts of a message to be
251
almost inaudible. It is also referred to as „creaky voice‟ and it is produced in the glottal
region with the arytenoids firmly pressed together while the front portions of the vocal
chords slowly vibrate. It is therefore a unique characteristic feature of male RP speech
which involves a glottal setting whereby the front vocal folds vibrate slowly while the
back vocal folds vibrate rapidly. Both articulatory settings produce a croak + voice
characteristic which the Nigerian listeners found difficult to understand. Most of the
listeners responded with devices such as space and omission marks.
To conclude this sub-section, the phenomenon of reduction which is the general term for
the vowel and consonant reduction processes discussed above will be now be discussed
in relation to intelligibility.
Reduction
This general term refers to a process in which a form or set of forms undergoes change
with respect to certain phonetic features. Two main types of reduction were observed in
RP speech and both affected intelligibility. They are phonetic and phonological
reduction.
Phonetic reduction
Phonetic reduction refers to changes which are not language specific but cut across all
languages. The effects of this phenomenon were noticed in two major ways in the
connected speech test. Firstly, it was observed that repeated words were extensively
reduced and hence led to many instances of lack of intelligibility. This observation
252
implies that the same word may exhibit different levels of importance in connected
speech. This observation is important because it relates to words that were initially focal
(at the first mention) but subsequently passed into lower information status (at second or
subsequent mention): the first time the word was used its articulation was more precise
and the resulting acoustic signal more distinct than in subsequent tokens of the same
word. This is related to the theory of least effort which is associated with vocal tract
inertia and which essentially says that since the topic is known, it is not necessary to
make the effort to achieve a maximal pronunciation after the first token. The consequence
of this was that repeated words were generally extensively reduced as they were not
stressed. A few examples are presented below (see Appendix 5):
FRP1: war (line 26)
cakes (line 33 & 34)
FRP2: friends (line 2)
camp (line 5)
MRP4: four (line 2)
culture (line 25& 26)
MRP8: prison (9&10)
café (line13&14)
Secondly, how common or otherwise a word is, also influenced speech intelligibility as it
was observed that more common items were reduction-prone and therefore led to many
instances of intelligibility failure while less common items were less prone to reduction
253
and therefore led to fewer instances of intelligibility failure. This is linked to the
explanation that the brain appears to process words of high frequency more quickly than
their infrequent counterparts. In this regard, it was generally observed that nearly all the
grammatical words - conjunctions, prepositions, pronouns etcetera – were affected by
various reduction processes, while lexical words – nouns, main verbs, adjectives and
adverbs – kept their stress and were not generally extensively reduced. The fact that the
stress patterns of these words were maintained probably ensured that they were highly
recognizable to the Nigerian listeners. A few examples from the data are presented
below (see appendices):
MRP5: and then (line 3)
and a (line 10)
MRP7: you can‟t (line 19)
so I (line 11)
MRP 10: are on (line 6)
FRP16: more of (line 16)
in their (line 7)
Phonological Reduction
Phonological reduction may be described as the changes which systematically occur to
certain sounds and in certain parts of words and syllables in a specific language/accent.
Phonological and phonetic reductions are similar but not completely similar as two major
differences exist between them. Firstly, unlike phonetic reductions which are sensitive to
254
previous mention, phonological reductions are not. Therefore, the frequency of a word
does not determine reduction or otherwise. Apart from this, phonological reductions are
considered to be accent/language specific while phonetic reductions are not considered to
be because phonetic reductions usually cut across many accents/languages while
phonological reductions do not. Phonological reductions are said to be accent/language
specific in the sense that specific reduction processes apply to particular languages rather
than to languages in general. In other words, the changes which were observed to
systematically occur to certain sounds and certain parts of words and syllables of the RP
speakers‟ speeches are specific to RP alone. Thus, the fact that words undergo reduction
in connected speech is a universal fact of human language but the nature and range of
reduction varies from one language/accent to another suggesting that accent continua
showing locations of languages/accents may be useful in the differentiation of accents.
Such a continuum will show the range of reduction in various accents varying from low
reduction to high reduction.
In the present study, it was observed that various types of phonological reductions
accounted for about 80% of the instances of intelligibility failure. This implies that there
is a high degree of variation between RP and NEA regarding reduction. This high degree
of variation probably led to the high degree of intelligibility breakdowns since variation
is a determinant of intelligibility. With regards to reduction, RP and NEA vary in two
major ways:
(a) NEA is known to exhibit low reduction rates while RP exhibits high reduction
rates.
255
(b) Going by the explanations above on the language/accent specificity of reduction,
NEA and RP most likely have different patterns of reduction.
These major aspects of variation between the NEA and RP may have led to the high
percentile rate of intelligibility caused by reductions. These reductions mainly affected
the vowels ,, (which are commonly known to be the weak vowels of RP) and to a
lesser extent ,. Among the consonant sounds, the alveolars ,,,, the
fricatives h,, and the dental were the most vulnerable to change. Besides
this, it seems that the immediate phonetic/phonological environment of a segment also
influences whether or not it will undergo reduction. This is because it was observed that
segments that were members of a syllable- or word-final cluster and those at the end of a
word/syllable changed in a variety of ways. These reductions probably resulted in the
Nigerian listeners thinking that unstressed segments and syllables had disappeared.
Instances are presented below:
similar
all the holes
quite (grumpy) t
now and then
bare
few of v
balanced
even that
256
that land
buy a house
and I think my
mint so
as though
5.8 ANALYSIS OF DATA: II
CAUSES OF INTELLIGIBILITY FAILURE IN RELATIVELY
DECONTEXTUALIZED SPEECH (TESTS 2 – 4)
The discussions in this section are based on the causes of intelligibility failure identified
in Tests II, III, and IV which are scripted tests. The results of these tests correspond to the
results of the connected speech test (Test 1) and therefore enhance our understanding of
the negotiation of intelligibility in the face of accent variation.
Causes of intelligibility failure: Test 2
Vowels and level of intelligibility
Test 2: Minimal pairs
Number of intelligibility
failures
Vowel Contrast
1. 294
2. 256
3. 230
257
4. 224
5. 93
6. 81
The high intelligibility failure rate for this test is probably due to two reasons:
(a) Test II is a phoneme test which consists of minimal pairs. It is without any of the
contextual clues of the connected speech test; and
(b) The test was deliberately constructed to include items drawn from areas of
segmental divergence between RP and NEA.
Bearing in mind these reasons, it was not very surprising that there was a high percentage
of intelligibility breakdowns. Generally, the items that led to the highest instances of
intelligibility failure with the vowels were those involving:
(a) the central vowels
(b) phonologically absent segments and,
(c) vowel length
The causes of intelligibility failure for individual phonemes are considered below.
;
Contrasts involving and were responsible for the highest instances of
intelligibility failure. The RP speakers produced these vowels for the pair of words: fur
and for and many Nigerians responded with devices such as space and omission marks.
Other responses included: for in place of fur and vice versa, from, fun and fall. For both
vowels, the RP speakers produced qualities articulated with the center of the tongue and
258
with considerable length. Although there were individual variations, all the realizations
were between close-mid to open-mid; the only difference between the two vowels being
that one is accompanied with lip rounding while the other is not.
The most likely cause for lack of intelligibility for this pair is the absence of these vowels
from the phonology of many Nigerians. They are both central vowels and the RP central
vowels are known to constitute problems for Nigerians. Awonusi described this pair as
„marginal vowels‟ in reference to their rarity in the speech of most Nigerians (221). ,,
may be the closest vowels that most Nigerians use instead of this pair.
These contrasts also led to a high number of instances of intelligibility failure. The RP
speakers produced the lexical set of said and sad but the Nigerian listeners could not
differentiate between the two words as they responded with writing one instead of the
other. Other responses were even farther off the mark as they included responses such as
send and sell. Some of the listeners also responded by writing said twice, probably
because both items sounded alike in terms of perception. This may have occurred because
the realization of produced by the RP speakers was rather more open that the usual RP
pronunciation. Rather than the tongue being raised to a position midway just above open,
the realization was with a more lowered tongue position and a more open lip position,
which was close to Cardinal . This realization sounded similar to RP . The
lowering of into is a recent development in RP but it brings this vowel closer to the
259
NEA equivalent of which is . is also a marginal sound in that it is not used by a
majority of Nigerians.
These pair of vowels were responsible for a high number of intelligibility failures. This
may be because is also a central vowel and vowels at this position tend to constitute
problems for Nigerians. It is also a marginal vowel as only a minority of Nigerians have
the vowel in their phonologies. The RP speakers articulated this vowel in the word cut
with the center of the tongue raised just above the fully open position and with the lips
neutrally open. The quality was similar to that of a centralized and slightly raised
Cardinal which is perceptively similar to. Therefore, most of the Nigerian listeners
responded with cat and car. The confusion with car may be because syllable final /t/ is
usually with no audible release in RP. Thus, to the Nigerian listeners who are accustomed
to full release and aspiration of /t/ in this position, it would have sounded as if the
consonant was missing.
The lengthening that accompanied should have assisted the Nigerian listeners to
differentiate between this pair of vowels but Nigerians like most L2 learners of English
are not accustomed to using length as a feature for sound differentiation. Lack of
differentiation in vowel length is one of the most striking features of the NEA. Thus,
where length is often employed by native speakers for vowel differentiation, hardly is this
feature used by L2 speakers. Therefore, the responses written by most of the Nigerian
listeners for cart was cat and a few even responded with can’t.
260
The items for this pair were often and orphan. The two are very close in phonetic space
being articulated by the RP speakers with the back of the tongue. The difference was that
while / was articulated with the back of the tongue in the fully open position, / was
articulated with the back of the tongue raised between the open-mid and close-mid
positions. This closeness in vowel space coupled with the phonologically absent /t/ in
often led to a majority of Nigerians responding with often and often or orphan and
orphan. As with the other long vowels discussed above, the length which accompanied
/ did not seem to assist the Nigerians to discriminate between this pair of vowels as
non-differentiation of vowel length is a major characteristic of NEA. Moreover, the most
common realization of / among Nigerians is /. This may have also added to the
confusion.
Apart from this, it is possible that the phonological absence of [t] in the pronunciation of
some of the RP speakers may have further added to the confusion experienced by the
Nigerians. The /t/ in often has variable realization which led to realizations of the item as
either or . However, the realization without „t‟ was more common among
the RP speakers. This was probably responsible for the confusion observed with the two
items.
261
This pair of vowels caused a considerable lower number of intelligibility failures than the
others which have been discussed earlier. They were represented by full and fool. The RP
speakers articulated the short vowel sound in full with a part of the tongue nearer to
center than to back raised just above the close-mid position. However, a degree of
fronting and lack of rounding were observed to accompany the realization of this vowel
and this gave the articulation a quality similar to Cardinal . This may be why some
of the Nigerians responded with fill. But others responded correctly. / was articulated
by the RP speakers with varying degrees of centralization, lowering and rounding ranging
from to to . While some Nigerians responded with fill, others responded
correctly. The relatively fewer instances of intelligibility failure that occurred with this
pair may be because these vowels exist in the phonologies of most Nigerians although
there is a blurring of the distinction between RP and .
The articulation of by the RP speakers was with the front of the tongue raised to a
height slightly below and behind the front close position with the lips spread. /is
somewhat similar with a part of the tongue nearer to center than to front raised just above
the close-mid position with the lips also loosely spread. The lexical items for these
vowels were fill and feel and most of the Nigerians responded correctly probably because
these vowels have similar representations in Nigerian indigenous languages. Thus, the
vowels were familiar to the Nigerians. Only a few instances of intelligibility failure were
observed with responses such as feel instead of fill and vice versa. This was probably due
262
to the characteristic of lack of discrimination in vowel length which is common with
Nigerians.
Table 2b: Diphthongs and level of intelligibility
Phoneme (Diphthongs) Number
Vowel contrast
1. 289
2. 266
263
3. 228
4. 162
5. 76
Diphthongs are glides containing two vowels within one syllable. They are usually
described as having a first element (the starting-point) and a second element (the point in
the direction of which the glide is made). RP diphthongs have as their first element
sounds in the general region of ,,,, and for their second element ,,. The
following generalizations were also observed with the diphthongs produced by the RP
speakers for the phoneme test:
(1) Most of the length and stress associated with the glide is concentrated on
the first element, the second element being only lightly rounded in some
cases.
(2) The process of monophthongization was observed with most of the RP
diphthongs.
The segment by segment analysis of the RP diphthongs and the responses by the Nigerian
listeners are discussed below.
The glide of the diphthong began with a close-mid and centralized from front and
moved in the direction of the more open variety of /. However, the glide to the position
of / was never completed and this led quality which is similar to. The
264
monophthongization of this diphthong led to intelligibility failure due to the peculiar RP
process.
was also observed to begin in the open mid-front position and moved in the
direction of . The lips were neutrally open throughout. However, the glide to the
second element was also not completed. Thus, the realization was that of a long
monophthong which was similar to . The items for these vowels were beer and bear.
The responses included bee, be, bin, bill etcetera for beer and beg, bell, bend, bad
etcetera for bear. Incidentally, both diphthongs have the realization // in NEA which is
quite far in phonetic space from the RP realizations of these two diphthongs.
Besides the reasons which have been discussed above, another reason why these
diphthongs may have led to so many instances of intelligibility failure is that the second
elements for both of them is the schwa sound and this vowel is very uncommon in the
speech of most Nigerians and this makes many Nigerians unfamiliar with it.
The RP speakers produced / by moving from a tongue position similar to that used for
// towards the position for // which forms the endpoint of all the three centring
diphthongs. The lips were weakly rounded. However, a dual process of lowering and
monophthongization of this diphthong was observed and this led to the realization of a
vowel which was close to /. Responses to the item tour included tore and torn. This
265
item led to many instances of intelligibility failure despite the fact that this diphthong is
usually monophthongized into // by many Nigerian speakers.
The glide for / was observed to begin at a central position between close-mid and
open-mid and moves in the direction of //, the lips were neutral for the first element but
rounded for the second element. This particular diphthong did not exhibit the
monophthonging process which was observed for the other diphthongs, but fronting of
the second element into was observed. The same characteristic fronting was also
observed with and but with the diphthong , it led to . This fronting
characteristic probably led to instances of intelligibility failure that were recorded for this
diphthong. Responses included shore, shone, show, shaw, and shawl.
The realization for / moved from a point between the back and front open positions,
and moved in the direction of //. The glide for this falling diphthong was observed to be
more extensive than the centring diphthongs that have been observed earlier. The lips
also changed from a neutrally open to a weakly rounded position. The item for this
diphthong was town and there were only a few instances of intelligibility failure with this
particular word. As for /, the glide began at a central position, between close-mid and
open-mid, and moved in the direction of //. The lips were neutral for the first element
but rounded for the second. This particular diphthong did not exhibit the
monophthonging process which was observed for the other diphthongs, but fronting of
266
the second element into was observed. The same characteristic fronting was also
observed with and but with the diphthong , it led to . This fronting
characteristic probably led to the various instances of intelligibility failure that were
recorded for this diphthong. The item for was tone and responses given by the
Nigerian listeners include: tune, twin, tow and tone.
The lexical items for these centring diphthongs were toys and ties respectively. / was
observed to begin at a point between the open-mid and open-back positions and moved in
the direction of . The lips were open and rounded for the first element but changed to
neutral for the second. NEA has a diphthong which is very similar in quality to this RP
diphthong, therefore, this diphthong did not present difficulties for the Nigerian listeners.
This is attested to by the high intelligibility score for this item (toys). The same high
intelligibility score was observed for the diphthong / which glided from the front open
position to the direction of . The Nigerian listeners also responded positively to ties.
Table 2c: Triphthongs and intelligibility
Phoneme (Triphthongs) Number
267
The triphthongs presented here are viewed as the combination of a diphthong and a
monophthong []. They are all diphthongs followed by [] within the word, either as an
inseparable part of the word, e.g. fire, tower and bower, , , or as a
suffix (morpheme) appended to the root e.g. layer, payer, blower and slower ,
,, . It was observed that there was a tendency to omit the second
elements of these triphthongs in the realization of the items. This implies that [] and []
were omitted in the realizations of these items. This smoothing process led to the
following realizations: ,,,, and . The responses
provided for these items by the Nigerian listeners include far and from for fire, tar and
tad for tire, lay, and lane for layer, pay and pain for payer, blow and plough for blower
and slow, slope for slower.
Table 2d: Consonant contrasts and level of intelligibility
Vowel contrast
1. 289
2. 278
3. 272
4. ; 267
268
Phoneme (consonants) Number
Consonant contrasts
1. Final 125
2. Medial 119
3. Initial ; 79
4. Initial 12
With reference to the intelligibility figures in the table on consonantal above, it was
observed that the sibilants and fricatives led to many instances of intelligibility failure.
The realizations of the consonants are examined below:
For , the items were lose and loose. The responses involved the confusion of the
two consonants. What may have caused this confusion is the fact that the two consonants
/s/ and /z/ have the same place of articulation. In other words, they are both realized at
exactly the same point in the mouth, using the same mobile and immobile articulators i.e.
the tongue and the alveolar region of the mouth. The only difference is that for [z] , the
vocal folds in the larynx are drawn together, thereby closing off the glottis and leading to
an accompanying vibration which is known as a voiced quality. Whereas, for [s], the
vocal folds are pulled back and drawn apart, leaving a free space through which air can
flow without any vibration which is known as a voiceless quality. It is these two
phonation settings that differentiate between these two sounds.
269
However, the absence of [z] in the phonology of most Nigerians may be responsible for
this. It may be said that while RP has a phonological /z/ representation in the word lose,
NEA has a phonological representation of /s/.
The items for these consonants are leisure and ledger. These two consonants also have
close areas of articulation points in the mouth. Both being articulated using the alveolar
ridge/hard palate and the tongue. The distinguishing feature of the two is that while the
former is a fricative sound, the latter is an affricate sound. This is in reference to the
manner in which both are articulated. These consonants are frequently substituted for one
another and this may have led to the confusion of these sounds. The responses were
mainly substitutions of / for / and vice versa.
;
The items for these in the test are hill and ill. The responses were majorly the substitution
of ill for hill and vice versa. Dropping of h is very common among Nigerians as h is a
sound which has complexities of realizations among Nigerians. These realizations
include h-dropping (non-articulation of h) and h-insertion (use of h in a h-less word).
This phenomenon may have been responsible for the confusion observed with these
items.
270
The items for these consonants are fail and veil. The few instances of lack of
intelligibility involved the substitution of veil for fail. It is possible that the neutralization
of opposition between / and / which characterizes the speech of many Nigerians. The
two consonants are labio-dental fricatives which are distinguished by voicing. However,
while the two consonants are phonologically significant for Nigerians, the voicing
opposition is occasionally neutralized and this may have led to the few number of errors
observed for these items.
The results presented below are the listener scores for Test 3 (Sentence stress test).The
scores were averaged over ten (10) listeners for each speaker and then converted to
percentages.
Causes of intelligibility failure for contrastive/emphatic stress (Test 3)
271
Table III: Mean intelligibility scores for contrastive/emphatic stress
Speaker ID Scores in (%)
FRP1 98.0
FRP2 98.0
MRP3 96.0
MRP4 98.0
MRP5 96.0
MRP6 98.0
MRP7 96.0
MRP8 100
FRP9 98.0
MRP10 96.0
FRP11 98.0
FRP12 94.0
FRP13 96.0
FRP14 96.0
MRP15 94.0
FRP16 96.0
Average 96.7
The average score for this test is 96.7%. The high score indicates that the Nigerian
listeners were able to correctly identify almost all the emphatic stress positions in the
speech of the RP speakers.
In this test, the British speakers indicated contrast on five utterances. Each consisted of a
single high falling nuclear tone on the first, second and third stressed syllable. This high
falling tone was used by the RP speakers on the significant units of information in order
272
to focus on the most important element that was to be highlighted in the sentence. Thus,
the focusing was achieved through the tonicity/pitch change which the RP speakers gave
to the significant words in the following utterances:
Utterance 1: Tom, fried and dinner in the utterance: “No, Tom had fried chicken for
dinner”.
Utterance 2: English, crazy and language in the utterance: “No, English is a crazy
language”.
Utterance 3: green, book and table in the utterance: “No, the green book is on the table”.
Utterance 4: jealous, woman and slapped in the utterance: “No, the jealous woman
slapped her husband‟s lover”.
Utterance 5: president, flew and night in the utterance: “No, the president flew into the
airport at night”.
The RP speakers were consistent in their use of the high falling nuclear tone to indicate
contrast in the five sentences presented to them and the Nigerian listeners seemed quite
proficient in the ability to identify the contrast being made in the sentences. This was no
doubt responsible for the high intelligibility score obtained in this test.
Causes of intelligibility failure for intonation (Test 4)
This test was designed to test the understanding of various attitudes expressed by the RP
speakers through the use of various intonation patterns. The intelligibility level for this
test was calculated to be 53.2%. The percentage value for each speaker is presented
below in a table.
(The scores are averaged over the ten listeners involved for each particular speaker.)
273
Speaker ID Scores in %
FRP1 53.2
FRP2 61.9
MRP3 58.7
MRP4 55.2
MRP5 55.6
MRP6 58.5
MRP7 68.0
MRP8 46.1
FRP9 48.0
MRP10 47.6
FRP11 54.0
FRP12 46.1
FRP13 67.1
FRP14 43.1
MRP15 47.2
FRP16 42.0
Average 53.2
The individual items in the test are discussed below:
Item 1: Where are you going? Attitude: Questioning.
All the RP speakers produced this item using a rising nucleus and the Nigerian speakers
all successfully interpreted it as a question. The rising tone is a simple unidirectional tone
274
which the Nigerian listeners are highly familiar with. This was perhaps responsible for
the success of the Nigerian listeners in correctly interpreting this item.
Item 2: I‟m starting exams today.
Attitude: Declarative; a statement.
A simple unidirectional falling tone was expected with this item which may be
structurally classified as a declarative sentence. However, instead of the falling tone
which is usually used for RP declarative sentences, some of the RP speakers used a rising
tone (HRT- high rising terminal contour) in the final tone unit of the item where a falling
tone would usually be used.
This particular intonation feature of British English is sometimes referred to as „upspeak‟
and its increasing use has been ascribed to the influence of Australian soap operas which
have a large viewing influence in England. The high rising terminal contour is common
in Australian English.
The effect which was produced by the use of the rising tone for the declarative sentence
in item 2 instead of a falling tone is that it gave the statement “I‟m starting my exams
today” a questioning attitude. The Nigerian speakers who are used to associating a rising
tone with a question misinterpreted the declarative statement as a question and responded
thus.
The high rising terminal contour used by the RP speakers is indicated below in item 2.
I‟m starting my exams today
275
The other items which many of the Nigerian listeners again misinterpreted were items 3
and 4 which involved the use of two main types of listing intonation tones. The two types
of intonation tones used by the RP speakers for these items were the high level and the
half-completed rise. The items are as follows:
Item 3: Would you like a Coke or Fanta or Sprite? Attitude: Complete list of choices
Item 4: Would you like a Coke or Fanta or Sprite?
Attitude: Incomplete list of choices
The RP speakers produced item 3 using either a high-level contour or a half completed
rise and fall contour. However, many of the Nigerians chose the wrong option as they
could not identify whether the list of items presented to them was complete or not. This
was probably because of the use of a bidirectional tone which is a complex intonation
contour rarely used by Nigerians.
For item 4 which was an incomplete list of choices, the RP speakers produced the
incomplete list of items using a half completed rise or low rising final tone. Again, many
of the Nigerians misinterpreted the item and they could not correctly identify that it was
the list was incomplete.
Items 5, 6 and 7 were to convey the attitudes of sincerity, mere formality and boredom
respectively with the statement “thank you”. The options presented to the Nigerian
listeners for these items were as follows:
276
(a) With genuine sincerity
(b) On a boring matter
(c) Mere social pleasantry
The items which were read out by the RP speakers and the various attitudes which the
statements were meant to express are presented below:
Item 5: Thank you
Attitude: Sincerity
Item 6: Thank you Attitude: Mere formulaic pleasantry
Item 7: Thank you Attitude: Boredom
The RP speakers produced item 5 using a high falling tone to show sincerity. “Thank
you” may be said with this contour when a favour has been done and one genuinely
wishes to show appreciation. However, many of the Nigerians could not correctly
identify the attitude expressed with this intonation tone. The Nigerian listeners chose the
wrong options and this contributed to the average scores obtained for this item.
Item 6 was produced by the RP speakers using a high rising tone which indicates the
“thank you” is merely the expression of a formulaic/social response and not that a
genuine favour has been done. The Nigerian listeners also could not correctly identify the
appropriate attitude being conveyed by the RP speakers. They responded with the other
options such as “boredom” and “with sincerity”.
277
Item 7 was produced by the RP speakers using a low falling tone. This tone was used to
imply that the matter at hand was boring. Again, the Nigerians could not correctly
identify the attitude which was being conveyed by this tone. Many of them chose
incorrect options.
Accent Familiarity and Intelligibility
Sixteen (16) out of the one hundred and sixty (160) Nigerian listeners involved in this
study were Nigerian undergraduates in the UK. These 16 are representative of Nigerian
diasporic dwellers (UK). The scores indicate that they correctly identified a higher
number of units than the other one hundred and forty four (144) Nigerians who are
resident here in Nigeria.
Table V: The effect of listener’s degree of familiarity on intelligibility
Listener ID Test 1
Connected Speech
Test 2
Minimal Pairs
Test 3
Sentence Stress
Test 4
Intonation
278
L10 77 50 100 75
L20 85 55 100 75
L30 67 54 100 73
L40 71 51 100 70
L50 77 57 80 82
L60 74 63 100 83
L70 73 57 100 79
L80 81 58 100 69
L90 82 50 100 74
L100 64 52 100 59
L110 83 57 100 58
L120 93 67 100 73
L130 83 59 100 79
L140 82 52 100 64
L150 75 54 100 43
L160 80 54 100 48
Average 77.9% 55.6% 98.7% 69.0%
The table above shows that the Nigerian listeners that had high familiarity with the
Standard British English accent (Nigerian Diasporic dwellers) obtained higher
intelligibility scores in all the tests than other Nigerians who have relatively less
familiarity with the Standard British English accent. This implies that the degree of
familiarity of the listener influences/determines the level of intelligibility in interaction
279
involving Nigerians and Britons. With a range of 93% and 64%, almost all of the
Nigerians in this group (Nigerians living in the U.K) scored higher than the general
intelligibility average of 62% in the connected speech test (Test 1), implying that high
familiarity with an accent correlates with high intelligibility.
The Nigerian listener (Listener L120) with the highest score is a female student at the
University of Leeds who has worked part-time for about three years as a telephone
operator. The job is similar to a public relations job which involves a lot of interaction
with the members of the public. This particular listener seems to have developed the
requisite skills for highly efficient listening. This efficiency was particularly on display in
certain areas of the listening test such as the areas of ability to correctly interpret some of
the intonation contours and ability to correctly identify the unfamiliar vocabulary items
such as interailing, veg, and swanky which occurred frequently in the discourse of the
British speakers.
Conclusion
In conclusion, the examination of the intelligibility scores obtained in the tests and the
subsequent analysis has corroborated the fact that accent variation is a determinant of
intelligibility. The phonetic and phonological analysis of the data showed that the areas of
divergence between RP and NEA were also the areas which led to instances of lack of
intelligibility. The identification of the features of RP which were responsible for lack of
intelligibility and the subsequent arrangement of these features into a hierarchy of
280
intelligibility has demonstrated the impact of phonological divergence on intelligibility
which in one of the major aims of the study.
CHAPTER SIX: CONCLUSIONS
281
This final chapter presents the conclusions of the study under the following sub-sections:
summary of main findings, contributions to knowledge, recommendations and directions
for further research.
6.0 SUMMARY OF MAIN FINDINGS
This phonetic/phonological study involving speakers of RP and Nigerians has focused on
accent variation and how it impinges on intelligibility in interactional communication
between educated native speakers and non-native speakers of English. Our corpus of data
comprising both contextualized and relatively decontextualized data has provided
insights into the dynamics of intelligibility negotiation between educated Britons and
Nigerians. The exploration of the quantitative and qualitative procedures to the analysis
of the British speakers‟ utterances and the Nigerian speakers‟ responses has equally
provided evidence of the phenomenon of accent variation as it occurs between speakers
of the Standard British English accent and Nigerians. Presented below are the main
findings of this research against the background of native and non-native English settings
on which the study is based.
1a. The average intelligibility level of the Standard British English accent (RP) to
Nigerians is calculated to be 62.2%. This percentage value is based on the
calculation of the number of items correctly received by the Nigerian assessors in
the continuous speech test (Test 1). This implies that the Nigerian listeners
involved in the study were able to correctly receive about 62% of the British
speakers‟ utterances in spontaneous unscripted speech which is considered to be
highly similar to real life communication. This average is comparable to Tiffen‟s
282
test of 1974 where British listeners obtained an average score of 64% with
speakers of Nigerian English.
1b. The features of RP which were observed to determine intelligibility are arranged
in a hierarchy of influence as follows: smoothing, schwa suppression, schwa
absorption, -reduction, nasal relocation, large drops in volume and salience at
utterance end, t-alteration, t-glottalling and glottal reinforcement, unilateral
idiomaticity and r-liaison.
2. The segmental features of the Standard British English accent (RP) determine
intelligibility as the percentage values of intelligibility were: 39%, 36%, 14%, and
74% respectively for the realizations of vowels, diphthongs, tripthongs and
consonants in the phoneme test. Causes of lack of intelligibility include peculiar
RP features such as monophthongization, smoothing, realizations involving the
central vowels, certain sibilants and fricatives, phonologically absent segments
and vowel length.
3. The nuclear stress patterns of the Standard British English accent (RP) affect the
degree of intelligibility of the speakers of this accent as the average degree of
intelligibility was calculated to be 96.7%, implying that the Nigerian listeners were
able to correctly identify the contrasts made in almost all the utterances.
4. Certain types of intonation contours employed by the RP speakers affected
intelligibility as an average intelligibility rate of 53.2% was observed for the
intelligibility level of RP intonation contours. This implies that the Nigerians were
able to correctly identify just about half of the various attitudes conveyed by the
various intonation contours used by the RP speakers. Generally, the Nigerians were
283
able to identify the intonation patterns conveying a positive statement and a simple
question in which the British speakers used falling nuclei. These utterances
involved relatively simple, unidirectional intonation patterns. However, utterances
involving more complex bidirectional tones left many of the Nigerians at a loss with
only about a tenth making the correct responses.
5. The degree of familiarity of the Nigerian listener with the Standard British English
accent influences the level of intelligibility. It was observed that Nigerian listeners
who had high familiarity with the Standard British English accent (Nigerian
diasporic dwellers) obtained higher intelligibility scores in all the tests than other
Nigerians who have relatively less familiarity with the Standard British English
accent. With a range of 93% and 64%, almost all of the Nigerians in this group
(Nigerians living in the UK) scored higher than the general intelligibility average of
62% in the connected speech test (Test 1), implying that high familiarity with an
accent correlates with high intelligibility.
This corroborates previous evidence that accent familiarity correlates highly with
intelligibility. Our juxtaposition of the performances of the Nigerians resident at home
with that of the Diasporic dwellers showed the relevance of degree of familiarity to
intelligibility. The Nigerian listener who obtained the highest score is a female student at
the University of Leeds who has worked part-time for about three years as a telephone
operator. The job is similar to a public relations job which involves a lot of interaction
with the members of the public. This particular listener seems to have developed the
requisite skills for highly efficient listening demonstrating that calibration of an accent
sufficiently assists the listener to achieve greater listening proficiency.
284
6.2 CONTRIBUTIONS TO KNOWLEDGE
Although various studies into the intelligibility of speech have been conducted,
limitations can be observed in terms of perspective as they have basically focused on the
measurement of the intelligibility of non-native varieties to native speakers based on the
premise that native speaker speech is inherently intelligible. In contrast, we have
attempted to examine the intelligibility of a native speaker accent (RP) from a non-native
(Nigerian English) perspective. This approach tends to provide a better framework for
identifying features of the Standard British English accent which present difficulties to
Nigerians, and thus was able to establish a hierarchy of intelligibility based on the
features of the Standard British English accent (RP). In line with our approach, the major
contributions are outlined below:
1. The examination of the spontaneous unscripted speech recordings (connected
speech) provided fresh insights into the nature of RP conversational speech
processes and their impact on intelligibility to Nigerians. The investigation of
these processes enabled us to establish that the intelligibility level of the RP
accent to Nigerians is 62.2% and even more importantly, it provided a detailed
characterization of the features of RP which constitute intelligibility problems to
Nigerians.
2. The exploration of the relatively decontextualized speech of the phoneme test
provided evidence that there is no necessary correspondence between the
285
intelligibility indicators in contextualized speech and relatively decontextualised
speech. Also, phonemic contrasts involving vowels, diphthongs and triphthongs
which tend to create problems of intelligibility were identified on the segmental
level of the Standard British English accent.
3. A significant contribution of the study is in the area of phonetic/phonological
investigation of RP consonant contrasts. The study revealed certain recurrent
patterns of RP speech were found to affect the intelligibility level of the RP
speakers to Nigerians. By establishing a hierarchy of the intelligibility of RP
consonant phonemes, a significant contribution has been made to the
accumulation of data on problematic areas of RP speech especially as it concerns
communicative interaction with Nigerian non-native speakers of English.
4. The examination of the use of nuclear tones and intonation contours by speakers of
the Standard British English accent and the Nigerian listeners‟ responses
demonstrated how intelligibility is negotiated in the face of accent variation. This
provides the basis for us to distinguish between RP nuclear tones and intonation
contours that are easily identifiable and those that are not and thus lead to
problems of intelligibility. Besides providing supporting evidence on the gradable
nature of intelligibility, these conclusions also provide data on the reactions of
Nigerians to native speaker RP nuclear tones and intonation contours.
5. The role of familiarity in the process of intelligibility negotiation was established in
this study and insights were provided on its crucial position. The dynamics
286
introduced into intelligibility through the manipulation of the variable of
familiarity were observed on both the RP speakers and their Nigerian listeners,
enabling us to establish that familiarity is a viable variable in native and non-
native speaker communication.
The findings of the research will provide data in three major ways:
a. As a contribution to the Lingua Franca Core (LFC), an ongoing
compilation of essential pronunciation features for a global English
language teaching syllabus aimed at international intelligibility.
b. To enhance the performance of Nigerian candidates in the listening
comprehension component of the International English Language Testing
Scheme (IELTS). The IELTS is a compulsory examination conducted by
the British Council for Nigerians who wish to undertake
undergraduate/postgraduate study or wish to work in the UK.
c. To provide data for the L2 pronunciation and listening comprehension
syllabus.
6.3 RECOMMENDATIONS
Our recommendations are for three major groups: Variationists, Curriculum Planners and
Teachers/Learners.
Recommendations for Variationists:
This study has demonstrated that the study of native speaker accents to non-native
speakers is a productive area of intelligibility research. Therefore, a departure from the
previous view of intelligibility as a phenomenon which may be studied only from a native
287
speaker perspective is recommended. A broader view that perceives speech intelligibility
as a two-way process that is without prejudice to the supremacy of native speaker accents
over non-native speaker accents is therefore encouraged.
Recommendations for Curriculum Planners:
Scholars have often commented that a remarkable feature of spoken English in Nigeria is
that it often falls short of the user‟s ability in written English. This has been ascribed to
the fact that English is learned by most Nigerians in school and the educational system
has emphasized the skills of literacy more than those of oracy. The result is that the
system has produced in many cases individuals with an impeccable ability to
communicate in written English without a commensurate ability in spoken English. This
study has demonstrated the necessity of incorporating spoken English and to a greater
extent, listening comprehension into the syllabi at both primary and secondary levels.
This is in view of the fact the fact that native speaker speech allegro and particularly
reduction phenomena are hardly ever taught. What suffices for this in the syllabus is
„listening comprehension‟ which is limited to the level of semantics and pragmatics. A
broader curriculum which gives due cognizance to casual speech phonology is therefore
necessary.
Recommendations for Teachers/Learners:
While it is not recommended that learners should be encouraged to ape the native speaker
through the use of casual speech reduction forms, it is necessary that these forms be
288
learnt given that reduction is likely to continue and be even be more ubiquitous in future.
This is because it is a recurrent theme in phonology that changes move in the direction of
simplicity, naturalness, or unmarkedness, echoing the pervasiveness of the theory of least
effort. Therefore, the teaching should be started in Nigeria if higher levels of lack of
intelligibility are to be avoided in future. The limitations of exposing students to only
citation forms of speech as contained in regular dictionaries has been demonstrated as
many Nigerian students find native speaker speech rather problematic. Rather, it is
recommended that natural forms such as the data generated in this research should be
exposed to students to facilitate greater listening proficiency.
6.4 Directions for Further Research
This study is an intensive exploration of normal native speaker pronunciation which some
may see as trivial or deleterious, but our argument remains that a knowledge of normal
pronunciation as it is used daily by native speakers is important.
Our area of study is relatively new because of the previous focus on the intelligibility of
L2 phonology to native speakers. An attempt has been made in this study to demonstrate
the usefulness of examining the intelligibility of RP from a Nigerian non-native speaker
perspective. However, more work needs to be done in terms of segmentation of
Nigerians into ethnic/language groups. This will reveal the level and nature of
intelligibility of RP vis-à-vis the Nigerian listeners‟ mother tongue.
289
Another potential area of study is the pragmatic intelligibility of native speaker speech to
non-native speakers. Given the role of cultural and situational context in the
determination of intelligibility, this seems to be a viable area of research which has not
been previously explored.
6.5 Conclusion
This research has attempted a thorough examination of the intelligibility of the speakers
of the standard British English accent to educated Nigerians. In the course of the
research, we have attempted to determine the level of intelligibility of the RP accent to
Nigerians as well as to identify features of the accent which lead to intelligibility
problems with Nigerian listeners. Our over-riding objective has been to assist Nigerian
listeners improve their level of understanding of RP speech and thereby enhance their
effectiveness in interactional communication with the British. It is hoped that the results
presented have made some contributions in this regard.
Finally, given the international linguistic imperialism of the English language in global
relations and coupled with its attendant expansive linguistic roles in Nigeria and other
nations, there is a need to achieve mutual intelligibility between all the speakers of the
language. This is even more so as the existing imbalance in terms of economic power and
influence is undeniably in favour of the native speaker of English. The implication of this
is that in interaction involving Nigerians and Britons greater “accommodation” may be
required of the Nigerian listener due to the inequality of the speaker/listener
communication hierarchy. Presently, the features of the emerging international English
290
called the Lingua Franca Core (LFC) – a variety which is supposedly a potpourri of all
Englishes – is distinctly identical to native speaker varieties of English. While this
imbalance exists, the necessity of learning native speaker phonology cannot be
overemphasized. It is the fulfillment of this need which research hopes to achieve. It is
hoped that a modest contribution towards the achievement of mutual intelligibility in
communication involving Britons and Nigerians has been made.
REFERENCES
Abercrombie, David. Fifty Years in Phonetics. Edinburgh: Edinburgh University Press.
(1991) 48-53.
291
Adam, Brown. Pronunciation Models. Singapore: Singapore University Press, 1991.
Adegbija, Efurosibina. “Lexico-Semantic Variation in Nigerian English”. World
Englishes. Vol. 8, 2. (1989): 165-177.
Adetugbo, Abiodun. “Nigerian English Phonology: Is there any Standard?” Lagos Review
of English Studies, IX (1987): 64-84.
Akinjobi, Adenike. “Vowel Reduction and Suffixation in Nigeria”. English Today 22 (1)
(2006): 54-72
Amayo, Airen. “Tone in Nigerian English”. Paper presented at the meeting of the
Chicago Linguistic Society. (1980)
Anderson-Hsieh,J. and Koehler, K. “The Effect of Foreign Accent and Speaking Rate on
Native Speaker Comprehension”. Language Learning. (1988): 38, 561-613.
Atechi, Samuel. The Intelligibility of Native and Non-Native English Speech: A
Comparative Analysis of Cameroon English, British and American English. PhD
Dissertation: Technischen Universitat Chemnitz. 2004
Awonusi, Segun. “Some Characteristics of Nigerian English Phonology”. Eds. Segun
Awonusi and A.B Dadzie. Nigerian English: Influences and Characteristics. Lagos:
Concept Publications. (2004) 203-225
________ Sociolinguistic Variation in Nigerian English.Ph.D Dissertation, University of
London. (1985)
Bailey, C-J N.“Variation and Language Theory”. Paper presented at the Linguistic
Institute in Suny.1973.
292
________ Essays on Time -Based Linguistic Analysis. Oxford. Oxford University Press.
(1996).
__________ What is Developmental Linguistics? Orchid Land Publications. (2000).
Bailey, Guy. “Real and Apparent Time”. The Handbook of Language Variation and
Change. Eds. Jack Chambers, Peter Trudgill and Natalie Schilling-Estees. USA:
Blackwell Publishing. (2004) 312-331
Bamgbose, Ayo. “The English language in Nigeria.” Ed. John Spencer. The English
language in West Africa. London: Longman. (1971) 9-26.
________“English in the Nigerian Environment”.Eds. New Englishes: A West African
Perspective. Mosuro. Ibadan. (1995) 9-26
________“A Recurring Decimal: English in Language Policy and Planning”. The
Handbook of World Englishes Eds. Braj Kachru, Yamuna Kachru, and Cecil
Nelson. USA: Blackwell Publishing.(2006) 645-660.
Banjo, Ayo. “Towards a Definition of Standard Spoken Nigerian English”. Annales
d’Universite d’Abidjan (1971): 24-28.
_______ “On Codifying Nigerian English: The Research So Far”. Eds. Bamgbose et al.
New Englishes: A West African Perspective. Ibadan: Mosuro (1995) 203-31
Bansal, R. The Intelligibility of Indian English. Hyderabad: Central Institute of English,
1969.
293
Bent, Tess and Andrew Bradlow. “The Interlanguage Speech Intelligibility Benefit”.
Journal of the Acoustical Society of America. 114 (2001) 1600-1610.
Bokamba, E. “The Africanization of English”, Kachru, B. ed. The Other Tongue: English
Across Cultures. Oxford. Pergamon Institute of English. (1985) 79-113.
Britain, David. “Space and Spatial Diffusion”. The Handbook of Language Variation and
Change. Eds. Jack Chambers, Peter Trudgill and Natalie Schilling-Estees. USA:
Blackwell Publishing (2004). 603-637
British Council. English in the World: the English 2000 Global Consultation. London:
The British Council, 1995.
Brown, Adam. “Some thoughts on Intelligibility”. Retrieved June 23, 2007 from
http://www.melta.org.my\ET1189\main4.html. 1989.
Brown, Gillian. Listening to Spoken English. London: Longman. 2nd edition. 1990
Brown, Roger. “The Intelligibility of Ghanaian English”. Language in Culture and
Society. Ed. Dell Hymes: New York: Harper and Row.(1969) 440-44
Brosnahan F. “English in Southern Nigeria.” English Studies 39, 97-100, 1958
Cameron,B and Coates, J. Women in their Speech Communities. London.Longman. 1988.
Catford, John. “Intelligibility”. English Language Teaching”. 5: (1950) 7-17.
Chomsky, Noam. Aspects of the Theory of Syntax . Cambridge. Massachusetts. 1965.
294
Chesire, Jenny. “Sex and Gender in Variationist Research”. The Handbook of Language
Variation and Change. Eds. Jack Chambers, Peter Trudgill and Natalie Schilling-
Estees. USA: Blackwell Publishing, 2004. (423-443)
Christopherson, P. “Towards a definition of International English”. English Language
Teaching. (14):127-38.
Collins, Beverly and Inger Mees. Practical Phonetics and Phonology: A Resource Book
for Students. USA: Routledge. 2008
Coulmas, Florian. The Handbook of Sociolinguistics. Ed. Florian Coulmas. Oxford.
Blackwell Publishers. 1997: x-xi
Chambers, Jack. Sociolinguistic Theory: Linguistic Variation and its Social Significance.
Great Britain: Blackwell. 1995.
Craig, Douglas. “Towards a Description of Caribbean English”. The Other Tongue:
English Across Cultures. Ed. Braj Kachru. Oxford: Pergamon Institute of English,
(1985) 334-57.
Cruttenden, Allan. Gimson’s Pronunciation of English. Malta: Hodder Education, 7th ed.
2008.
Crystal, David. “How Many Millions? The Statistics of English Today”. ET 1 (1985) 7-9.
_______ English as a Global Language. Cambridge: Cambridge University Press. 2nd
ed.1997.
Deterding, David and Andrew Kirkpatrick. “Emerging South-East Asian Englishes and
Intelligibility”. World Englishes 25 :3/4 (2006) 391-409.
295
Dittmar, Norbert. A Critical Survey of Sociolinguistics: Theory and Application. New
York: St. Martins Press. 1976
Downes, William. Language and Society. Cambridge.Cambridge University Press. 1984.
Eckert, Penelope and Jane Rickford. Style and Sociolinguistic Variation. Cambridge.
Cambridge University Press. 2001.
Elalani, M.I. An Assessment of the Intelligibility of Jordanian English to Educated
Speakers of British English. University of Leeds. M.Phil. Thesis. 1968.
Fasold, Ralph. “Variation Theory and Language Learning”. Ed. Peter Trudgill. Applied
Sociolinguistics. England: Academy Press. (1984) 89-119.
Ferguson, C.A. “Diglossia”. Word 15 (1959) (2) 325-40.
Gaskell, E. North and South. Easson, A (ed). Oxford, 1973.
Gillian, Brown. Listening to Spoken English. London:Longman. 1990
Gimson, A.C. A Practical Course of English Pronunciation: A Perceptual Approach.
London: Arnold. 1975.
Gorlach, Manfred. Even More Englishes: Studies 1996-1997: Amsterdam: John
Benjamins. 1997
Gupta, Anthea. “Inter-accent and Inter-cultural Intelligibility: A study of listeners in
Singapore and Britain”. English in Singapore: Phonetic Research on a Corpus. Eds.
David Deterding, A. Brown and E. Low. Asia: McGraw-Hill Education.(2006) 15-
31
296
Gussenhoven, Carlos. The Phonology of Tone and Intonation: Research Surveys in
Linguistics. Cambridge University Press.2004.
Gut, Ulrike: “Nigerian English Phonology.” The Handbook of Varieties of English. Ed.
Edgar W. Schneider, Kate Burridge, Bernd Kortmann, Rajend Mesthrie and Clive
Upton. New York: Mouton de Gruyter. (2004) 813-829
Hayes, Bruce. Metrical Stress Theory: Principles and Case Studies. Chicago and
London: The University of Chicago Press. 1995.
Hughes, Arthur, Peter Trudgill, and Dominic Watt. English Accents and Dialects: An
Introduction to Social and Regional Varieties of English in the British Isles. 4th ed.
London: Hodder Arnold, 2005.
Honey, John. “Sociophonology”. The Handbook of Sociolinguistics. Ed. Florian
Coulmas. Oxford: Blackwell Publishers. (1997) 81-117.
Jenkins, Jennifer. The Phonology of English as an International Language. Oxford:
Oxford University Press, 2000.
Jibril Munzali . “Sociolingusitic Variation in Nigerain English.” English Worldwide,7:
(1986) 47-75.
Jowitt, David. Nigerian English Usage: An Introduction. Lagos: Longman, 2nd ed. 2003
Kachru, Braj. “English in South Asia”. Current Trends in English. V. Sebeok, V. Ed.
The Hague: Mouton. (1969) 41-58
________”The Pragmatics of Non-Native Varieties of English.” English for Cross-
Cultural Communication. London. Macmillan.(1981) 45-64
297
________ “Models for Non-Native Englishes”. The Other Tongue: English Across
Cultures. Ed. Braj Kachru, Oxford: Pergamon Institute of English, 1985 (102-121)
________ “Models of English for the Third World: White Man‟s Linguistic Burden or
Language Pragmatics”. TESOL Quaterly 10 (2) 22 1976: 22 -39.
Kerswill, Paul “Koinezation and Accommodation”. The Handbook of Language
Variation and Change. In Jack Chambers, Peter Trudgill and Natalie Schilling-
Estes Eds. USA: Blackwell Publishing. (2002) 669-701
Labov, William “Sociolinguistic Patterns”. Philadelphia University of Pennyslavania
Press. 1972.
_________ “Quantitaive Reasoning in Linguistics”. Report presented at the annual
meeting of the American Linguistics Society at the University of Pennsylvania. 2008.
Le Page, R. “The Evolution of a Sociolinguistic Theory of Language”. The Handbook of
Sociolinguistics. Ed. Florian Coulmas USA: Blackwell Publishing, 1997. 15-32
Lindblom B. Phonetic Variation: A Sketch of the H&H Theory. In W.J. Hardcastle and
A. Marchal (eds), Speech Production and Speech Modelling. Kluwer, (1990).403-
39.
McArthur Thomas. “The English languages”. Cambridge, Cambridge University Press,
1998.
Major Roy, Susan Fitzmaurice, Ferenc Bunta and Chandrika Balasubramanian. “Testing
the Effects of Regional, Ethnic and International Dialects of English on Listening
Comprehension”. Language Learning, 55:1 (2005) 37-69.
298
Mesthrie, Rajend. “The Phonology of English in Africa and South and Southeast Asia.”
Handbook of Varieties of English. Edgar Schneider, Kate Burridge, Bernd
Kortmann, Rajend Mesthrie and Clive Upton Eds. New York: Mouton de Gruyter,
2004, 1099-1109.
Milroy, James and Leslie, Milroy. “Varieties and Variation”. The Handbook of
Sociolinguistics. Ed. Florian Coulmas USA: Blackwell Publishing( 1997). 47-63
Milroy, Leslie. “Social Networks”. The Handbook of Language Variation and Change.
Eds. Jack Chambers, Peter Trudgill and Natalie Schilling-Estees. USA: Blackwell
Publishing (2004). 549-572
Mugglestone, Lynda. Talking Proper: The Rise and Fall of the English Accent as a
Social Symbol. Oxford: Oxford University Press. 2007
Munro, Murray, Tracy Derwing and Susan Morton L. The Mutual Intelligibility of L2
Speech. Cambridge: Cambridge University Press. 2006
______and Derwing, T.M. “Foreign Accent, Comprehensibility and Intelligibility in the
Speech of Second Language Learners”. Language Learning, 45. (1995) 73-97.
Nair-Venugopal, S. “Intelligibility of English: Of What Relevance to Intercultural
Communication”. Language and Intercultural Communication (2003) 3, 36-47.
Nelson, Cecil. “Intelligibility and Non-Native Varieties”. The Other Tongue: English
Across Cultures. Ed. Braj Kachru. Oxford: Pergamon Institute of English. (1982)
68-83
299
Nelson, Gerald. “World Englishes and Corpora Studies”. The Handbook of World
Englishes. Braj B. Kachru, Yamuna Kachru and Cecil L. Nelson Eds. Singapore:
Blackwell Publishing Ltd, 2008: 733-49.
Osoba, Sola. “Word Formation Processes in English and Nigerian Pidgin”. The
Domestication of English in Nigeria. Eds. Segun Awonusi and E. Babalola. Lagos:
University of Lagos Press. 2004, 231-247.
Prator, C. “The British Heresy in TESL”. Fishman, J. ed. Language Problems of
Developing Nations. New York. John Whely & Sons. 1968.
Rajadurai, Joanne. “Intelligibility Studies: A Reconsideration of Empirical and
Ideological Issues”. World Englishes.26 (1) (2007) 87-98.
Ramsaran, Susan. “RP: Fact and Fiction” Susan Ramsaran Ed. Studies in the
Pronunciation of English: A Commemorative Volume in Honour of A.C. Gimson.
London:Routledge (1990) 38-59.
Roach, Peter. English Phonetics and Phonology. Cambridge: Cambridge University
Press. 2000
Romaine, Suzanne. Language in Society. Oxford. Oxford University Press. 2000.
Sapir, Edward. Language. New York. Harcourt Brace, 1921.
Schmidt, Jon. “Die Verwandtschaftsverhaltnisse der indogermanischen Sprachen”
Weimar: Hermann Bohlau (1872) 147-55. Rpt. in Accents of English: An
Introduction. Cambridge: Cambridge University Press, 1982.
300
Schneider, Edgar. “Variation and Change in Written Documents” In J. Chambers, P.
Trudgill and N. Schilling-Estes (eds.) The Handbook of Language Variation and
Change. USA: Blackwell Publishing, (2002) 67-96.
Seidlhofer, Barbara. “Closing a Conceptual Gap: The Case for a Description of English
as a Lingua Franca”. International Journal of Applied Linguistics, 11 (2), (2008)
133-58
Shockey, Linda. Sound Patterns of Spoken English. Oxford: Blackwell Publishing
Limited, 2003.
Smith, Larry E. and Cecil L. Nelson. “International Intelligibility of English: Directions
and resources”. World Englishes, 3, (1998) 333-442
_______“World Englishes and Issues of Intelligibility”. The Handbook of World
Englishes. Eds. Braj B. Kachru, Yamuna Kachru and Cecil L. Nelson. Singapore:
Blackwell Publishing Ltd, 2008, 428-445
Smith, Larry E. and Elizabeth Christopher. “Why can‟t they understand me when I speak
English so clearly?” The Three Circles of English. Ed. Edwin Thumboo. Singapore:
UniPress, 2009. 91-100
______and J. Bisazza. “The Comprehensibility of Three Varieties of English for College
Students in Seven Countries”. Language Learning.32, (1982) 259-269.
______ and Khalilullah Rafiqzad. “English for Cross-Cultural Communication: The
Question of Intelligibility”. TESOL Quaterly. 13, (1979) 371-80.
Strevens, Peter. “Testing Intelligibility among Ghanaians”. IJAL,23, (1965) 185-205.
301
Tiffen, Brian. “The Intelligibility of Nigerian English”. PhD Thesis. University of
London. (1974)
Trudgill, Peter. Sociolinguistics: An Introduction to Language and Society. England.
Penguin Books. (1995)
_____„Received Pronunciation: Sociolinguistic Aspects‟. Studia Anglica Posnaniensia
International Review of English Studies. Ed. Jacek Fisiak, (2001) Vol 36.
_____„The Sociolinguistics of Modern RP‟. Reprinted in: Practical Phonetics and
Phonology. Collins Beverly and Mees Inger. London: Routledge. (2008) 264-271.
Udofot, Inyang. “Pausing in Spoken Nigeria English”. Journal of Humanities. (1996) 4,
41-49.
________ “The Rhythm of Educated Spoken Nigerian English”. Journal of Nigerian
English and Literature. (2000) 1: 11-35.
________ “Stress and Rhythm in the Nigerian Accent of English”. English Worldwide,
(2003) 24(2) 201-220.
van der Walt, Christa. “The International Comprehensibility of varieties of South African
English”. World Englishes, 19 (2), (2000) 139-262
Waller, P. “Democracy and Dialect, Speech and Class”. ed. P.J. Waller. Politics and
Social Change in Modern Britain. Essays Presented to A.F. Thompson.
Brighton.1987: 1-28.
Wardhaugh, Ronald. An Introduction to Sociolinguistics. Blackwell: Blackwell
Publishing Ltd (1986)
302
Wodak Ruth and Gertraud Benke. Gender as a Sociolinguistic Variable: New Perpectives
on Variation Studies.Ed. Florian Coulmas. The Handbook of Sociolinguistics.
USA: Blackwell Publishing, 1997. 127-149.
Wells, John Christopher. Accents of English. 3 vols. Cambridge: Cambridge University
Press, 1982
Wolfram, Walt. “Dialect in Society”. Coulmas, F. ed. The Handbook of Sociolinguistics.
Oxford.Blackwell Publishers. (1997) 65-81.