What is Language?
Language usually refers to the human system of units of sound (phonemes) compounded into words, in turn combined through grammatical rules (syntactically) to form a mode of communication that may be realized in both speech and writing. Saussure suggested that linguistics, the study of language in this narrow sense, was part of a wider field of investigation of signs and signification in general which he called ‘semiology’. A notable instance of the application of a semiological perspective is Lévi- Strauss’s analysis of mythology, though terms such as ‘grammar’ of myth, or of clothing, should be understood as analogies with language in the narrow sense, and do not mean that all human sign systems necessarily share common principles of organization.
Many anthropologists have adopted this broad view, studying all the channels and modes of communication that humans use to organize and convey meaning, including paralinguistic features such as gesture, facial expression, tone of voice and so on. It has the advantage of revealing the controversy over whether other animals have language as, in many respects, misplaced. Detailed studies of bees, birds, apes and dolphins, among many others, have conclusively demonstrated that they have very complex systems of intra-species communication which, with considerable difficulty, can be decoded by human observers. That cross-species communication, not least between humans and other animals, also occurs will be confirmed by anyone familiar with domesticated beasts. Apes, certainly, dogs and horses, up to a point, cats and sheep, barely, can all be trained to interact quite meaningfully with human beings, and to an extent vice versa (i.e. with convergence on a mutually satisfactory channel and style of communication). Nonetheless, experiments with primates reveal that the faculty for language narrowly defined seems to be confined to humans: it is a species-specific characteristic.
All human groups have a language in the narrow sense, and each of the many thousands of languages can, eventually, be learned and understood by speakers of other languages. Although, therefore, anthropologists take a broad view of communication, they share with linguists the perception that human language is sufficiently distinct, complex and wide-ranging to require understanding in its own right. This does not mean that anthropologists and linguists agree on what constitutes the nature of language and how it should be studied.
Anthropologists versus Linguists
Linguists sometimes complain that other academics treat their subject as if it were a social rather than a cognitive science. Like most disciplines linguistics is very diverse, but the transformational revolution associated with Chomsky led to acceptance of a view of language as an abstract system, which for theoretical and practical reasons may be studied in isolation from its social and cultural context. For Chomsky, the core subject matter is grammar, and the universal human ability to generate and understand grammatical utterances: linguistic competence. Chomsky has remarked that other disciplines are ‘presumably concerned not with grammars … but rather with concepts of a different sort, among them, perhaps, “language”, if such a notion can become an object of serious study’ (1979: 190).
This conception of an ‘autonomous linguistics’, as it is sometimes called, poses fundamental problems for anthropologists, and indeed some linguists, who believe that the cognitivist emphasis marginalizes language’s role in human communication. A similar point had been made much earlier against Saussure, when, in the 1930s, Malinowski, like the Soviet linguist Voloshinov, expressed serious reservations about the distinction between langue (the abstract linguistic system) and parole (actual speech). Since the 1960s, one of Chomsky’s most vociferous opponents has been the linguistic anthropologist Hymes (e.g. 1977). In the USA, there was a long and fruitful association between pre-Chomskyan linguists and anthropologists working in the tradition of Boas, Kluckhohn, Kroeber and Sapir, who attached considerable importance to the study of language. Hymes sought to reaffirm that tradition, and rescue the study of language from the transformationalists.
There are four main points on which Hymes diverges from Chomsky: speech (or parole) is accorded priority over grammar (or langue); competence is redefined to mean communicative competence in general, and treated as a behavioural rather than a cognitive phenomenon; universal forms of speech and language must be discovered by research in specific cultures and in cross-cultural comparison, not assumed in advance; and, crucially, language must be investigated in its social and cultural context. There is a very substantial body of work in anthropology and other disciplines which explicitly or implicitly shares these and similar assumptions. It includes, for example, research on class, both historically and in contemporary, especially urban, society; on education, ethnicity, gender relations, law, literacy, politics; and on the language of and in literature, both written literature and oral literature.
Language in Context
Social linguistics, as it may be called, is by no means a uniform field, theoretically or methodologically: functionalism, structuralism, Marxism (structural and other), feminism and not least post-modernism have all shaped how social scientists and others have conceptualized the relationship between culture, society and language. American linguistic anthropology illustrates one influential approach.
In the 1940s, the dominant issue was the relationship between language and worldview. The abandonment of the so-called Sapir-Whorf hypothesis, that the structure of a language determined our conceptualization of the world, led to a period in which folk categories and taxonomies and their organization became a distinct specialism. How different cultures classify flora and fauna (ethnobotany, ethnozoology, ethnoscience generally) yields rich data with practical implications for rural development programmes needing to take into account indigenous technical and environmental knowledge (Posey et al. 1984). Folk linguistics (how people talk about their own language and speech) has also attracted attention. Stross (1974), discussing what he calls the ‘metalinguistics’ of speakers of the Tzeltal language of Chiapas Province, Mexico, identifies a key term, k’op, which may roughly be glossed as ‘speech’. He lists 416 phrases in which k’op is modified to refer to various kinds of speech, speaker, or styles of speaking: bolobem k’op, ‘drunken talk’; pawor k’op, ‘asking a favour’; niwak k’op, ‘speech of adults’, and so on. This extensive lexicon reflects the propensity to evaluate people by reference to their speech styles. Tzeltzal themselves divide k’op into two main categories: ‘recent speech’ and ‘traditional speech’, which also includes prayers, flute music and drum sounds. Both have their own highly valued style: rason k’op, eloquent, sensible, slow and deliberate, associated with old men; and ?ista k’op, humorous speech used by ‘clever and witty younger men’.
In the early 1960s, seeking to go beyond uncovering indigenous systems of classification and develop a more dynamic approach to language as social process, Hymes suggested that attention should be focused on ‘ways of speaking’. This idea generated a great deal of research (of which Stross’s work may be considered part) under the broad heading of the ‘ethnography of speaking’ or ‘communication’. In this approach, speech is treated as the property of persons and social groups. People who share ways of speaking (i.e. who have a common set of linguistic practices) are said to be members of the same speech community. This is not the same as a society within which there are likely to be numerous speech communities. The totality of languages or linguistic varieties in a society, or those available to an individual speaker, constitute the verbal or linguistic repertoire. At the micro-level are speech events, activities involving verbal exchanges which participants may well recognize as distinct, and duly label: a funeral oration, a diagnosis by a physician. Speech events are in turn composed of speech acts: greeting, explaining, apologizing, commiserating, etc. This definition is similar to that found in a separate field of ‘speech act theory’, associated with linguistic philosophers such as Austin, Grice and Searle.
The ethnography of speaking is important because it emphasizes language in use, and locates that use within a social and cultural context. Its approach also forces attention on higher order linguistic practices (i.e. above the level of the phrase). In linguistics, these are what constitute discourse, though that term has a number of other meanings. Discourse includes conversation, and for anthropologists, it is axiomatic that all conversation is culturally embedded, and only understandable through what Moerman (1988) calls ‘culturally contexted’ investigation. Many linguists say they accept this, but in practice ignore its implications. What those implications are becomes apparent if we consider what participants in a verbal exchange need to know in order to understand the meaning of references to persons. The answer is cultural knowledge, and the anthropological understanding of such exchanges demands ethnography. This is illustrated in Moerman’s analysis of conversation in Thai, showing just how much background information may be crammed into a single utterance whose significance can only be revealed by a complex process of cultural unpacking. Moerman demonstrates this in a chapter entitled ‘Society in a Grain of Rice’, where he discusses a conversation with a Thai District Officer on a village visit. Overtly it is about arrangements for supplies, hence the rice; more subtly it is about ethnicity and bureaucracy. Moerman claims that conversation analysis is doubly important for anthropology: methodologically because most fieldwork is talk, and theoretically since culture ‘gets done’ in and through conversation. He therefore criticizes anthropologists such as Geertz whose interpretation of cultural symbols, he says, deals with concepts as category labels, omitting the practices through which symbols are given meaning. ‘Thick description’ must, so to speak, be grounded in thick, interactive, data.
Language and Social Differentiation
Contextualization, then, is crucial. So is linguistic heterogeneity. In most societies there co-exist different languages and dialects, and different modes of speaking which linguists call ‘registers’, ‘styles’, or ‘codes’. These languages and modes of speaking are often hierarchically ordered and their speakers of unequal status, power and authority. This is common in modern nationstates with their ‘standard’ languages, but also occurs in traditional societies. Contextualizing language means understanding heterogeneity in terms of social differentiation at large.
The ‘sociology of language’, pioneered by Fishman, Haugen and Ferguson, has specialized in analyzing linguistic differentiation within the nation-state. Ferguson (1959) devised the term ‘diglossia’ for situations where two varieties of a language (e.g. a standard language and a dialect) are spoken by members of the same community, with each variety having its own function and situationally defined range of usage. One variety he termed ‘H(igh)’, the other ‘L(ow)’ by reference to the generally perceived status of the variety’s functions. For example the H language might be used for education, the L for family conversation. Diglossia is therefore associated with a division of social life into sets of institutions or activities (domains) in which, generally, one language (say the H variety) is expected or appropriate or obligatory.
Mapping the domains of language use is, of course, important, but analysis must do more than summarize the results of statistical investigations showing that the H language tends to be spoken in this context, the L language in another. There are invariably social and political reasons for the restriction of the L language to a limited range of domains, and there is need for a ‘political economy’ of language to understand how and why such restriction occurs. One approach is to see it as an effect of processes of nation-state formation which create powerful centres or ‘cores’, with subordinate, dependent, peripheral regions. This has the danger that language may be construed only as an epiphenomenon of economic processes of marginalization, whereas culture may be crucial in the creation and experience of dependency. Writers on ‘Occitanie’, the region of southern France in which the Langue d’Oc was historically spoken, see the way in which that region came under the economic and political hegemony of a French-speaking centre as fundamentally entailing a cultural, and hence linguistic, imperialism. This perspective is widely shared by proponents of minority or subordinate languages, not least in the Third World. The novels and critical essays of the Kenyan writer, Ngugi wa Thiongo, for example, respond extensively to the cultural and linguistic effects of colonialism and neocolonialism.
Macroscopic perspectives of this kind, however, need to be complemented by micro-level studies which show how people handle diglossia in their daily lives. By contrast with sociolinguists such as Labov and Trudgill, who seek statistical correlations between linguistic features (often phonological) and their use by specific social groups, the work of Gumperz provides just such a dynamic, interactional view. An account of a small commercial and industrial town in northern Norway, where a dialect of Norwegian (Ranamal) is spoken alongside one of the Norwegian standard languages (Bokmal), examines the social meanings attached to the use of these varieties in particular contexts (Blom and Gumperz 1972). Ranamal, learned at home, is the language of the family, of friendship, of local loyalty, and of equality. Bokmal, learned at school and in church, was historically associated with non-local landowners, business people, and administrators. It is thus the language of outsiders and of inequality, and its use by locals is interpreted as ‘putting on airs’. Language use is not solely determined by social position, however. Setting, too, is important, and even within the same setting participants may change their definition of the event in which they are engaged, and indicate this by changing the language they use. This is called ‘situational switching’. There is also ‘metaphorical switching’, for example when two people conduct business in an office in Bokmal, and then move to dialect when the topic of conversation turns to personal matters, or when one or other manages to define the encounter in a different way. The participants are engaged in what Gluckman called ‘multiplex’, or many-stranded, relationships, different aspects of which may be emphasized at various points in an interaction.
In modern nation-states it is usually assumed that the standard language is a superior mode of communication, a view completely at odds with that of both linguists and anthropologists who hold that no language is inherently inferior to any other. Hierarchies of language are social phenomena; they have nothing to do with intrinsic linguistic features. The pre-eminence of the standard often means that speakers of nonstandard varieties and dialects are put at a serious disadvantage when obliged to interact in the dominant code. Other work by Gumperz has shown how the different cultural conventions that structure conversational exchanges in multilingual societies may give rise to the drawing of incorrect inferences, misunderstandings, and communicative breakdown. For example nonnative English speakers, even those with a high level of formal competence in the language, may be thus affected, when under stress in a job interview or in a courtroom. In contemporary society, access to jobs, housing, social services, etc. often depends on the communicative resources that can be brought to bear, and lack of competence in the appropriate language or code has serious consequences, especially for those who already occupy a subordinate social, cultural, economic and often political status. Since the late 1980s, linguistic anthropologists have pursued the stratifying consequences of language ideologies with even greater empirical and theoretical rigor (Woolard and Schieffelin 1994; Kroskrity 2000), bringing the detailed analysis of Gumperz’s interactional sociolinguistics to bear upon broader relations of power.
Many studies of ethnicity and class concern the disadvantages experienced by speakers of non-standard varieties, such as creoles, notably in education. Anthropologists and linguists have frequently been caught up in the political debates surrounding the value of such languages and dialects. In the USA, for example, researchers into what is called the Black English Vernacular (BEV), have argued strenuously, and with considerable success, against the idea that such varieties are in some sense ‘deficient’ (Labov 1982). In Western Europe, the USA and Canada, bilingualism and multicultural education have, relatedly, raised major questions of policy. Since World War II, many European countries have seen the growth of immigrant populations, with origins in Asia, Africa, the Middle East, and the Caribbean, whose different cultural traditions and languages often put them at variance with the receiving societies. The Linguistic Minorities Project (1985) provided extensive documentation of what this change meant for linguistic diversity in British cities. Holding that Britain had become an inescapably multilingual society, they urged (in vain) that bilingualism be considered a valuable resource, and promoted within mainstream education. The USA, whose cities have, in general, a population which is more varied ethnically, and often more influential, than those of European cities, has probably gone further in accepting this than most other countries, and many communities in the USA are effectively bilingual.
Subordination through Language?
The study of linguistic heterogeneity therefore rapidly leads to questions of disadvantage, and thus to politics. There is, however, another, some would argue more profound, way in which language and power are related. Drawing on theories of ideology and discourse, it is suggested that language, rather than simply reflecting or reinforcing non-linguistic structures of domination, itself fashions subordination. Bloch, for example, found that among the Merina of Madagascar speeches in village councils are highly formalized, using a restricted syntax and vocabulary. One formalized speech must be followed by another whose content is shaped by the first. Thus, ‘communication becomes like a tunnel which once entered leaves no option of turning either to left or right’ (1975: 24). For the Merina, the only acceptable reply is one which acknowledges the previous speech and its premises. This ethnographic example leads us to a central concern of linguistic anthropology from its origins with Malinowski’s detailed consideration of the Coral Gardens: i.e. how does language in use both reflect and construct social life? (Malinowski 1935).
The example of gender and language use is particularly salient with regard to this central issue. Is ‘gendered’ language use a reflection of gender inequality and stereotyping, or a site for the construction of gender inequality? Since the 1990s, linguistic anthropologists have rejected essentialist conceptions of ‘women’s’ language, including the unsupported attribution of particular speech forms (or silence) to ‘women’ across various cultures and times. Some have characterized the study of gender and language as a debate over ‘difference’, represented by Deborah Tannen’s (1996) approach, or dominance (Cameron 2007). More importantly, several significant theoretical interventions emerged in the 1990s, reformulating the study of gender and language in anthropology. Elinor Ochs (1992) argued that activity mediates between language and social identity, pointing out that language use rarely directly indexes gender identities, but instead that language has a constitutive relationship to gender. M.H. Goodwin’s (1990) important monograph, He-Said-She-Said, used conversation analysis to develop a sophisticated understanding of how gender is enacted in everyday interaction. Similarly, Penelope Eckert and Sally McConnell-Ginet (1992) advocated an activity-based approach by applying the concept of community of practice to language use (Eckert and McConnell-Ginet 1992: 464). Recent work has attempted, with somewhat limited success, to push toward integrating the linguistic anthropology’s rigorous empiricism with Judith Butler’s less empirically grounded, but theoretically congruent, work on gender performativity (McIlvenny 2002). Linguistic anthropologists, in keeping with the broader discipline, have moved the study of what was ostensibly gender and language beyond what was practically the study of women and language, exploring the multiplicity of gendered forms and practices, as constructed in face-to-face interaction (e.g. Kulick 2000).
Beyond the study of gender, linguistic anthropology’s approach to power relations does not take speakers as passive victims of language, or language ideology, nor as ‘subjects’ constructed in and through discourse (Ahearn 2001). Like anything else, language, and linguistic practice is a site of contestation, and of creativity, in which a multiplicity of ‘voices’ struggles for attention. As a Bakhtinian approach has come to the fore in linguistic anthropology, his concept of dialogism has informed nuanced studies of how power relations are contested (Hill 1986). This is illustrated in Hill and Hill’s account (1986) of the experience of diglossia by peasant farmers in the Malinche area of Mexico. Mexicano refers to Nahuatl, an Indian language which co-exists with Spanish. In the colonial period bilinguals were rare, but nowadays most people ‘speak two’, as they say. Historically, Mexicano was the code of the ‘inside’, Spanish of the ‘outside’, and in contemporary society Spanish still represents external power: it is dominant in religion, law, work (though not agriculture), commerce and the media, and exerts a strong influence on the way Mexicano itself is spoken. Using data mainly from interviews designed originally to collect examples of linguistic usage to be later counted and correlated with social factors, the Hills demonstrate what these codes signify, who uses them, and how they are realized in daily speech. But they discovered that the interviews were themselves significant speech events. Interviewers and interviewees engaged in verbal duels embodying the ‘symbolic strategies’ through which individually and collectively the peasants manage their changing relationship with national society.
The Linguistic Minorities Project, too, found that surveys were not inert fact-finding devices. They could and did change or create linguistic consciousness, as when pupils were asked to set down their own language practices, or simply to state the languages they spoke. There is a bewildering variety of labels which parents, children, teachers, administrators and linguists attach, for example, to a single South Asian language, and negotiating those labels is a fundamental part of the experience of speaking ‘other’ languages in predominantly monolingual England.
The need for a dynamic perspective, giving prominence to creativity and contestation, is further illustrated by work dealing with the ‘performative force’ of language. This concept, which influenced Bloch and Parkin, refers to the idea that the social significance of a word or phrase is to be found less in its propositional content (what it says) than in its effect (what it does). An early attempt to show this was Malinowski’s notion of ‘phatic communion’, a way of speaking which serves a social function; that of establishing human solidarity, Malinowski thought. Greetings in most cultures have this form: elaborate enquiries into an interlocutor’s health should not be taken literally. Developed by the linguistic philosopher, Austin, the concept of the performative force of language has been highly influential in studies of oratory, especially of political and religious rhetoric, and of the persuasive nature of devices such as metaphor. Expressive language generally has proved a very fertile field of inquiry, and much of what is written about language and identity, especially ethnic identity, explores the creative use of language. Such use need not, however, be solidaristic. Evans-Pritchard once demonstrated how a veiled and ambiguous style of speaking known as sanza enables Azande to indulge, relatively safely, in personal and social criticism.
There is an enormous range of data, oral and written, available to the social linguist, and this review has not attempted comprehensive coverage of the whole field. What has been emphasized is the way that anthropologists depart from linguists, first by emphasizing speech, rather than the language system, and then by studying speech in its social context. The Azande example offers a good illustration of this. With sanza, where the surface meaning of what is said may be exactly the opposite of what is intended (something not entirely unknown to speakers of English), an account which confined itself to specifically linguistic features such as phonology and syntax would be pointless, though linguistic data in the narrow sense may well, of course, give crucial clues for a social interpretation.
The anthropologist, therefore, studies linguistic, or more broadly, communicative, practice, the social activities through which and within which language or communication is produced, as well as the form that they take. This entails working in and through detailed ethnography, balancing structure and construction, actors’ interpretations and observers, judgements, micro and macro levels of analysis. These challenging tasks have become increasingly important for linguistic anthropologists, for since the 1980s many social anthropologists (following Foucault) have borrowed categories and terminologies from their subfield. Such work has spurred linguistic anthropologists engage their work with contemporary theoretical debates (Woolard 1985; Irvine 1989; Bauman and Briggs 2003). However, the primary aim of linguistic anthropology has not shifted: locating speech within prevailing systems of social differentiation and of power relations which both shape language and are shaped by it.
- Ahearn, L.M. (2001) ‘Language and Agency’, Annual Review of Anthropology 30: 109–137.
- Bauman, R. and C.L. Briggs (2003) Voices of Modernity: Language Ideologies and the Politics of Inequality, Cambridge: Cambridge University Press.
- Bloch, M. (ed.) (1975) Political Language and Oratory in Traditional Society, London: Academic Press.
- Blom, J.P. and J.J. Gumperz (1972) ‘Social Meaning in Linguistic Structures: Code- Switching in Norway’, in J.J. Gumperz and D. Hymes (eds) Directions in Sociolinguistics: The Ethnography of Communication, New York: Holt, Rinehart and Winston.
- Cameron, D. (2007) TheMyth ofMars and Venus: Do Men and Women Really Speak Different Languages? Oxford: Oxford University Press.
- Cameron, D. (1998) ‘Gender, Language, and Discourse: A Review Essay’, Signs: Journal of Women in Culture and Society 23(4): 945–73.
- Chomsky, N. (1979) Language and Responsibility, Brighton: Harvester Press.
- Eckert, P. and S. McConnell-Ginet (1992) ‘Think Practically and Look Locally: Language and Gender as Community-Based Practice’, Annual Review of Anthropology 21: 461–90.
- Ferguson, C.A. (1959) ‘Diglossia’, Word 15: 325–40.
- Gal, S. (1995) ‘Language, Gender, and Power: An Anthropological Review’, in Kira Hall and Mary Bucholz (eds) Gender Articulated: Language and the Socially Constructed Self, New York and London: Routledge, 169–82.
- Goodwin, M.H. (1990) He-Said-She-Said: Talk as Social Organization Among Black Children, Bloomington: Indiana University Press.
- Hill, J.H. (1986) ‘The Refiguration of the Anthropology of Language’, Cultural Anthropology 1(1): 89–102.
- Hill, J.H. and K.C. Hill (1986) Speaking Mexicano: Dynamics of Syncretic Language in Central Mexico, Tucson: University of Arizona Press.
- Hymes, D. (1977) Foundations in Sociolinguistics: An Ethnographic Approach, London: Tavistock.
- Irvine, J. (1989) ‘When Talk Isn’t Cheap: Language and Political Economy’, American Ethnologist 16(2): 248–67.
- Kroskrity, P. (ed.) (2000) Regimes of Language: Ideologies, Polities, and Identities, Oxford: James Currey.
- Kulick, D. (2000) ‘Gay and Lesbian Language’, Annual Review of Anthropology 29: 243–85.
- Labov, W. (1982) ‘Objectivity and Commitment in Linguistic Science: The Case of the Black English Trial in Ann Arbor’, Langauge in Society 22: 165–201.
- Linguistic Minorities Project (1985) The Other Languages of England, London: Routledge and Kegan Paul.
- Malinowski, B. (1935) Coral Gardens and Their Magic, London: Kegan Paul.
- McIlvenny, P. (ed.) (2002) Talking Gender and Sexuality, Amsterdam: John Benjamins.
- Moerman, M. (1988) Talking Culture: Ethnography and Conversation Analysis, Philadelphia: University of Pennsylvania Press.
- Ochs, E. (1992) ‘Indexing Gender’, in Alessandro Duranti and Charles Goodwin (eds) Rethinking Context: Language as an Interactive Phenomenon, Cambridge: Cambridge University Press, 336–58.
- Parkin, D.J. (1984) ‘Political Language’, Annual Review of Anthropology 13: 345–65.
- Posey, D.A., J. Frechione, J. Eddins, L.F. Da Silva, D. Myers, D. Case and P. Macbeth (1984) ‘Ethnoecology as Applied Anthropology in Amazonian Development’, Human Organization 43(2): 95–107.
- Stross, B. (1974) ‘Speaking of Speaking: Tenejapa Tzeltzal Metalinguistics’, in R. Bauman and J. Sherzer (eds) Explorations in the Ethnography of Speaking, Cambridge: Cambridge University Press.
- Tannen, D. (1996) Gender and Discourse, Oxford: Oxford University Press.
- Woolard, K. (1985) ‘Language Variation and Cultural Hegemony: Toward an Integration of Sociolinguistic and Social Theory’, American Ethnologist 12(4): 738–48.
- Woolard, K. and B. Schieffelin (1994) ‘Language Ideology’, Annual Review of Anthropology 23: 55–82.