1 - Why Language Assemblages?

Summary

This book unsettles regular accounts of knowledge about language in several ways. The idea of assemblages allows for a flexibility about what languages are, not just in terms of having fuzzy linguistic boundaries but in terms of what constitutes language more generally. Languages are assembled from different elements, both linguistic elements as traditionally understood as well as items less commonly included. This is to approach language not as a pre-existing or circumscribed entity but rather as something produced in social action. Language from this point of view is embedded in diverse social and physical environments, distributed across the material world and part of our embodied existence. Once we ask whose version of language counts, we have to base any practical theory of language on what language is within a local ontology. This necessitates an engagement with language ideological assemblages and an understanding that languages are inevitably locally made assemblages (linguistic, semiotic and material). Applied linguistics as a practical assemblage is one way we can start to address the needs for a practical theory of language that can remain both plural and political.

Thinking of language as an assemblage, according to Wee (2021), has a number of advantages over other views of language, particularly those that suggest that language is a system with defined boundaries. Thinking in terms of assemblages can help us understand how languages are constantly under construction, how they are put together through social processes and why it is better to start with an understanding of social action than an assumption about pregiven languages. The idea of assemblages also allows for a flexibility about what languages are, not just in terms of having fuzzy linguistic boundaries (languages blend together) but in terms of what constitutes language more generally. Languages are assembled from different elements, both linguistic elements as traditionally understood (words and grammar, for example) as well as items less commonly included (bodies and things). An assemblage approach to language thus raises questions about what constitutes the linguistic as well as giving us ways of thinking about language as dynamic, constructed, open-ended and in and of the world. This is to approach language not as a pre-existing or circumscribed entity but rather as something created, produced in social action. Language from this point of view is embedded in, indeed part of, diverse social and physical environments, distributed across the material world and part of our embodied existence.

In this book I want to unsettle regular accounts of knowledge about language in several ways and for several reasons. Something of an ontological panic seems to have gripped some areas of linguistics recently. There is nothing new in questions about the ontological status of language and languages. In one of the earliest introductions to applied linguistics, when the field was still heavily reliant on formal linguistic accounts for an understanding of language, Corder (1973, p.27, emphasis added) warned of the dangers of following a ‘linguistic approach to language’ since it is the ‘most objectivizing. But language is not, after all, a thing with real existence.’ This caution was already pointing to the problem that approaches to language developed within the field of linguistics tended towards the reification of its object. The ontological status of language and languages, and thus the subject matter of linguistics, has always been a topic in need of serious discussion (Santana, 2016), though for obvious reasons linguists have tended to tiptoe around this problem (the discipline defines itself in no small measure around the idea of separate and comparable languages).

Questions about what language is, or what languages are, or how the two are related, are a necessary part of any ontologically curious position (Wee, 2021). Yet such curiosity has been met recently with a rather panicked response, a concern that if the status of languages is questioned, so too are the possibilities of language policy, language maintenance, bilingualism, second language acquisition and much more (MacSwan, 2020, 2022a). All this seems rather alarmist, as well as mistaken: the obvious problems with the status of ‘languages’ as commonly conceived doesn’t mean that language learning practices and policies will somehow cease to occur; it simply challenges the terms with which these are described. Integrational linguists have long raised questions about the status of languages in the field: ‘linguistics does not need to postulate the existence of languages as part of its theoretical apparatus’ (Harris, 1990a, p.45). Questions that have pushed this concern further, however, asking where the boundaries around language should be drawn – whether we can study animal communication with sociolinguistic tools (Cornips, 2022, in press) or what roles artefacts may play in social relations (Kell and Budach, 2024) – have brought warnings of a ‘neo-pagan apocalyptic linguistics’ (Pablé, 2022, p.6).

As Sinfree Makoni and I (Makoni and Pennycook, 2007, p.2) pointed out some years ago, languages do not exist as pre-formed entities in the world; they are, by contrast, ‘the inventions of social, cultural and political movements’. This is to acknowledge that languages are social creations, and to warn against the reification of languages that comes from treating them as bounded systems. Like others who have pointed to the obvious problems with the ontological status of languages within linguistics (Otheguy et al., 2018), we also emphasized that languages nonetheless exist as social entities with very real effects, and that once we have raised questions about the status of languages, we need to develop alternative ways of thinking about this focus of our work. We talked about this in terms of ‘a project not only of critique but one of reconstruction’ (Makoni and Pennycook, 2007, p.3), while García and her colleagues have taken this up, arguably more productively, in terms of translanguaging. As we shall see in the discussion in Chapter 3, translanguaging itself is also a term that needs to be handled with a degree of caution, but these projects are by no means an abandonment of language, language education or language planning. They do not, as some suggest, undermine minoritized speakers’ possibilities for social justice (Tannenbaum and Shohamy, 2023) so much as shift the grounds on which such campaigns are fought.

The question of whether and how languages exist is an ontological one. The ontological panic that follows ideas such as disinvention, integration or translanguaging is also about whose knowledge counts. As Jakobs and Hüning (2022, p.46) warn, a rejection of the concept of (different) languages is not necessarily a very useful direction for linguistics because it goes against both the disciplinary investment in different languages and everyday concepts of language. This should not mean, however, that it is not worth trying to find better ways of grasping linguistic realities than are currently presented in both domains, or of trying to understand how common concepts of language and linguistic understandings of language are related. There is a tension in linguistic approaches to this question: on the one hand a not unreasonable position that, as academic linguists, we know best; on the other, an egalitarian acknowledgement that other people’s views matter. As discussed in this and the next chapter, how everyday knowledge about language is understood without falling into the descriptive–prescriptive dichotomy or labelling popular views as myths or errors in need of correction is another theme of importance here.

Corder’s (1973) observation about the existence of language was also an applied linguistic concern.¹ It has been common in applied linguistics to see our work as applying existing linguistic knowledge to real-world contexts. This might seem fairly obviously to be what we do – inherent perhaps in the disciplinary terminology we inherited – and this has traditionally been the way that linguistics and applied linguistics have operated, with the one providing theories of language and the other putting them into practice, or at least mediating between the two. Alongside the hierarchical understanding of knowledge distribution that this view espouses, it also raises the question of what it means to ‘use linguistics’ for applied purposes. For Kramsch (2015, p.455), applied linguistics is not so much ‘the application of linguistic theory or any other theory to the real-life problem of language learning and teaching’ as ‘the practice of language study itself, and the theory that could be drawn from that practice’. That is to say, the work we do in applied linguistics – translation, language in the workplace, language education, language policy and so forth – is itself the study of language, not the application of someone else’s version of language.

Like Kramsch’s (2015) call for a theory of language practice and Li Wei’s (2018) call for a practical theory of language, this book makes central an applied linguistic view of language based in practice. Both views are part of a broader critical orientation that argues that practice and theory should not be separated and certainly not placed in a hierarchical relationship. Also known as praxis, this view suggests that theory derives from practice and that theory therefore needs to be practical: ‘The process of theorization, or knowledge construction, involves a perpetual cycle of practice-theory-practice’ (Li Wei, 2018, p.11). There are a number of reasons why it is important to develop practical theories of language: the knowledge about language drawn from linguistics may not be fit for purpose; if we are trying to deal with real-world contexts, it doesn’t really make sense to draw on theories of language that haven’t emerged from such contexts.

This book therefore sets out to look at what language is and what languages are (noting these may be quite different questions) with a view to arriving not at one practical theory of language, but rather at ways of assembling practical ways of thinking about language or, as I discuss in Chapter 6, understanding applied linguistics as a practical assemblage (Pennycook, 2018c). Rather than thinking about applied linguistics in disciplinary or interdisciplinary terms (approaches that keep structures of knowledge in place), this view suggests the coming together of language-oriented projects (social or educational endeavours that involve language), practical theories of language (different ways of approaching linguistic questions) and critical appraisals (ethical, material and political concerns). As applied linguists, we may not be interested so much in developing theories about language as in doing stuff with language: language policies, translations, language education, language in aviation and so on. To do so, however, we have to start to take responsibility for the ways we think about language. Approaches to language that derive from attempts to describe language structures or to account for language use in structural terms may be counterproductive. The terrain has changed from when applied linguistics was first seen as the application of linguistic knowledge to real-world contexts. We can now start to think seriously about practical theories of language or ways of thinking about language that derive from contexts of practice.

1.1 Relational Ontologies, Social Epistemologies and Critical Appraisals

Following Latour’s (2004) warning that critical work has focused for too long on the construction of reality (regimes of truth, orders of discourse, discursive production) rather than also making its own claims to reality, I extend the critical realist position developed by Block (2022), though with a number of twists. Drawing on Bhaskar’s (1989) work, Block (2022) makes a case for a form of critical realism based on an argument for an external reality (there is an external, real world), a relativist position on epistemologies (there are different ways of getting at this reality) and a rationalist mode of judgement (we need to be able to decide between these competing takes on the world). Drawing on Haslanger’s (2012) critical social realism, the implications of the ontological turn in the social sciences (see later in this chapter and Chapter 2) and a concern about how to ground any critical project in the field (Pennycook, 2021a), I take a slightly different view. I argue for a form of relational ontology (Barad, 2007; Escobar, 2016, 2018) that emphasizes both the multiple and the relational qualities of existence. A key argument throughout this book will be that there is not just one ontology: the world is plural. This ontological stance has implications for the discussions of languages in other chapters, not as different understandings of the same thing but as different things. This focus derives from various quarters: the ontological turn in the social sciences (and particularly anthropology), a response to the recent ontological panic about what it means to question the existence of languages and an increased interest across the field in raising ontological questions (Demuro and Gurney, 2021; Hall and Wicaksono, 2020; Kell and Budach, 2024).

This perspective cuts across the book but is discussed in greater detail in the next chapter, where I outline what it means to look at questions of being from a pluralist (relational) stance and questions of knowing from a social stance. In line with Haslanger’s (2012) critical social realism, and her scepticism about the usefulness of positing some kind of independent reality (neither a dependent nor an independent reality is very plausible), I am interested in the implications of social construction, or rather the importance of understanding different kinds of social construction, the ways things may be socially distinguished, constituted or caused (Haslanger, 1995; Sveinsdóttir, 2015). Epistemologies are obviously plural – we have many different ways of thinking about the world – but rather than Block’s epistemological relativism, following Haslanger (2012),² I insist on social epistemologies, entailing a focus less on the relativism of epistemologies than on their social nature.³ This view of epistemology aims to understand the social, cultural and political interests of different epistemological positions, how social epistemologies work, how forms of knowledge derive from social orders. It also, as will be argued later, by no means suggests that things that are social constructs are not real.

Critical social realism needs ways of deciding between alternatives. From Block’s (2022) point of view, a form of rational judgmentalism enables the critical realist to link a critical project to an ontological realism by looking analytically at the different epistemologies. I take a slightly different approach in this book: while we should be cautious not to throw rationalism out just because of its ties to particular modes of so-called Enlightenment thought (rationalism can be salvaged from this history), we need a clearer set of ethical and ideological principles on which to evaluate ways of thinking. Once epistemologies are viewed in social rather than relativist terms, the seeds of critical evaluation have already been sown: we can look at different understandings of the world in terms of the interests they serve. By assuming relational ontologies, the goal is no longer to use this judgemental position to decide on which epistemologies best account for a given reality, but to explore how ontologies, epistemologies and ideologies are intertwined (or assembled).

A key framework for this book, therefore, is a form of critical social realism that allows for more than one reality, grounds epistemologies in social relations and takes a critical-ethical position on choosing between different versions of the world. The next chapter lays out these basic concerns, explains why they matter and discusses questions of ontology – what language is – and the ontological turn in the social sciences. In Chapters 3 and 4, in line with the thinking of Demuro and Gurney (2021), I focus on language as structure, language as practice and language as assemblage. These by no mean exhaust possible language ontologies, as will be discussed in Chapter 5. If we ask what ontological position has been taken on language by mainstream linguistics over the last century or so, the answer is, fairly uncontroversially, language as structure (or object or system). For much of its modern history, linguistics has taken an ontological stance on language as a structural entity, with a wide set of implications for how languages are understood as bounded entities. Linguists might immediately object to this, suggesting that structuralism was a passing phase of linguistics, and that things have moved on since then, but this is to confuse ontological and epistemological stances. Structuralism was a particular epistemological position, related of course to the ontological position on language as a structure, but the ontology of language as a structure has outlasted structuralist epistemologies.

This is not about the different epistemological approaches to a structural version of language – various schools of linguistics, or whether structuralist linguistics was superseded by generativist or even functional schools of linguistics – but about the basic ontological assumptions about what language is. A structural ontology made it possible to treat language as an object amenable to scientific study, enabling descriptions of languages around the world and facilitating many advances in our understandings of languages as structural entities. Yet this very tendency towards seeing languages as autonomous systems has enabled those forms of thinking that emphasize boundedness. A significant argument in this book is that this ontology – language as structure – has remained a cornerstone of linguistic analysis and may only be helpful in limited sociolinguistic and applied linguistic cases. It is this ontology – with its underlying assumptions about bounded systems – that often leads to confusion when linguists are criticized for assuming languages to be discrete, countable entities, a position they may also disavow.

This discussion will be of particular importance in Chapter 3, where I try to disentangle some of the translanguaging debates: simply put, the two sides of the discussion are often talking about different things, language as structure and language as practice (though without always being clear about their own ontological assumptions). Because structural and social (practice) language ontologies are so different, the debates about translanguaging have become mired in misunderstandings. The idea of ‘a language’, Blackledge and Creese (2014, p.1) suggest, ‘may be important as a social construct, but it is not suited as an analytical lens through which to view language practices’. The discussions around codeswitching versus translanguaging often hinge on this problem: people are talking about different things, some focusing on language as structure (How do we account for one language or another being used in a particular context?), others on language practices (What are people doing with different linguistic elements?). Language as a social practice (not to be confused with sociolinguistics) puts the emphasis on language as something we do.

The idea of language as something we do is not always easily expressed in English, hence either the addition of the term ‘practices’ to language, literacy and so on, or the creating of variants of ‘languaging’, a term with a longer history than we might expect (Cowley, 2019). Proponents of polylanguaging (Jørgenson, 2008) and translanguaging (García and Li Wei, 2014) have insisted not only on the poly or trans aspects of this terminology but also on the languaging, while others have opted for translingual or translinguistic practices (Canagarajah, 2013; Lee and Dovchin, 2020). This focus on practices has a long history in sociology and anthropology, and it is revealing to reflect that the common linguistic assumption that systems or structures produce processes or practices, rather than the other way around (systems are the products of rather than the precursor to what we do), renders linguistics something of an outlier in the social sciences on this score (Ahearn, 2001; Van Leeuwen, 2008). From a standard (socio)linguistic point of view, languages as entities pre-exist their instantiation, so it is possible to think in terms of ‘language use’ or ‘language in context’ or ‘codeswitching’ where the language systems come before the social activity. From an integrational linguistic point of view, by contrast, ‘first-order’ activity is seen as communicative practice, while languages as structures are only ‘second-order’ concepts (Thibault, 2011). As discussed in Chapter 3, a practice ontology turns the tables on the language-as-structure perspective and makes languaging or social practices primary.

Taking different views of language seriously may mean entertaining the possibility that languages are different things to different people. What a linguist means by language may not be at all what a non-linguist means by language. This is a question of ontology rather than epistemology. The shift from questions of knowledge to questions of being urges us to consider not so much that there is one reality that we cut up differently from different perspectives (knowledge, culture, worldview, ideology) but rather that we are dealing with different realities. For Van Dooren (2019, p.8), it is important to escape the dominant Western belief in a single reality over which are layered various perspectives and cultures that provide different takes on this otherwise consistent world. ‘Our worlds’, he suggests, ‘are not preexisting, static entities’. We live in a world of many worlds, or a pluriverse (Escobar, 2020), a means for thinking about ‘ecologies of practices across heterogeneous(ly) entangled worlds’ (de la Cadena and Blaser, 2018, p.4). Such a view rejects the assumption that there is a given, independent world cut up by different worldviews, moving instead towards an understanding of entangled relationships, or assemblages.

Ontological questions cut across the book but are discussed in greater detail in the next chapter. As will be discussed in the following two chapters, the recent, rather panicked reactions to discussions about whether or how languages exist need to be seen within a much longer history of sceptical appraisal, and the concern, particularly from an applied linguistic point of view, that linguistics has tended to reify its objects of inquiry, to lose the connections between language, people and the world (Corder, 1973), and to fall into the trap of the methodological nationalism with which languages are associated (Schneider, 2018). If we acknowledge that languages are ‘social constructions, artifacts analogous to other constructions such as time’ (Makoni and Pennycook, 2007, p.1), this does not mean the end of all language learning, activism or politics, but rather a need to think through the implications of the ‘social’ and the ‘construction’ more carefully.

What flows from the observation that languages are social constructs? If, as Cummins (2021) suggests, in the context of arguments about translanguaging (see Chapter 3), there is no dispute about the fact that languages are socially constructed, the question is what is actually therefore under dispute? If it is no longer controversial to see monolingualism and multilingualism as inventions (Gramling, 2016, 2021), then what is at stake here? The problem in part is that social construction can be understood in multiple ways and is often seen as implying that something constructed is not real. Hence MacSwan (2022a) assumes that a critical constructivist position on language (what he calls deconstructivism) implies that languages are fictions, while Cummins (2021) understands this in terms of languages being social artefacts with unclear edges that can nevertheless also be described in terms of their linguistic reality (we will return to this). Yet if we consider Haslanger’s (2012) point that gender and race are both social constructs and real, and if we draw an analogy with language, there is clearly more at stake here.

If ‘we decide that languages exist’, argues Hutton (2002, p.121), we would likewise have to concede that ‘races exist’. This point is not necessarily under dispute if we understand that to concede the existence of language or race is to concede that they are social constructs: they do not exist as anything other than socially created entities that pull together certain features – words, morphemes, hair colour, facial features – into a supposed unifying construct. As Haslanger (2012) makes clear, if we want to resist reality (oppose racism or sexism), we first have to acknowledge that race and gender are socially constructed realities. If ‘any attempt to classify and characterize different races is unscientific’, Hutton (2002, p.121) continues, ‘then any attempt to classify and characterize different languages must similarly be pseudo-science’. That is to say, it is one thing to accept that language, gender and race are real as socially produced constructs, but it is quite another to turn a scientific gaze onto each as objects of study beyond the social.

As recent raciolinguistic work has made clear, language and race are deeply entangled in many contexts (Alim, 2016; Rosa and Flores, 2017). Linguistics is ‘both the parent and the child of race theory’; the parent in the sense that linguistic categories were crucial in the development of physical anthropology in the nineteenth century, the child ‘in the sense that linguistics has reclaimed its role as the premier science in the classification of human diversity, elaborating a “characterology” or “typology” of the world’s languages, and therefore of the world’s ethnic groups’ (Hutton, 1999, p.3). On this score, we have to appreciate the ‘contribution of linguistic theory and linguists to the murder and mayhem of twentieth century ethnic politics’ (Hutton, 2002, p.137). For these and other reasons it is incumbent on linguists of whatever sort to consider carefully what kind of realism we want to pursue: a realism that insists that languages are scientifically analysable entities or a critical social realism that insists on relational ontologies, social epistemologies and critical appraisals.

1.2 Whose Version of Language Counts?

Alongside ontological questions about what language is and what languages are, a related concern is whose version of language counts. A practical theory of language surely needs a strong relationship to how language users think about language. For Bauer and Trudgill (1998b), like Pinker (1994), the important linguistic distinction between descriptive (linguistic) and prescriptive (lay or pedagogical) approaches to language enables a focus on language myths, or popular but erroneous views on language that need to be rectified. There are, to be sure, many mistaken views about language across different social worlds, some that evidently matter – that women speak too much, or that some languages are primitive, for example – and others that may appear less consequential but may still have serious implications – that ‘Eskimos’ have a hundred words for snow, for example, or that there are 7,117 languages in the world. The descriptive/prescriptive distinction itself, however, is something of a myth. It assumes that scientific descriptions of language are somehow above and distinct from social norms; it overlooks the ways that attitudes towards language, or language ideologies, are part of the social world of language and cannot therefore be dismissed as outside the linguistic purview. Above all, however, it conceals the point that linguists’ attitudes to language are similarly value-laden, and their pronouncements about what is and is not possible in language can be equally normative (Cameron, 1995). As will be discussed further in Chapter 3, once decisions have been made about how a language system works, or how two such systems can work together, linguists are as capable of prescriptive pronouncements as those they denounce.

It is certainly true that people are interested in language: like the weather or families, it is something many people like to comment and have views on. Whether this fits a descriptive/prescriptive dichotomy, however, is another matter. Take this recent comment by a young woman to her friends on a tram in Melbourne: ‘Isn’t it weird the way English doesn’t have a word for, like, the day before yesterday but has some, like, really complicated word for throwing someone out a window.’⁴ Such everyday commentary on language is far more descriptive than it is prescriptive, and suggests both a general interest in how language works, as well as intimated comparisons of languages that can express the day before yesterday more easily – avant-hier in French, 一昨日(ototoi) in Japanese, 前天(qiántiān) in Chinese, vorgestern in German and so on – and languages such as English that struggle to do so. Common utterances about language along the lines of ‘words cannot express how I feel’ likewise have little to do with prescription and more to do with a perceived gap between emotional states and our capacity to express them: ‘Words are not enough, mate’; ‘There are no words to explain what we’ve been through’; ‘Words can’t describe what they’re going through’; ‘Words cannot express how much she will be missed’; ‘Words cannot express how devastated we have been by this’; ‘Words cannot express the depth of our sorrow.’⁵ The idea that words can’t do some of this emotional work for us seems commonplace and arguably reveals a popular attitude towards language and its limits (though I have not yet pursued this interest across other languages). It has become interestingly formulaic (words cannot express), suggesting a popular discourse that expressing grief, anguish, distress, amazement or shock cannot be done well with language (defined in terms of words). Maybe language can do more than people think, but I certainly have more respect for this view than to try to question it, and it is more useful in any case to understand this in pragmatic rather than propositional terms. It might therefore have the potential to be seen as a language myth, though not in terms of being untrue so much as a common way of thinking about language.

If we look at newspaper or other media, where people write in to comment on language matters, it is true that a more normative attitude is common, though as discussed in Chapter 3 such normativity may be equally shared by linguists. In letters to the Sydney Morning Herald on the last day of 2022, people wrote to suggest that the term ‘hero’ was being overused, that a medical receptionist’s use of ‘gorgeous’ to confirm an appointment was a bit over the top, or to inquire when ‘snuck’ (rather than ‘sneaked’) became the past of sneak, and so on.⁶ As (applied) linguists, we like to step back and look at such views as examples of everyday language ideologies, though as daily language users we may also agree with the comment about heroes (how did health workers, firefighters, teachers and many others all become heroes by doing their commendable jobs?), feel some sympathy with the reaction to ‘gorgeous’ (how have common terms of verification now become ‘fantastic’, ‘gorgeous’, ‘terrific’ and so on?) and feel we may need – a little hypocritically – to look up ‘sneak’ (I thought the past tense was ‘snuck’). Such letters, of course, don’t necessarily give us insight into popular views on language, since this is a self-selecting group who choose to write to a particular newspaper about language. They perhaps have prescriptive overtones in their comments on change but they are also descriptive in their observations about contemporary language use, and as both linguists and language users we often have rather mixed reactions to such commentary.

The language myths position (common views about language are mistaken) can be seen as a form of the ‘error correction’ approach that, it has been suggested, fails to engage with the reasons and interests in particular views on language (Lewis, 2018). If we simply try to disprove arguments, we overlook the investments people have in their views about language, the wider political and ideological contexts from which such ideas emerge, and the material conditions they support. Arguments aimed to counter raciolinguistic or homophobic ideologies that attempt to put people right about language fail to address the material, institutional and historical aspects of racism and homophobia from which such views emerge. This is also a question of what we mean by ‘myths’. If we assume myths are simply falsehoods, we fail to understand their power as ‘communally shared narratives told in the construction of an ideological set of beliefs’ about structures and functions of language (Watts, 2011, p.10). Since these myths are the shared properties of groups – stories about language that are one of the ways people make sense of the world – ‘deconstructing language myths is unlikely to have much effect on how people, on an everyday basis, view language’ (Watts, 2011, p.17).

Drawing a distinction between popular and scientific views of language – where the former are seen as myths and the latter as truths – fails to acknowledge both the social nature of academic work (particularly when something like language is at stake) and the social embeddedness of beliefs about language. It suggests that one set of beliefs are social, cultural, political or ideological while denying such elements in linguistic knowledge. As discussed in Chapter 3, the liberal egalitarian beliefs common in linguistics – that all languages are equally complex, for example – may be estimable but they also need to be seen as ideological positions. Various alternative positions on everyday views on language have suggested that rather than dismissing popular views about language, they should be the starting point for any useful linguistics. Integrational linguistics, for example, claims a lay-oriented position, arguing that any study of language needs to take into account common views about communication (Pablé, 2019b). Indeed integrational linguistics turns the tables on the science versus myth position, suggesting that everyday views on language should be our starting point and that the real language myth is the one upheld by mainstream linguistics (Harris, 1981).

Folk and citizen linguistic projects, by contrast, may be less critical of orthodox linguistic principles but insist nonetheless on the importance of everyday views of language and even on putting scientific analysis in the hands of the participants (Rymes, 2020; Svendsen, 2018). A linguistic anthropological approach, meanwhile, takes language ideologies seriously as local ways of understanding language. As Blommaert (2013) makes clear, we can no longer assume static languages in a landscape that can be interpreted by linguists. We need instead an appreciation of mobility, complexity and unpredictability, and we will only be able to approach an understanding of language by ‘close ethnographic inspection of the minutiae of what happens in communication’ and by ‘keeping in mind the intrinsic limitations of our current methodological and theoretical vocabulary’ (2013, p.8). The idea of language ideological assemblages (Kroskrity, 2021) urges consideration of the entanglements of people, ideology, place and material arrangements in any consideration of what languages are. How these different approaches may contribute towards a practical theory of language is discussed in much greater depth in Chapter 2.

The question as to whose version of language counts takes on a much sharper political focus when placed in the context of global knowledge production and distribution. In light of the decolonial demand to question Western or Northern ways of thinking about language and to take seriously not just different contexts around the world but different ways of thinking about languages, and indeed different language ontologies, the focus is on all that has been dismissed and denied in contemporary linguistic approaches. Language activist-scholars have asserted their own community ways of thinking about language, questioning the power linguists hold to define and describe Indigenous languages, and calling instead for the need to decolonize standard ways of considering what language is (Leonard, 2021). This links to a similar call to decolonize orthodox views of multilingualism in the Global South, challenging mainstream understandings of multilingual education, mother tongue education, language policies and so on (Ndhlovu and Makalela, 2021). For MacSwan (2022a, p.1), this kind of questioning ‘implies that multilingualism and a vast array of related topics on linguistic diversity are fictions’. This is not what is being argued, however. Rather, the point is to try to understand how multilingualism is not the same thing in different contexts. Such moves, discussed further in Chapter 5, not only raise questions about whose version of language counts but pose major challenges for how the field thinks about language.

1.3 Language as Assemblage

A central interest of this book (given the title) is the notion of language assemblages. On one level, the entire book is about language assemblages: languages are not pregiven entities but rather are assembled, gathered from a range of different elements and experiences, the products of social and ideological processes. Languages are made by both linguists and non-linguists, and the question about whose version of language counts asks whose language assemblage prevails in which contexts. On another level, however, there are more specific ways of thinking about language as an assemblage, discussed in depth in Chapter 4. This perspective eschews assumptions about languages as structural entities, focusing instead on the spatial gathering of linguistic and other material elements. A focus on languages as assemblages reconfigures what counts as language and how social, spatial and material worlds interact.

This understanding of assemblages as entangled groupings of different elements allows for an appreciation of the ways in which different trajectories of people, semiotic possibilities and objects meet at particular moments and places. This emphasis draws on the wider interest in assemblages and entanglements in the social sciences to understand how different kinds of things – bodies, words, artefacts, space, emotions, policies and so on – come together at particular moments and in particular configurations, creating a dynamic arrangement that is greater than the sum of its parts. Thinking in terms of assemblages points to ways that social life happens as an unfolding set of uneven practices. It insists that we explore social life not through broad abstractions about language, society or culture but in terms of local combinations of things that become happenings. This does not mean turning our back on the possibility of thinking about political economy or structural racism, for example, or indeed language in some of its more traditional senses, but insists instead that these only make sense when looked at in their local entanglements.

There are three slightly different ways that language and languages can be considered in relation to the idea of assemblage: assemblages as combinations of linguistic items (language assemblages), assemblages as semiotic gatherings (semiotic assemblages) and assemblages as material arrangements that involve language (sociomaterial assemblages). While these three different approaches often overlap, the different implications of each are discussed further in Chapter 4. Thinking about language as an assemblage, Wee (2021) suggests, can account for how languages are assembled through varied experiences with language in the world. Looking at language in terms of assemblages emphasizes the processes of communication as people draw on their prior linguistic encounters to create meaning. The notion of semiotic assemblages opens up ways of thinking that focus not so much on language use in particular contexts – as if languages pre-exist their instantiation in particular places – but rather on the ways in which particular assemblages of objects, linguistic materials and places come together. This is to approach language not as a pregiven or circumscribed entity but rather as something that is constantly being compiled from a range of semiotic possibilities. Sociomaterial assemblages similarly focus on the gatherings of things, places and linguistic elements, and consider language to be embodied, embedded and distributed, where language is not so much an abstract system of signs as changing sets of material relations.

The idea of language as assemblage therefore suggests an ontological commitment that differs in a number of ways from structural or practice-oriented ways of thinking. In structural ontologies, language is rarely seen in material terms, any relation to the material world being largely symbolic or representational. This interest in materiality does not propose that matter is all there is, as if language, thought, consciousness and so on can be reduced to material explanations, nor that material relations in terms of political economy or worldly circumstances define all other concerns. The point is to take matter seriously and to find ways of understanding language and its connections in material terms. This is of particular concern for a practical theory of language, for while practice cannot be reduced to material processes, it makes little sense to extract practice from its material surrounds. Whether language is understood as an assemblage of linguistic items, semiotic gatherings or sociomaterial arrangements, an assemblage ontology is a move away from autonomy and towards complexity.

Language cannot be separated from human or other life but provides potential meanings that participate in social and material events. An assemblage focus emphasizes the dynamic relations among people, things, places and artefacts, enabling a view that languages may be socially, materially and politically reassembled. While emphasizing human action in the assembling of languages, it also downplays the centrality of the individual actor, drawing attention to the ways human, non-human, technological and material actants combine. An assemblage ontology is a much better candidate for a practical theory of language than a structural ontology, since it urges us to understand how language operates in the world. It can give us a better handle on language learning, social interaction, semiotic landscapes and much more by opening up ways to see how language is not so much an abstract entity confined to our minds that escapes now and then when we talk and write, but rather is part of the symbolic and material world we inhabit.

1.4 Other Language Ontologies

If a focus on language assemblages is central to this book, an approach based on relational ontologies also emphasizes the importance of considering how different language ontologies are related. This is not just the liberal egalitarian focus of folk or citizen linguistics that asks how people ordinarily think about language, but rather of taking alternative views of language much more seriously. As argued in the discussions of whose version of language counts, this is a political question, a concern about how some versions of language have been discounted, disparaged and dismissed at the expense of others. Of particular importance here are Indigenous and minoritized languages. Taking seriously other ways that languages can be understood is of significance for both practical and political reasons. If language revival or other applied projects need practical theories of language, they have to be drawn from concerned communities rather than imported from elsewhere. A difficulty with such work is that community activists and language experts are not always trying to reclaim the same thing.

The ‘very idea of “language”’, as identified and delimited in Western ideological frameworks, corresponds, as Dias (2019, p.90) reminds us, ‘to an invention arising from seventeenth- and eighteenth-century coloniality and nationalism’ and an ‘assumption that such a notion, as a preconceived, independent object, is readily transposable to all locations and populations’. If this is particularly true of the notion of language as object or structure, it is also important to ask what structure, practice and assemblage as ontologies may exclude. An argument can be made that in their emphasis on social and material or human and non-human relations, practice and assemblage ontologies are closer to Indigenous and other Southern ways of thinking. The focus on land, Country and a more-than-human world that runs through many accounts of what language is from different Indigenous perspectives echoes a number of the themes discussed in relation to assemblages. Yet such relations need to be drawn with great caution lest we reduce Indigenous cosmologies to Western ontologies or assimilate a diverse range of ways of thinking about language into Northern ways of thinking. The focus on ontologies that runs throughout the book derives in part from attempts to grasp these different worlds.

The material focus developed in the discussion of assemblages can shed light on connections to land, water and surrounds, yet it runs the danger of overlooking what are often deeply spiritual relations. Assemblage-oriented thinking may point to more grounded views of language, but the material can also be connected to the spiritual. At the same time, to focus on Indigenous ways of knowing is not to cast such knowledge in terms of spiritual at the expense of rational, or local at the expense of extendable, or other such troublesome binary thinking. Approaches to radical Indigenism (Garroutte, 2003) or Indigenous standpoints (Nakata, 2007) emphasize the importance of understanding Indigenous philosophies of knowledge as coherent logics for understanding the world and to appreciate that these have also developed in relation to colonial forms of knowledge that surround them. The issue is far more than one of exclusion or distortion – Indigenous languages have been overlooked or misunderstood – but of taking marginalized knowledge seriously for what it can bring to contemporary thinking and applied projects.

For many Indigenous people, language is deeply connected to land, or what is commonly known as Country in Australia. Country encompasses far more than earth, dwellings and place. It can include not only rocks, trees and many physical features (and the particular significance they may have) but also water – sea, rivers, water holes – as well as animals – linked to people and their stories – and many other things such as wind and other beings (including humans) (Bawaka Country et al., 2022). Language within these ways of thinking is not connected primarily to people but to land or Country. It is because these ways of thinking about language are so different from a consideration of language as structure, as object, as separate from people and the world that many language revival projects have foundered. As long as Indigenous languages are thought of in terms of non-Indigenous ontologies, there will always be at best misunderstanding, if not appropriation and extractivism. On these grounds, Indigenous language activists have called for local control of language reclamation projects and the need to decolonize what is meant by language (Leonard, 2017).

1.5 Conclusion

Taking up ontological questions has implications for both what we think the world is and for what we think language is. As Grace (1987, p.9) pointed out long ago, an orthodox linguistic standpoint is ‘implicitly committed to the strong ontological assumption that there is an objectively given world common to all people which defines for all time what can be talked about’. It is on these grounds that it is commonly assumed that despite various differences, the same thing can be said in different languages: there is, after all, one world to be spoken about and all languages are ultimately cut from the same linguistic cloth. Part of that one world is a general property called language, which can then be subdivided into different but equal entities called languages. Apart from lacking philosophical and anthropological curiosity, such a view does not account either for the possibility that the world may be more ontologically diverse than this, or that the languages that are the focus of linguistic inquiry are as much objects of our own making as they are entities waiting to be described (Jakobs and Hüning, 2022). Above all, such a view does not do enough to question the interests – national, ethnic, political and economic – behind these linguistic divisions.

What if we listen to other worlds and start to consider both that the world may be plural and that languages may not necessarily be comparable things? Once we ask whose version of language counts, taking into account people’s investment with language as well as the demand to decolonize language, once we engage with local language accounts, we have to base any practical theory of language on what language is within a local ontology. This necessitates an engagement with language ideological assemblages and an understanding that languages are inevitably locally made assemblages (linguistic, semiotic and material), and that applied linguistics as an epistemic assemblage is one way we can start to address the needs for a practical theory of language that can remain both plural and political (Pennycook, 2021c). This will be developed further in Chapter 6. In the next chapter, I will discuss questions of whose version of language matters and what it means to talk in terms of language ontologies.