Interview with Rajend Mesthrie

When we started this interview by email, I immediately jumped in and asked you about theory. I wanted to find out how it was that you came to work across the fields of language contact, second language acquisition, and variationist sociolinguistics, all which informs your take on World Englishes. I’ve always valued the way that you combined those in your work. But my question was putting theory before data! When we met up and started talking about your trajectory, the conversation was about how in South Africa you can study all of these things.

At the crucial time when I was undertaking graduate studies in South Africa and the USA, there was a lot going on in terms of multilingualism in South Africa. There were so many unstudied or understudied phenomena that I was spoiled for choice. It seemed to me that the most urgent of these was to record whatever I could on the receding Bhojpuri language in its post-plantation context. But I was also aware of varieties of L2 or L2-turning-into-L1 English in daily use that were largely undocumented. I initially didn’t feel competent enough to make forays into the linguistics of African languages, but was familiar with the pidgin Fanakalo and realised that sociohistorical work on its KwaZulu-Natal settings was overdue.

You told me about how you were surrounded by all of these varieties in early childhood growing up on a farm in KwaZulu-Natal.

With only a slight exaggeration I can be said to have grown up with no named human language. We were descendants of third generation migrants on my mother’s side and fifth generation on my father’s side. Most of my grandparents and great grandparents had been indentured labourers—the two exceptions were grandmothers who were born in the colony, worked on the farms without being indentured and later married incoming indentured migrants from Uttar Pradesh, India. The language of the home was an offshoot of Bhojpuri, much changed by the experiences of migration by ship and plantation life—so much so that while we referred to it as “Hindi,” people familiar with the standardised language called ours “kitchen Hindi.” At the same time us younger ones started using some English, brought home from school by our two elder sisters. It was very much an L2 variety turning into L1—and would have been considered “broken English” in earlier records of the colony. I was familiar with the standard written language from reading anything I could lay my hands on from my elder siblings’ books—but didn’t speak it. Looking back, I think this is one of the reasons why we were so shy in talking to strangers (silence was easier for rural people than talking in standard English and its effete forms like good morning, how are you, etc.). And we didn’t speak Zulu, though we thought we did: it was a surprise to me later on to see the way we spoke described as a “pidgin” (technically true of Fanakalo, as it was a simplified and mixed lingua franca and no one’s mother tongue). This might not seem the most promising start for an eventual career in linguistics. However, I did have an early flair for reading, writing, and speechmaking in standard English.

Was there anything else at school that sparked your interest in studying languages and linguistics?

Another factor which boosted my entry into the world of “named languages” was that I belonged to the second last cohort in South Africa allowed to take Latin (as a second language) for four years in high school. It had been a compulsory language in previous times, but in the apartheid era (post 1948, and certainly in the late 1960s) was starting to be phased out in favour of Afrikaans as a second language. Our primary school had been so remote that it didn’t have a teacher with a knowledge of Afrikaans (all teachers had to be of Indian descent). And so, when we had to take a second language at high school the apartheid desideratum that this be Afrikaans had to be set aside. I probably owe my career in linguistics to this fact, since—unlike most of my friends—I found Latin entirely congenial. It gave me entry into a world of technical linguistic terminology, copious vocabulary, and rigorous structure.

Were you able to pursue those interests as an undergraduate?

Apartheid laws determined that I undertook undergraduate studies at the university reserved for “Indians.” The teaching bursary I took required a major in education and two teaching subjects (in my case mathematics and English—I had hoped to squeeze in Latin as a minor at least, but it clashed on the timetable with mathematics). The bursary was a four-year allowance that covered most costs except accommodation, in return for which one had to teach for the provincial department for a period of five years. While constrained by the dictates of apartheid, the University of Durban-Westville (UDW) turned out a number of graduates who gained prominence in the post-apartheid era in fields like law, drama, education, politics, psychology, the sciences, economics, and so forth. Indian languages were taught on the campus, but at that stage I couldn’t see myself as a scholar of Indian languages. Their overriding association was with religion and traditional culture, and English was considered much more attractive intellectually and economically. No African language was offered on the campus prior to and during my years of study there.

And then what happened after Durban-Westville?

It took an alienation from my familiar KwaZulu-Natal environment and the Indian communities there to kindle an interest in the linguistics of Indian and African languages. After a two-year stint of high-school teaching, I won a scholarship to study for a year-long honours degree at UCT (University of Cape Town) then reserved for Whites mainly (with perhaps a 2% of the student body allowed in, if their bona fide subject of choice was not offered at a Black/Coloured/Indian university). In my case that was linguistics, which I had thoroughly enjoyed as part of the English curriculum and which started me thinking about alternate perspectives on language than hitherto offered by classical Latin and prescriptive English grammar. At UCT I took five year-long modules on English language and the history of English.

I know that you were taught by John Coetzee when you did your honours (4th year) at UCT. What was that like?

J. M. Coetzee—later Nobel Laureate in Literature—taught the linguistics and literature module and gave me the freedom to write term papers on language rather than literature. I realise now that those two term papers (on a structural analysis of Norse myths and of Tamil kinship semantics) were of credible quality—though there was no pressure on students to publish then. It was J. M. Coetzee who recommended that I think of going to Austin, Texas, given my interests in both historical linguistics as well as logic and mathematics. Coetzee was one of several inspiring teachers I have had (others being Poobalan Pillay, of the mathematics department at UDW, Winfred Lehmann of the University of Texas (Austin), and Roger Lass, who never taught me—but was my PhD supervisor at a distance from UCT while I taught at UDW and my head of department when I joined UCT subsequently).

I remember when I was an honours student at UCT you told me about your MA year in Texas and how that was pivotal for you. How did your experience on that programme lead to your PhD research on Bhojpuri-Hindi and to your subsequent work?

Doing an MA at Austin, Texas was an incredible experience. At that time it was rated second in linguistics in the USA (to MIT), though I had settled on it because of its low fees, even for international students. (I also finished the course within three full semesters instead of the expected four, for financial reasons). It was here that I studied Hindi and Sanskrit for the first time in my life and encountered the names Bhojpuri and Fanakalo (in my library reading). And the compulsory courses in syntax, semantics, historical linguistics, and the like gave one the tools for a lifetime of language study. Sociolinguistics was on the curriculum, but I didn’t take it on account of having done that at summer school (Albuquerque, New Mexico) in the middle of my MA studies. But I knew that I was set for future studies of language contact, variation, and social history in South Africa.

This takes me to my first question by email which was how your early work on South African Indian English (SAIE) was informed by language contact on the one hand and second language acquisition on the other. This combination is evident in all your thinking on World Englishes, for example, your text book with Rakesh Bhatt (Mesthrie & Bhatt 2008). I recommend this text to students because few World Englishes textbooks (which are often aiming for geographic coverage) engage with these areas in depth. What was your first experience of these two fields?

I think the most important influences come from the context I’ve worked within. I was uneasy that linguistics was largely theoretical, and second language acquisition (SLA) was mainly considered applied and practical. (One could also say fashionable versus unfashionable; exciting versus ordinary, etc.). Yet in South Africa of the 1980s it was clear that the acquisition of English as well as of African languages as additional languages was a desideratum. The MA in Austin, Texas, featured an eclectic range of courses: historical linguistics, generative linguistics, phonology, Montague semantics—with anthropological linguistics, and sociolinguistics being options with something of a lesser profile. SLA was treated as a poor relative, with students having to move to education to gain a qualification in it. There wasn’t a course in language contact per se, though it turned up frequently in discussions of historical linguistics. I took summer school in linguistics at Albuquerque, New Mexico, thanks to a Linguistic Society of America fellowship and in fact took my first courses in sociolinguistics (with John Edwards) and pidgin and creole linguistics (with Derek Bickerton) there. Bickerton’s classes were riveting, not least for having scholars sitting in on them who vehemently disagreed with him. During a consultation after class Bickerton indicated to me that he had no special interest in second language acquisition and even with language attrition. But it was thanks to the setting of native American languages including Navajo in New Mexico, plus the range of visiting scholars and talks, that I knew that my prior intention of working on Bhojpuri-Hindi as a plantation language in South Africa (Mesthrie 1991) was a viable one from the perspectives of language contact. Overall, language endangerment was a relatively new field to which I thought I could contribute.

I resumed my junior lecturing post in Durban (at UDW) and completed my PhD on Bhojpuri as intended on a part-time basis. I was fully aware that the flip side of language loss on the plantations was the acquisition of English. I was also aware of the role of Fanakalo pidgin as the first lingua franca there. These were interests I had to postpone till later.

 So this means that I am largely “self-taught” in the area of SLA. My reading in the area came after my PhD, which was on koineization (the formation of a new dialect out of contact between different antecedent dialects) and language decline rather than new language acquisition.

On completing my PhD in 1986 I more or less immediately started a new project on the rise of Indian English in the plantations of KwaZulu-Natal (as a kind of “habilitation,” or “post-doc” though I didn’t know either term at the time). I completed extensive interviews and undertook analysis of the data in fairly quick time. I used Kachru’s (1983) book The Indianization of English for the background to English in India. And I was fairly “up” in contact linguistics. The original draft of my 1992 book didn’t have a dedicated chapter on SLA (Mesthrie 1992), but it did have a chapter on parallels between pidgin and creole formation and the rise of Indian English in KwaZulu-Natal. Wolfgang Klein who had read the manuscript suggested a chapter on SLA. And so I started reading in earnest. I was lucky that some readable compilations were just coming out—thanks to the interlanguage concept and SLA programme starting at the University of Edinburgh. But above all Rod Ellis’s book Understanding Second Language Acquisition had just come out in 1985. He is a great synthesiser and textbook writer. I liked the emphasis on data, no matter how makeshift, and the longitudinal perspective of speakers’ learning trajectories. In fact I was enthusiastic enough about it to start a course of twelve lectures at second year undergraduate level on SLA. I had first to convince my senior colleagues (Roger Lass and Nigel Love) that this was linguistics and not about language teaching—a taboo subject to linguists!

I returned to the States in 1989 on a six-month fellowship, taking a hot summer school in Phoenix, Arizona, and then a month at Urbana-Champaign, Illinois, in Braj Kachru’s department and reading up on new work on World Englishes. This was followed by a semester at Penn, where I sat in on William Labov’s fieldwork class, Sherry Ash’s acoustics and phonetics class, George Cardona’s class on Sanskrit grammarians, and Terry Pica’s SLA class. The SLA class attracted students from the Language in Education programme more than from linguistics per se; and provided a very up-to-date take on all the latest theories and ideas. But there was indeed a paradigm gap between World Englishes studies at Illinois and Pica’s SLA at Penn. World Englishes studies emphasized the multilingual reality of interference or transfer, while SLA studies focused on the monolingual norm as ideal and target. I was happy to learn things from both perspectives.

I took that course with you, and I still have my copy of Ellis (1985)! You said you were up on contact linguistics when you wrote the book on SAIE. Apart from your experience of pidgin and creole linguistics with Derek Bickerton, what exposure did you have to pidgin and creole studies? I always felt you were quite influenced by John Platt. He seemed, in his work on Singapore and Malaysia (Platt 1975), to be drawing on creolistics in the study of postcolonial Englishes, and you picked up on that. Did Platt inform your understanding of SAIE? Do you think that Platt played down substrate influence in favour of universals?

That’s an interesting observation about these diverse views and how they relate to my work as it unfolded, which I haven’t thought about much. I never met John Platt, but we do seem to have had similar interests—him in Malaysia and Singapore, and me in South Africa. I think he was an older scholar who got interested in first creole studies and then World Englishes (or the term he preferred—“New Englishes”). An early article of his on Singapore English as a “creoloid” (1975) resonated very much for me in terms of SAIE falling between the cracks of creole and non-creole (as more than one reader noted in connection with my manuscript for the 1992 book). Platt’s analysis of similarities across New Englishes in South and South-East Asia, Africa, and elsewhere (Platt, Weber & Ho 1984) certainly caught my interest in the mid 80s. And of course, I used it in the World Englishes book I co-wrote with Rakesh Bhatt (Mesthrie & Bhatt 2008), adding other varieties in the mix, including Native American Englishes and even (with Ray Hickey’s blessing) Irish English. Platt was obviously taken by Bickerton’s presentation of twelve features of creole grammar and—as you say—did take a universalistic stance on the New Englishes. But I never took him to be a substrate denier, somehow. Following up on his work it was easier to take a Mufwenian perspective (Mufwene 2008) from creolistics: that there were complementary roles for (a) SLA universals, (b) substrate influences, (c) the particularities of the target language, and (d) processing universals relating to economy of production of morphemes, etc. I came across the University of Pennsylvania thesis and 1987 article by Jessica Williams in English World-Wide (Williams 1987) on those parallel processing possibilities in that semester at Penn.

It was only once your major project on the “creoloid” SAIE was complete that you were able to properly tackle the pidgin Fanakalo. How did you approach Fanakalo? Growing up in South Africa I remember all the prejudices about this variety. Was it even considered suitable for academic study when you began this work?

I got interested in Fanakalo as I was familiar with it from my rural upbringing in KwaZulu-Natal. Indeed, most of the Indian users on the farm considered it to be a “low” form of Zulu, but Zulu nevertheless. I was intrigued—again while a graduate student at Austin—to come across an article by Cole (1953) on why Fanakalo was not really a Bantu language, but technically a pidgin. It had none of the agglutinative and alliterative concordial richness of a Bantu language and was relatively restricted in terms of its vocabulary compared to a “full” Bantu language. For example, in Zulu one might say “Abafana bami abakhulu bayasebenza bonke” (‘My big sons are all working,’ where the subject noun has the prefix aba- with which subsequent elements—possessive, adjective, verb, and adverb respectively—show agreement or concord with a similar sounding prefix. Had the initial noun been an animal, not a human, like izinkomo (‘cows’), then the concords would all contain a z- prefix). Cole’s article remains a classic today, even though I was able to prove it wrong in one respect. He had been aware that, although Fanakalo was a mining pidgin par excellence at the time of his writing, its origins were likely to be in the coastal areas of KwaZulu-Natal. He conjectured that indentured Indians were the originators of the pidgin in the late nineteenth century, with their need to undertake work in setting up the sugar, tea, and tobacco plantations, but without a knowledge of English or Zulu. I found this a very interesting hypothesis, but felt that it wasn’t convincing linguistically. So I set out to show why. I argued that whereas the basilectal and more makeshift “pre-basilectal” varieties of South African Indian English did show the influence of Indian languages at all turns and in lexis, this was not at all true of Fanakalo. I also showed a lack of correspondence between items from English in Fanakalo and the English of Indians on the plantations. But to really close the argument I decided to look at early (nineteenth century) manuscripts pertaining to communication in the colony. I struck gold with almost the very first unpublished manuscript I examined at the Killie Campbell Museum in Durban. In it a British settler of the 1840s described the fraught communication situation of the times, and while describing what he took to be Zulu, was clearly describing Fanakalo, a decade or two before the arrival of Indian indentured workers. I was subsequently able to push back the earliest examples of Fanakalo to the Eastern Cape of 1816, showing it to be in use by at least one missionary. Some people felt I was making an ethnic point in all this; but the truth is my interest was in pidgin origins versus second language acquisition (Mesthrie 1989) and the fraught nature of communication outside areas where Afrikaans had been earlier established as a lingua franca. Later on I wrote fairly extensively on Fanakalo as a mining language (Mesthrie 2019), focussing on its vocabulary; and on its split modules: lexis largely from Zulu and syntax a simple Subject-Verb-Object which looked a lot like basic English. When I started writing on Fanakalo, it was still denigrated by linguists and other intellectuals as a colonial tool and as an abomination compared to Bantu languages. There are, however, people with more positive attitudes than this, and for whom it enabled a living on plantations and in post-plantation society. I was at least able to show that it was more than interesting linguistically. The pidgin is slightly on the decline now. The mines do not accept it as the sole medium of communication with the labour force; and English is spreading in contexts where Fanakalo used to be the main medium of work in KwaZulu-Natal.

You mentioned that in 1989 you sat in on William Labov’s fieldwork class at Penn. Was this the start of your lifelong interest in variationist sociolinguistics? I remember that your sociolinguistics course which I took (and taught briefly in 1994, using your handwritten overhead transparencies) was mostly Labov! Yet it seems like it took some time before you were able to carry out variationist sociolinguistics. Is that fair?

Your observation about my relationship with variationist sociolinguistics is an astute one. In my PhD and post-doctoral work I was more interested in the effects of language contact and shift and the very salient repercussions for morphosyntactic structure rather than phonetics. Processes of morphological koineization evident in Bhojpuri out of the experiences of migration and plantation contact were far more salient than minor changes in phonetics. Thus the main dialect in coastal KwaZulu-Natal has endings that are clearly from Bhojpuri of India for the present tense of ‘I see’ (ham dekhi-lā), but the second and third persons have a form taken from Eastern and Western Hindi dialects (e.g., tu dekhe he ‘you see’). Gambhir (1981) had found similar koine effects in the Bhojpuri of Guyana. Gambhir was in fact a student of Labov and completed a fine PhD (1981) on structural outcomes of koineization mainly of a morphological and syntactic nature. While it wasn’t a classic variationist study, it adhered to the principle of accountability, always specifying proportions of usage of those morphological variants. It was also sociohistorically astute. Thus I had a viable, broadly variationist guide for my PhD work. It has been remarked—fairly—that even with SAIE I seem to have done little phonetic work in comparison with syntax and lexis. While I’ve provided basic outlines in the 1992 book and in a 2004 chapter for the Mouton Handbook of varieties of English around the world (Mesthrie 2004), the fine-grained variation by age, gender, ancestral languages (no fewer than six), and social class of the shifting generation would require a decade’s work that I’ve never quite found the time to manage.

Still, I always imagined myself to be doing a kind of social syntax arising from contact-related factors but responsive to syntactic variation. In fact, I read everything I could find on syntactic variation (which wasn’t really much). At Penn, Gillian Sankoff and Bill Labov encouraged me to write a paper on variation in the relative clause in SAIE. As a shifted variety it had almost everything a typologist, discourse specialist, and variationist could want, and I had great pleasure relating the topic to both the structural typology of Keenan and Comrie (1977) as well as to the variationism expected of LVC (the journal Language Variation and Change, in which the article was published in 1990). Showing that the correlations hold is much more challenging for syntax given the paucity of tokens per interview compared to phonetics. So, I enlisted the help of Tim Dunne of UCT’s statistics department who was able to vet and verify the major correlations (Mesthrie & Dunne 1990). He was a great supporter of our quantitative efforts in sociolinguistics at UCT and couldn’t believe it when two students in a later cohort mastered the programme R, after workshops from visiting scholar Thomas Hoffmann, then of Regensburg, and now Chair of Linguistics at the Catholic University of Eichstätt-Ingolstadt.

So, I was doing variationist work, mainly on syntax and therefore somewhat on the periphery of things. One paper that I think deserves more attention despite being published in LVC was a 2005 paper on illuminating the nature of a stereotype via variationist principles (Mesthrie 2005). This time it was more a morphological than syntactic variable—the use of bare -ing forms, by the earliest SLA learners (“I working in the mill,” denoting present habitual). Only a few speakers (whom I labelled “pre-basilectal” in the sense that theirs was an early interlanguage and not characteristic of the more focussed basilectal SAIE) used this form. But a booklet based on a radio series of the 1940s (The adventures of Applesammy and Naidoo, a series perhaps modelled on the US Amos and Andy series, stereotyping African Americans) gained full mileage out of the construction. By a complete quantitative analysis comparing actual excerpts from pre-basilectal users and the representations of modern insider writers from this booklet, I was able to show how the gross stereotype worked via the suppression of variation. There was quantificational generalisation at all levels: structural over-generalisation to contexts in which bare -ing was not used, stylistic over-generalisation, as well as ethnic over-generalisation.

Your work on the English accents of the first black South Africans to attend elite (formerly white) “Model C” schools is well known to many readers. It uses sociophonetic methods to capture the rise of a Black middle class post-apartheid. In fact by that point you had been working with speakers of Black South African English (BSAE—a variety and term you discuss extensively in your work) for some time, and you had published earlier studies of syntactic variation in BSAE. How did you first begin to collect this data?

My sample of forty of these students—the first in their families at university and the first in their communities to be at a “White” university—comprised long “Labovian” interviews, even though the 1997 study was ultimately geared toward second-language syntax and the building of a grammar (Mesthrie 1997). My interviews from that time are currently being digitised and form an important part of the social history of the times.

 Years later with the rise of large-scale corpora Bertus Van Rooy was able to monitor my methods and findings via the ICE Corpus of South African English (e.g., Van Rooy 2006) and learner corpora (Van Rooy 2008) of his own painstaking efforts. He was able to verify that my sociolinguistic approach, albeit with a small sample compared to his, gave the same robust results. He has since built a sociohistorical corpus of BSAE which has provided new insights into the evolution of the variety.

From about 2008 you were able to start the sociophonetic work. How did that come together?

My entry into sociophonetics was indeed relatively late. One reason is that I was quite content to leave the analysis of WSAE (White South African English) to Roger Lass. He was a first-class historical linguist and phonetician and within a couple of years of arriving in South Africa had written incisively on the phonetics of the variety including a chain shift of the short front vowels (Lass & Wright 1985). Roger was an inveterate talker and I a reasonably good listener so I picked up a great deal in passing and from reading his pre-publications. A lot of the minute details were new to me. Remember I grew up under apartheid, so I had no social contacts with Whites until the age of about twenty-four, and even then mainly via English honours at UCT classes and colleagues at UDW, rather than outside the “ivory towers.” Thanks to Roger I developed a reasonably good ear for minute variation, though I was something of a Pickering to his Higgins.

By about 2005 Roger was edging closer to retirement and I had the chance of a large research grant—my first really. (My first decade of research was funded by scraps from the research authorities, and I had to borrow funds and dig into my shallow pockets to attend conferences abroad during my first job at UDW—all good for the soul; and funding opportunities did improve after I joined UCT). What motivated me to promote sociophonetics to the top of my research agenda was the large-scale accent and dialect changes evident among young people of colour, who were matriculating from the more prestigious schools that now accepted (or had to accept) them. The accents of Black students entering UCT in the late 1990s was already quite different from those I had studied a decade earlier. There was even a contestation over the authenticity of the accents of those labelled “coconuts” by the mid-1990s by those who had been to apartheid-based schools for Blacks, a label and identity that continues to be controversial in the South African context. The term has been widely known since then referring to those who are allegedly “black on the outside, white on the inside.” So I began to do extensive interviews with the new generation, paying attention to their experiences and dilemmas while being immersed in a “white” or multiracial world that was a far cry from their parents’ generation. The goose vowels of these UCT students are reported on in the 2010 paper—with the /u:/ vowel in words like goose, true, move, and rude considerably fronted in keeping with global trends (Mesthrie 2010).

After 2008 I consolidated this research taking the tape recorder into five cities interviewing people of four historical ethnicities (White, Coloured, Black, and Indian) mainly in their homes. I was the main interviewer with some graduate students being trained to undertake interviews of their own. For the first time in my career I incorporated acoustic work with large data sets and had to train myself in advances in acoustics, way ahead of my initial training in Texas as an MA student with the huge Kay sonograph in a small lab. Funding from the NRF (National Research Foundation) was generous, and we managed to attend workshops and conferences abroad and even invite some scholars to run workshops for us (Paul Foulkes and Thomas Hoffman especially). I also benefitted immensely from repeat attendance at NWAV (New Ways of Analyzing Variation) conference acoustic workshops (often arranged by Marianna Di Paolo, a former student-friend from the Austin, Texas days). In the period 2008 to about 2018 we had a golden generation of young scholars working on PhDs and publishing papers locally and internationally on the new sociophonetics of English in South Africa (see Mesthrie, Chevalier, Ribbens-Klein, Toefy & Wileman [2023] and references therein for that period). An article published in Language (Mesthrie 2017) bears testimony to the excellence of my research team.

We met recently in Stony Brook for the IAWE conference where you gave the opening plenary. As always IAWE is a very inclusive conference that brings together scholars in linguistics and applied linguistics. There have been many developments in World Englishes as a discipline, but perhaps a big change that we can talk about is the shift away from the nation state that was prominent in the Kachruvian model and subsequent versions of it. There are many reasons for re-examining the idea of nationally bounded Englishes: a re-examination of substrate influence; internet cultures which are not nationally bounded; corpus projects which are supra-regional rather than national (cf. Bernaisch’s [2023] plenary about epicentres at this conference). How do you think about these developments?

For World Englishes nation-based approaches are historically understandable. Many of these “new nations” attained that designation only from the 1960s onwards. But it was always implicit that there were zones of—dare I say—“proximal development.” No one queried that nation labels could co-exist meaningfully with broader labels like West African English, South Asian Englishes, and Caribbean Englishes. And these are not purely geographical, but pertain to common histories and/or shared indigenous languages. It all depends on how fine-grained or general one wants to be. I’ve become interested in thinking of southern Africa as a linguistic area, prompted by having to contribute to Hickey’s (2017) Handbook of areal linguistics. In this area old language commonalities persist, despite language shift and death and old lingua francas continue to be influential (notably Afrikaans). I’ve been able to show that processes of SLA play a bigger role in the making of at least this incipient linguistic area than commonly believed. But “nationism” can’t be avoided, given slightly different language policies and influential languages in different countries within southern Africa. Corpus linguists can play a role in showing large-scale effects, possibly beyond the boundaries of individual countries; and while there are reservations about the edited nature of written texts that form much of the foundations of corpora, the results from corpus linguistics have been promising. But corpora have also been used for fine-grained analyses which can reveal the influence of different substrates: I’m thinking of the work of Davydova (2019), Van Rooy (2006) and Paulasto (2014) on the progressive in World Englishes. They have been able to fine tune earlier claims about certain universals of New Englishes (or “New Englishisms” as Simo Bobda [1998] called it, or “Angloversals” as others prefer, following Mair [2003]).

Of course, globalization and the new media have altered the Circles models (Kachru 1982), but not I believe irrevocably. Inner Circle speakers of English are now more familiar with New Englishisms originating in the Outer Circle than two generations ago, thanks to large scale immigration and resulting superdiversity. I see this a lot on Sky Sports channel (being a football fan all my life), where sports heroes are now quoted verbatim in the closed captions, rather than being edited. I also note that British sportcasters and commentators easily understand Outer and to some extent Expanding Circle Englishes on live broadcasts and in live international zoom or similar-based call-in programmes. But to some extent the more things change the more they stay the same—meaning that it’s still possible to differentiate between the circles, broadly speaking. There is no magic wand for SLA, even though AI might change the communicative profile of things in the not-too-distant future.

I want to ask about the aims of your IAWE plenary, which fit well with the different aspects of your career that we have been talking about. You began by saying that you wanted to reconcile Kachruvian and Labovian approaches through developing a framework for the variationist study of multilingual data. My own observation is, although the variationist sociolinguists that you would find at NWAV, for example, are committed to representing a wider range of languages (not just North American English!), this data is often monolingual. Can you explain how a variationist approach to multilingual data works, and why you are thinking about this now?

This has worried me for some time. Variationism works best in societies with a clearcut dominant (or hegemonic) language in which there is some consensus about education, status, prestige, and style shifting. In multilingual societies variation seems to involve the degree of substrate influence and/or code switching. In my South African work, I’ve been content to do variationist sociophonetics with varieties of English, and contact linguistics with other languages. Though of course the two strands of this work do connect and inform each other. But lately I’ve become a bit bolder (and this time I have another handbook editorial team to thank: Asahi, D’Arcy & Kerswill [forthcoming]) in trying to see all of this as one package. The trick is not to characterise variation language-by-language in relative isolation, but to view the multilingual’s entire repertoire as a system with principles that are mostly parallel to those of monolingual variation elucidated in western variationism. I start by identifying “S” and “P” codes within such a repertoire—S for solidarity and P for power/prestige. (Granted that covert prestige a la Peter Trudgill fits under S not P). (“H” and “L” signify another kind of relationship within diglossia studies, but they are close enough—though it’s perhaps a good time to avoid high versus low). I ask how factors like ethnicity, class, gender, and style relate to varying degrees of use of the S and P codes. (There might be lingua francas that are neither S nor P, by the way, but effective in work related communication—the pidgin Fanakalo comes to mind). I’m starting to explore changes from above and below for specific parts of the repertoire. Remember that the repertoires of some may include, say, Xhosa as L1 and S, and English as L2 and P. So a change from below in P would be equivalent to substrate influence. A change from above in S would involve English influence over Xhosa, which we’re seeing more and more of. I believe this aspect has relevance to a number of situations as young multilingual speakers become more and more proficient in English globally. A change from above in P would involve the influence of L1 English features over traditional L2 P, as is happening with some speakers starting to adopt goose fronting, and some features relating to new front vowel variants. Finally, we might ask if changes from below in S are possible: the answer seems to be that youth language practices play a role here. I’m still at an early stage in developing these ideas, but it would be good to have a variationist sociolinguistics of multilingual variation. I see that Julia Davydova is working along similar lines on Indian English and Hindi.

In your plenary you referred to the code-switching that you demonstrated as existing in the “third space”—it is by and for fluent bilinguals rather than speakers of either language. Does a version of this idea already exist in code-switching research? I’m thinking of Myers-Scotton’s claim about code-switching as a code in itself. Are you extending this idea, or do you have a different idea?

Yes, I think code-switching is a challenge to the idea that monolingual and multilingual variation are exactly parallel, though Myers-Scotton (1995, 1997, 2002) did make a good case of style shifting in monolingual situations having a social motivation parallel to her important and convincing work on this theme in East African multilingual situations. The third space idea—from Bhabha (2004) of course—comes to mind because of the new multilingualism showing the effects of a good knowledge of S and P with P as educational language. Here I believe speakers are keeping P active while using S as main medium of in-group communication (active to what extent is still to be carefully researched psycholinguistically). This results in practices that go beyond not only traditional borrowing and new nonce borrowing, but code-switching as well. So, yes Myers-Scotton’s characterisation of “mixing” as a code in itself is more than relevant; but the influence of P on speaking S is not a theme I recall from either of her important books on the topic. But I think in one or more subsequent papers with Janis Jake (e.g., Myers-Scotton & Jake 2017), the idea of a composite matrix was starting to take shape. A South African colleague of mine, Ron Simango, who was supervised for his PhD in North Carolina by Myers-Scotton, and who has followed her subsequent work more closely than I have, has recently confirmed (Simango 2021) that the switching we’re starting to see in South Africa goes beyond that earlier work. While Simango does not call this a third space effect, he has noted the syntactic convergence between Xhosa and English in similar multilingual speech situations among young people.

Your latest thinking has clearly been stimulated by some of the recent work you have been doing on urban youth languages in South Africa and beyond. Your work with colleagues ( Mesthrie, Hurst-Harosh & Brookes 2021 ) has shown how in South Africa among younger speakers there is a dominant African matrix language depending on the city, which takes English, Afrikaans, and innovative slang lexical insertions. Does this model apply in African contexts beyond South Africa? Is there a similarity between Isicamtho and Sheng? And does this have relevance beyond African contexts?

For a long time—at least since the early 1980s—there has been an insistence that new African languages were emerging in urban centres in Africa. Scholars pointed to the degree of multilingual contact that produced a great deal of innovation. Not being really familiar with these situations, I was content to leave their linguistic description to emerging scholars who wrote on the topic of varieties that went by names like “Flaaitaal,” “Tsotsitaal,” “Isicamtho,” etc. I did caution as early as 1992 that one should be cautious about reifying these varieties (in a famous conference on African linguistics at University of the Witwatersrand, the fourth and last day of which was totally changed by the parliamentary announcement that the ANC was to be unbanned and Mandela to be released from prison). I was particularly sceptical about the claim that there were about sixteen such new “languages” in South Africa. Speaking of Mandela, I still remember his first visit to Soweto after incarceration. The late Chris Hani in a televised event introduced him, saying, “Heita Com. Pres. Mandela, heitha” (‘Greetings, Comrade President Mandela, Greetings’). By welcoming him in a language of the youth, many things were being acknowledged: the role of young people in the struggles leading up to political change, a sense of change against the norms of older people and a recognition of the role of language secrecy and innovation in the struggle eras. What got me thinking about the underlying unity of all the youth varieties (for want of a better term) was that two different authors writing on this welcome claimed them to be “pure Isicamtho” and “pure Tsotsitaal” respectively (Makhudu 1995; Ntshangase 1995). Years later I started writing on this underlying unity, mainly under the impetus of a new PhD student I was supervising—Ellen Hurst. I think most people now accept my formulation of “tsotsitaal” in lower case as the underlying phenomenon, subject to many local instantiations, depending on youth affiliations, different urban L1s of different cities and parts of cities, etc. I argued that the semblance of multiple new languages arose from a false comparison of traditional, rural-based standard African languages and the performative practices of mainly urban, male youth. The missing link that went against this analysis was the rise of urban varieties of Zulu, Xhosa, Sotho, etc., which most analysts would accept as partially restructured but not really “new languages.” The base syntax of the “tsotsitaals” was almost entirely that of the urban version of Zulu, Sotho, etc. The vocabulary varied slightly from area to area, and was subject to immense creativity, most of which fell into categories described by Halliday (1976) in other contexts: phonological inversions, semantic disguise, etc. These youth languages didn’t really involve switching between two antecedent codes—and there wasn’t, say, a matrix and embedded language. Nor was there evidence of pidginization and creolisation. There was an embedded, innovative slang lexis which didn’t involve switching to another code, since some of the lexis was innovated on the spot and usually involved deliberate semantic or phonological disguise if coming from an antecedent code. I am very pleased that our work (Mesthrie 2008; Mesthrie & Hurst 2013) was able to uncover a great deal of underlying similarities, across “tsotsitaals” and with urban varieties, while in no way lessening our appreciation of the creativity of youth. And yes, I believe our analysis applies very closely to Sheng of East Africa—where claims about autonomy and entry as a fully-fledged language miss the mark of what youngsters are doing. Sure, there are signboards in Sheng and so forth (and perhaps creative writing in it), but Sheng occurs—and derives its status from being—in complementary sociolinguistic distribution (so to speak) to urban forms of Swahili, and standard Swahili. Ellen Hurst-Harosh and Heather Brookes have built decisive links on our continent (see Mesthrie, Hurst-Harosh & Brookes 2021), aiding our understanding of other youth varieties (or forms of “male youthspeak”) in Africa.

You emphasized in your plenary that adherence to psycholinguistic principles and structural constraints is key to our understanding of the behaviour of these fluent bilinguals, and you expressed some concern that translanguaging as a framework might inhibit us from doing that. Could you expand on this?

I am somewhat sceptical of most of the tall claims of translanguaging scholars (that languages can no longer be said to exist in the modern world, that all that exists are bits of languages that are meshed together). I have learnt a few things in relation to the pervasiveness of phenomena we’ve taken for granted: viz., the use of bits of another language without being fully proficient in it. Take the upper-middle class penchant for using Latin or French phrases in conversation (though less so than in writing, and perhaps less so overall today). I also like the distinction implicit in the term “named” language. (But didn’t Trudgill prime us long ago for the looseness of real world usage in relation to official language designations?). I absolutely disagree with one prominent scholar’s frequent assertion that in his view, knowing one word of another language made a speaker bilingual (see Blommaert 2010). If “language doesn’t exist” (as we are frequently told), then how do we know that such an item comes from another “language”? How can we “translanguage” without acknowledging at least two different antecedent languages?

Of course, we do know that language is far more diffuse than prescriptivists and educationists make out. And it’s good that this diffuseness is starting to be acknowledged by them under pressure from the applied linguists espousing translanguaging. But the dialogue with linguists and sociolinguists is another matter. I have seen some scholars critiquing a rule or approach within sociolinguistics, and trying to give a counter-example entirely based on another facet of language—for example, a lexical item, which is subject to cognitive and conscious needs and therefore quite different to the more tantalising sociophonetic variation we see in P codes. Or finding an error in a multilingual road sign and concluding that rules don’t exist in (spoken) language.

The more I look at code-switching, the more I see the interplay between structure and creativity. Creativity works against the backdrop of antecedent structure, and may even result in new structural patterns. My current work raises questions about how such creativity is “managed” psycholinguistically. And here sociolinguists may have a lot to learn from other branches of linguistics. The work of the contact-cum-socio-linguists whom I respect the most does connect with these other branches (e.g., Ray Hickey, Pieter Muysken, John Rickford, Miriam Meyerhoff, John Singler, Salikoko Mufwene, Carol Myers-Scotton, Gillian Sankoff, and still others).

So, through this systematic analysis, it seems that you are better able to capture the creativity of these speakers, and their communicative and expressive aims, than if you gave up on the idea of a system or structure. In a way this resonates with the approach that you have always taken to documenting difficult and painful histories of the South African context: capturing the language loss of Bhojpuri speakers through the linguistic measurement of koineization; uncovering communication with the coloniser or the lack of it by reconstructing the origins of Fanakalo pidgin. In some environments it’s hard to stick to linguistics.

I think working somewhat on the periphery of the world in a city closer to the South Pole than most sociolinguistics-practising cities, I have had of necessity to be eclectic in my work. I’ve also had to pay attention to debates in other disciplines, viz., literary studies and history. I have been positively influenced by such work, though I refuse to be blown off my feet by scholarship in other disciplines. I am thus rather pleased to have recently done a dialect study of Gujarati in Cape Town, with the assistance of two post-doctoral scholars from India—Vinu and Mrunal Chavda—and identifying a consonantal chain shift brought over from one of the dialects (Surti). This work has contributed to and perhaps enriched the study of colloquial dialect speech in Gujarat within a socio-historic framework (Mesthrie 2021). In such a minority discipline as ours, it saddens me to see young scholars being encouraged to do everything but linguistics. Sociolinguistics in linguistics departments is being weakened locally by an overemphasis on politics and cultural theory, even anthropology and an attempt to delink altogether from core linguistics or even core sociolinguistic concepts. I think it’s fair to say that I’ve struck a balance in my work between these disciplines and the close analysis of language and language practices. Disarming power and prescriptivism by such close and sustained analysis underlies much of my work and I hope it is a tradition that won’t be lost in South Africa.

More broadly, where do you see the field of sociolinguistics in ten, or twenty years’ time?

Sociolinguistics is playing catch up to the forces of globalisation, social and electronic media, decolonisation, and new forms of domination and multilingualism. No doubt there will soon be a sub-field of style and variation in artificial intelligence communication. These areas call for a stronger commitment to using and modifying established tools of linguistics, not wilfully disregarding them. I personally continue to be intrigued by and inspired by the tools provided by the sociolinguistics of variation, change, and contact.