Where is Language?

An Anthropologist's Questions on Language, Literature and Performance

Ruth Finnegan

• Publisher: Routledge
• Year: 2020
• ISBN:
  9781000189759
• Edition: 1
• Topic: Social Sciences
• Subtopic: Anthropology
• Index: Social Sciences

Language is central to human experience and our understanding of who we are, whether written or unwritten, sung or spoken. But what is language and how do we record it? Where does it reside? Does it exist and evolve within written sources, in performance, in the mind or in speech? For too long, ethnographic, aesthetic and sociolinguistic studies of language have remained apart from analyses emerging from traditions such as literature and performance. Where is Language? argues for a more complex and contextualized understanding of language across this range of disciplines, engaging with key issues, including orality, literacy, narrative, ideology, performance and the human communities in which these take place. Eminent anthropologist Ruth Finnegan draws together a lifetime of ethnographic case studies, reading and personal commentary to explore the roles and nature of language in cultures across the world, from West Africa to the South Pacific. By combining research and reflections, Finnegan discusses the multi-modality of language to provide an account not simply of vocabulary and grammar, but one which questions the importance of cultural settings and the essence of human communication itself.

1

What is the art of language?

It is common to assume we know what language is and what is needed to capture and describe it; hence, by implication, what ‘language’ in essence is. But there are many contending theories, too easily forgotten in the understandable rush to document and describe. These need to be considered at the outset, above all the performance approach to linguistic action entailed in pragmatic perspectives, and the issue of how and for whom linguistic accounts are constructed in the first place.

I too was once confident of what ‘language’ was, where its boundaries lay, and hence what might count as data for documenting it. But I am no longer sure. Nor am I clear where information about a given language should be found, or how, by, and for whom a language should be documented.

My uncertainties are founded in my own puzzles over the many years that I’ve worked, mainly as an anthropologist, on aspects of unwritten literature, performance and communication, based both in comparative reading and fieldwork in Africa and Britain (Finnegan 1967, 1970, 1977, 1988, 1989, 1992, 1998, 2002, 2007). Within that limited experience, I find that the issues I have confronted are unexpectedly (to me) relevant for the understanding of the nature of language and how to capture it, whether in our contemporary world or in the so-called ‘vanishing’ cultures.

What I offer here are some informal reflections, not any pretence of a scholarly or theoretical disquisition.¹ I write not as a specialist linguist nor as someone with any expertise in documenting languages, but merely about my experience of becoming increasingly doubtful of my initially confident assumptions about just where in the great spectrum of human communicating and expression we are to find ‘language’.

1 Given the personal setting of this introductory chapter there are many references to my own work, unclothed furthermore by the decencies of systematic citations throughout. But since my personal experience is of course interrelated with changing and contending approaches to language and communication, let me mention that works I have at various times found especially illuminating include Austin 1962, Bakhtin 1986, Bauman 1977, Bauman and Sherzer 1989, Bauman and Briggs 1990, 2003, Clark 1996, Cummings 2010, Dalby 1999/2000, Duranti 2004, Gippert et al. 2006, Hanks 1996, Harris 1987, 1998, Harris and Wolf 1998, Hodge and Kress 1993, Hymes 1977, Robinson 2006, Tracey 1999, Verschueren 2009, Verschueren and Östman 2009. Some issues touched on here are considered in more fully referenced framework in Finnegan 2002/14, 2007.

My first degree was in Classics – Greek and Latin. At that point I ‘knew’ what language was – or rather, I didn’t need to know because it seemed self-evident. It was what came in written texts. Written texts were the prime sources that had come down to us from classical antiquity, transmitted in the manuscript tradition and with, of course, no audio records of speech. The texts we read and studied were wonderful and enriching, covering a wide range of genres: literary, historical, epistolatory, oratorical, lyrical and much else. Both drawn from and supporting this corpus of texts was the extensive apparatus of vocabulary, of grammar and of syntax, all once again encapsulated in writing in the form of dictionaries of words (usually offering equivalencies in some European language) and accounts of grammatical and syntactical rules. The written words, organized in the correct classic formulations – that was ultimately what language was.

This emphasis on the textual and written was not totally unqualified. Archaeology – the study of material remains – played a part, and some scholars (like Eduard Fraenkel from Germany) went beyond the printed page to read aloud a Catullus love poem or (W. B. Stanford from Ireland) engaged with the acoustic dimensions of Greek lyric meters. There was an established tradition, though not within the examination curriculum, of live performances of Greek plays or of reading Homer aloud. But the paradigm was indubitably of the centrality of written text both as the object of what was studied and the medium in which such study was appropriately expressed.

From this viewpoint, documenting a little-known language (i.e. one unwritten-about philologically) would entail finding and pinning down its essential constituent: texts that could be read, analysed and form the basis for identifying underlying rules. The texts might have to be snared by transcribing spoken words into writing. But ultimately those resultant scripts, together with a similar scholarly apparatus as for classical languages, would form the necessary documentation data. Language was capturable and realized in the communication technology dominant in the mid-twentieth century and earlier – writing – and it was ultimately there that the data could be recognized.

Emerging doubts ...

Things began to look different when, as a graduate, I embarked on anthropological studies, followed ineluctably by my first piece of fieldwork. This was in the early 1960s among a people called the Limba, in northern Sierra Leone. My focus came to be on their stories and story-telling, an interest that followed on well from my enthusiasm for literary texts in my earlier studies. I was hugely impressed by the many story-telling performances I experienced there and wanted to make that aspect of Limba culture the central core for my thesis and subsequent publications.

My initial presupposition was that the way to study these stories – and most certainly the way to present them in my doctoral dissertation – was to capture them as written text. That, after all, was surely where their reality lay, and the medium in which I and other scholars possessed the necessary analytic tools. There seemed no other proper way to pin them down for scholarly study.

So some of the stories I transformed, directly, into script by taking them down from dictation. Many others I recorded on one of the (relatively) portable tape recorders then available. The obvious next step was to transcribe from tape into written lines on a page in similar format to the classical texts I and others were accustomed to.

My thesis could then take the familiar form of introductory background and analysis followed by the key data: parallel texts in Limba and English translation. It consequently ran to three large volumes (I still remember their weight as I lugged the required three copies of each through Oxford by bicycle, then up the steps to the examinations schools). I assumed – as, apparently, did my examiners – that the substantive data, the corpus of texts, had to be there in my presentation.

But there was a problem. I had been greatly struck by the richness and subtlety of these narrations, and in my thesis tried to convey something of their artistry. But that had somehow melted away in the stories I presented. At one point, trying to demonstrate why I was so enthusiastic, I showed one of the texts to a friend from my classical days, expecting him to be impressed. He read through and rejoined – politely – ‘Oh yes, another of those charming African animal tales’, to my mind missing all its wonders.

The point is of course only too obvious, though it took me some time to appreciate it fully. The reality lay in the performance. It was this that the written texts had failed to capture. They missed the subtle characterizations, the drama, the way the tellers used volume, pitch, tempo, repetition, emphasis, dynamics, silence, timbre, onomatopoeia, and a whole plethora of non-verbal indications to convey humour, pathos, irony, atmosphere … The written forms could never replicate the ideophones that peppered the tellings – vivid little mini-images in sound and more than sound. Nor could unilinear textual layout show the many-voiced interaction and co-construction by the audience as they joined in songs led by the narrator or reacted with horror or laughter to key turns in the tale. Nor, either, did it capture the Limba practice of picking out one among the audience as the ‘replier’ – a second voice to give special support, prompting, echoing and, where needed, exaggerated reactions and response. Compressing this multidimensional and multi-participant performance within the narrow one-voiced medium of writing was to miss its substance.

I soon discovered that similar patterns were found elsewhere – obvious once you look, but for long concealed from me (and others) by the presupposed centrality of written text. The study of oral poetry, performance, and ‘oral literature’ more generally, hammered home the same point. Both in Africa and further afield, those creating performed literary art deploy not just writable words but a vast range of non-verbalized auditory devices of which those conventionally captured in written text, such as rhyme, alliteration and rhythm, are only a small sample. The wondrously varied expressive resources of the human voice are exploited for multifarious delivery modes, varying with genre, situation or performer: spoken, sung, recited, intoned, shouted, whispered, carried by single or multiple or alternating voices. Not just in faraway places but in the spoken and sung forms nearer home too, there turned out to be near-infinite combinations of vocal expression and auditory resources of which most escape from view on the written page.

I had to conclude, then, that the core lay not in written text after all but in the performance. And that included the setting, the delivery, and not just the ‘lead’ speaker but the full range of participants. All this showed up the contentious nature of my earlier ‘language-as-written-text’ model. This was reinforced by ongoing trends in the study of verbal expression, not least the performance-oriented approaches and ethnography of speaking in folklore and anthropology – stressing performance and process rather than text and product – as well as more recent developments in linguistic anthropology, sociolinguistics and performance studies. At the same time interdisciplinary interests in oral performance and in ‘orality’ more generally were, and are, flourishing, opening up a new vision of the nature of human communication and expression previously concealed by the focus on the written.

This turned me towards seeing language as ultimately something spoken, performed, oral. It no longer seemed to be existent essentially in written text but in active performance and interaction. And if so, language documentation would have to be approached very differently than from the familiar written-text perspective. For it would have to focus on audio, not just written, materials, and to include records and analyses of oral performances and (where relevant) their multiplicity of overlapping participants. Such data would not only count, but be essential.

Plunging into the 'oral'

Acknowledging the limitations of a written-text model of language is hopefully by now scarcely problematic. Audio recordings are nowadays widely accepted as a regular (though perhaps not universal) part of serious language documentation. I would like to add two further comments, however, about the implications.

First, a qualification. The move away from the written to the ‘oral sometimes jumps to the opposite extreme, envisaging the spoken as somehow the bedrock, natural, traditional, to be set against the artificial imposition of writing. A seminal Western myth sometimes lurks behind this, constantly challenged but also constantly recycled. In ways more fully explored in the next chapter, this posits a fundamental opposition between two mutually exclusive types of social and cognitive organization: the one literate, rational, scientific, civilized, Western, modern; the other communal, emotional, non-scientific, traditional, primitive – and oral. This has underpinned a trend to mystify ‘orality’ and the ‘oral’ as if something distinctive and separate: characteristic of a culture belonging prototypically to the ‘them’ of far away or long ago and one in which writing, even if in certain respects present, is intrinsically alien (and to be ignored). This is a set of assumptions I have long found myself struggling against and one which no doubt also crops up – controversially – in certain approaches to language documentation.

In other ways, however, the analysis of the oral and performed dimensions of language has, paradoxically, not been taken far enough. The vocabulary to capture the amazing use of the voice with its huge range of subtleties is relatively little developed, and the sonic elements of language are still often sidelined. But if we are to document the auditory practice of language, then the data to count would need to cover not just rules about phonetics, word forms or (limited elements of) prosody but its active sonic realization in such features as, for example, pacing and speed, volume, pitch, melody, rhythm, onomatopoeia, voice quality, timbre, mood, mix with other voices and sounds – or silences, distancing, vocalized sounds like sobs, sighs or laughter … and so much else. Data about tone or prosody would have to include not just smaller units like words, phrases or sentences but also the sonic patternings of larger chunks and of speech genres more widely. It’s true that such elements sometimes get mentioned under the heading of ‘paralinguistic’ or ‘extralinguistic’ elements, but in an oral-performance model of language these are not supplementary extras but intrinsic. A Martian anthropologist might well be puzzled by a demarcation which included some auditory elements in the delineation of language but excluded others which can equally form part of both the conventions and the unique personality communicated through human vocal utterance.

So though the importance of audio features may now be increasingly taken for granted in documenting languages, helped by the audio technologies which now facilitate the recording, storage and accessing of such data, has this yet been fully followed through? Documenting the oral is inevitably enormously complex; nor, despite the wizardries of modern technology, have we really developed adequate techniques, vocabularies or perhaps concepts to fully capture and analyse these inevitably more fleeting and temporal performed features.

Small wonder, perhaps, that the written model of language is so extraordinarily persistent, with its implicit suggestion that data doesn’t quite ‘exist’ until it is reduced to, transcribed as, transformed into, or analysed through the spatial solidity of writing and print. As Hodge and Kress well put it: ‘The distinctive resources of spoken communication which are not transcribed are eliminated from linguistic theory’ (1993: 11). Even when we accept a view of language as sounded and performed, we still too often fall comfortably back into a model in which the true reality – and the key data – reside in visually written textualizations rather than vocal enunciation.

Cognitive models

My Limba fieldwork brought me face to face not just with story-telling performance but also with the active way that Limba speakers used vocal utterances to do things. This, I gradually discovered, ran counter to a further implicit model of language that, if only in a vague and muddled way, I had also had at the back of my mind.

This was a set of somewhat contradictory and elusive assumptions, which could indeed be split apart but which nevertheless tended to come together in a kind of general mindset which I’d sum up under the label of ‘cognitive’. Basically I pictured language as something essentially mental, rational, decontextualized. Language was to do with mind and meaning, and its central function was referential. Artistry and rhetoric were secondary embellishments only to be considered once the core prose and information-bearing elements were grasped. Language might or might not constitute an independent rule-governed system existing autonomously in its own right – I vaguely assumed that it did – but it certainly could be assumed to have a structure that could be abstracted from the messiness of context, usage and social action or experience.

Of course I should already have known that this was not the whole story, both from my own experience and from my encounter with the multiplicity of classical genres. Even so, I was still somehow steeped in that set of preconceptions. It had been reinforced in part by the legacy of logical positivism still influential in my undergraduate years at Oxford (though tempered by Austin’s lectures on ‘performative utterances’ which were much to influence me subsequently). More radically, as I came to realize, it was a continuance of an ideology powerful in Western thought over several centuries which asserted the rationality of language and its relation to science, objectivity, civilization, literacy and, ultimately, the achievements of the West.

In some ways it was a serviceable model for a field situation. My language learning had indeed initially relied on the presupposition of a systematic vocabulary and grammar that I could learn independently of the pressures of spoken situations. There was a short missionary-compiled Limba dictionary, a couple of translated gospels, and two short articles based on data from an overseas Limba visitor, elicited and analysed by a linguist (Jack Berry) then at the School of Oriental and African Studies in London, all of which I found hugely helpful.

This all fitted both my preconceptions about the systematized and meaning-carrying nature of language and where to find data about it, and my conviction that meaning could be conveyed cross-culturally and out of context. Language as the repository of thought offered the potential for its ‘translation’, a channel by which minds could be brought into contact across space and time. It was through language that Limba stories could be transported to others as text – something which I indeed aspired to do through my verbal translations.

My aim was not to document language as such, whether that of Limba speakers or any others. But if it had been, I would doubtless have started from the assumption that the core data would be found in the information-carrying forms, in ‘plain prose’ sentences and the logical structure underlying them; also that I would have to produce clear translations and word-for-word equivalences to enable the direct transference of meaning from this lesser-known culture into some accessible European language.

Greater experience of Limba life somewhat undermined that set of preconceptions. I could not really miss the way Limba speakers used speaking as organized action and performance rather than, or as well as, for conveying meaning. They used language to do things rather than just describe them: to recognize and forge relationships, ratify contracts, issue orders, assert a position, strike an attitude, show off as performer.

Further, in some interchanges, and even in some Limba stories, the cognitive ‘content’ as it were – the meaning I had assumed I could transfer – was not after all the only, or in some cases even apparently the most important, element. I think, for example, of one ridiculous short story I recorded about a fictional character called Daba, an incorrigible snuff-taker. All that happened in it was that Daba went round the local chiefs badgering them to give him vast quantities of snuff, then finally overreached himself by taking a huge sniff and falling down dead: nothing to it really. And yet this was hugely successful with the audien...