Which pronunciation norms

and models for English as an

International Language?
Jennifer Jenkins
The recent growth in the use of English as an International Language (EIL)
has led to changes in learners pronunciation needs and goals. The
acquisition of a native-like accent is no longer the ultimate objective of the
majority of learners, nor is communication with native speakers their
primary motivation for learning English. Instead, what they need above all
is to be able to communicate successfully with other non-native speakers
of English from different L1 backgrounds. This article proposes that with
English assuming the position of the worlds major lingua franca, a radical
rethink is called for in terms of the role ofpronunciation and its aims within
the ELT curriculum. In particular, there is an urgent need to consider the
question of which pronunciation norms and models are most appropriate
for classes aiming to prepare learners for interaction in EIL contexts, and to
raise teachers awareness of the issues involved.
Introduction: the Some years ago, Kachru (1988: 3) described six fallacies about the users
basic conflict and uses of English across cultures. According to his sixth fallacy, the
diversity and variation in English are indicators of linguistic decay;
restriction of the decay is the responsibility of native scholars and of ESL
programs. This fallacy, argued Kachru, has resulted in the position that
deviation at any level from the native norm is an error.
Many of us working in the field of pronunciation have moved on
conceptually from this position. We no longer regard English as being
taught mainly for communication with its native speakers (the goal of
EFL), or the target of pronunciation teaching as a native-like accent,
with the eradication of all traces of a foreign accent, however
unrealistic that target always was. We acknowledge that the EFL-ESL
distinction is beginning to blur as the two merge into English as an
International Language (EIL). Nowadays English most frequently
serves as a worldwide lingua franca for its vast numbers of non-native
users, and as Widdowson (1994) so forcefully argues, it is no longer the
property of its native speakers. However, faced with a lack of clear-cut
alternatives, we have not been able to move on in any practical way; and
this situation has been compounded by the relative neglect that
pronunciation teaching has suffered in ELT curricula since the advent
of communicative approaches, within whose paradigms it does not sit
comfortably. As a result, the position that Kachru described is still
largely that adopted by teacher educators around the world, and
consequently by classroom teachers themselves.
ELT J ournal Volume 52/2 April 1998 Oxford University Press 1998 119
articles welcome
Possible solutions
120
Two main obstacles prevent the conceptual progress outlined above
from being translated into classroom practice. The first is the difficulty in
resolving the basic conflict between the practical need to harmonize
pronunciation among L2 varieties of English sufficiently to preserve
international intelligibility; the second is the social and psychological
need to respect the norms of the largest group of users of English, i.e.
non-natives. So while it is now becoming conceivable for us to dispense
with the idea that local non-native norms are wholly inappropriate, and
that every pronunciation which differs from a native variety is deviant,
clear specific pronunciation goals for teaching EIL are thin on the
ground.
Two main approaches to the conflict have emerged. The first is an
attempt to establish some sort of simplified, neutral, universal
pronunciation variety, intelligible and acceptable to both native and
non-native users of English - the phonological equivalent of Quirks
Nuclear English (1981), which endeavoured to do this for syntax and
morphology. An example of this is Gimsons rudimentary international
pronunciation (1978: 51). Gimson reduced the phonemic inventory of
English, i.e. 24 consonant sounds and 20 vowel sounds, to 14 and 15
respectively. The following is typical of the result:
which, transcribed, is:
Dense fog now covers the whole of southern England, with visibility at a
maximum of eighty metres.
Of course, few English speakers, native or non-native, would find this
sort of thing either intelligible or acceptable, particularly as regards the
loss of the voicing distinction for consonants such that, for example,
dense is pronounced tense. Equally problematic is the fact that
although such schemes cannot be totally dismissed, it would be
extremely difficult to impose constraints in this way. Without the help
of international pronunciation police, one could not force people either
to acquire such forms in the first place, or to maintain them without
elaboration in their subsequent interactions. Influences such as the
British and American media, local norms, and group identity are likely
to intervene to varying degrees to prevent success among either native
or non-native users of EIL. Any neutral, universal forms of English
pronunciation, simplified or otherwise, will therefore probably have to
be unplanned, developing naturally from below rather than being
imposed from above, as seems to be happening at present, albeit on a
smaller scale, among the different English accents of Singapore.
The second and potentially more promising approach to solving the
conflict is that of variationists such as Pennington, whose recent (1996)
book on English phonology, subtitled An International Approach,
advocates the provision of multiple models, both native and non-native,
J ennifer J enkins
articles welcome
and the teaching of generalized norms according to individual learner
need and choice. rather than a narrow focus on a standard British or
American accent.
For the present, at least, neither type of approach solves the growing
conflict between global intelligibility and local norms in EIL. This
situation is exacerbated by the fact that remarkably little research has
been conducted into the intelligibility of English among its non-native
speakers from different L1s, be they fluent bilinguals (for whom the
epithet bilingual speaker of English seems rather more appropriate than
non-native: see Jenkins 1996a) or, far more commonly, those speakers
who have interlanguages (hence non-bilingual speakers of English) and
are, by definition, engaged in interlanguage talk when they use English
to interact with one another. These latter speakers are, of course, my
major concern, since they are the ones who fill our classrooms, and whose
pronunciation is likely to provide an obstacle to successful international
communication.
Intelligibility in Because of the dearth of research on English used among its non-native
EIL vs. EFL speakers, the majority of published materials on pronunciation, whether
contexts theoretical or for classroom use, tend to focus exclusively on
intelligibility for the native rather than the non-native receiver. This
research focus on the native receiver provides important insights into
the way native speakers of English structure and decode spoken
messages via the suprasegmental system (stress, rhythm, and intona-
tion) and, as such, constitutes an essential source of information for
those learning English in order to communicate with its native speakers.
However, the assumption that these findings can be applied wholesale
to non-native lingua franca contexts is dubious. While it is almost
undoubtedly true that the suprasegmentals contribute far more than the
segmentals (sounds) to intelligibility for the native receiver, complicat-
ing factors in interlanguage talk make it necessary to maintain a balance
between the segmentals and suprasegmentals in teaching where the
learners goal is to be effective in EIL rather than in EFL (native/non-
native) contexts.
A compromise For the purposes of promoting intelligibility through teaching EIL, while
solution at the same time allowing speakers the freedom to express themselves
through their own pronunciation norms, I suggest the following
approach: we should concentrate the productive focus of pronunciation
teaching on the three areas that appear to have the greatest influence on
intelligibility in EIL, i.e. certain segmentals, nuclear stress (the main
stress in a word group), and the effective use of articulatory setting, to
the extent that it underpins the first two areas. Other aspects of
pronunciation can then be dealt with purely at a receptive level.
We will now look in greater detail at this proposal for an EIL
pronunciation syllabus, taking each of the three productive areas in turn
and then going on to consider the receptive areas.
Pronunciation norms and models 121
articles welcome
122
Segmentals The crucial difference between a non-native speaker of English and a
native speaker (or fluent bilingual) as regards segmentals, is that the
non-native may deviate from native models in precisely those sounds
that are considered to be core sounds of English, and therefore
essential to approximate closely because they figure in all native
varieties. This core quality applies to most consonant sounds and to the
distinction between long and short vowel sounds, and can be extended
to methods of cluster simplification, particularly consonant deletion - for
example, the deletion of /t/ when it is the middle of three consonants,
such that postman is pronounced /-/. (See Jenner 1989, Jenkins
1996b for further discussion of the common core.) As well as deviating
from native models in these areas, non-native speakers often deviate
differently from each other because of Ll transfer.
Where non-core sounds are concerned we are, for EIL purposes, in the
realm of acceptable L2 regional norms - non-standard, but
nevertheless non-deviant. This applies generally to vowel quality
(provided it is consistent), which varies widely among native varieties,
and to the consonants /-/ and /-/, which do not occur in the majority of
the worlds languages, or even in some native English varieties. Core
sounds are a different matter altogether. While accurate production
does not of itself guarantee intelligibility or signal meaning, in
interlanguage talk lack of accurate production often has the opposite
effect, that is, it obstructs meaning by distracting or even opening up
possibilities of a different message. This is chiefly because the receiver
in interlanguage talk tends to use more bottom-up processing, and to
have a narrower band of allophonic (phonetic) tolerance and a reduced
lexical repertoire, as compared with a native speaker or fluent
bilingual; paradoxically, the effect can be compounded by the correct
placement of nuclear stress on the syllable containing the deviant core
sound or consonant deletion.
Nuclear stress Accurate use of nuclear stress is essential in EIL because learners seem
to acquire the native English system relatively quickly for receptive
purposes, but do not acquire it productively until considerably later, if at
all. This applies particularly to contrastive stress, through which English
typically highlights extra meanings by moving the nucleus from the last
content word to another word in the group. (For example, in My
husband drinks a lot of wine, the most usual position for nuclear stress is
on wine, whereas if this stress was shifted to husband, some other
contrastive meaning, such as but I dont, would be established.) The
thorough, systematic teaching of the nuclear stress system would
undoubtedly go some way to rectifying the learners receptive-
productive mismatch. However, until pronunciation is given a higher
priority in the classroom, such an outcome is unlikely to be widespread,
and the effect of a misplaced nucleus, particularly in combination with a
deviant core sound, will continue to be potentially disastrous for EIL
communication.
J ennifer J enkins
articles welcome
Articulatory setting Mastery in this area will both facilitate the production of core sounds
and allow the speaker to manipulate these sounds to produce nuclear
stress, i.e. to lengthen sounds, change pitch on them, and utter them with
greater volume. If the articulators are not comfortably positioned, such
manoeuvres will prove difficult, if not impossible, whereas concentrat-
ing on this holistic aspect of pronunciation . . . makes it easier to allow
suprasegmental and segmental aspects to work in unison (Dalton and
Seidlhofer 1994a: 142).
Areas open to The three phonological areas discussed above have the advantage of
variation being not only teachable but also learnable: systematic, and not riddled
with complicated exceptions and fine distinctions, or dependent on
individual learners and contexts (for more on teachability and
learnability see Dalton and Seidlhofer 1994a: 72-4). On the other
hand, most other aspects of phonology are neither easily learnable nor
necessary for most EIL contexts. Thus, EIL should be able to embrace
non-native local norms and/or disregard native norms with no threat to
intelligibility for the (non-native) receiver in at least the following
areas:
Word stress The rules are highly complex, containing manifold exceptions and
differences among L1 varieties and according to syntactic context. Some
words, e.g. controversy, ' ice-cream, even have optional stress patterns
within Received Pronunciation (RP), the standard British pronuncia-
tion. Reliable rules therefore cannot be easily formulated, let alone
learnt.
Features of These include elision, assimilation, linking, and weak forms. The
connected speech suggestion that learners can safely abandon weak forms (the production
of words like to and from as /-/ and /-/ in fast speech) will no
doubt prove controversial. However, I would argue that it is possible to
highlight some syllables without necessarily reducing others, and that
some native varieties do precisely this, along with the vast majority of
fluent bilingual speakers. Lack of weak forms may prove disconcerting
for some native receivers, but is unlikely to do so for non-natives.
Rhythm While English is relatively stress-timed, with stresses occurring on
important syllables in the speech stream at roughly equal intervals, most
of the worlds languages tend towards varying degrees of syllable-timing.
While it will always be necessary to retain nuclear stress on the most
important syllable of a word group, because English is relatively
restricted syntactically and morphologically in its capacity to highlight
important aspects of a message in other ways, this does not mean that
stress-timing is essential in any strict sense. In fact, rigid stress-timing is
no more than a convenient fiction for classroom practice and, if David
Crystals (1996) tentative prediction is correct, English may be moving
towards the syllable-timed end of the stress/syllable-timing continuum,
under the influence of other world languages in general and of rap music
in particular.
Pronunciation norms and models 123
articles welcome
Thus, while approximation to the native model is probably essential for
intelligibility in non-bilingual EIL contexts as regards core sounds,
nuclear stress, and relevant articulatory setting, local non-native norms
are likely to be both acceptable and intelligible in many other
phonological areas. This solution to the conflict involved in the selection
of pronunciation goals for EIL offers teachers and learners the best of
both worlds: a universal, realistically teachable and learnable core, based
on the native model (model being singular in the sense that the
designated areas are common to all native varieties), which can then be
fleshed out according to a wide range of acceptable, local non-native
norms.
Implications for If such an approach to the teaching of pronunciation is to be adopted in
teacher education the EIL classroom, teachers will require substantial awareness-raising
and practical preparation. In particular, teacher education programmes
will need to prioritize two issues: first, the difference between a model
and a norm and, second, the fact and acceptability of L2 sociolinguistic
variation.
124
The difference According to Dalton and Seidlhofer (1994b: 27) if we treat RP and/or
between a model General American as a norm, we connect them strongly with ideas of
and a norm correctness. The norm is invariable and has to be imitated independently
of any considerations of language use. The aim, however unrealistic, is
100 per cent attainment of the norm, which is regarded as an end in
itself. On the other hand, if we treat RP and/or General American as a
model, we use them as points of reference and models for guidance. We
decide to approximate to them more or less according to the demands of
a specific situation.
Thus, instead of treating a native norm as the goal for production, as has
generally been the case hitherto, teachers should be made aware that
this is neither a desirable nor, in fact, a likely outcome. They can be
shown how to use a native model as a point of reference to prevent local
non-native varieties from moving too far apart from each other, as well
as to promote receptive competence in interaction with native speakers.
However, for active use and correction, they can be directed towards
focusing on the comfortable production of core sounds and nuclear
stress, and on the rules governing nuclear placement, but otherwise to
accept, indeed, to promote the use of L1 norms.
Nevertheless, the demands of the specific situation of classroom
teaching are such that non-native teachers will, themselves, still be
required to develop the ability to approximate more closely than their
students to a standard native model. This will enable them to provide the
classroom exposure that is necessary to provide points of reference and
models for guidance, thus preventing local norms from diverging too far
from each other and resulting in international unintelligibility. In
addition, teachers, whether native or non-native, will also be required
from time to time to satisfy the minority of learners who desire to follow
the EFL rather than the EIL route, and achieve more or less native-like
J ennifer J enkins
articles welcome
proficiency. Again, the teachers ability to approximate closely to a
standard native model will be essential to this end. In this context, it is
important to emphasize that we should all guard against political
correctness, in the sense of telling our learners what their goals should
be: in particular that they should not want to sound like native speakers
if they clearly wish to do so.
All teachers, native and non-native, will need to be well educated in the
three core phonological areas, i.e. sounds, nuclear stress, and articu-
latory setting. They will need a thorough grounding in how and where
sounds and stress are produced, in the rules for elision as they relate to
acceptable and non-acceptable consonant deletion, and in the nuclear
placement system. They will also need to be well informed as to how
their learners do these things in their L1s, so that they can introduce
contrastive work into the classroom as a means of enhancing productive
competence. It is a current irony that although pronunciation teaching .
tends to be marginalized throughout the ELT world, it is non-native
teachers who are generally the better versed in all these areas, and thus
the better prepared to embark on teaching pronunciation for EIL.
L2 sociolinguistic Phonological variation, whether L1 or L2, can be both inter- and intra-
variation speaker. Before teachers are likely to promote L2 inter-speaker
variation in the classroom, they will need to experience a change of
attitude towards it and, in turn, be equipped with the means of changing
their students attitudes (and this includes native-speaker students,
possibly at secondary school level). Such an attitude change among
teachers will best be effected by the introduction into teacher education
courses of a sociolinguistics component dealing specifically with the
social and psychological issues involved in accent variation. Translated
into pedagogy, inter-speaker variation implies the inclusion of extensive
exposure to different L2 varieties of English, particularly in the form of
contrastive work, to make the differences salient for learners, and thus
to enhance their receptive competence for EIL.
Intra-speaker variation in an EIL context refers to the fact that non-
native speakers may be motivated to accommodate their non-native
receivers when their own desire to be understood is especially strong.
This accommodation is expressed particularly in the suppression of
those features of L1 transfer (often relating to core sounds) that have
the potential to obstruct meaning for a specific receiver (cf. Jenkins
1996c). It can be elicited especially by means of information exchange
tasks between learners from different L1 backgrounds. However,
teachers need to be made aware that suppression of L1 transfer through
accommodation cannot be expected to occur in any task type, or
between learners from the same L1 background.
Clearly, then, both types of variation can be embraced more easily in
multilingual than in monolingual classrooms, which involve high
exposure to a single non-native variety and lack any genuine
communicative need to use English or improve pronunciation for
Pronunciation norms and models 125
articles welcome
intelligibility. Moreover, these last two factors sometimes seem to
engender a degree of embarrassment about approximating closely to the
L2 model in front of the peer group, especially in the case of adolescent
learners. Paradoxically, it is also more difficult to convince both teachers
and their learners from monolingual settings that L2 variation is
acceptable for EIL. Lacking an international context for learning, the
inclination tends to be towards EFL rather than EIL, regardless of the
future uses to which the learners will put their English. In addition to
reforming the pronunciation syllabus, two major tasks for EIL over the
coming decades are, therefore, to reposition the crucial pedagogical area
of pronunciation centre-stage rather than in the wings, and to make
multilingual classes available wherever in the world English is taught
and learned in order to serve as an international lingua franca.
Received February 1997
References
Crystal, D. 1996. The past, present and future of
English rhythm. Speak Out! 18: 8-13.
Dalton, C. and B. Seidlhofer. 1994a. Pronuncia-
tion. Oxford: Oxford University Press.
Dalton, C. and B. Seidlhofer. 1994b. Is pronun-
ciation teaching desirable? Is it feasible? in
T. Sebbage and S. Sebbage (eds.). Proceedings
of the 4th I nternational NELLE Conference.
Hamburg: NELLE.
Gimson, A.C. 1978. Towards an international
pronunciation of English in P. Strevens (ed.). I n
Honour of A.S. Hornby. Oxford: Oxford Uni-
versity Press.
J enkins, J . 1996a. Native speaker, non-native
speaker and English as a Foreign Language:
time for a change. I ATEFL Newsletter 131:
10-11.
J enkins, J . 1996b. Changing priorities for success-
ful communication in international contexts.
Speak Out! 17: 15-22.
J enkins, J . 1996c. Changing priorities. Speak
Out! 18: 33-40.
J enner, B. 1989. Teaching pronunciation: the
common core. Speak Out! 4: 24.
Kachru, B. 1988. Teaching World Englishes.
ERI C/CLL News Bulletin 12/1: 1-4.
Pennington, M.C. 1996. Phonology in English
Language Teaching. London: Longman.
Quirk, R. 1981. International communication and
the concept of nuclear English in L. Smith
(ed.). English for Cross-cultural Communi-
cation. London: Macmillan.
Widdowson, H.G. 1994. The ownership of
English. TESOL Quarterly 28/2: 377-89.
The author
J ennifer J enkins is Head of Teacher Education
and Lecturer in Applied Linguistics in the English
Language Teaching Centre and the Department
of English Language and Literature, Kings
College London. She has a PhD in TESOL from
the Institute of Education, University of London.
Before taking up her present position, she taught
EFL, ESL, and EAP for many years in a range of
language schools, adult education institutes, and
universities. Her current interest is the role of
interlanguage phonology in English as an Inter-
national Language, the subject of a forthcoming
book with Oxford University Press.
E-mail: <[email protected]>
126 J ennifer J enkins
articles welcome