Voices of English: language and the construction of religious identity amongst Ismaili Muslims in Pakistan and Tajikistan

1 Introduction

This paper brings together notions of heteroglossia (Bakhtin 1981) and religious identity to explore how English becomes indexical of Ismaili distinction in local communities in Pakistan and Tajikistan. Bakhtin’s (1981) heteroglossia can be seen to provide “a conceptual entrée to social meanings of bilingual speech and related identity constructions” (Bailey 2007: 257). My use of heteroglossia in this paper is similar, in that I view it as a “site” or “entry point” (Heller et al. 2018: 47) to the social meanings of English for Ismaili Muslims and related identity constructions. Drawing on Bakhtin (1981), I understand his focus on the “multi”, “hetero”, “poly” (274) and “vari” (261) as foregrounding the need to move away from “the dungeon of a single context” (Bakhtin 1981: 274). Heteroglossia provides a framework for the study of language as involving and invoking a “multiplicity of social voices” that are linked and interrelated in a more or less dialogical fashion (Bakhtin 1981: 263).

In this paper, I am particularly interested in heteroglossia as an entry point to religious identity. This is an analytically productive but not yet well explored angle in the wider sociolinguistic or religious studies literature. My focus on religious identity stems from the importance of religion for the unity of Ismaili Muslims across space and time. Ismaili Muslims are geographically dispersed across South Asia (e.g., Pakistan and India) and Central Asia (e.g., Tajikistan, Kyrgyzstan and western China) as well as in West Asia (e.g., Iran and Syria), East Africa, North America, Europe and Australia. The Ismaili community is challenging to define, as indicated by the use of wide-ranging terms to describe it, including “religion”, “polity”, “corporate structure” and “bureaucratic network” (Steinberg 2011: 207). At the same time, Ismailis around the world are predominantly united in their allegiance to the Aga Khan IV, the community’s Imam since 1957, and its social, political and religious leader. Religion is thus central to ideas of what it means to be Ismaili.

The community is organized hierarchically, with the Aga Khan governing Ismailis in social and religious matters, together with members of National and Regional Councils and Tariqah and Religious education boards. A highly centralized institutional infrastructure serves to regulate and provide access to services, and to facilitate movement across social and geographical spaces. As argued by Steinberg (2011: 1), “Ismaʿili institutions penetrate deeply into participants’ lives; they suffuse the fabric of their daily activities. In this way, the complex of Ismaʿili forms, processes, and structures seems to represent a new possibility for transnational social organization, for sociopolitical participation beyond the nation-state, for citizenship without territory.”

Unity is actively forged to create an awareness and sense of belonging to a formation which transcends local, national and regional boundaries. Examples include visits of the Aga Khans (predominantly the Aga Khan III, Imam from 1885 to 1957, and the Aga Khan IV, Imam since 1957) to Ismailis around the world; the creation and maintenance of a centralized infrastructure and global development network – the Aga Khan Development Network (AKDN); the launching of a system of Aga Khan Schools, Academies and institutes of tertiary education; a global Ismaili constitution; a series of online sites, mailing lists, blogs and archives run for and by (members of) the Ismaili community; and an official language, English. This status of English is made manifest in various ways, many of which I engage with in the course of the paper. The most explicit perhaps is the Aga Khan’s reference to English as the community’s official “second language” as part of a language policy (Aga Khan interview with Lebanese Broadcasting Corporation International Interview Aleppo, Syria and Lebanon 2011, published on NanoWisdoms). The Aga Khan IV also elevates English by using it in his speeches and farmans (‘edicts’),^[1] and he calls upon Ismailis around the world to learn English. For example, during his first visit to Hunza, northern Pakistan in 1960, he is said to have ordered his followers to “Think in English, speak in English and dream in English” (see also Bolander 2016, 2018).

Despite its official status, English is not widely used in the day-to-day lives of Ismailis living in Hunza, northern Pakistan and Khorog, eastern Tajikistan, where I conducted ethnographic fieldwork in 2012 and 2013. Yet my data shows English having a range of functions, including, as a means to construct unity (across nation-state borders) and draw distinction between Ismailis and Shia (Pakistan) and Sunni (Pakistan and Tajikistan) Muslims in the nation-states of Pakistan and Tajikistan. These two processes are interconnected, since drawing locally meaningful distinctions facilitates the forging of unity across nation-state boundaries, and thus the creation of a transnational community (see also Bolander 2017).

To demonstrate the social meanings of English for Ismaili Muslims in this paper, I adopt a “discourse analysis across the speech event” approach (Wortham and Reyes 2015). This analysis is applied to transcriptions of largely narrative interviews, taken from a corpus of interviews conducted during my fieldwork. After developing the theoretical bases of this analysis in Section 2 and outlining my methodology in Section 3, I analyze a range of examples in Section 4 with the aim of showing how English becomes a means to claim religious identity, and thereby indexical of Ismaili distinction.

2 Heteroglossia and religious identity

Bringing together theories of heteroglossia and religious identity necessitates understandings of language and religion that are epistemologically compatible. More specifically, this warrants a perspective on language and religion that emphasizes doing, performativity and emergence whilst at the same time acknowledging history, power and perspective. Heteroglossia, as it is explicated by Bakhtin (1981: 263) and used and applied by others (e.g. Baily 2007, 2012; Wilce 1998), is “the social dialogue among languages”. This theoretical lens foregrounds multiplicity, layering and complexity. It embraces a perspective of language as framed by and in turn serving to construct contexts which are particular and situated, whilst at the same time transcending singular scales of space and time. From this vantage point, language is active; it is performative and agentive. In the words of Bakhtin (1981: 276–277),

[t]he living utterances … cannot fail to brush up against thousands of living dialogic threads, woven by socio-ideological consciousness around the given object of an utterance; it cannot fail to become an active participant in social dialogue. After all, the utterance arises out of this dialogue as a continuation of it and as a rejoinder to it – it does not approach the object from the sidelines.

Dialogue within and between utterances foregrounds this intersection of “threads”, which are woven together by speakers in ways that are socially and ideologically meaningful. This process of weaving involves the coming together of different voices. As underscored by Wilce (1998: 230), “[m]ultivocality characterizes everyday discourse […] and particularly our conversational propensity to cite and recycle others’ words”, with “every word or utterance” exhibiting this tendency.

Reviewing research from within a critical tradition in anthropology, sociology and the science of religion highlights an approach to religion which is compatible with and analogous to the understanding of language underlying heteroglossia. This is a perspective which foregrounds how religious practices and discourses are related to particular historical traditions, how knowledge is generated, authenticated and legitimated, and how religion is negotiated in day-to-day social life, within but also beyond the bounds of organized religion. The focus thereby is on religion as social (Beckford 2003), human (Jensen 2016) and interactional (Van der Veer 2001), as intricately connected to power (Asad 1983, 1993) and as both reflecting and being shaped by perspective and positioning (Mostowlansky and Rota 2016). As stated by Beckford (2003: 2),

whatever else religion is, it is a social phenomenon. Regardless of whether religious beliefs and experiences actually relate to supernatural, superempirical or noumenal realities, religion is expressed by means of human ideas, symbols, feelings, practices and organisations. These expressions are the products of social interactions, structures and processes and, in turn, they influence social life and cultural meanings to varying degrees.

As suggested, what is foregrounded, is not the question of whether a transcendent exists. The focus is rather placed on the expression and negotiation of religion in human practice, and thus on processes of meaning-making, contestation and negotiation. Indeed, despite different perspectives, critical viewpoints on religion encourage its study as spatially and temporally situated interaction, authentication and knowledge, with scholars variously analyzing how the expression of religion in and as ideas, symbols, feelings, practices and organizations is related to broader social interactions, structures and processes.

Scholars working in this critical tradition attempt to elicit “the historical conditions (movements, classes, institutions, ideologies) necessary for the existence of particular religious practices and discourses” (Asad 1983: 252). They thereby foreground the creation and generation of knowledge (Taylor 1976: 16), and hence questions of how religion is “configured and transformed throughout history in relation to a conceptual field” (Van der Veer 2007: 253), and how it is negotiated, contested and authenticated by human actors on a day-to-day basis. The starting point for such analyses is the contention that the narrative of religion is not linear, but subject to struggle and fragmentation, with these processes shaping the ways particular meanings become dominant whilst others are marginalized (Van der Veer 2007). From this vantage point, then, the study of religion is also the study of power (Asad 1993: 45) – of how the power to form, label and negotiate is relevant to understandings of what counts as religion and what is variously viewed as authentic or legitimate by individuals, within and across communities. Within such work, interaction is seen to play a key role, as both a means and site for the negotiation of the meaning of religion in social practice.

Within sociolinguistic and linguistic anthropological studies invoking heteroglossia, there is a small body of research that brings together heteroglossia with religion and religious identity. This includes Wilce’s (1998) research on multivocality and identity among Muslims in Bangladesh; Eisenlohr’s (2007) work on linguistic plurality, diaspora, religious identity and ethnolinguistic belonging on Mauritius; Sawin’s (2013) paper on the importance of the superaddressee for gender performance in an American evangelical men’s bible study and Catedral’s (2018) scholarship on language and religious identity in relation to LGB+ exclusion in two Christian institutions in the United States. In this paper, I build on Wilce (1998) and Catedral (2018) to make a contribution to this developing area. I thereby engage with two ideas in particular.

The first is the critical argument made by Wilce (1998: 229) which posits that whereas “[c]ommunities arise out of the diversity of interacting voices […,] widely distributed representations of Muslims reduce their diverse voices to a monologue”. Wilce (1998) provides evidence for this diversity of voices based on an analysis of a series of distinct, discursive examples amongst Muslims in Bangladesh, including an interview, letter to the editor, writings, a lament and a meeting. He thereby highlights the interplay of Bakhtin’s (1981: 272) co-presence of both “centrifugal and centripetal forces”; of the central and the decentral, the shared and the individual. In Wilce’s (1998: 230) research, this manifests as “kaleidophonic”, i.e. through a metaphoric “kaleidoscope” of speech. Whilst “the kalimah (‘word’ or creed) of Islam, bearing witness to faith” is centralizing (Wilce 1998: 230, emphasis in original), Wilce’s analysis depicts how codeswitching, reported speech and pronoun play, in particular, highlight a multiplicity of ways that individuals negotiate their identities as Bengali and Muslim.

Building on this, the second idea I wish to highlight here as integral to my own work is from Catedral (2018). It is the argument that the scope of the multivocality of religious language and identity is wide. In Catedral’s (2018: 109) words,

[s]cholars of ritual and religious language have noted that this type of speech is multivocalic in that it contains both the speaker’s voice and the voice of divine authority […]; however, this study goes further to argue that religious language is also multivocalic because of its use of ‘we’ as different and simultaneous ways of conceptualizing the religious community.

In this vein, then, a heteroglossic approach to analyzing religious identity facilitates studying multivocality as scalar (Catedral 2018), with language and linguistic difference becoming relevant to the manifold ways religious identities and religious distinction are constructed. And indeed, whilst it is certainly important to pay attention to how religious authorities enter into individuals’ positioning, it is equally pertinent to extend the scope, as Catedral (2018) does, to incorporate multivocality at the level of individuals and communities. This aligns well with the Bakhtinian focus on diverse rather than (solely) authorized voices.

This theoretical perspective is particularly warranted for two main reasons. Firstly, because of a dominant perception in the sociolinguistic literature that “religious language” is readily identifiable, presumably because of a parallel focus on particular practices (e.g., liturgy, preaching), domains and institutions (e.g., the Church), or texts (e.g., the Bible), which are themselves viewed as inherently religious (for example, see Fishman 2006; Liddicoat 2012; Samarin 1976). The emphasis on particular practices, texts and domains which are deemed to be religious in and of themselves suggests that these offer a kind of container for religion and hence also for religious language. Whilst the focus on what Beckford (2003: 26) has called “organized religion” is not without its merits, an overemphasis on the study of religion in institutions and organizations has “tended to reinforce the image of religion as a generic object by reducing it to its most visible expressions in the form of institutionalized beliefs and practices”. This focus, in other words, naturalizes the relationship between religion and organizations potentially at the expense of recognizing that organizations which might typically be viewed as religious are home to a range of activities; that religion is also practiced outside of organizations and institutions and; more generally, that agency cannot be located at particular clearly delimited sites, but is instead shifting and defiant of reification.

The second reason relates more specifically to linguistic scholarship on English and Islam, which is suggestive of an ideology that English is somehow “extraordinary” (compare here Dovchin and Lee’s 2018 convincing work on the importance of “ordinariness”; see also Bolander and Sultana 2019), in the sense of English being imagined as variously “Christian”, “foreign”, “colonial” and “Western” (for examples see the papers in Karmani 2005). This ties in with discussions on English in Asian contexts which underscore a perceived opposition between “instrumental” and “identity” functions of English. From the vantage point taken in such studies, English is either “not considered to have any identity related function or is at any rate not supposed to have any such function” (Wee et al. 2013: 304). This emphasis on and orientation towards the instrumental, then, can be understood as a facet of English being viewed as a foreign language with western, colonial and/or Christian roots. However, as attested both by a body of research on English in Asia (see, e.g., the chapters in Wee et al. 2013) and the aforementioned scholarship which uses heteroglossia as an entry point to religious identity, “the language of the Other can be appropriated in order to serve as a resource for the formation of identities” (Park and Wee 2008: 242–243). Moreover, and as I go on to posit in the remaining sections of this paper, fusing heteroglossia with religious identity enables me to complicate the very meaning of this notion of “the Other”.

3 Data and methodology

To explore the social meanings of English for Ismaili Muslims, I predominantly draw on ethnographic fieldwork data. As mentioned previously, Ismailis are dispersed globally, with communities in over 25 countries around the world. These include Ismailis living in the land mass bridging northern Pakistan, eastern Tajikistan, western China and eastern Afghanistan. My fieldwork data comes from seven months of ethnographic fieldwork I conducted in that area in 2012 and 2013, specifically in Kyrgyzstan, Pakistan and Tajikistan, with the majority of time spent in a village in Central Hunza, northern Pakistan and the city of Khorog in eastern Tajikistan.

Ismailis form a numerical majority in both Hunza and Khorog (Bliss 2006; Kreutzmann 1996; Steinberg 2011), despite being a minority in the nation-states of Pakistan and Tajikistan. Linguistically, Ismailis in Central Hunza tend to speak Burushaski as their first language, an as-yet unclassified variety which is learnt at home and largely oral. Urdu and English – the two official languages of Pakistan (with Urdu also being a national language) – tend to be learned at school. In Khorog, Pamiris, who form the largest ethnic group and the majority of whom are Ismaili, generally speak Shughni as their first language. Shughni is an unwritten southeastern Iranian language, which despite not being taught at school is the “lingua franca of this micro region” (Bliss 2006: 98). Pupils also learn Tajik (Tajikistan’s sole official language) and Russian (Tajikistan’s former official language of interethnic communication), with Russian being more popular than Tajik, despite having been demoted from its status as the language of interethnic communication in 2009 in connection with Tajik President Emomali Rahmon’s “Law on Language” (cf. Kellner-Heinkele and Landau 2012). Schools in Khorog tend to offer English as a foreign language; and English is one of three languages of instruction (together with Russian and Tajik) in Khorog’s Aga Khan Lycée, the only school to offer this type of education in the area. English thus typically has third or fourth language status in Hunza and Khorog, although both localities have seen an upsurge in institutions providing English-medium education and English-language teaching, largely as a result of efforts of the Aga Khan Education Services (AKES), which runs over 200 schools around the world (Aga Khan Development Network, Aga Khan Education Services), many in areas with sizeable or majority Ismaili populations.^[2] In Hunza these institutions began to spread in the early 1990s (Benz 2013, 2014; Harlech-Jones et al. 2005), and in Khorog in the late 1990s and 2000s.

During my fieldwork, I lived with local families, took fieldnotes and was a participant observer. I also conducted qualitative interviews (N = 21) and two focus group discussions, and I spoke to a mixed audience including a large proportion of Ismailis at the American University of Central Asia in Bishkek, Kyrgyzstan, about my preliminary impressions at the end of my fieldwork. Since the majority of my interviews were in English^[3] and I was not able to learn either Burushaski or Shughni, living with local families and attending all kinds of events with these families and their friends provided me with important insight into language practice. More specifically, I was able to get insight into the relationship between what my interlocutors reported themselves and others as doing and what I observed them doing. The people with whom I lived and spoke knew that I was conducting research and that I was interested in the role of English, but in accordance with my understanding of religion as situated, discursive and interactive (see Section 2), I tended to refrain from making links between English and Ismaili identity when talking with Ismailis. I did this to be able to see if and how my interlocutors made English relevant to Ismailism.^[4]

To analyze the data collected through this fieldwork, I adopt Wortham and Reyes’ (2015) linguistic anthropological approach to discourse analysis. This “discourse analysis across the speech event” approach emphasizes the importance of focusing on “pathways across linked events” (Wortham and Reyes 2015: 1). To do so, it draws on Jakobson’s (1971 [1957]) distinction between a “narrated event” (that which is being spoken about) and a “narrating event” (the activity of speaking about this event) (Wortham and Reyes 2015: 3). In the case of my data, I thus conceptualize instances where my interlocutors talk about English as the “narrated event”, which form the starting point for analysis of how they thereby position themselves as Ismaili (“narrating event”).

In Wortham and Reyes’ (2015: 41) approach, the first stage of analysis, that of “mapping narrated events”, involves the researcher drawing on “semantic, pragmatic and grammatical regularities” to describe the contents. In a second stage encompassing the activity of speaking about the narrated events, the authors suggest a focus on “selecting indexicals” (signs relevant to contextualizing the event in question), “configuring indexicals” (analyzing what these signs presuppose), and “construing indexicals” (attempting to infer “models of voicing, evaluation, positioning and social action” that align with this context) (Wortham and Reyes 2015: 41). This process is followed by a third stage, where available evidence is used to interpret the “social action” that the narrating event is accomplishing (Wortham and Reyes 2015: 42). Wortham and Reyes’ (2015) approach is epistemologically compatible with the social constructionist and heteroglossic perspectives on language and religion which my research takes and which I have explained above; particularly in that it foregrounds ideas of dialogue, voice, identity, evaluation, action and context. I also believe it offers a response to the challenge of operationalizing heteroglossia (Bailey’s 2007: 263).

4 Multi-Englishness and religious identity

5 Conclusion

In this paper, I build on a small body of scholarship that brings together heteroglossia and religion to illustrate the value of heteroglossia as an entry point to examining how English can become indexical of religious identity. Engaging with data collected during ethnographic fieldwork amongst Ismailis in Hunza and Khorog, I am able to thereby demonstrate the need to eschew an “either/or” perspective on the role of English (instrumental or identity) and advocate instead for a “both/and” approach (instrumental and identity).

Working with Wortham and Reyes’ (2015) “discourse analysis across the speech event”, my analysis shows the key role of the Aga Khan for the generating, authenticating and legitimating of individualized perspectives on English. The Aga Khan’s voice is incorporated into my interlocutors’ situated “appropriations” or “creative interpretive practices” (Levinson et al. 2009: 768), whereby English is cast as relevant for their own and/or their community’s identity. Through this process, unity is created across Ismailis living in very different geo-political and linguistic environs. The examples thus highlight how ideologies of English can be made relevant to the negotiation of religious identity in social life, as something which encompasses unity as well as processes of linguistic and social distinction.

Working with heteroglossia facilitates being able to see English as both potentially bound to but also not solely bound to religion. English can be about religious identity and distinction, as well as something else; with the something else, in my data, typically surfacing as ideas of economic value and English for social and geographical mobility. This simultaneity (Woolard 1998) adds weight to Beckford’s (2003) argument on the need to explore religion as a social practice outside of the bounds of organized religion. It also cautions from mapping particular languages onto particular religions, or particular communities. Religion is not linear, but dynamic and fragmented. English then, can, but must not, be made relevant to the processes of “how, when and why” (Bailey 2007: 258) Ismailis construct religious identity. From this perspective, it is rendered “part of complex language chains, mobilized as part of multiple acts of identity; [… and] caught in a constant process of semiotic reconstruction” (Pennycook 2007: 122), with these chains and processes of reconstruction integrating different perspectives and suggestive of a multiplicity of dialogue.

These acts of constructing religious identity are locally relevant and meaningful whilst also providing evidence of a kind of “speaking-into-being” of transnational community. Studying heteroglossia and religious identity thus offers a window into how local distinctions can become relevant for the construction of broader social formations, even in instances where neither diaspora nor physical mobility is at the forefront of these acts of identity. This vantage point is operationalized through a discourse analysis across the speech event approach. Coupled with an understanding of identity as difference, this allows for the operationalization of heteroglossia as expressed through dialogue on English and religious identity.