Table Of ContentPOINT AND COUNTERPOINT
The spread of EIL: a testing
time for testers
Jennifer Jenkins
Thisarticleargues that recent changesin bothusersand usesofEnglish have
become so far-reaching that a major rethink of English language teaching
(ELT)goalsiscalledfor.Itgoesontoclaim,however,thatthiswillfirstrequire
a substantial overhaul of English language testing, given that teachers and
learners alike will be reluctant to embrace any curriculum change that is not
reflected in the targets set by the major examination boards.
Introduction Sincethesecondhalfofthetwentiethcentury,theEnglishlanguagehas
spread around the world to an extent hitherto unknown in any other
historicalperiodorforanyotherlanguage.Ontheonehand,Englishhas
developed into a nativized language in many countries ofthe Outer
Circle, i.e. countries such as India, Nigeria, and Singapore, where it
performs important local roles in the daily lives of large numbers of
bilingual and multilingual speakers. On the other hand, it also serves
as a lingua franca among non-native speakers (NNSs) of English from
all over the world, many of whom come from the countries of the
Expanding Circle (i.e. countries for whom English does not perform
internal roles), whose members are more likely to communicate in
EnglishwithNNSsfromotherfirstlanguagesthantheirown,thanwith
either native speakers of English (NSs) or with people who share their
first language.
One result of this spread of English is that many sociolinguists have
beguntotalkof‘Englishes’or‘WorldEnglishes’ratherthan‘English’,in
recognition of the fact that the language now has a growing number of
standard varieties and not only two globally useful or appropriate
versions (standard British and standard American English). This view
of English recognizes that local linguistic and cultural influences have
affected the way it is spoken in its different L2 locations around the
world: its characteristic accents, its syntactic structures, its lexis, its
pragmaticfeatures,andthelike.Stillmoreimportantly,theviewaccepts
thattheseinfluences,throughnaturalevolutionaryprocessesoflanguage
contact (see Mufwene 2001), have led and are continuing to lead to the
emergence of a range of educated L2 English varieties which differ
legitimatelyfromstandardNSEnglish.Inotherwords,supportersofthis
view are able and willing to distinguish between NNS language variety
and interlanguage, that is, between acceptable NNS variation from NS
42 ELT JournalVolume60/1January2006;doi:10.1093/elt/cci080
ªªTheAuthor2006.PublishedbyOxfordUniversityPress;allrightsreserved.
English norms and NNS error caused by imperfect or incomplete
language learning. The logical extension of this position is that there
seems to be no good reason for speakers from the Outer or Expanding
Circles to continue to defer to the NSs of the Inner Circle, to bow to
exonormative (i.e. externally-dictated) standards and ‘conform to norms
which represent the socio-cultural identity of other people’ (Widdowson
in Howatt with Widdowson 2004: 361).
Ontheotherhand,largenumbersofELTprofessionalsincludingmany
(and possibly the majority) involved in language testing still appear to
regard any differences from British or American NS variants as
deficiencies—aserrorscausedmainlyby‘L1transfer’(orworseinterms
ofthe implicit attitude, ‘L1 interference’). There is, for many ELT
professionals, no possibility that an L2 speaker, however proficient, can
departfromNSnormsandyetberegardedascorrect;nopossibilitythat
NNSs of English as an International language (as contrasted with ENL,
EnglishasaNativeLanguage)canbecreativewiththestandardlanguage
inwaysthatarepermittedtoitsBritish,itsAmericanandrecently(ifstill
begrudgingly), its Australian L1 speakers.
Therestofthisarticlelooksattheseissuesinmoredetailandconsiders
the implications of moving away from NS norms for the testing and,
because of the washback effect, for the teaching of English as an
International Language.
Variation across Around twenty years ago, Quirk contended that ‘[T]he relatively narrow
Englishes: NNS range of purposes for which the non-native needs to use English ... is
varieties arguably well-catered for by a single monochrome standard form that
looksasgoodonpaperasitsoundsinspeech’(1985:6).Evenifweleave
asidethehighlyquestionableassumptionthatNNSshaveonly‘anarrow
range of purposes’ when they use English, Quirk’s claim has two
weaknesses: firstly, it takes no account ofthe well-documented
differencesbetweenspokenandwrittenEnglish;andsecondly,itignores
the vast amount of inter- and intra-speaker variation according to social
context that occurs within each channel, and particularly the spoken
channel.Inbothcases,variationperformsimportantlinguisticandsocial
functions regardless of whether the user is native or non-native.
What happens in It seems, then, unreasonable to expect NNSs to produce a more rigidly
examinations? consistent kind of English than is typical or expected of NSs. However,
almosttwodecadesafterQuirkmadehisclaim,thereisstillaninsistence
on ‘correct’ grammar and pronunciation in ELT examinations (and
therefore classrooms). The only rather limited progress that has been
made has been in the recognition of different levels of formality and
some of the ways in which these impact on lexico-grammatical choice.
Even in this respect, however, candidates in ELT exams are expected in
the main to produce informal lexico-grammatical items entirely in
accordance with NS norms (so-called ‘real’ English), and even then to
clothe informal lexical items in standard NS English written grammar.
For example, ‘there’s five cars in my picture’ or ‘I’ve got less cars in my
picture’, if noticed by an examiner, would not be likely to meet with
approval, despite the fact that both ‘there are’ and ‘fewer’ plus plural
ThespreadofEIL:atestingtimefortesters 43
count noun are rare in informal NS speech, and that these ‘errors’
generally pass unnoticed in NS spoken English.
The only exceptions to the ‘written-grammar-for-speech’ rule seem to
occur when a NNS reproduces an instance of NS lexico-grammatical
creativity that has passed into accepted (as opposed to unnoticed)
informalNSspeech.Forinstance,acandidateinanELTspeakingexam
wouldberewardedfortheirknowledgeof‘real’Englishiftheyweretosay
‘three teas’ or ‘two coffees’ instead of ‘three cups of tea’ or ‘two cups of
coffee’. On the other hand, ifthey extended this use of uncountable
nouns to ‘wine’ and referred to ‘two wines’ instead of ‘two glasses of
wine’,theycouldbepenalizedforlackofcompetencewiththecountable/
uncountable distinction. The same is likely of any NNS use of
uncountable nouns as countable (‘a staff’, ‘four furnitures’, etc.), even
though these forms are standard in many ofthe nativized Englishes of
the Outer Circle and used by many speakers ofthe Expanding Circle.
And the same is also true of pronunciation where, for example, the
majority ofthe world’s (NNS) English speakers have extended the use
of/t/ and /d/ or /s/ and /z/ to the voiceless and voiced dental fricative
sounds for ‘th’, so that ‘thin’ and ‘this’ are pronounced ‘tin’ and ‘dis’ or
‘sin’ and zis’. Again, though widely intelligible in EIL, these forms are
penalized in ELT exams, and consequently discouraged in ELT
classrooms.
As Lowenberg (2002) demonstrates, the creative processes involved in
NSand NNS linguistic innovation tend to be the same. But while both
types of innovation often start life as forms that are widely perceived
as errors in the standard language, the NS ‘error’ gradually becomes
acceptedasanewstandardform(forexample,theuseof‘data’toreplace
‘datum’ in the singular), whereas the NNS ‘error’ is likely to be
categorizedassuchforperpetuity—despitetheonlydifferencebeingthat
the first reflects NS creativity and the second NNS creativity. What we
seemtobefacedwithhere,then,isanaspectoflinguicism:thevaluing
of NS English language forms above those of NNSs even though the
former do not lead to greater communicative efficiency for the majority
in international contexts of use (Ammon 2000).
WhereELTexamsareconcerned,thestatusquoleads,thus,toabizarre
stateof affairs,with candidates examinedfor qualifications whichclaim
tohaveinternationalcurrency(TOEIC,IELTS,andsoon),butpenalized
forusinginternationally-communicativeformsofthelanguage.Thereis
nothing‘international’aboutdeferringtothelanguagevarietiesofamere
twooftheworld’sEnglishes,whosemembersaccountforatinyminority
of English speakers. Nor is there reason to suppose that the study of
British or American English will promote international understanding.
This, as Matsuda (2002) points out, is more likely to occur through
a more equitable representation of World English varieties in ELT. The
merefactofhavinganearlierplaceinthechronologicaldevelopmentof
the English language does not confer everlasting rights of ownership.
Sociocultural theory FurthersupportforNNS varietiesofEnglishintestingandteachingcan
be found in sociocultural theory. Here social context is paramount, the
44 JenniferJenkins
construct of mediation is central, and language is seen as being learnt
through the medium of interaction in context. As Pavlenko and Lantolf
pointout,secondlanguagelearningisnotmerelyaquestionofacquiring
new grammatical, lexical, and phonological forms, but ‘a struggle of
concretesociallyconstitutedandalwayssituatedbeings toparticipatein
the symbolically mediated lifeworld ... of another culture’ (2000: 155).
Learners, according to sociocultural theory, ‘undeniably belong in their
second self-chosen world, not as observers but as fully-fledged
participants’ (op. cit: 159). In the case of English as an International
Language, most of the meaningful interaction occurs between NNSs
rather than between a NSand a NNS. Learners are present and future
membersofaninternationalcommunityconsistinglargelyofNNSs like
themselves, and in line with Donato (2000), are entitled, through their
contact with each other and with the L2, to transform their linguistic
world rather than merely to conform to the NS version presented
to them.
Thus, fromasociocultural perspective,NNS creativityistobeexpected.
NNS English variants have legitimacy and hence the right not to be
automatically relegated to the status of error. This in turn means that
accounts of variation across Englishes (interspeaker variation) need to
includealltheNNS Englishesforwhichwehaveinformation.(Currently
thesearemainlyOuterCircleEnglishesbutalso,increasingly,Expanding
CircleEnglishes.)Testersfortheirpartneedtorespondbytakingaccount
ofthis variability and—at the very least—not penalizing candidates for
employingitwithcommunicativesuccess,whetherinspeechorwriting.
However,justasimportantforsuccessfulEIL interactionisintraspeaker
variation. Whereas interspeaker variation concerns the influence of the
widerEIL socialcontext,intraspeakervariationinvolvesthecontextofthe
specificinteractionandthewayinwhichindividualsadjusttheirspeech
to accommodate to the needs of their interlocutors. It is to this kind of
variation, accommodation, that we now turn.
Variation within In the discussion that follows, I will be focusing specifically on spoken
Englishes: English and speaking tests, although some of the points could
accommodation in conceivably apply to written English and especially to electronic
NNS interaction communication.
Accommodation is a major factor in almost all spoken interaction
regardless of whether it involves NSs or NNSs. When speakers adjust
their speech to make it more like that of an interlocutor, they are
employing a strategy known as ‘convergence’ and when they do the
opposite,thestrategytheyuseistermed‘divergence’.Therearetwomain
motivations for convergence: an affective motivation (the desire to be
liked),andacommunicativeefficiencymotivation(thedesiretobeeasily
understood). In NNS–NNS communication between speakers from
different first languages, the latter motivation is thought to be
particularly salient. (See Jenkins 2000: Chapter 7.)
ConvergenceinEIL communicationhasbeenshowntomanifestitselfin
three main ways: speakers may converge on one another’s forms, they
may converge on a more targetlike form, or they may avoid a NS form.
ThespreadofEIL:atestingtimefortesters 45
Regardless of the communicative outcome, however, testing procedures
are likely to penalize the first and third of these manifestations of
convergence and reward the second.
Converging on one Thisfirsttypeofteninvolvesonespeakerreplicatingthe‘error’ofanother.
another’s forms This is also common in NS–NSinteraction where, for example, one
speakermaysubstituteaword-finalormedial/t/withaglottalstop,asin
thewords‘right’and‘water’,andaconversationpartnermaythendothe
same. Ultimately, though, this kind of convergence is thought to lead
to language change, depending on both how widespread the use of
aparticularformandtheextenttowhichthelanguageispredisposedto
moveinthisdirection.For,asMufweneargues,‘[T]heagencyofchange
lies definitely within the behavior of individual speakers, and causation
partlyinthemutualaccommodationstheymaketoeachotherwhilethey
are more intent on communicating effectively than on preserving
idiolectal, dialectal, or language boundaries’ (2001: 24).
When NNSs communicating with one another engage in this kind
of convergence, however, the outcome is regarded not merely as
astigmatizedform(asintheNSexampleabove),butasan‘interlanguage
error’. So, for instance, if a German speaker of English substitutes /w/
with /v/ when conversing with a Turkish speaker (for example, in the
word ‘wind’), or an Italian omits the sound /h/ in communication
with a French L1 speaker (as in the word ‘hotel’), the result is
a phonological error. But this ignores the fact that in so adjusting their
pronunciation, the German and Italian render their English
pronunciationmoreratherthanlessintelligibletotheirinterlocutor.The
sameistruewhereinterlocutorsarebothfromthesameL1.TwoKoreans
taking part together in an oral test may both substitute /f/ with /p/
(for example, ‘pamily’ for ‘family’) while two Germans may substitute
final voiced consonants with voiceless consonants (for example, ‘roat’
for ‘road’). Again, the result is that they are able to communicate more
efficiently within the context of the interaction.
Converging on Unlike the first type of convergence, this second type occurs when
a more target-like speakers do not share either the same L1 or a mutual L2 English form.
form Instead of converging on a NNS variant, speakers in the presence
of certain conditions may converge on a more target-like form. The
conditions are:
n the target-like form is in their linguistic repertoire
n they perceive their L1-influenced rendering of a form as threatening
intelligibility for a specific interlocutor
n they regard being understood as important to the outcome of the
interaction (for example, some kind of task completion)
n theyhavetheopportunitytomaketheadjustment(especiallyfreedom
from processing overload).
In the pronunciation data I collected from students practising for the
UCLES (now Cambridge ESOL) Certificate of Advanced English
speaking exam,clearpatterns arose in the use andextent ofthis type of
convergence:itoccurredstatisticallymorefrequentlywheninterlocutors
46 JenniferJenkins
were engaged in information-exchange tasks than in social interaction
tasks (informal chats); it often did not involve adjustments to certain
sounds shown elsewhere to be inconsequential in EIL, particularly
substitutionsofvoicedandvoiceless‘th’;anditdidnotoccuratmoments
of processing overload, for example, when a candidate was searching
their mentallexiconfor aforgottenword. Candidates from different L1s
produced many more target-like pronunciations in the information-
exchangephasesoftheexamthaninthechats.Ontheotherhand,when
these same candidates were paired with interlocutors from their own
L1, they reverted to the first type of convergence and produced many
moreformsinfluencedbytheirL1,asdescribedabove.Thisdemonstrates
that the use of more target-like forms in the different-L1 pairings
was born of a desire to be understood rather than a desire to be
‘native-like’—something that was confirmed by questionnaire
responses (see Jenkins 2000).
The problem is that the second type of convergence is rewarded in
speaking examinations not because ofthe use of convergence to bring
about a successful communicative outcome, but purely because the
resulting form is close to the standard NS variant. Conversely, in the
first type of convergence, if a NNS variant is used to promote mutual
intelligibility, candidates may be penalized. In other words, testing is
here penalizing rather than rewarding appropriately-used
accommodationstrategies.Testingpracticeinturninhibitstheteaching
of accommodation strategies, by discouraging the use of forms (written
as well as spoken in the case of lexicogrammar) which are not standard
in NS English but are nevertheless communicatively efficient in NNS
interactions. This is particularly shortsighted for the teaching of EIL,
where students are likely in their future use of English to interact with
speakers from a wide range of NNS varieties and need to develop
accommodation skills appropriate to this kind of interaction.
Avoiding certain Thethirdtypeofaccommodation,theavoidingofcertainforms,involves
forms idiomaticlanguage.Itisoften‘unilateralidiomaticity’(Seidlhofer2001),
where one speaker employs a NSidiom that is not known to the other,
which poses the greatest threat to intelligibility in EIL. In the following
typical example,1 a Korean speaker of English has just asked a French
speaker of English what he likes to do in his spare time:
French L1: I like er... I like chilling out.
Korean L1: Hmmm?
French L1: Doing nothing.
Korean L1: Ah.
The Korean was evidently not familiar with the expression ‘to chill out’
(and nor was the Japanese student to whom it was repeated a few days
later). If successful communication was the primary object of the
interaction, the French speaker would have been more effective had he
usedtheneutralverb‘torelax’.However,inanexamsituation,hewould
have been rewarded for his knowledge of the idiomatic NSexpression,
whilehisunfortunateinterlocutormayhavebeenpenalizedforthe‘gap’
inhisknowledge.YetasfarasEIL isconcerned,theFrenchspeakerwas
ThespreadofEIL:atestingtimefortesters 47
responsibleforthebreakdownofcommunicationbecausehedidnotpay
attentiontotheintelligibilityneedsofhisinterlocutorandaccommodate
by avoiding NSidiomaticity.
It seems, then, that testing (and therefore teaching) at present
discourages the development of vital accommodation skills by, on the
one hand, penalizing their use whenever the result is not native-like
productionregardlessofanybeneficialeffectoncommunicationand,on
the other hand, rewarding their use whenever the result is native-like
production, regardless of any negative effect on communication. Until
the examination boards are able to conceive of ‘correctness’ as being
communicativelymotivatedandcontextuallyrelevant,thereislittlehope
that classroom practice will change in this regard. However, given that
accommodation has a major role in language change and that the
majority of English speakers are now NNSs, the indications are that
ultimately little can be done to halt the progress of NNS forms in EIL.
The examination boards would be well advised later if not sooner to act
on developments in World Englishes.
Taking change Some ofthe examination boards, it has to be acknowledged, are already
on board: confronting the issue of how to make English language testing more
recommendations relevant to the international needs of many test-takers, rather than
for testing EIL continuing to consider relevance purely in relation to what is real and
grammatical for the relatively small number of ENL and EFL speakers.
Taylor reveals that ‘[a]s a major worldwide provider of English language
tests, Cambridge ESOL has been grappling with these issues for some
years’(2002:19).Unfortunately,though,practicaloutcomesaretrailing
badlybehindtheoreticalgoodintentions.Thisisnotentirelythefaultof
the examination boards. Lack of descriptions of EIL varieties is a major
obstacletoprogress.WhileseveralOuterCirclevarietiesofEnglishhave
been codified, descriptive work has only just begun in the Expanding
Circle.Atatimeofmajorchange,Englishlanguagetestingstillhastobe
conducted in relation to some sort of standard. And ELTexaminations
also have to provide for those students whose preferred goal remains,
despite EIL developments, a near-native variety of English.
Thereisneverthelessmuchthattheexaminationboardscoulddointhe
meantimetodemonstratetheirwillingnesstoembraceNNS-ledchange
in practice. It is not a question of asking them to ignore standards
altogether, to move to a situation where ‘anything goes’. Instead, I
recommendaninterimphaseinwhichtheybaseEIL testingcriteriaon
empirical evidence from EIL interactions as it becomes available, and
avoidsettingcriteriaforwhichthereisnosuchevidence.Examplesofthe
sortofevidencetheymightconsiderincludeNNS–NNS corporasuchas
Seidlhofer’s Vienna-Oxford International Corpus of English (Seidlhofer
2001) and Mauranen’s Corpus of Academic English (Mauranen 2003),
and for pronunciation, the Lingua Franca Core and research into
phonological accommodation (Jenkins 2000). In all these cases, the
focus is on successful communication between NNSs rather than
between a NNS and NS.
48 JenniferJenkins
While it is too early to suggest that the examination boards should be
establishingEILcriteria,theycouldatleastmovetoapositionwherethey
refrain from penalizing the use of those NNS variants which are
emergingthroughtheirfrequentandsystematicuseaspotentialformsof
futureEILvarieties.Thiswouldsendastrongmessagetoteacherswhose
students’aimisEIL,thattheycouldsafelystopcorrectingitemssuchas
substitutions of ‘th’, uncountable nouns used as countable, omission of
articles,ortheuseofanall-purposequestiontag(forexample,‘isn’tit?’).
Instead, exams—and therefore teaching—could turn their attention
to rewarding the successful use of accommodation strategies and
penalizingtheirabsence,andtofocusingforerrorcorrectionontheuse
of forms that are not mutually intelligible in EIL, such as NSidioms.
Iendwithtworequeststotheexaminationboards.Thefirstconcernsthe
washback effect. A few years ago, Saville and Hargreaves claimed that
UCLES examinations had ‘kept pace with changes in English teaching,
so that modifications to the examinations have taken place in an
evolutionary way’ (1999: 42). This claim fails to recognize the order of
events.For,asDavidsonhasindicated,‘thedeterminationofwhatisand
is not an error is in the hands of the linguistic variety that sets the test’
(1993: 116). It is changes in teaching which keep pace with changes in
testing and not vice versa. This is why it is so very crucial for the
examination boards to engage with EIL. Unfortunately their apparent
inaction in this time of shifting sands means that they risk seeming to
bury their heads in them.
My second request concerns the nature of EIL. If and when the
examination boards finally embrace EIL, they need to guard against
producing testing criteria that differ only in name from their existing
criteria.Daviesetal.arguethatexistinginternationaltestsofEnglishare
‘biased’ because they ‘represent the old colonial Standard English of
the UK, USA, etc.’ (2003: 571). An EIL approach to English should be
a pluricentric approach—one which recognizes that while speakers of
English around the world need sufficientin common to enable them to
communicate,theyarealsoentitledtouseEnglishvarietieswhichproject
their identities and protect their language rights in international
communication.SomeoftherecentorientationstoEIL appeartobelittle
different from traditional monocentric orientations: ‘World Standard
English’, ‘World Standard Spoken English’, ‘International English’,
and ‘Global English’ on closer examination all turn out to bear an
unmistakable resemblance to standard NS English. If there is to be any
major change for the better as far as EIL is concerned, the examination
boards should beware of moving forward only to find they are back
where they started.
Notes native speakers?’ in R. Phillipson (ed.). Rights to
1 DataprovidedbyMartinDewey,PhDstudentat Language. Equity, Power, and Education. Mahwah,
King’s College London. NJ: Lawrence Erlbaum and Associates, Inc.
Davies, A., L. Hamp-Lyons, and C. Kemp. 2003.
References ‘Whose norms? International proficiency tests in
Ammon, U. 2000. ‘Towards more fairness in English’. World Englishes 22/4: 571–84.
International English: linguistic rights of non-
ThespreadofEIL:atestingtimefortesters 49
Davidson, F. 1993. ‘Testing English across Pavlenko, A. and J. Lantolf. 2000. ‘Second
cultures: summary and comments’. World language learning as participation and the
Englishes 12/1: 113–25. (re)construction of selves’ in J. Lantolf (ed.).
Donato, R. 2000. ‘Sociocultural contributions to SocioculturalTheoryandSecondLanguageLearning.
understanding the foreign and second language Oxford: Oxford University Press.
classroom’.In J.Lantolf(ed.).SocioculturalTheory Saville, N. and P. Hargreaves. 1999. ‘Assessing
and Second Language Learning. Oxford: Oxford speaking in the revised FCE’. ELT Journal 53/1:
University Press. 42–51.
Howatt, A. P. R. with H. G. Widdowson. 2004. A Seidlhofer,B.2001.‘Closingaconceptualgap:the
History of English Language Teaching 2nd edn. case for a description of English as a lingua
Oxford: Oxford University Press. franca’. International Journal of Applied Linguistics
Jenkins, J. 2000. The Phonology of English as an 11/2: 133–58.
International Language. Oxford: Oxford University Taylor, L. 2002. ‘Assessing learners’ English: but
Press. whose/which English(es)?’ Research Notes 10.
Lowenberg, P. 2002. ‘Assessing English Cambridge: University of Cambridge ESOL
proficiency in the Expanding Circle’. World Examinations.
Englishes 21/3: 431–35.
Matsuda, A. 2002. ‘International understanding The author
throughteachingworldEnglishes’.WorldEnglishes Jennifer Jenkins is Senior Lecturer in the
21/3: 436–40. Department of Education and Professional
Mauranen, A. 2003. ‘The corpus of English as Studies,King’sCollegeLondon,wheresheteaches
lingua franca in academic settings’. TESOL World Englishes, sociolinguistics, and English
Quarterly 37/3: 513–27. phonology/phonetics. She is the author of The
Mufwene, S. 2001. The Ecology of Language Phonology of English as an International Language
Evolution. Cambridge: Cambridge University (Oxford University Press 2000) and World
Press. Englishes (Routledge 2003), and is currently
Quirk, R. 1985. ‘The English language in a global writing a book about perceptions of English as
context’ inR.QuirkandH. G.Widdowson(eds.). an International Language.
English in the World: Teaching and Learning the Email: [email protected]
Language and Literatures. Cambridge: Cambridge
University Press.
50 JenniferJenkins