anthro as writing

Anthropology as a Kind of WritingAuthor(s): Jonathan SpencerReviewed work(s):Source: Man, New Series, Vol. 24, No. 1 (Mar., 1989), pp. 145-164Published by: Royal Anthropological Institute of Great Britain and IrelandStable URL: http://www.jstor.org/stable/2802551 .

Accessed: 20/01/2013 04:58

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Royal Anthropological Institute of Great Britain and Ireland is collaborating with JSTOR to digitize, preserveand extend access to Man.

http://www.jstor.org

This content downloaded on Sun, 20 Jan 2013 04:58:16 AMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=rai

http://www.jstor.org/stable/2802551?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


ANTHROPOLOGY AS A KIND OF WRITING

JONATHAN SPENCER

University of Sussex

This article reviews the recent interest in the literary aspects of ethnographic writing, concentrating on the work of Geertz, Sperber and the authors associated with the collective volume Writing culture. While it is argued that serious questions are raised in some of this work, it is also argued that recent fashions in literary critical theory may prove unhelpful in addressing those questions. In particular, the tendency to read texts with little or no consideration for the social and historical context in which they were written seems an especially barren approach. Instead it is argued that anthropology is as much a way of working-a kind of practical activity-as it is a way of writing. Acknowledgement of the personal element in the making of ethnographic texts may help the reader to a better assessment of the interpretation on offer; more radical change requires a change in anthropological practice as well as in anthropological writing.

Writing and interpretation This article is about recent anthropological questioning of anthropologists' own use of language-the way in which anthropologists use language in representing other cultures, that is the writing of ethnography. It was originally provoked by a recent collection of essays on anthropology as literature entitled Writing culture (Clifford & Marcus 1986), but I swiftly found that this was merely the tip of a highly self-conscious and reflexive iceberg, mostly to be found in American anthropology, but with voices intruding from Europe. I shall review developments up to the end of 1986, but not attempt to cope with the subsequent torrent of publications on this theme.

My position is that of an interested but agnostic spectator: I agree that the way in which ethnography is written is a great deal more important than has usually been recognised in the past and that some review of the literary procedures of anthropology is long overdue; my scepticism concerns the gains to be made from an undue concern on the part of anthropologists with recent trends-and I use the word trend advisedly-in literary theory. If Writinig culture is to be taken as evidence for a more wholehearted subjection of anthropology to literary theory the possible gains look meagre indeed. I would go so far as to suggest that, despite its trappings of political and intellectual radicalism, it is in some of its presup- positions a depressingly reactionary document.

The thrust of my argument concerns the relationship between text and context. For various reasons the context of anthropological representations-the actual work of enquiry and the material on which generalisations are based-has been omitted from much ethnography. This

Man (N. S.) 24, 145-164



146 JONATHAN SPENCER

context can be restored in two ways: by re-reading ethnography in terms of some wider historical context we may learn a great deal about the past of our discipline; while the effort to incorporate some self-consciousness about such matters within anthropological writing promises to improve the usefulness of new ethnography. But if we want to effect more significant change in the writing and reading of ethnography, then we shall have to reconsider not just anthropological writing-most of which takes place at considerable remove from ethnographic experience but anthropological practice as a whole.

The recent florescence of literary self-consciousness in American anthropology can be conveniently traced to an apparently innocuous footnote in Clifford Geertz's 1973 essay 'Thick description':

Self-consciousness about modes of representation (not to speak of experiments with them) has been very lacking in anthropology (Geertz 1973: 19 n.3)

The essay in which this is embedded is a dense and allusive text which is accordingly difficult to summarise. It makes a number of assertions: that anthropology is what anthropologists do; that what they do is ethnography; and that ethnography is (or at least should be) writing of a very particular sort. To characterise this peculiar sort of writing Geertz borrows a term and an example from the philosopher Gilbert Ryle. This is of a boy winking; to describe this as 'a contraction of the eyelid' is what Ryle calls 'thin description'; to unravel the significance of it-the boy may be winking, he may be parodying a friend winking, he may be imitating a friend parodying a third party winking and so on-requires interpretation, what Ryle, and Geertz after him, call 'thick description'. Ethnography is, then, an interpretative exercise in 'thick description'.

Ethnography moreover should not be assessed by the amount of un- digested information it contains but rather by the clarification it offers. But one apparent advantage of 'information' as a criterion of ethnographic worth is, of course, that it is relatively tangible. The disadvantage, for many, of Geertz's 'clarity' is that it sounds subjective on the one hand, while, as a final arbiter of ethnographic success, it has its own peculiar dangers: Margaret Mead's account of Samoa (1928) was if nothing else beautifully clear; it seemed to generations of American readers to correspond to Geertz's criterion of interpretative success: 'the power of the scientific imagination to bring us into touch with the lives of strangers' (1973: 16). The problem, of course, was that the strangers themselves disagreed not with the power, or even the imagination, but with the content of Mead's representation of their lives.1

A further problem occurs when the ethnographer-and this is something that much absorbed me in trying to write about Sri Lanka in the early 1980s, a place which, in many respects, was on the brink of political disintegra- tion-is concerned to represent areas of cultural incoherence and confusion. It is after all a recurring aspect of change in the modern world, perhaps especially in those areas of it where anthropologists have been thickest on the ground, that old answers prove inadequate, old cultural cloth no longer stretches to cover uncomfortably new and worrying ex-



JONATHAN SPENCER 147

perience. Yet ethnographers are understandably reluctant to report that some things may not make sense in any particular cultural context. 'If law is anywhere' as Tylor ordained 'it is everywhere', and if you couldn't find it you can't have looked hard enough or in the right places (Levi-Strauss 1969:xi).

Geertz, it is true, acknowledges these problems, complaining that 'Noth- ing has done more... to discredit cultural analysis than the construction of impeccable depictions of formal order in whose actual existence nobody can quite believe' (1973: 18). But this is precisely what Geertz himself can be accused of doing; it may not necessarily be true of his ethnographic analyses but it is impossible to tell because he so often denies his readers the opportunity to assess for themselves the material from which he has constructed his accounts. His justification for this way of working lies in the distinction between thick description and thin description: one cannot assess an ethnographic interpretation against some sort of raw data, 'radi- cally thinned description' as he puts it, because this is itself already an interpretation, a construction:

What we inscribe (or try to) is not raw -social discourse, to which, because, save very marginally or very specially, we are not actors, we do not have direct access, but only that small part of it which our informants can lead us into understanding (1973: 20).

The problem is that Geertz ignores two things-that interpretation itself can be situated socially, and that different forms of life vary in the kind and degree of interpretation they can or should receive. Without denying the real methodological problems involved, it is obvious that something like raw figures for paddy ownership or demographic change is less dependent on informants' constructions than, say, the changes in tenancy patterns and justifications for those changes that follow demographic change. As well as interpreting and writing, many ethnographers do a great deal of counting or weighing or surveying, not to mention reading documents in archives and in the writings of their predecessors.

But let me now return to Geertz's first evasion-that interpretation is a socially determined activity. It is surely palpably obvious that, for example, a paddy-farmer's explication of decisions over the hiring of labourers on his field is likely to be different from an anthropologist's; the anthropologist should certainly use the farmer's account, and the labourer's too if it is accessible. A good anthropologist will also allow his or her readers to assess the differences between the two or three versions, differences which we can expect to correspond to the different purposes and positions of the explicators. Indeed, in skilled hands, these differences can become the centre of the whole analysis.2

But this is what Geertz refuses us. In his ethnographic writings, especially those from the mid-1960s onward, there is less and less space allowed for readers to agree or disagree or make their own connexions. His charac- teristic strategy is to seize on a metaphor-likening a peasant economy to a style of baroque decoration, describing the pre-colonial Balinese state as a theatre, talking of the Balinese cockfight as a text in which the Balinese can, as it were, read about themselves-and then sustain it through flashes




of description, before climaxing in a kind of adjectival blizzard. On the cockfight:

Any expressive form lives only in its own present-the one it itself creates. But, here, that present is severed into a string of flashes, some more bright than others, but all of them disconnected, aesthetic quanta. Whatever the cockfight says, it says in spurts (1973: 445).

You may find such writing either exciting or enervating according to taste or academic inclination; much of the time I incline to the former view. What you will have difficulty doing is sorting out the kind of evidence Geertz could possibly adduce to support it. What, one wonders, is the Balinese for 'aesthetic quanta' and what sort of statements, what informants' explications, what entries in sweaty notebooks, could have been synthesised into the account Geertz presents?

Geertz's answer would fall back on the impossibility of using uninter- preted data in anthropological work: 'what we call our data are really our own constructions of other people's constructions of what they and their compatriots are up to' (1973: 9). This may well be true. But it would seem the merest politeness to acknowledge the source of a particular construction. One may not inscribe raw discourse; one does take down a lot of quotes, explications, constructions and any half-decent fieldworker has some idea of who it is who has provided the quote, explication, or whatever. The idea that there is no dividing line-because all is interpretation-between the high literary gloss of Geertz's ethnography, and what one assumes are the drabber, more mundane jottings in his notebooks may be a useful excuse for the exercise of a particular kind of literary style; but the style in question presupposes a passive readership. In Geertz's world ethnographic accounts are assessed on a take-it-or-leave-it basis; one study rarely replaces an earlier deficient study, different accounts of the same place tend to run in parallel rather than building directly on each other. The ethnographer provides a finished product and never anything less than a finished product.3

James Clifford in his essay 'On ethnographic authority' (1983a) glosses this move of Geertz's in terms of Ricoeur's (1971) discussion of text and discourse. Discourse, says Ricoeur (following Benveniste), is to be found in the specific moment of its production, in the I-and-you of its referents; textualisation removes discourse from these specific conditions of production so that it can speak to other people at other times. So ethnographers take away from the field texts which are by definition already freed from the conditions of their own production, and the turning of these texts into ethnography further eliminates the specificities of the original context. The losses in such a process-and Clifford's catalogue (1983a: 132) of such losses is similar to the one I have already provided-are, it seems, the inevitable result of the process of textualisation.

But are they? It seems to me that Clifford (who merely describes but doesn't endorse this position) is following Geertz in confusing Ricoeur's argument-a feat easily accomplished as may become apparent. Ricoeur in the paper they both cite ('The model of the text' (1971)) is concerned to establish an analogy between the interpretation of a text and the inter-




pretation of what Weber called 'meaningful action'. A given action may be subject to competing interpretations, just as a given text is the subject of competing interpretations; but in both cases some interpretations are more probable than others: 'It is always possible to argue for or against an interpretation, to confront interpretations, arbitrate between them, and to seek for an agreement, even if this agreement remains beyond our reach' (Ricoeur 1971: 550). These possibilities are greatly reduced with Geertz's work because he insists on filling the dual role of author-producer of the text which is Bali-and interpreter. The text is, in Ricoeur's phrase 'a limited field of possible constructions' (1971: 550); but an assessment of competing interpretations of a given text presupposes access to the text itself, not merely another critic's interpretation of it. The 'text' of Geertz's interpretation is the Bali of his experience and his notebooks-this is what he is interpreting. The irony is that this most hermeneutical of anthropologists adopts a literary practice which tries above all to close the hermeneutic circle by limiting his readers' access to that which he wants to interpret for himself.

Geertz's argument in 'Thick description' has important implications for the relationship between theory and practice in anthropology. The conventional view is pretty straightforward (which isn't to say that anyone would accede to it when presented as starkly as this): there are facts, found in variable quantities in different ethnographies, and there are theories which attempt to make general statements based on those facts. Facts which don't fit can disprove a theory; odd facts can be used for new theoretical synthesis. Of course it has been long recognised that theoretical precon- ceptions determine what does or doesn't count as a fact to the ethnographer; Malinowski, for example, used this as the criterion to mark off scientific anthropology from the work of enthusiastic amateurs (1922: 9). But for Geertz anthropological theory is found in specific interpretations in specific ethnographies: 'Theoretical formulations hover so low over the interpretations they govern that they don't make much sense or hold much interest apart from them' (1973: 25).

Description and interpretation Some anthropologists, especially in Britain, may be ready to dismiss Geertz's discussion of anthropology as representation, feeling it to be no more than the personal preoccupation of one of the discipline's foremost literary dandies. But similar points have also been made from the point of view of a would-be generalising anthropologist, unafraid to use outre' words such as 'science' and 'epistemology'. Dan Sperber, in his essay 'Interpretive ethnography and theoretical anthropology' (1985), acknowledges the limited nature of anthropological theory but argues that this is because what we call our theory is, by and large, nothing of the sort; it is in fact a rag-bag of vague generalisations that provides a sort of intermediate language that is useful for the task of interpretation and translation, but useless for the real task of a scientific anthropology-the building of generalisations .(a task which he disguises behind the construction of an 'epidemiology of cultural representations'). As an example of genuinely scientific anthropo-




logical generalisation he proffers Berlin and Kay's (1969) celebrated work on colour classification.

Sperber's strictures are, though, of relevance, even for those of his colleagues who are sceptical of his broader project. Ethnographies, he says, deal in representations. Representations can be divided into two kinds: descriptive and non-descriptive. Descriptions are a kind of representation which are 'adequate when they are true'; that is to say, they can be refuted by observation. Truth and falsity are properties of propositions; propositions are utterances; therefore descriptions can only come in the form of utterances. Moreover, they are the kind of utterance which can be used in a logical argument: 'The Nuer are transhumant pastoralists...', 'If the Nuer are transhumant pastoralists...', 'Therefore the Nuer are transhumant pastoralists...'.

Unfortunately for Sperber (but not, I suspect, for the rest of us) only a small part of ethnography comes in the form of descriptions. Non-descriptive representations come in two forms: reproductions and interpretations. Interpretations involve a combination of objective and subjective ele- ments-characteristically they are what the interpreter makes of an experience and offers to an interlocutor. For Geertz, remember, ethnography is, from notebook to monograph, a seamless web of interpretation: our own constructions of other people's constructions of what they and their compatriots are up to' (1973: 9). Sperber, on the other hand, is concerned to unpick the stitches that hold it all together. Rather than settle for the finished ethnographic product, he wants to ask-indeed his overall project requires him to ask-whose construction of what?

If we are to use anthropological interpretations as the materials for building empirical generalisations, they need a particular kind of qualifica- tion-what he calls a 'descriptive comment':

A descriptive comment identifies the object represented and specifies the type of representation involved. It thereby makes it possible to draw non-empirical inferences from a non-descriptive representation. It provides, so to speak, the directions for its use (Sper- ber 1985: 12).

Obvious examples of descriptive comments include captions to pictures and keys to maps. Less obvious examples-like, for example, what would count as an adequate descriptive comment in an ethnographic account-are a little harder to come by. Sperber, unfortunately, does not offer his readers an example of what an anthropologically-useful ethnographic account might look like.

We can, though, get the general idea which is not in itself especially wild-eyed or radical. Consider how a historian constructs a historical monograph. The language in such cases is likely to be quite similar to the language of the typical anthropological monograph, and to contain a similar mixture of description and interpretation. Where the two tend to differ is in the way in which the reader is made aware of the raw material upon which the account is based. The raw data of an historical account, apart from the occasional direct quotation, are no more present than the raw material of an anthropological account. They are, however, made explicit




through footnotes and documentary citations. Most readers will be content to read the surface of the text and ignore the fine print which details the conditions of production of the main text, but the fine print is there for specialists and the sceptical to scrutinise. Above all it allows the possibility of empirical challenge to both the description and the interpretation found in the main text.4 That this leaves us no closer to an impossible contact with 'what really happened' in no way detracts from the importance of this rule of the game of historical discourse. It is still adequate to its purpose of limiting the 'field of possible constructions'.

The scholarly apparatus of footnotes in a work of history is, I suggest, an example of a highly developed system of what Sperber calls 'descriptive comment'. Compare this with Sperber's remarks on an example from Evans-Pritchard's Nuer religion. The chosen passage is an account of an incident when a man had been accused of practising too many sacrifices. Of the account itself Sperber notes that, while it seems 'about as raw a factual statement as you will ever find in most ethnographic works... not a single statement in it expresses a plain observation' (1985: 14). Of the generalisation which the anecdote arid its gloss are called forth to support by Evans-Pritchard ('Through the sacrifice man makes a kind of bargain with his God') Sperber asks the sort of questions that generations of bright undergraduates have asked of standard ethnographies: whose interpretation is this? the anthropologist's? the Nuer's? all Nuer's or just one or two? In fact, he concludes, the interpretation in question seems to be an attempted 'compromise between Nuer thought and the ethnographer's means of expression' (1985: 16). And much of what passes for anthropological theory is, in fact, 'interpretative generalisation' of a low and rather uninformative kind, employing terms such as 'sacrifice', 'shaman', 'ritual', which have been long since cut adrift from any original and specific denotation and instead act as intermediaries in the interpretation of ethnographic examples.

Now if I were Evans-Pritchard (to coin a phrase) my answer to these strictures would, I imagine, be something like this. The task of the anthropologist is the translation of culture; our first priority is to render intelligible the ideas and actions of people in another culture; it is therefore quite reasonable that we should attempt to do so by working away from a specific utterance or incident, through various intermediary interpretations, such that the content of the original is rendered as faithfully and as coherently as possible. It is true that someone like Sperber, interested in gaining access to a wide range of more-or-less unmediated representations, may be dis- appointed by this procedure; but his is a minority interest, and an ethnography that would satisfy him would probably frighten off all but the most dedicated of readers. I imagine that a similar argument would be advanced by quite a few other ethnographers, especially those-probably now a majority-with little or no commitment to the building of general models of the variability of human social and cultural existence.

I think, though, that the force of Sperber's critique is not limited to its implications for what he sees as a properly scientific anthropology. Al- though he approaches ethnography with very different assumptions and




intentions he nevertheless, like Geertz, has to concede the problematic status of interpretation in anthropological work. Unlike Geertz, though, he would have us make all possible effort to separate interpretation from description. This can be seen in the second part of his argument where he examines the use of 'free indirect speech'. Free indirect speech is 'the style which allows the author to tell a story "from the point of view of the actors", and the reader to identify with them' (1985: 19). 'Through the sacrifice man makes a kind of bargain with God' is a representation which allows the reader to see things as if he or she were a Nuer. The relationship between this and any utterance provided by a Nuer to Evans-Pritchard is, as Sperber's analysis demonstrates, unclear, as is the relationship of Geertz's 'disconnected, aesthetic quanta' to any real or imagined Balinese representation of a cockfight.

The problem of naturalism In other words, a great deal of ethnographic writing carries little or no explicit reference to the ethnographic work on which it is based. Why should this be so? The most compelling reason would be the uneasy status of ethnographic work itself, in particular the relationship between individual experience and 'scientific', or, if you prefer, 'objective', generalisation. Because ethnographic experience is so specific as to be unrepeatable-a fact which in itself removes ethnographic evidence from most understandings of scientific data-generalisation is peculiarly problematic. A male ethnographer learns different things from a female ethnographer, and countless contingencies intervene during the time in the field, from world historical eruptions such as elections and droughts and wars to such apparent trivia as chance meetings, illness and missed buses. Obviously, good ethnographic practice involves the attempt to make methodical what may have been first discovered by chance; but there is no denying the idiosyncrasy of individual ethnographic experience. In addition, the tradition of the lone fieldworker (occasionally supplemented by spouse and children) magnifies the personal anxieties faced by all researchers. The anthropological habit of writing at arm's length is not to be dismissed as an act of simple bad faith; it is as often a tactic of emotional self-defence.

In that crucial period of professional consolidation between, say, 1940 and 1962-marked in British anthropology by the publication of Evans- Pritchard's The Nuer (1940) at one end, and the polemical attacks of Needham (1962) and Leach (1961) at the other-it is possible to discern the growth of a style of ethnographic writing which I shall call 'ethnographic naturalism'. 5 I use 'naturalism' by analogy with dramatic theory, to refer to the creation of a taken-for-granted representation of reality by certain standard devices.

My choice of terms comes in particular from Raymond Williams's discussion of Brecht's dramatic theory; this is because Williams's discussion is imbued with a recognition of the power and importance of some kinds of naturalism. The- danger of naturalism, though, is 'the exclusion, by particular conventions of verisimilitude, of all direct commentary, alternative




consciousness, alternative points of view' (Williams 1971: 278). For Brecht the effect of such naturalism was to lull the audience and render it passive; in its place he proposed the use of various techniques which would make the audience aware of the conditions of production of the play itself, and also of the circumstances of the action within the play. For modern anthropology in its period of professional consolidation one effect of naturalistic devices was to deny the particularity of ethnographic experience by literary means rather than confront the implications of such particularity. Against this we have to chart the gains of the style, not least the success of classic ethnographies in establishing the potential intelligibility of what had hitherto been dismissed as 'savage', 'primitive', or 'superstitious'.

Free indirect speech-the replacement of 'An old man told me at a sacrifice, "This is a kind of bargain with God"' with 'Man makes a kind of bargain with God'-is but one feature of 'ethnographic naturalism'. The devices of ethnographic naturalism do not serve as one more-or-less adequate way amongst others to represent a chosen object. Rather they serve to constitute a particular sort of object-homo ethnographicus-and, in the process, other possible understandings of the ethnographer's material are eliminated. Take for example the way in which this object of discourse is homogenised: 'The Nuer is a product of hard and egalitarian upbringing, is deeply democratic, and is easily roused to violence' (Evans-Pritchard 1940: 181). We know that some of our neighbours and colleagues are more democratic than others, we know that different kinds of English people are more easily roused to violence than others (men for example); but the Nuer, as represented by Evans-Pritchard, do not appear to vary in this way. 'In the normal course of things, the Balinese are shy to the point of obsessive- ness of open conflict' (Geertz 1973: 446)-except, one presumes, those that are not, or moments when the course of things is palpably abnormal (1965 for example).6 It may well be the case that Balinese society and Nuer society are culturally homogeneous in a way that Britain and the United States are not; but given that consensus has been taken to be a defining feature of primitive society, at least since Durkheim's mechanical solidarity, while difference is read as the sign of the modern, it seems probable that this feature is as much a product of our stylistic repertoire as it is of any particular observation. Certainly, my own field experience in Sri Lanka was of a cultural setting characterised by argument, scepticism and dispute about all sorts of aspects of everyday culture; yet it is nonetheless quite possible to read recent ethnographic accounts of a curious homogeneous thing called 'Sinhalese culture'.

Most spectacularly of all, a few writers have performed the same levelling process in the west, for example David Schneider whose account of Amer- ican kinship and American culture (1968) eliminates differences of class and ethnicity and presents instead a disturbingly seamless description of key American symbols and their interpretations. It is at this point, in my experience, that students start to give voice to their worries about what it is that they are supposed to be reading about. Ethnographic naturalism, while working with ostensibly unproblematic literary devices, in fact con-




structs a kind of object-a world robbed of its idiosyncrasies and foibles- which is foreign to the experience of its readers; and while the readers can accept such foreignness if the object is said to be from a distant time or place, the use of similar devices in describing a known area of experience provokes considerable resistance. Defenders of Schneider (e.g. Marcus & Fischer 1986: 149-51) might argue that his true purpose is defamiliarisa- tion-the rendering strange, and thus new, of the commonplace and unquestioned-and there is an element of truth in this. But the most telling lesson of Schneider's work concerns ethnography rather than America-it is anthropological writing that it puts in question far more than American kinship.

Another aspect of ethnographic naturalism is the absence of any tangible point of view. The narrator is invisible and omniscient-an effect much enhanced by the use of free indirect speech. The reason most often put forward for the habit of ethnographic effacement-the removal of the ethnographer from the scene of writing-is that without it ethnography will descend into subjectivity and autobiography. This is indeed a danger, but the alternative, the denial of ethnographic presence and the specificity of ethnographic experience, is equally dangerous: it substitutes an unchal- lengeable subjectivity for a challengeable subjectivity.

Experimental ethnography A brief examination of three recent American ethnographies will demon- strate some of the gains, and some of the limitations, of the kind of approach I have been advocating. They are all accounts of Morocco based on fieldwork in the late 1960s and early 1970s-fieldwork carried out in the shadows of Clifford, Geertz and, I suspect, Franz Fanon as well. The three books in order of publication are Paul Rabinow's Reflections on fieldwork in Morocco (1977), Vincent Crapanzano's Tuhami (1980), and Kevin Dwyer's Moroccan dialogues (1982).

Rabinow's book is the shortest, in many respects most conventional, and probably the most successful of the three. It is a companion to a more conventional ethnography Symbolic domination (1975). It is a fairly straightforward account of fieldwork, written in an autobiographical mode, differing from what is rapidly becoming a popular genre only in its some- what greater frankness and self-consciousness.

The chief value of Rabinow's account is the description of his relationship with his various informants. These can be classed into two rough categories-the dispensers of 'official' information, and the marginals. In the town where he first settled he spent an uncomfortable time trying to learn Moroccan with an ex-colonial interpreter who, he gradually sensed, was providing him with predigested gobbets of language and culture, packaging Arabic 'as if it were a tourist brochure' (Rabinow 1977: 27); he then found himself at the other end of the social spectrum, hanging out with a small- time pimp. Once installed in a village, his first useful informant was a young man of dubious reputation, who, as the relationship became a growing source of tension between the anthropologist and the rest of the village,




was replaced by a pillar of respectability in the community, a young married landowner who simply presented himself and offered to work as informant. What this man provided over a long period of time was another kind of 'official' version; together he and Rabinow systematically constructed full genealogies, details of landholding and agricultural arrangements and the like. What Rabinow could not elicit was the detail of past political tension which he wanted; it required a visit from his friend the pimp-who was happy to volunteer his version of these events-to provoke the other villagers to unbutton.

Although the book is autobiographical in mode, and otherwise largely addressed to problems of anthropology, it does provide an occasionally sharp and often convincing portrait of Moroccan society. In particular, the ways in, which people responded to the anthropologist-as an object of suspicion, or fear, or as- a resource to be exploited, or as a medium to be managed-are themselves important data about the society, quite apart from whatever light they may shine on Rabinow's other ethnography. That there was a general concern to present the ethnographer with an official or respectable version of the place reveals something of the nature of cultural self-awareness, as well as of the politics of fieldwork. There is, on the other hand, a feeling of completeness, of parts gradually fitting into a jigsaw, in the description of the research itself which I cannot help finding suspiciously neat and tidy.

What Rabinow emphasises is the way in which ethnographic knowledge is the result of a complex process in which the informant and the anthropologist, in trying to establish a common ground of understanding, are both forced to reflect on their own assumptions and cultural preconcep- tions:

Whenever an anthropologist enters a culture, he trains people to objectify their life- world for him. Within all cultures, of course, there is already objectification and self- reflection. But this explicit self-conscious translation into an external medium is rare. The anthropologist creates a doubling of consciousness. Therefore, anthropological analysis must incorporate two facts: first, that we ourselves are historically situated through the questions we ask and the manner in which we seek to understand and experience the world; and, second, that what we receive from our informants are interpretations, equally mediated by history and culture (1977: 119).

This also clarifies the apparent elective affinity between anthropologists and informants who are themselves marginal to the culture being studied; one thinks, for example, of Victor Turner's (1967) famous account of his long exegetical sessions with Muchona and the schoolteacher, all three of them in their different ways on the fringe of Ndembu society. The marginal figure is more likely to ponder on what is going on and why, precisely because his or her partial detachment from the centre of things is if not a problem, at least pause for thought.

What is created in this kind of encounter is a kind of intermediate ground between cultures, 'the beginnings of a hybrid, cross-cultural object or product', a 'liminal world' as Rabinow (1977: 153) describes it. This is the same basic process that Sperber describes; but where Sperber locates it at the level of ethnography as text-as a product of the ethnographer's attempt




to interpret their representations-it seems to reside rightly at the level of ethnography as practice. While a 'descriptive comment' (specifying what Rabinow calls the historical situation of anthropologist and informant) can place such interpretations back in the fieldwork situation, the passage from there to unmediated, or unreflective, representations seems blocked.

A similar insight lies at the heart of Crapanzano's Tuhami (1980) (subtitled Portrait of a Moroccan). Tuhami himself is a markedly marginal character, an unmarried tilemaker living on the fringe of the urban proletariat in a Moroccan town, who first came to Crapanzano's attention as a potential informant on saints and demons, not least because of his considerable personal experience of encounters with them. The book is more self-con- sciously experimental in its writing than Rabinow's, taking the form of a series of dialogues between Crapanzano and Tuhami interspersed with passages of context about Moroccan society, written in a conventional ethnographic voice, and passages of reflection on the relationship between Tuhami (the Other with a capital 0) and the anthropologist. Like Rabinow, Crapanzano points out that the relationship with the informant created a new situation for both of them. 'I became,' Crapanzano reflects, 'I imagine, an articulatory pivot around which he could spin out his fantasies in order to create himself as he desired' (Crapanzano 1980: 140). A natural storyteller, Tuhami began to use the sessions as an opportunity to create a life for himself, weaving together what we would call fantasy and reality; as a psychoanalytically-inclined anthropologist Crapanzano began to see himself in a therapeutic role, pushing Tuhami towards some imagined release.

The element of self-conscious experiment is to be found in the juxtaposition of different 'voices' and styles, the text's abrupt shifts between anthropological detachment and snatches of raw dialogue. The intention is to force readers into making their own connexions between the different parts of the text, while Tuhami's own mixture of the fantastic and the down-to-earth is allowed gradually to subvert Crapanzano's (and, I guess, the readers') expectations of a 'realistic' life-history.

The result, though, strikes me as unsatisfactory on a number of counts. The carefully selected and edited snatches of dialogue with Tuhami are too oblique-too little context is provided for the reader to cope with them as they stand and the result, perversely enough, is to force a greater reliance on the author's subsequent commentary. The commentary itself is, though, also oblique. It is most revealing not about Morocco, nor about Tuhami or his milieu, nor even about Crapanzano himself-it tells us most about the contents of Crapanzano's library. The increasingly fragile structure of Tuhami's life groans, and all but collapses, as it has to bear the weight of Sartre and Simmel and Lacan and a great deal of serious and committed head-scratching over the nature of The Other.

The most radical experiment in simultaneous authorial presence and effacement is Kevin Dwyer's Moroccan dialogues (1982). The book is the product of a second period of fieldwork, in which Dwyer, troubled by the nature of anthropological representations, returned to the family he had earlier stayed with and, in the course of one summer, took to recording




long conversations he held with the head of the household. The book is make up of transcriptions of these conversations, with a certain amount of ethnographic scene-setting, and passages in which Dwyer reflects on the context of the particular conversation and what, at the time, he was trying to get at in asking particular questions. After a couple of hundred pages in this vein the book climaxes with two chapters of reflection on anthropology, self, other and the dialogic.

Dwyer is concerned to avoid the bad faith involved in the presentation of seamless, polished accounts of other cultures. In this respect he perhaps succeeds only too well. One learns very little about Morocco, about Moroc- can peasants, or even about the particular Moroccan peasant who dominates the book. The impression given of anthropological work is that it is a rather long-winded device for keeping innocent people up all night talking. In this case far too much of the conversation is too particular; as with Tuhami, the use of unmediated material-material that is as full of allusions to particular people and local events as fieldnotes always are- leaves the reader helpless. The anthropologist as writer cannot help intervening, however much his whole strategy depends on his non-intervention. The sorry conclusion that can be drawn from the book is that 'thin description' by itself-the attempted refusal of textual mediation between the actual field encounter and the reader of the ethnography-is as much a barrier to ethnographic representation as is the more conventional practice of authorial omniscience.

A number of conclusions can be drawn from this brief survey, conclusions which refer back to points made by Geertz and Sperber. Much of what I have described would seem to bear out Geertz's insistence on the interpretative nature of ethnographic data as well as of ethnographic texts: that anthropologists explicate other people's explications. The attempts by both Crapanzano and Dwyer to eliminate the anthropologist as mediator only serve to render the reader more helpless and dependent on whatever eventual interpretation is going to be offered. But as Rabinow and Cra- panzano both make clear the alternative elimination-the removal of the anthropologist and informants from the finished text-also deprives the reader of important information. The conditions of production of anthropological knowledge-who told whom? what? when? and why?-are themselves data of considerable importance. Something like Sperber's descriptive comment can begin to situate ethnographic interpretations back in the context of their own production, even if that context is already one of interpretation rather than one of unreflecting representation. Most important of all, if we want to present culture as an area of contest or dispute in which people have different points of view, in which power and politics affect the way in which different people make sense of their world and represent it to others, then we have to employ some sort of literary practice which allows the historical specificity of ethnographic experience to appear as an integral part of the final analysis.8




Text and context This brings me finally to Writing culture. There has been in the last decade a rising interest in America in what, with apologies, I shall call meta- anthropology-the anthropology of anthropologists. Key texts in this area are, apart from Writing culture itself, the series of History of anthropology volumes edited by George Stocking (Stocking 1983a; 1984; 1985; 1986), and two important essays on ethnography-one by James Clifford (1983a) in the journal Representations, and one by Marcus and Cushman (1982) in Annual review of anthropology. But within this trend there are in fact two discernible tendencies; I shall describe their exponents as Formalists and Historians. Formalists are concerned with analysing the internal structure of anthropological texts as things in themselves; Historians on the other hand are concerned with the 'external' relations of texts, with siting particular work in its own social, cultural and historical context. Formalists do not refer to themselves as such, preferring on the whole sexier designations such as post-structuralist or experimentalist; Historians seem to be happy to be thought of as historians.

Some people manage to be both-James Clifford, for example. In a review of Clifford's splendid biography of the French missionary ethnographer Maurice Leenhardt Person and myth (Clifford 1982), Paul Rabinow points to an area of tension within Clifford's work; Clifford's book, he points out, hovers 'between a successful standard historical approach and a more dangerous post-modern one whose claims to traditional standards are less secure, but whose claims to creativity are stronger' (Rabinow 1983: 196-7). This can be seen in Clifford's essay in Representations: 'On ethnographic authority' (Clifford 1983a). The essay starts with an unexceptional summary of the historical development of ethnography as the central activity of modern anthropology, and after a middle section on Geertz's interpretative ethnography and the model of the text, proceeds to assess Dwyer's and Crapanzano's experiments, as well as the possibility of more radical departures from standard modes of authorship. But what we are left with in a rather flat conclusion is a typology of modes of authority ('ex- periential, interpretive, dialogical, and polyphonic' (1983a: 142)), and the observation that ethnographers have to choose between them.

The lameness of this conclusion-which shows Clifford in Formalist mood (and not a little-baffled as to where to go next)-contrasts strongly with, for example, his essay on Marcel Griaule in History of anthropology (Clifford 1983b) in which Griaule's ethnographic practice, as well as his ethnographic writing, are assessed against a subtly invoked background of iolonial and post-colonial society. Similarly, Marcus and Cushman's essay

on ethnography is bland and unremarkable compared to Stocking's (1983b) recent reconstruction of the historical context of Malinowski's Argonauts, with its playful reading of Malinowski's precepts on fieldwork as a 'mythical charter' for subsequent anthropology, a reading which derives its force from the strength of the historical analysis which surrounds it. On the basis of Clifford's own work, I think we are bound to reverse the terms of Rabinow's judgement: the 'dangerous post-modern' approach may look




more exciting but, so far, has yielded rather ordinary results; the 'standard historical approach', on the other hand, has proven remarkably fruitful.

My unhappiness with the post-modern turn in American anthropology is twofold. First there is the abandonment of any consideration of problems of validation: these are instead subsumed under the rubric of 'authority', which is itself portrayed as a literary rather than a practical issue.9 It is worth remembering that both Geertz (1973: 30) and Ricoeur (1971: 549-51) expend considerable energy on the problem of the validation of differing interpretations, proving, if nothing else, that such questions are not solely the province of naive empiricists and other sorts of Anglo-Saxon.

The most easily imaginable answer to the problem of the validation of competing interpretations is not literary but practical. The case of Margaret Mead, to which I have already alluded, suggests one simple strategy for 'limiting the field of constructions'. We live in a world where the possibility of ethnographic subjects contesting the anthropologist's description is more and more likely; not only is this a thoroughly good thing, we should make every effort to encourage it. The moment of writing is a rather late stage for the interpreter to reveal his or her interpretations to the interpreted. Instead it would seem more fruitful to try to devise ways in which ethnographic subjects are actively involved in their own self-representation. The anthropologist's skills could then be used as an enabling device in a complex and quite possibly acrimonious process of interpretation and argument. This is not to insist on the subjects' ratification of a particular ethnography-if anthropology is indeed a 'long conversation' the anthropologist also has a right to his or her own say-but simply to suggest that they be included in the 'interpretative community' to which the text is addressed.

I can sense what remains of the reader's patience expiring at this apparently naive and utopian point. But there are plenty of examples of the kind of research I am talking about. The Mass Observation movement in the Britain of the 1930s and 1940s-currently undergoing a publisher's revival-relied upon a huge number of volunteer ethnographers reporting on various social and cultural nooks and crannies (see Calder & Sheridan 1984). The group of historians associated with the journ,al History workshop in the 1970s again deliberately sought to involve people in the production of their own histories (Samuel 1981). Still more recently in France the sociologist Alain Touraine and his collaborators have devised what they call the strategy of 'sociological intervention', in which the division between researcher and researched is again blurred, for their studies of social move- ments like Solidarity or the French anti-nuclear movement (Touraine 1981; Touraine et al. 1983a; 1983b). More modestly there is the work of Richard Thorn and his colleagues in Visual Anthropology at Bristol Polytechnic on the recording of contemporary culture. None of these is offered as an unproblematic model for emulation; but all of them provide valuable examples of the potential strengths and weaknesses of different and more open research practices.

Ethnographic writing is but one stage in a complex process of cultural production. But the pervading assumption in Writing culture seems to be




that texts can be compared as formal objects, regardless of the conditions of their own production. Derrida's slogan 'There is nothing outside the text' may have its strategic value in the style wars currently raging in the literature faculty. If nothing else, it can serve as a constant reminder that, for example, we are reading Evans-Pritchard's account of his interpretation of what he took to be religion among some people then classified as the Nuer at a time of extended political crisis in the southern Sudan in the 1930s; we are not, for all that, confronting some transcendent reality called 'Nuer religion' to which Evans-Pritchard's text may correspond more or less successfully. The force of the slogan in this reading may not be especially novel (cf Rorty 1982: 154) but its effect may still be salutary.

A slightly different reading, which runs implicitly through a number of the contributions to Writing culture, is that texts can be wholly decontextu- alised and compared as formal objects, stripped of history and living in a social vacuum. Crapanzano (1986), for example, compares Goethe's description of the Roman carnival with Geertz's essay on the cockfight without feeling any need to justify such an anachronistic juxtaposition. Context is ignored in other ways too. Writing of Evans-Pritchard's introduction to The Nuer, for instance, Renato Rosaldo asks 'Why did the government of the Anglo-Egyptian Sudan request his report? How much did it pay for the research and the publication of its results?' (Rosaldo 1986: 88). To this I would answer with some exasperation: why don't you find out? the relevant papers are presumably sitting in some colonial archive or other and are available for inspection (cf. Geertz 1988: 49-72). It is hard to avoid the conclusion that *'intertextuality' is used as a chic excuse for abandoning primary research altogether. I am put in mind of the definition of a critic attributed to the poet Robert Frost: 'A critic is someone who pisses into a river and says look what a big stream I made'.

A number of important questions about ethnography are omitted from the discussion in Writing culture. These include, for one, the relationship between institutional structures and styles of writing: here a relevant comment might be that what remains of social science funding in Britain is in the hands of Gradgrinds of a distressingly factual sort, few of whom are likely to be impressed by a hundred pages of heartfelt reflection on Self and Other. The relationship between anthropological work and anthropological writing is also undiscussed. When Geertz answered his own question 'what do anthropologists do?' with 'anthropologists write' he managed to single out the one activity that does not differentiate anthropology from any other kind of intellectual work-anthropologists wade into paddy fields, get sick and read bad novels rather than confront another day of mounting misapprehensions; they also take photographs, make films and tape record- ings; a privileged few even get to teach students. The fact that they mostly do it by themselves in strange places is another oddity that passes unre- marked upon in Writing culture. There is also the question of who reads ethnography 10 and the ways in which anthropology texts are written so as to exclude all but the committed fellow professional from the exclusive circle of understanding. (Writing culture, with its lumbering use of terms




from the dead critical discourse of the seventeenth and eighteenth centuries, will never be accused of popular vulgarisation.)

Some of these issues are raised in the penultimate-and best-paper in the book, Paul Rabinow's essay (1986) which is tellingly entitled 'Repre- sentations are social facts' and which delivers its own critique of the assumptions on which the whole project is based. Not least among these is the discernible shift from trying-however imperfectly-to represent other people in other places, to an exclusive concern with representing people who try to represent other people in other places; at one point Rabinow describes this as 'the voice from the campus library'. The scene of post-modern anthropological writing is revealed in that phrase: it is an institution of higher education, preferably in the USA, usually somewhere in the sunbelt (although student-free zones such as Princeton's Institute for Advanced Study have their appeal). The libraries are well-stocked, the students well concealed. Time hangs heavy. Ambition nags away at the back of the mind. A new product appears in the consumer paradise: it is exclusive, well-packaged, evoking the radical spirit even as it establishes the magic of authority; it can be applied to anything at all; and its application, after some preliminary intellectual discomfort, is quite painless; you don't even have to set foot outside your front door. We are in the world of what Michael Silverstein (1985) has described as 'yuppie anthropology'.

There is something in the tone of Writing culture which encourages such ungenerous fantasy. Much that I have argued earlier-about the literary denial of anthropological presence in ethnographic writing for example-is echoed again and again by the contributors. Why then am I so critical? First of all, it seems more than likely that the book will provoke a trend away from doing anthropology, and towards ever more barren criticism and meta-criticism. Secondly, there is every chance that it has set the agenda for self-criticism in American (if not British) anthropology for the next few years; the omission of crucial questions-who does anthropology, in what context, and to whom are the results made available-is all the more regrettable. Thirdly, there are already signs that literary criticism will cross the boundary between the descriptive and the prescriptive as, for example, in Marcus and Fischer's construction of a canon of 'experimental' ethnography in Anthropology as cultural critique (1986).

The great strength of fieldwork in Anglo-American anthropology has lain in the fieldworker's ability to relate specific representations to a rich context of social, cultural and historical knowledge. There can be no prior justification for refusing the same context for anthropological representations themselves; the work of Stocking and Clifford on the history of anthropology not only instructs us about the past, it also promises to make us more profitably self-conscious in the future. Similarly, in arguing for a more open style of ethnographic writing, I have suggested that it is in the interest of both writer and reader that explicit attention should be paid to the specific historic and social sources of anthropological representations. I should add that there is every sign that this is what a large number of anthropologists are doing.




A critic once likened the music of the great Afro-American composer Thelonious Monk to the feeling of 'missing the bottom step in the dark', that brief moment of vertigo when the familiar and unexpected drops away. I suspect most anthropologists aspire to a similar ability to upset the apple-cart of cultural expectation. What I hope this article has demonstrated is that this ability cannot be reduced to recipes and formulas from the light industry that is American literary criticism; as Sperber says, 'ideally... each ethnographer should rethink the ethnographic genre' (1985: 33). In the end, though, we may also need to rethink the style of anthropological work as well as the style of anthropological writing.11

NOTES

I should like to thank Mark Whitaker, Elizabeth Nissan, Richard Fardon, and audiences at Sussex and Durham for their comments on a first draft of this article. I must also thank Ralph Grillo for his encouragement. I hope that Terry Turner will also recognise his (necessarily unacknowledged) contribution.

1 The furore over Freeman's (1984) attack on Mead was, I suspect, far more traumatic to American anthropology than it was in Britain; the search for alternatives to Mead's 'scientistic' ethnographic epistemology, of which Writing culture is obviously one, can be partly traced to this trauma.

2 The example is in fact based on Scott's (1985) brilliant account of class in everyday peasant life in a Malay village, a text which modestly embodies much that I am recom- mending. It is all the more ironic that it is the ethnographic contribution of a political scientist rather than an anthropologist.

3 It is interesting to note that Geertz's non-ethnographic writings show a growing irrita- tion with the usual trappings of scholarly attribution, preferring instead the knowing allu- sion and the buried half-quote, while in Negara (1980) and Islam observed (1968), footnotes are replaced by a parallel scholarly commentary, citing references and quibbles of detail, printed at the back of the volume and loosely tagged to the pages of the main text.

4 Public questioning of the empirical content of ethnography is extremely rare, and, tellingly, almost always confined to cases where an ostensibly anthropological text has won a wider public audience-Coming of age in Samoa, The mountain people, The teachings of Don Juan, Shabono. Such questioning seems as much a product of the patrolling of disciplinary boundaries as of anything more high-minded.

5 Marcus and Cushman describe a similar style which they call 'ethnographic realism'; I have problems with their claim that this has been dominant for 'approximately the past 60 years' (1982: 25) i.e. since the publication of Malinowski's Argonauts. This is to gloss over the considerable stylistic differences between Malinowski, Mead, Firth, Evans-Pritchard, Fortune and Benedict, let alone the differences between them and recent figures such as Geertz or the ethnoscientists of the 1960s.

6 In fairness it should be pointed out that Geertz himself alerts the reader to the danger of this practice in the earlier 'Person, time and conduct in Bali' ([1966] 1973: 368 n.7), without in any way modifying the main text to take account of possible exceptions and qualifications. By the time of 'Deep Play', seven years later, he seems to have felt no such qualifications necessary (cf. n.3 above).

7 Dwyer's book was also presented as a BBC radio documentary in 1985 under the title Making a noise about life. I have listened to a recording of this with two very different groups of students; the majority in both groups described it as both baffling and boring. This strikes me as a pretty fair evocation of what much anthropological fieldwork is actually like.

8 This seems to be in agreement with Fabian's (1983) denunciations of anthropological 'allochronism'; Fabian also recognises that the problem of anthropological representation is a problem of anthropological practice. Unfortunately, the specification of desirable changes in anthropological practice goes missing in his breath-taking display of polysyl-




labic metatheory. 9 Clifford murmurs in a footnote 'Modes of authority based on natural-scientific

epistemologies are not discussed' (1983a: 142 n.1). 10 Marcus and Cushman devote two pages of their article to a limited discussion of the

readership of ethnographies (1982: 63-4); although students and the general public are men- tioned as categories of readers, all the important author-reader relations are seen to inhere within a world of professional academics.

1 My description of American literary criticism as a 'light industry' seems especially prescient in view of the deluge of publications on this theme which have appeared since this essay was originally drafted. The most important of these is probably Geertz's slim volume on anthropologist as author (Geertz 1988), which uses a number of the examples discussed above, makes a number of the same points of detail, but heads off in a quite incompatible direction.

REFERENCES

Berlin, B. & P. Kay 1969. Basic color terms. Berkeley, Los Angeles: Univ. of California Press.

Calder, A. & D. Sheridan 1984 (eds). Speak for yourself London: Jonathan Cape. Clifford, J. 1982. Person and myth: Maurice Leenhardt in the Melanesian world. Berkeley,

Los Angeles: Univ. of California Press. 1983a. On ethnographic authority. Representations 1, 118-46. 1983b. Power and dialogue in ethnography: Marcel Griaule's initiation. In Stocking

(ed.) (1 983a). & G.E. Marcus 1986 (eds). Writing culture. Berkeley, Los Angeles: Univ. of California

Press. Crapanzano, V. 1980. Tuhami: portrait of a Moroccan. Chicago: Univ. Press.

1986. 'Hermes' dilemma: the masking of subversion in ethnographic description'. In Clifford & Marcus (eds) (1986).

Dwyer, K. 1982. Moroccan dialogues. Baltimore: Johns Hopkins Univ. Press. Evans-Pritchard, E.E. 1940. The Nuer. Oxford: Univ. Press. Fabian, J. 1983. Time and the other. New York: Columbia Univ.Press. Freeman, D. 1984. Margaret Mead and Samoa. Harmondsworth: Penguin. Geertz, C. 1968. Islam observed. Chicago: Univ. Press.

1973. The interpretation of cultures. New York: Basic Books. 1980. Negara. Princeton: Univ. Press. 1988. Works and lives. Stanford: Univ. Press.

Leach, E.R. 1961. Rethinking anthropology. London: Athlone Press. Malinowski, B. 1922. Argonauts of the Western Pacific. London: Routledge & Kegan

Paul. Marcus G.E. & Cushman R. 1982. Ethnographies as text. Ann. Rev. Anthrop. 11, 25-69.

& M. Fischer 1986. Anthropology as cultural critique. Chicago: Univ. Press. Mead, M. 1928. Coming of age in Samoa. New York: Morrow. Needham, R. 1962. Structure and sentiment. Chicago: Univ. Press. Rabinow, P. 1975. Symbolic domination. Chicago: Univ. Press.

1977. Reflections onfieldwork in Morocco. Berkeley, Los Angeles: Univ. of California Press.

1983. Facts are a word of God: a review essay. In Stocking (ed.) (1983a). 1986. Representations are social facts: modernity and post-modernity in

anthropology. In Clifford & Marcus (eds) (1986). Ricoeur, P. 1971. The model of the text: meaningful action considered as a text.

Social Res. 38, 529-62. Rorty, R. 1982. Consequences of pragmatism. Brighton: Harvester. Rosaldo, R. 1986. From the door of his tent: the fieldworker and the inquisitor. In

Clifford & Marcus (eds) (1986). Samuel, R. (ed.) 1981. People's history and socialist theory. London: Routledge & Kegan

Paul.




Schneider, D.M. 1968. American kinship: a cultural account. Englewood Cliffs: Prentice-Hall.

Scott, J.C. 1985. Weapons of the weak. New Haven: Yale Univ. Press. Silverstein, M. 1985. Review of D. Parkin (ed.) Semantic anthropology. Am. Ethnol. 12,

795-6. Sperber, D. 1985. On anthropological knowledge. Cambridge: Univ. Press. Stocking, G.W. (ed.) 1983a. Observers observed: essays on ethnographicfieldwork (Hist.

Anthrop. 1). Madison: Univ. of Wisconsin Press. 1983b. The ethnographer's magic: fieldwork in British anthropology from Tylor to

Malinowski. In Stocking (1983a). (ed.) 1984. Functionalism historicized (Hist. Anthrop. 2). Madison: Univ. of Wisconsin

Press. (ed.) 1985. Objects and others (Hist. Anthrop. 3). Madison: Univ. of Wisconsin Press. (ed.) 1986. Malinowski, Rivers, Benedict and others (Hist. Anthrop. 4). Madison: Univ.

of Wisconsin Press. Touraine, A. 1981. The voice and the eye. Cambridge: Univ. Press. - Z. Hegedus, F. Dubet & M. Wieviorka 1983a. Anti-nuclear protest. Cambridge: Univ.

Press. - F. Dubet, M. Wieviorka & J. Strzelecki 1983b. Solidarity. Cambridge: Univ. Press.

Turner, V.W. 1967. The forest of symbols. Ithaca: Cornell Univ. Press. Williams, R. 1971. Drama from Ibsen to Brecht. London: Chatto & Windus.

L'anthropologie comme un genre d'ecriture

Resume Cet article fait la revue de l'interet recent dans les aspects litt6raires de l'6critur( ethnographique, se concentrant sur le travail de Geertz, Sperber, et les auteurs associes aL volume collectif Writing culture. Alors, est-il argumente, que des questions s6rieuses sont soulevees dans une partie de ce travail, iA est argumente aussi que des modes recentes dans la th6orle critique litt6raire peuvent se reveler inutiles pour s'adresser a ces questions. En particulier, la tendance a lire des textes avec peu ou pas de consideration pour le contexte social et historique dans lequel ils furent ecrits semble une approche specialement sterile. Plut6t, est-il argumente, l'anthropologie est aussi bien une facon de travailler-une sorte d'activite pratique-qu'elle est une feaaon d'ecrire. Une reconnaissance de l'element per- sonnel dans la r6alisation de textes ethnographiques peut aider le lecteur a une meilleure evaluation de l'interpretation proposee; un changement plus radical demande un changement dans la pratique anthropologique aussi bien que dans l'ecriture anthropologique.



anthro as writing

Documents

anthropological writing

way of writing

writing of ethnography

use of language

terms conditions of

jstor terms

jstor archive

amall use subject