Introduction
When the journal World Language English changed editors in 1985, the new editors changed its title to World Englishes. This was their explanation:
“Englishes” symbolizes the functional and formal variation in the language and its international acculturation, for example in West Africa, in Southern Africa, in East Africa, in South Asia, in Southeast Asia, in the West Indies, in the Philippines and in the traditional English-using countries: the USA, the UK, Australia, Canada and New Zealand. The language now belongs to those who use it as their first language and to those who use it as an additional language, whether in its standard form or in its localized form. (Kachru & Smith, Reference Kachru and Smith1985, p. 210)
The new editors thus gave themselves a generously inclusive map for their contributors to fill, encompassing all three of the interpretations of World Englishes (WEs) discussed by Bolton. The first of these is “as an umbrella label referring to a wide range of differing approaches to the description and analysis of English world-wide” (Bolton, Reference Bolton, Kachru, Kachru and Nelson2006, p. 240).
English worldwide is sometimes referred to as World English, to distinguish it from World Englishes. The second use of English worldwide refers to the “New English” found in the Caribbean and West African and East African societies such as Nigeria and Kenya and to such Asian Englishes as Hong Kong English, Indian English, Malaysian English, Singaporean English (Bolton, Reference Bolton, Kachru, Kachru and Nelson2006, p. 240). Bolton's third sense is the one most often associated with Kachru and his followers, their “world English paradigm.” This, writes Bolton, “is wide-ranging, and argues for ‘the importance of inclusivity and pluricentricity in approaches to the linguistics of English world-wide” (p. 240).
The Handbook of World Englishes (Kachru, Kachru, & Nelson, Reference Kachru, Kachru and Nelson2006), which contains the chapter by Bolton from which I have already quoted, is indeed wide-ranging and discusses the spread and diffusion of English both historically and currently, taking account of settler and postcolonial societies, but concentrating mainly on the inner and the outer circles of Kachru's model (Reference Kachru1992) and not the expanding circle; in other words, on English as a native language (ENL) and English as a second language (ESL) communities rather than on English as a foreign language (EFL).
For the purpose of the present discussion, I will concentrate on the postcolonial or ESL contexts, in other words, the so-called New Englishes (Brutt-Griffler, Reference Brutt-Griffler2002). The settler or ENL societies (Australia, Canada, Ireland, New Zealand, South Africa, the United Kingdom, and the United States) are closely associated with Standard British or American English (McArthur, Reference McArthur1992) and since British and American English are themselves considered very similar to one another, I shall not discuss these varieties here. Apart from their similarity to one another, always of course meaning the written language, it is also the case that descriptions of the standard language exist in each of these countries: United States, Canada, United Kingdom, Australia, New Zealand, and South Africa. It is therefore possible, in principle, for tests of Canadian English, American English, and so forth, to be developed. In practice, what we find is that there are tests of American English and of British English: Canadian, Australian, New Zealand, South African are all likely to use in their tests Standard British (or American) English.
McArthur (Reference McArthur1992) made the point:
Australian English: there are no syntactic features that distinguish Standard Australian English from Standard British English.
Canadian English: where Canadian English differs grammatically from British English, it tends to agree with American English.
Irish English: Standard Anglo-Irish is close to the Standard British English variety.
New Zealand English: Standard New Zealand English is to all intents and purposes the same as British English. (quoted in Davis, Reference Davis, Kachru, Kachru and Nelson2006, p. 517)
In all these cases, what we are considering is the written language. Most of the obvious variation that exists across language settings is found in the spoken language. The written language is more stable for two reasons: It is the goal and it is the product of institutionalized learning both for first language (L1) and second language (L2) learners.
Joseph (Reference Joseph2004) proposed two essential features of language: “communication and representation” (p. 19). Representation is typically conducted in the spoken language: “Language is our means of understanding the world and representing it in our minds, as well as of communicating with others.” Representing includes linguistic identity; “mother tongue is obviously the central component in the construction of . . . linguistic identity.”
Tests of Canadian, Australian, New Zealand, and Irish English (perhaps for asylum seekers and refugees) are likely to be based on the local variety, whereas tests intended for global use, focusing on communication, as opposed to in-country use, are more likely to be based on international norms, which, as I have just pointed out, are likely to be based on Standard British or American English norms. The great advantage of these two standards is that they are widely intelligible, precisely what is needed for communication purposes.
Expressions and demonstrations of identity, on the one hand, are largely made through a spoken L1, not through a written L1 (which may not be available) and not through an L2. Communication, on the other hand, is in both written and spoken language. In writing, the likelihood is that the participants will use the norms of Standard English. And in less formal settings, interlocutors across national boundaries cannot normally use their own L1 and so will use a hybrid—which operates as an internationally intelligible spoken language. Across boundaries, this cannot normally be any one speaker's L1, and for spoken communication, Standard Written English is likely to be overly formal. It is in this gap for spoken international communication that we may find some type of International English (also called World English) in use. This variety may be normed for international use, although it is much more likely that everyone will use their own version of (WEs), just as in interpreting, each speaker uses his or her code, which the interlocutor understands receptively but does not him- herself use productively. It is here, then, that there may be room for English as a lingua franca (ELF, which I discuss later in this article). It is here that we need to consider the demands of the assessment of WEs. For communication in the written language, as I have suggested, the demand is for Standard English, the variety that is used in the International English Language Testing System (IELTS) and the Test of English as a Foreign Language™ Internet-based-Test (TOEFL® iBT) (but see Lowenberg, Reference Lowenberg1993, Reference Lowenberg2002). And for identity demonstrations in the L1, as I have suggested, there is unlikely to be a major demand for assessment, except in the case of, for example, government tests for proof of refugee status, where what is being tested is not communication but, it is argued, a demonstration of identity.
These paradoxes are addressed in Table 1.
The Languages of Identity and Communication

To restate, assessment in Kachru's Inner Circle (United States, United Kingdom, Australia, etc.) is normed on the basis of the local variety of Standard English. Assessment in Kachru's outer circle will be as in Table 1. Assessment in Kachru's expanding circle is likely to make use of one or other of the inner circle Standard Englishes, depending on which inner circle country has had most local influence, although this is changeable. In Japan, for example, American English replaced British English as the main influence on English teaching after World War II (Honna, Reference Honna, Kachru, Kachru and Nelson2006). As noted earlier, to an extent this was not a major issue since the differences in terms of written communication between the inner circle Standard Englishes are minimal.
Communication across WEs is held to be unproblematic in terms of communicative competence (Berns, Reference Berns, Kachru, Kachru and Nelson2006) and intelligibility (Smith & Nelson, Reference Smith, Nelson, Kachru, Kachru and Nelson2006). Or perhaps I should say less of a problem than was suggested by, for example, Prator (Reference Prator, Fishman, Ferguson and Gupta1968) and Quirk (Reference Quirk, Quirk and Widdowson1985). Nevertheless, scholars closely involved in language testing are less optimistic. Davidson (Reference Davidson, Kachru, Kachru and Nelson2006) explained that he was favorably disposed toward the use of global varieties of English in international English tests, even though
There is a well-established and legitimate concern that large, powerful English language tests are fundamentally disconnected from the insights in analysis of English in the world contexts. These exams set forth linguistic norms that do not necessarily represent the rich body of English varieties spoken and used in contact situations all over the world. (Davidson, Reference Davidson, Kachru, Kachru and Nelson2006, p. 709)
Davidson argued that the fundamental disconnect he observed needs to be informed by “empirical data as a driving force in solving the World Englishes/testing issue” (p. 714). He cited work by Chalhoub-Deville and Wigglesworth (Reference Chalhoub-Deville and Wigglesworth2005) and by Davies, Hamp-Lyons, and Kemp (Reference Davies, Hamp-Lyons and Kemp2003) as contributing to the necessary empirical work. And he quoted with approval what he regarded as the key question in the World Englishes/testing debate, which Davies et al. used as the title for their 2003 article: “Whose Norms?” pointing out that what is needed to further the study of native speaker (NS) and nonnative speaker (NNS) raters by Chalhoub-Deville and Wigglesworth is extension into performance by raters who are themselves members of various WEs communities.
One such study is reported in Hamp-Lyons and Davies (Reference Hamp-Lyons and Davies2008). The authors conducted
a small empirical study to attempt to find support for—or evidence against—the view that international tests of English language proficiency are unfair to speakers of non-standard forms of English, since these tests privilege standard forms. We explore the questions of whose norms should be imposed in these tests and what the consequences for test-takers are if the norm imposed by the test is not the “normal” variety accepted in their own society. Data used for the study are written texts by English learners from six language backgrounds, scored by raters from their own language backgrounds as well as by native American English raters. (p. 26)
Ten essays (N = 60), representing a wide distribution of scores from six language groups (Arabic, Bahasa Indonesia/Malaysia, Japanese, Mandarin Chinese, Tamil, and Yoruba) composed the data.
All essays were written on the same or very similar topic from the MELAB (Michigan English Language Assessment Battery) pool: each essay received the same or very similar score from two official MELAB raters who were native speakers of Standard American English.. . . The essay sets were rated by pairs of raters: each pair shared their language background with one of the six sets of candidates: thus there were two native-speaking Japanese raters, two native-speaking Bahasa Indonesia/Malaysia raters, etc. We hypothesized that if there were bias, it would be reflected in a significant difference between our native speaking raters and the official MELAB scores. (Hamp-Lyons & Davies, p. 27)
Furthermore, Hamp-Lyons and Davies hypothesized that the two WEs groups (Tamil and Yoruba—labeled as WEs because both had experienced colonial imposition of English) were more likely to show difference than the Chinese and Japanese groups that had not had such influence. The other two groups—Bahasa Indonesia/Malaysia and Arabic—were regarded as in between the two extreme groups.
In the event, there was some indication of difference between the NS and NNS raters, but it was not possible to make any definitive claims. The study highlighted the difficulty of research in this area, in particular the selection of NNS raters, an issue discussed in an earlier article describing research in Indonesia (Hill, Reference Hill1996). There are two basic problems. The first is that in multilingual societies, which many WEs societies are, it is not always clear who is an NS of the target language. In the case of the Hamp-Lyons and Davies study. “Our ‘Yoruba’ raters turned out to be first language Ibo speakers whose Yoruba was near-native but not native” (Hamp-Lyons & Davies, Reference Hamp-Lyons and Davies2008, p. 30). The second problem is that those eligible to act as NNS raters, who of necessity must be near-native speakers of English, may well have internalized Standard English norms during their language acquisition. If that were the case, however sympathetic they may be to their local WEs’ variety, which they may themselves normally speak, their judgments, especially in a test situation, are very likely to be quite prescriptive.
At the end of their article, Hamp-Lyons and Davies referred back to the three questions they had posed at the outset. (p. 26)
How possible is it to distinguish between an error and a token of a new type?
If we could establish bias, how much would it really matter?
Does an international English test privilege those with a metropolitan Anglophone education?
They admitted that they were no closer to answers, but were becoming clearer as to how further research could help clarify whether they are the right questions to ask.
The celebration over the last 20 years or so of WEs has understandably put the status of the NS in question. Whose English is it? Who decides whether the English I deploy is correct? Whose norms do I appeal to? These are the questions demanding answers in the postcolonial literature. Ikome (Reference Ikome and Singh1998) maintained that NS is “foremost a political description for social empowerment or for peer recognition” (Ikome, Reference Ikome and Singh1998, p. 63). The postcolonial imperative is that, just as the Australian NS of English no longer takes account of the norms of British English, similarly the New Variety of English “native user” (as the proficient WEs speaker has come to be known) of, say, Singapore English, no longer takes account of these norms. There is a further, though related, challenge to the NS that comes from English as a lingua franca (ELF, the variety used by nonnative speakers of English to interact with one another in business, professionally, and so on (Jenkins, Reference Jenkins2007). Although the term native user does not appear in the ELF literature, it seems reasonable to use it to refer to those proficient in ELF. The ELF assertion is that English is increasingly being used as a vehicular language among NNS interlocutors who do not speak one another's language. As Crystal (Reference Crystal1995) and Graddol (Reference Graddol1999) pointed out, such speakers now vastly outnumber NSs. ELF speakers are said to be less focused on form, and more concerned with reaching for the message. What matters is intelligibility rather than correctness. Misunderstandings are, it is claimed,
resolved by communication strategies, interference from L1 interactional norms is rare, ELF talk is overtly consensus oriented, cooperative and mutually supportive and thus fairly robust. (Seidlhofer, Reference Seidlhofer2004, p. 6)
In their 2006 article, Elder and Davies proposed that: “Post-colonial—or world Englishes—such as Indian English could be classified as a subset of ELF” (p. 283). They point out that there are other ways of considering ELF, all of them assuming a dependence on Standard English. For testing purposes, it is essential to be clear as to the status of ELF. If it is viewed as dependent on Standard English, then tests of ELF need be no different from existing tests of Standard English, using, where appropriate, approximations and adjustments so that what would be classified as an error in Standard English is recognized as an acceptable accommodation in ELF. However, if ELF is to be regarded as nondependent on Standard English and as a new and discrete international variety, then it is necessary for testing purposes to know what it is: A language test requires a description of the language under test to establish the criterion against which to measure performance. (This is an argument that applies more generally to WEs). Without that description, it is not easy to see how assessment is possible. Such is the classical view of a language test where the emphasis is primarily on language form. There is another view, not unlike that of specific purpose tests (Douglas, Reference Douglas2000), where the criterion is not linguistic form but functional effectiveness, more concerned with getting the message across. Such a view is attractive and seems to deny the necessity for a language description. Seidlhofer's comment that ELF talk seems to be “overtly consensus-oriented, cooperative and mutually supportive and thus fairly robust” (Reference Seidlhofer2004, p. 218) is an appropriately communicative view of interaction, but of course there has to be some formal marker that the interaction is in ELF and not in two quite unconnected languages. That being so, the question remains: What is in and what is out for ELF? The answer may be that it is up to the participants to decide for themselves during an interaction, the conduct of which demonstrates just what counts as ELF. If that is so, as it may well be, there is a problem: The tester needs to know what counts as interaction success, not a successful end to the interaction (e.g., agreement on a sale or confirmation of an arrangement) but a satisfactory and intelligible conversation. After all, much successful interaction is not about winning or losing but about satisfactorily intelligible contact. If or when ELF can be demonstrated to constitute a new code in its own right,
then we need to use that code rather than Standard English as a basis for our test construction (Elder & Davies, Reference Elder and Davies2006, p. 290).
What are regarded as accommodations to Standard English from a Standard English view of ELF
would no longer be seen as accommodations but instead would become an integral part of the new test construct. . . . Test items or tasks used to elicit performance would be sampled directly from domains relevant to ELF communication and therefore would contain lexical, grammatical, phonological, prosodic and discourse features peculiar to ELF. (Elder & Davies, Reference Elder and Davies2006, p. 290)
Underlying the urgent quest for acceptability of WEs, including ELF, lie two ironies. The first is that calls for new approaches to the assessment of WEs appear to be “stronger on politics than on applied linguistic realities and represent a claim to the ownership of English” (Elder & Davies, Reference Elder and Davies2006, p. 296). No harm in that, of course, except that such a motive is about identity rather than about communication. But as I have suggested, for ELF speakers, indeed for WEs speakers more generally, identity is likely to reside in the L1: What WEs offer is a communicative opportunity. That is less a political than a linguistic means to a business-like outcome. There is a further irony: that if WEs (including ELF) were ever to provide their own basis for measurement, then they would exclude all nonstandard WEs speakers, just as Standard English is said to exclude WEs speakers.
The first of these ironies applies to the measurement of WEs quite generally, the second perhaps more particularly to ELF. In most WEs settings it does seem that what is at issue is not so much the spoken language in its everyday use where the use of WEs is normal and accepted; what is at issue are formal tests of speaking and of writing. It is these where the norms of Standard English still persist. If WEs are to strike out for their own norms, that is a choice that they must make. Only their own uncertainty prevents them.
The decision as to which norms to use in a formal WEs test is a language-planning decision, informing the selection of test instruments by national and international bodies and, if the decision is for the WEs variety, contributing to WEs standardization. But that decision can be made only if, as Haugen (Reference Haugen1966) made clear in his celebrated study of the relationship between dialect and language, between vernacular and standard, the speech community is prepared to accept the dialect or vernacular as its language, its standard.

