of scientific description and interpretation, comprising not only technical, methodical aspects (Malinowskian fieldwork) but also, for example, cultural relativism and behaviourist–functionalist theoretical underpinnings. Ethnography was the scientific apparatus that put communities, rather than human kind, on the map, focusing attention on the complexity of separate social units, the intricate relations between small features of a single system usually seen as in balance.3 In Sapirian linguistics, folklore and descriptive linguistics went hand in hand with linguistic classification and historical-genetic treatments of cultures and societies. Ethnography was an approach in which systems were conceived as non-homogeneous, composed of a variety of features, and in which part–whole relationships were central to the work of interpretation and analysis. Regna Darnell’s book on Boas (Darnell, 1998) contains a revealing discussion of the differences between Boas and Sapir regarding the classification of North American languages, and one of the striking things is to see how linguistic classification becomes a domain for the articulation of theories of culture and cultural dynamics, certainly in Boas’ case (Darnell, 1998: 211ff). It is significant also that as ethnography became more sophisticated and linguistic phenomena were studied in greater detail and nuance, better and more mature theories of social units such as the speech community emerged (Gumperz, 1968).
So there always was more than just description in ethnography – problems of interpretation and indeed of ontology and epistemology have always figured in debates on and in ethnography, as did matters of method versus interpretation and issues of aligning ethnography with one discipline or another (linguistics versus anthropology being, for example, the issue in the Boas–Sapir debate on classification). In fact, it is our conviction that ethnography, certainly in the works of its most prominent practitioners, has always had aspirations to theory status. No doubt, Dell Hymes’ oeuvre stands out in its attempt at retrieving the historical roots of this larger ethnographic program (Hymes, 1964, 1983) as well as at providing a firm theoretical grounding for ethnography himself (Hymes, 1972, 1996). Hymes took stock of new reflections on ‘theory’ produced in Chomskyan linguistics, and foregrounded the issue in ethnography as well, and in clearer and more outspoken terms than before. To Hymes, ethnography was a ‘descriptive theory’: an approach that was theoretical because it provided description in specific, methodologically and epistemologically grounded ways.
We will discuss some of the main lines of argument in Hymes’ work at some length here, adding, at points, important elements for our understanding of ethnography as taken from Johannes Fabian’s work. Fabian, like Hymes, is probably best known for his documentary work (e.g. Fabian, 1986, 1996), while his theoretical reflections have not received the attention they deserve.
To start with, a crucial element in any discussion of ethnography should be its history, for inscribed in its techniques and patterns of operation are numerous traces of its intellectual origins and background. Ethnography has its origin in anthropology, not in linguistics, nor in sociology or psychology. That means that the basic architecture of ethnography is one that already contains ontologies, methodologies and epistemologies that need to be situated within the larger tradition of anthropology and that do not necessarily fit the frameworks of other traditions. Central to this is humanism: ‘It is anthropology’s task to coordinate knowledge about language from the viewpoint of man’ (Hymes, 1964: xiii). This means that language is approached as something that has a certain relevance to man, and man in anthropology is seen as a creature whose existence is narrowly linked, conditioned or determined by society, community, the group, culture. Language from an anthropological perspective is almost necessarily captured in a functionalist epistemology, and questions about language take the shape of questions of how language works and operates for, with and by humans-as-social-beings.4
Let us immediately sketch some of the implications of this humanist and functionalist anthropological background to ethnography. One important consequence has to do with the ontology, the definition of language itself. Language is typically seen as a socially loaded and assessed tool for humans, the finality of which is to enable humans to perform as social beings. Language, in this tradition, is defined as a resource to be used, deployed and exploited by human beings in social life and hence socially consequential for humans. Further implications of this will be addressed below. A second important implication is about context. There is no way in which language can be ‘context-less’ in this anthropological tradition in ethnography. To language, there is always a particular function, a concrete shape, a specific mode of operation, and an identifiable set of relations between singular acts of language and wider patterns of resources and their functions. Language is context, it is the architecture of social behaviour itself, and thus part of social structure and social relations. To this as well we will return below.
Let us summarise what has been said so far. Central to any understanding of ethnography are its roots in anthropology. These anthropological roots provide a specific direction to ethnography, one that situates language deeply and inextricably in social life and offers a particular and distinct ontology and epistemology to ethnography. Ethnography contains a perspective on language which differs from that of many other branches of the study of language. It is important to remember this, and despite possible relocations and redeployments of ethnography in different theoretical frameworks, the fact that it is designed to fit an anthropological set of questions is important for our understanding of what ethnography can and cannot perform. As Hymes says, ‘failure to remember can confuse or impair anthropological thinking and research, setting up false antitheses and leaving significant phenomena unstudied’ (1964: xxvii).
Let us now get a bit deeper into the features identified above: the particular ontology and epistemology characterising ethnography.
Language is seen as a set of resources, means available to human beings in societies. These resources can be deployed in a variety of circumstances, but when this happens it never happens in a neutral way. Every act of language use is an act that is assessed, weighed, measured socially, in terms of contrasts between this act and others. In fact, language becomes the social and culturally embedded thing it is because of the fact that it is socially and culturally consequential in use. The clearest formulation of this resources view on language can be found in Hymes’ essay ‘Speech and language: on the origins and foundations of inequality among speakers’ (1996: Chapter 3). In this strident essay, Hymes differentiates between a linguistic notion of language and an ethnographic notion of speech. Language, Hymes argues, is what linguists have made of it, a concept with little significance for the people who actually use language. Speech is language-in-society, that is, an active notion and one that deeply situates language in a web of relations of power, a dynamics of availability and accessibility, a situatedness of single acts vis-à-vis larger social and historical patterns such as genres and traditions. Speech is language in which people have made investments – social, cultural, political, individual-emotional ones. It is also language brought under social control – consequently, language marked by sometimes extreme cleavages and inequalities in repertoires and opportunities.
This has no small consequences to the study of language. For one thing, studying language means studying society, more precisely, it means that all kinds of different meanings, meaning effects, performativities and language functions can and need to be addressed than those current (and accepted) in mainstream linguistics.5 Second, there is nothing static about this ethnographic view of language. Language appears in reality as performance, as actions performed by people in a social environment. Hence, strict synchrony is impossible as the deployment of linguistic resources is in itself, and step by step as sentences and utterances are constructed, a process. It is this process, and not its linguistic product (statified and reified sentences or utterances) that needs to be understood in ethnography. In order to acquire this understanding, as much attention needs to be given to what is seen from the statified and reified perspective
