DISCOURSE AND POWER: ON THE LIMITS OF ETHNOGRAPHIC TEXTS

Author(s): Paul Rabinow

Source: Dialectical Anthropology , JULY 1985, Vol. 10, No. 1/2, ANTHROPOLOGY AFTER
'84 - STATE OF THE ART, STATE OF SOCIETY PART II (JULY 1985), pp. 1-13
Published by: Springer

Stable URL: https://www.jstor.org/stable/29790142

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

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
https://about.jstor.org/terms

Springer is collaborating with JSTOR to digitize, preserve and extend access to Dialectical
Anthropology

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 1

DISCOURSE AND POWER: ON THE LIMITS OF ETHNOGRAPHIC TEXTS

Paul Rabinow

WRITERS AND AUTHORS

"language is . .. the institutionalization of to the nineteenth century (Barthes calls this

subjectivity" the "entire classical, capitalist period", dem?
An interest in the making of ethnographic onstrating his disinterest in sociology) the
texts - the rhetorical conventions of how person who wrote was the author. The liter?
anthropologists convey their material and ary profession (with its rigid rules of use,
establish their authority ? has become a genre and composition) sanctioned, protected
thriving cottage industry in certain quarters and surveilled by state institutions, was charg?
of anthropology in recent years [1]. These ed with the production of language.
attempts have varied from unpretentious The writer, Barthes tells us, emerged during
and highly successful attempts at integrating the time of the Revolution, when language
other voices, photographs and a certain was seized, and used for political ends. Barthes
ethnographic humility into a basically stan? should have added that the Academie Fran
dard ethnographic presentation [2], to more caise has always used language for political
ambitious programmatic calls, heavily influ? ends [4]. Perhaps Barthes means by polit?
enced by deconstructive practices, which seek ical the self-conscious use of language in ex?
(or so they proclaim) truly radical recasting plicitly instrumental ways. The replacement
of ethnographic writing. We already have of the "author" by the "writer" was a slow
several influential overviews of this material process. Although the revolutionaries began
[3]. What I intend to do here, is to begin the process of turning language into a trans?
from the given (although hardly widely parent instrumentality, revolutionary orators
recognized) that ethnographic texts are in? still employed the Classical rhetorical embel?
deed texts, and to question some of the lishments and perorations. Only with the
critical claims made for this insight. emergence of modern intellectuals, during
Barthes asks: Who speaks? Who writes? the Zola affair, did this historical transition
To answer this question adequately we would finally establish a new place and a new use
need a sociology of language. Barthes gives us for language. "The writer ... is a transitive
only a semiotics. But semiotics provides us man, he posits a goal (to give evidence, to
with commonplaces and commonplaces, as explain, to instruct), of which language is
any student of rhetoric who has read her only a means; for him language supports a
Cicerco knows, are necessary to advance praxis, it does not constitute one. Thus lan?
any and all arguments. During the sixteenth guage is restored to the nature of an instru?
ment .of communication, a vehicle of
Paul Rabinow is Professor of Anthropology at the University
of California, Berkeley.
"thought." Even if the writer pays some at

03044092/85/$03.30 ? 1985 Elsevier Science Publishers B.V.

This content downloaded from
129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
2

tention to style, this concern is never onto Barthes makes a similar point somewhat
logical [5]. This change in language was one differently. The wonderful paradox of the
element in the transformations occurring author's narcissistic labor is that, unlike the
throughout society. It is unlikely that modern writer, it ends up interrogating the world:
capitalism, democracy or social science would "... it is precisely when the author's work
have been possible without such a transforma? becomes its own end that it regains a mediat?
tion. ing character: the author conceives of liter?
Not all writers are the same. Writers, Barthes ature as an end, the world restores it to him
adds, produce recognizable dialects but rarely as a means: and it is in this perpetual incon
styles. The reason for this homogeneity is clusiveness that the author rediscovers the
that "what defines the writer is the fact world, an alien world moreover, since liter?
that his project of communication is naive: ature represents it as a question ? never,
he does not admit that his message is reflex? finally, as an answer" [8]. Although the
ive, that it closes over itself, and that we can writer-function and the author-function are
read in it, diacritically, anything else but what two poles, they are not symmetrical ones for
he means. . ." [6]. Rhetoric, interpretation, Barthes. Instrumentality as blind if often ef?
mediations: that honnete homme, the writer, fective; craft seeking self-enclosure, happily
will have none of these impedimenta. More stumbles.
precisely, the writer's desire is to have them We live, Roland Barthes told us in 1960,
disappear, but this obliteration occurs only in an age where we are condemned to be
on the level of desire; for have them all she neither authors, dwelling unselfconsciously
surely will. Swinging too far on the instru? in language, nor writers, saying "at once and
mental side of language's see-saw, the writer on every occasion what [one] thinks. . . To?
assures self-obfuscation but gains, at times, day, each member of the intelligentsia harbors
a certain efficacy and strength. both roles in himself, one or the other of
The author is about nothing if not style which he 'retracts' more or less well: authors
and self-reflexivity. Language is its own end. occasionally have the impulses, the impa?
"To write is an intransitive verb. Thus the tiences of writers; writers sometimes gain
author existentially forbids himself two kinds access to the theater of language. We want
of language. . . first, doctrine, since he con? to write something, and at the same time we
verts despite himself, by his very project, write (intransitively)" [9]. We anthropol?
every explanation into a spectacle: he is al? ogists should now know that we live in lan?
ways an inductor of ambiguity; second, guage. Yet in their role as writers, most
evidence since he has consigned himself to anthropologists continue to act mainly as
language, the author cannot have a naive writers, to treat language as a transparent
consciousness, cannot 'work up' a protest tool, and to attack as subjective those who
without his message finally bearing much seek to incorporate their authorial function.
more on the working-up than on the protest: The philistine temptation remains alive: a
by identifying with language, the author loses phobia of navel gazing, however, does not
all claim to truth, for language is precisely make language into a neutral medium.
that structure whose very goal (at least histor? Among those concerned with the pro?
ically, since the Sophists), once it is no longer duction of ethnographic texts we have a
rigorously transitive, is to neutralize the true spectrum of writers and authors. To men?
and the false" [7]. But as Foucault has shown, tion two of the best for purposes of com?
the myriad forms of embededness of lan? parison: Kevin Dwyer seeks to overcome
guage in fields of power and truth will not go language and become the writer; James
away through mere craft. Clifford, through writing, hopes to capture

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 3

the world. Dwyer, in his Moroccan Dialogues face up to the fact that we can never avoid
[10], is didactic to a fault, writing as white the author function. There are no transparent
and flat a prose as can be imagined. He ex? ethnographies. In different ways, Barthes,
horts us to vigilance against the temptations Foucault and Bourdieu have made it impos?
and sins of textuality. Yet, his monograph is, sible to write naively about the social world
quite unwittingly it seems, perhaps the and get away with it. Whether cast in terms
most radical post-structuralist text we have to of writers and authors, of discourses and
date. By attempting to achieve absolute trans? power, or the strategic politics of cultural
parency, eliminating all irony, Dwyer dem? production, the relation of texts and the
onstrates Barthes' claim about the futility of world has been put in question. If the work
such gestures. A melange of Beckett-like dia? of the recent past has put these questions on
logues, supplemented by the kind of home our agenda much remains to be discussed,
photos Pierre Bourdieu has analysed for us so thought through and written about.
well [11], replete with stylized 'natural'
handwritten captions under them (exactly AUTHORS: ON ETHNOGRAPHIC AUTHORITY
reproduced in a recent Ralph Lauren adver?
tising campaign for Safari wear), an academic If ethnographic writing still awaits its crit?
index, a serious preface functioning, Bourdieu ical theorist (although Steven Webster is
would remind us, to establish the writer's cre? trying hard to fill this role) it now has a
dentials, could easily have appeared with a strong candidate for its critical historian,
slightly different contextualization, as a spe? James Clifford. Following in the steps of
cial issue of Semiotexte. The zero point of George Stocking, although gazing down
textuality and style become not transparent French avenues and not Anglo-Saxon ones,
but radically deconstructive. and with an ironic meta-voice rather than a
In contrast, the works of James Clifford reconstitutive one (e.g., Stocking's important
signal themselves, emphatically and inces? rehabilitation of a more fully portrayed Boas),
santly, as all author: always nuanced, lan? Clifford has been engaged in what is in many
guage honed and presented for our delecta? ways a sophisticated "second-generation"
tion, almost every declarative statement methodological enterprise. Not an anthropol?
ironised or immediately qualified. Yet, flair ogist himself, Clifford is making a claim to
and talent aside, Clifford has produced a occupy the role of ex-officio scribe of our
book and a series of articles whose form scribblings. While Clifford Geertz, among
is much more academic ? and thereby others, may pause between monographs to
writerly - than Dwyer's. Ignoring Barthes muse on texts, narrative, description and
admonitions about authors who "work-up interpretation, Clifford takes as his natives,
protests," Clifford proclaims the political his colonial and post-colonial informants,
credentials of his texts as well. Not content anthropologists past and present. We are
to be an author, Clifford employs a form being observed and inscribed. What I want
which will enable him to be, at the same to do in this section is to return the gaze, to
time, a writer. His see-saw goes up and down look back at this collector of ethnographical
with amazing velocity. tropes, sitting at his cafe table, and, follow?
I will argue that the proposed relations ing his own prescriptions, examine his textual
between style and the political message in productions.
both Dwyer's and Clifford's works are about Clifford's substantive historical work has
as organic as Huysman's garden. Put another concerned the cultural "place" carved out for
way, those of us who produce texts must French anthropology between the older

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
4

figures of the traveler and missionary onanthropological

the authority of the author on
one hand, and the avant-garde artist and the
ex? "I was there" credential is accompanied
by a suppression of the inevitable dialogic
perimental writer during the interwar period
in France on the other. He has described construction
the of anthropological knowledge.
Clifford has shown us how these two ele?
shifting, uneasy and ambiguous jostling which
obtained between science, belief, politics ments
and have functioned together. As he says
about Clifford Geertz's cockfight paper:
writing in Maurice Leenhardt's transition
from missionary to anthropologist: how "The
the research process is separated from th
missionary turned fieldworker and then texts
in? it generates and from the Active world
stitutionally came to succeed Mauss in French are made to call up. The actuality of
they
anthropology [12]. Constantly juxtaposed discursive
to situations and individual inter?
Leenhardt, the man of institutions, howeverlocutors is filtered out. . . . .The dialogical,
ambivalently he related to them, is Michelsituational aspects of ethnographic interpre?
tation tend to be banished from the fina
Leiris, his experimental writing and subjective
obsessions floating at the edge of Parisianrepresentative text. Not entirely banished
institutions if always at the center of cheof course; there exist approved topoi for the
world of the arts. Leiris' shadow, or is itportrayal
his of the research process." ["O
Ethnographic
aura, is always present, if always slightly off Authority", Representations
center, of Clifford's frame. #2, Spring 1983, page 132]. Geertz's "appeal
Clifford has sought to demonstrate the ing fable" is paradigmatic: the anthropol?
devices through which anthropologists ogisthave establishes his unique authority b
showing that he was there and then disap
textually constructed a legitimate authorial
role for themselves. He shows us how the pears
con? from the text. The reason the anthro?
pologist
stitution of a specific authorial legitimacy was can do this leads us to the scientific
underpinnings of modern anthropology, or
forged out of a number of separate elements.
The fabrication of a scientific voice was,more
not accurately, the scientific rhetoric appro
surprisingly, one of the conditions for the by modern anthropology.
priated
Referring to Malinowski, Clifford argue
entry of the discipline into the academy. This
that: "Argonauts is archetypical of the gen?
required a number of steps. First, the anthro?
eration of ethnographies that successfully
pologist as fieldworker had to be separated
from and set in opposition to the arm chair
established participant-observation's scientif
validity"
anthropologist. This was done primarily by [13]. The complementary differ
stressing the experimental dimension of ences
the between Clifford's approach and Stock?
ing's are nicely juxtaposed in this quote
anthropologist's work: "The predominant
mode of modern fieldwork authority isStocking
sig? has shown [14] that the practice
naled: You are there, because I was there."
of fieldwork in British anthropology was no
["On Ethnographic Authority," Representa?invented by Malinowski; Clifford has shown
tion, #2, Spring 1983, page 118.] Cliffordthat what Malinowski actually achieved wa
has shown us several of the ways in whichfieldwork's exemplary textual enshrinement
this was achieved; we will never again be as a central element of the "ethnography" as
able
to read these texts of the Golden Age a new
of genre.
The third element in this story is the rise o
anthropology without seeing these rhetorical
devices at work. a theory of "culture," one assuming tha
there is a unified Dogon or Balinese cultur
The second element of this strategy is the
textual suppression of the first. From Malithere and that representing it is relativel
out
no wski on, a double step of foundingunproblematic.
the Clifford Geertz, once again

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 5

seems paradigmatic both in bringing this seeking to include in a self-conscious fashion,

tradition to its culmination and pushing it in anthropological texts themselves, more
almost to its breaking point ? and then stop? of the elements mentioned above. Thus, for
ping. Certainly a core assumption of the example, Clifford's claim of the necessary
Golden Age of anthropology is that there are presence of a dialogic element in the "pro?
discrete units out there (called either cultures duction" of anthropological knowledge. All
or societies) and that these are knowable and anthropologists work with informants. An?
representable. If this is so, then the gigue of thropological understanding arises out of the
erasure described above is consistent and give and take between them. How to bring
harmless. The erasure is what constitutes this dialogic dimension into the anthropol?
the traditional framing of ethnographies. ogical texts is an important problem. Re?
Pointing it out is a service but puts only cently a number of anthropologists have
peripheral matters in question. attempted to do just that.
The literary devices involved, "free indirect Before evaluating Clifford's reading of
style," have been well analysed by Dan Sper? these attempts I think that it would be help?
ber and need not be rehearsed here [15]. ful to introduce some further distinctions.
Sperber criticizes the use of these devices be? There seem to be at least four interconnected
cause they are not leading us in the direction but different dimensions at issue: (1) aes?
of a universal science. For many of us this is thetic (or formal): what devices could be used
not a criticism. Although interesting, once to introduce a more inventive or imaginative
recognized, there is nothing inherently crisis dimension into anthropological books and
provoking in the insight that anthropologists arguments? (2) epistemological: would bring?
unwittingly employ well described literary ing more voices into the text (however rep?
conventions. Thus, for example, Richard resented) produce a truer anthropology?;
Rorty recognizes and applauds a similiar (3) ethical: is it incumbent on anthropol?
insight about the work of Geertz [16]. For ogists to introduce a dialogic element into
Rorty, there are no interesting differences the text? Do we want to constitute ourselves
between science and fiction and he urges as the kind of subjects who are in dialogue
us to stop worrying about the problem. with other equal subjects? (4) political: will
Clifford and Geertz continue to worry; the field of either world or local power rela?
Geertz holds on to the science tag, Clifford, tions be effectively changed if we write dif?
following Hay den White [17], seems to ferent texts?
bracket the question of truth and attempts to Clifford proclaims the urgency of change
remain on the purely textual level. This on the epistemological and political levels.
move turns out to have its problems. We He makes his most sustained and convincing
will return to this point shortly. case for the epistemological dimension in
Advances have been made in our aware? his book on Leenhardt where he argues at
ness of the fictional, in the sense of the length that Leenhardt's long familiarity with
"made," "fabricated," quality of anthropol? New Caledonia, his deeply existentially in?
ogical writing and in the integration of its volved questioning of the Melanesian's an?
characteristic modes of production. The swers to the fundamental problems of life,
self-consciousness of style, rhetoric and his insistence that their answers must be
dialectic in the production of anthropol? understood as equally valid attempts to
ogical texts should lead us to a finer aware? comprehend and hence must be taken seri?
ness of other ways to write. This would in? ously in their truth claims, produced, among
clude a range of formal experimentation other things, a deeper understanding of the

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
6

culture than did a more complacent anthro? rarely find themselves in this type of situa?
pology. By putting himself in question the tion. As Clifford has shown, texts produced
anthropologist learned. But this questioning for specific academic audiences have their
was or always should have been more than particular coherence built in afterwards.
epistemological. Anthropology's truths arose When, in a few limited cases, as with Leen
in an eminently political situation. Although hardt's trained transcripteurs, there is a joint
Leenhardt rigorously put truth and self inproductions of texts, interesting results may
question he never rejected colonialism and the well be produced but they can hardly be
French mission civiliatrice. Clifford inter? synechdocically generalized. These Melan
prets this as an irony. esians were converted and traind by Leen
In his more recent work, Clifford seems to hardt. This does not negate their productions,
want to posit a link between representation but only situates them. They can hardly be
and politics. At times, he seems to equate taken as exemplary.
realism as a mode of representation with a In, "On Ethnographic Authority," Clif?
political affirmation of colonialism. The claim ford makes the claim that: "The present
for an elective affinity between realism and predicament [of representation] is linked to
colonial politics is further developed in Clif? the breakup and redistribution of colonial
ford's article on Marcel Griaule (typically, it power in the decades after 1950 and to the
is both asserted and denied in the "On Eth? echoes of that process in the radical cultural
nographic Authority" article). Griaule is theories of the 1960s and 1970s. . . ." While
juxtaposed with and opposed to Michel I don't think this claim is false, as it stands,
Leiris, one of the earliest and most con? too many mediating levels are missing. It
sistently vocal critics of colonialism and one covers over other changes, less significant on
of the most avant-garde practioners of writ? a world scale, but perhaps more telling in
ing: "Griaule's energetic confidence in cul? effecting the discursive changes in the anthro?
tural representation could not be farther frompological profession in the last thirty years.
Leiris' tortured, lucid uncertainty" [ 18]. But, One of the main jobs to be done is to fill in
again, as Leenhardt's case shows, politics, the sociological mediations through which
formal experimentation and epistemology the macro level events of decolonisation
can be independent variables. While Leen resulted in different modes of writing in
hardt refused to condemn colonialism, he ethnography. The vast explosion in the num?
produced innovative "multi-authored" texts; ber of anthropologists, new standards for
co-authored in a complex sense by fifteen publication, tenure, etc., would have to be
Melanesian transcripteurs. Conversely, many examined. Such an inquiry might well prove
of Leiris' "ethnographic" or "political"highly revealing. For the moment, let me
texts are themselves rather "realist." assert my ungrounded opinion that if there
This last point indicates an important was a loose connection between realism and
problem. Making textual production the colonialism in the past, no such ready equa?
guiding metaphor of the anthropological tion of form and politics is available to us
encounter risks serious distortion. The over? today. The equation of avant-garde exper?
whelming majority of fieldwork encounters iments in form with progressive politics
are not about the mutual production of remains questionable.
texts. As Marilyn Strathern has pointed out To return to Clifford's argument, having
[personal communication], the production of cast realist and interpretive modes of eth?
texts is based on common understandings nographic authority in basically negative
and a shared tradition. Anthropologists terms (at the very least their time has passed,

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 7

their "interest" is limited), Clifford moves on that its textual form should be that of a
to a much more positive portrayal (although literal dialogue" [20]. Exactly what the
highly elusive) of the dialogical and poly? formal characteristics of the genre are remain
phonic modes of authority. He says: "dia? unclear.
logic and constructivist paradigms tend to Perhaps they can not be specified uniquely
disperse or share out ethnographic authority, in terms of form. Perhaps they require a dia?
while narratives of initiation confirm the re? logue intent or message available to the acute
searcher's special competence. Paradigms of critic. However, if the supposed aim and ad?
experience and interpretation are yielding vantage of ethnographic dialogue is to deepen
to paradigms of discourse, of dialogue and intercultural understanding and to produce a
polyphony" [ 19]. The claim that such modes medium of exchange, Clifford owes us more
are triumphing or that older modes are on how the laconic refusals to engage in such
"yielding" is empirically dubious. That a exchange on the part of Dwyer's informant
small group is experimenting with them is demonstrates how "interlocutors actively
true and it would be worth while trying to negotiate a shared reality." Dwyer's record
locate them sociologically. However, in my of his exchanges with the Berber farmer
experience, the older canon is being rigor? demonstrate how thin and unreciprocal a rela?
ously defended in most anthropology depart? tionship was established between them.
ments. Defended, more often than not, Dwyer is being consistent and courageous in
through reproduction of older paradigms presenting this lack of mutual understanding.
strongly reinforced by a wide variety of As Marilyn Strathern has suggested this is
sanctions. I will argue below that since Clif? probably quite typical. The interlocation, the
ford seeks to remain on the textual level Faqir, is quoted as saying that he was never
he can not incorporate the kind of social interested in a single question posed to him
structural and political variables which dif? by Dwyer. I think Dwyer was making a polit?
ferentiate between "yielding" and "triumph? ical and ethical point by showing this. The
ing." Faqir's tolerance certainly demonstrates the
Nonetheless, the more interesting (although strength of Moroccan canons of hospitality
unsupported) claim is that a dialogic text rather than the inherent mutual involvement
(for the moment assuming there are such of ethnographer and informant in the produc?
texts and that we can identify them) inher? tion of ethnographic texts. Dwyer's Berber
ently presents a dispersal of authority and was clearly not interested in this project;
thereby is preferable because (1) it is truer, calling this "negative dialogue" [21] hardly
(2) politically superior. Clifford, at first helps.
blush, seems to be using the term dialogic But lest we get too involved in these is?
in a literal sense, a text which presents two sues, Clifford quickly moves on to remind
subjects in discursive exchange. Kevin Dwyer's us: "But if interpretive authority is based on
"rather literal record" of exchanges with a the exclusion of dialogue, the reverse is also
Moroccan farmer is the first example cited true: a purely dialogical authority represses
of a "dialogic" text. Dwyer's condemnation the inescapable fact of textualization" [22].
of textuality is not commented on by Clifford. The opposition of interpretive and dialogic
Rather, he lauds Dwyer's "considerable frankly escapes me for a variety of reasons
sophistication," although a page later he not the least of which is that, several pages
adds "to say that an ethnography is com? later, Clifford praises Gadamer, certainly
posed of discourses and that its different com? the most renowed representative of her
ponents are dialogically related, is not to say meneutics whose texts certainly contain no

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
8

direct dialogues, for aspiring to "radical retro" of sentimentality for the authentic
dialogism." The term seems to be little (interpreted by some as neo-fascist), so
more than a badge of approbation. Finally common in France in the 1970s, should be
Clifford asserts that dialogic texts are, after guarded against.
all, texts, "representations" of dialogues. Clifford immediately takes this option
Clifford plays each of these textual modes away by quite correctly reminding us that,
in two registers: there is unquestionably a "quotations are always staged by the quoter.. .
positively evaluated progression from mode a more radical polyphony would only displace
of authority to mode of authority; but then ethnographic authority, still confirming, the
the film is run backwards, the outstretched final, virtuoso orchestration by a single
hand drawn back. Clifford is uneasy here, author of all the discourses in his or her text"
offering suggestions, proposing a progression [Ibid., page 139]. New forms of writing, new
of modes, quoting with approval the "right" textual experiments would open new possi?
sources, then taking them away. Read gener? bilities ? but guarantee none. Having opened
ously, this indicates an openness and a salu his essay with promises of new textual free?
tory tentativeness; read more skeptically, a doms, we end having returned, enriched, to
postured eclecticism privileging the critic as no promised land. Realism, perhaps about to
cultural hero. Today, in a certain academic make a comeback, is no more or less liberat?
culture in which experimenting with rep? ing than polyphony. "The modes of authority
resentations of dialogue is valued, it would be reviewed in this essay - experiential, inter?
important to attempt to specify sociologically pretive, dialogical, polyphonic - are available
where and by whom these discursive forms to all writers of ethnographic texts, Western
are carried. and non-Western. None is obsolete, none is
If dialogic texts fall prey to the evils of pure: there is room for invention within each
totalizing ethnographic adjustment then per? paradigm" [Ibid., page 142]. So, despite the
haps even more radical polyphonic ones "recent questioning of colonial modes of rep?
might not: "Ethnography is invaded by resentation" or the closing peroration, "polit?
heteroglossia. If accorded an autonomous ical and epistemological assumptions are built
textual space, transcribed at sufficient length, into this [realistic] and other styles, assump?
indigenous statements make sense on terms tions the ethnographic writer can no longer
different than those of the arranging ethno? afford to ignore" [Ibid., pages 141/2] we are
grapher . . . This suggests an alternative free to use each of these modes for, in and of
textual strategy, a Utopia of plural author? themselves, they offer us no guarantees, con?
ship that accords to collaborators, not merely tain no secret powers, afford no textual pass?
the status of independent enunciators, but words to truth or politics.
that of writers" [23]. The authenticity trap Clifford is uneasy about this. Not content
lurks here: why is it that a Melanesian voice to advocate aesthetic experimentation for its
is superior? Only a view of culture as an own sake, or an ethical imperative to place
integrated (epistemologically and politically) oneself in a dialogic situation, he seems to
whole, unambiguously representable by a seek a ground, a further reason to do these
speaking subject would guarantee any higher things. It seems the ground must in some way
truth value or innate political superiority to be political, be on the right side ? but he
such a formal attempt. After all, the inser? can't find such a ground. He moves on. Tem?
tion of extended native texts might well porarily enthusiastic for dialogic, rhetorically
function as a false "realist" or "interpretive" adroit, on the verge of swaying us, Clifford
authentication; the potential for a "mode immediately qualifies his praise. He leads us

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 9

on to the wilder shores of polyphony: se? rhetorical forms of a literary genre" [24].
duced - for a paragraph - until we see that, Ethnography is also a set of social and his?
polyphonia in a text, is, unfortunately, once torical practices located in institutions.
again, only writing. The modernist impulse Without returning to the transparency of
towards seeking a ground in the self-referen language the task ahead would have to center
tiality of form and the objectivity of condi? on attempts to connect genre and tropic
tions, once again, founders in the friable soil clusters with the social world. The tension
of the world. between text and world can only be rethe
I think we would all agree that we should matized because it was never absent.
struggle self-consciously to avoid portraying If we begin from the proposition that there
an abstract, ahistorical other. . .and that the are no a priori homologies between discourse
development of an ethnographic science can and power we open up a much greater wilder?
not ultimately be understood in isolation ness to explore. What have been the relations
from more general political-epistemological in specific instances, at particular historical
debates about writing and representation of and institutional conjunctures, of discourses
otherness. But I do not think it follows that and power? There would seem to be a number
"difference is an event of inventive syn? of different levels at which discourse and rela?
cretism." Literary critics and anthropologists tions of power intersect.
did not invent the world, or cultural differ? To schematize: there are micro-relations of
ence, even if, for a certain segment of our power and macro-relations and their intercon?
society, they have come to interpret it. I nections. On the level of micro-relations we
agree with Stanley Diamond that anthro? can distinguish two different locations for the
pology arose out of political and cultural generation and production of discourse and
crisis. If indeed it has "become necessary power in ethnography: relations in the field
to conceive of a world of generalized eth? and relations within the discipline or "inter?
nography," this indicates a crisis of mammoth pretive community" [25]. We now know a
proportions not a solution or a salutory certain amount about relations of power
opening. between the anthropologist and the people
with whom she works. For a variety of histor?
ical, political and disciplinary reasons the level
DISCOURSE AND POWER of micro-relations between informants and
ethnographers - although absent from the
The recent work on the tropes and rhet? earlier Golden Age ethnographies once the
oric of ethnographic writing is unquestion? "I was there" authority was established - has
ably of value. We are now well on our way to exploded in the last six or seven years and
having a Kenneth Burkean Rhetoric of there is now a new sub-genre of these books.
Motives for ethnography. However, we have Although hardly universally accepted as valid
also suggested some of the dangers and limi? within anthropology they have benefited
tations inherent in this development. Steven from the fragmentation of the discipline in
Webster makes the valid point that Clifford, the last decade, from some of the boundary
and Marcus and Cushman, stop just at the crossing within the university system and
point where the ethnographic genre itself from political movements which have de?
would be put in question. "Such respect for manded some accountability by the anthro?
the strictures of writing convention is even pologist to the peoples she describes. The
more paradoxical than the discovery that a creation of a visible (and hence problem
social science genre operates through the atized) ethnographic subject - ethically,

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
10

politically, epistemologically and aesthet? rich and linguistically diverse survival tech?
ically - is at least on the agenda. More ex? niques created by the inmates to cope with
perimentation and more analysis of why the endless brutalization of their jailors, in
this has happened (and the reasons are various) The Mouroir [27]. Many more examples
would probably be worth while. Contextual could be cited of formal experimentation
pressures ranging from the impact of the growing out of political sensitivity or struggle.
"new journalism," the federal government's The work of Pierre Bourdieu can be help?
policy guidelines on treatment of human sub? ful in this regard [28]. Bourdieu has taught
jects, to explicit formulation of conditions us to ask about any author: in what field of
under which anthropologists could enter into power, from what position in that field,
a community by a large variety of peoples using what cultural strategies from among
have all played their part. those historically available, does an author
While actually less extensive than one operate? Posing these simple question to
might have hoped, there is some analysis of ourselves, we would become aware that
the macro-relations of ethnographic discourse current proclamations of anti-colonialism, for
and historical macro-conditions. Ranging example, while admirable, must be under?
from the almost purely discursive pole of stood as part of a more contemporary strat?
Edward Said's Orientalism, to Eric Wolfs egy as well. We are not writing in the late
materialist Europe and the People without 1950s, after all, and our audiences are neither
History and passing through the more medi? colonial officers nor those working under
ated attempt of Talal Asad's Anthropology the aegis of a colonial power. The political
and the Colonial Encounter or Dell Hymes' field is more immediate and mundane: the
Reinventing Anthropology, we are devel? academy in the 1980s. [Clifford: "I write
oping a more complex understanding of with '60s idealism, plus '80s hesitation (or
how anthropology and other descriptive put more positively, ironic historicism").]
sciences of the Other arose and flourished Hence, situating the crisis of representation
in the world. Michael Taussig's The Devil within the context of the rupture within
and Commodity Fetishism in Latin America colonialism and decolonialism is both true
and Kevin Dwyer's Moroccan Dialogues and false. It is true to the extent that anthro?
attempt to bring world conditions and local pology is certainly parasitic on the course
conditions together, although it is fair to say of larger world events. It is false in that,
that the immense complexity of the task at least without a large number of important
has yet to find either a political or textual mediations, one can not convincingly assert
form adequate to it. It may well be that that new ethnographic writing emerged
formerly subjected peoples will lead the directly out of decolonisation or the Viet?
way. Knowing that many cultural issues nam war. Rather, it makes sense to begin
are eminently political and that certain with the institutional setting in which it
political issues are eminently cultural, writers emerged: the American Academy of the
as different as C.L.R. James in his analysis late 1970s and early 1980s, for reasons
of West Indian cricket, British education, still to be pinpointed.
ethical constitution, political context and If there has been an explosion of literature,
writing offers one brilliant example [26] debate and action about ethnography and
while Breyten Breytenbach, a South African ethnographized micro-relations, there has
Afrikaner poet, jailed for nine years, has been almost a total silence about power/
given us a very innovative novel/ethnography knowledge relations within the disciplines
of South African jail life, the fantastically in recent years. The micro-relations among

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 11

the interpretive community have simply refer to "corridor talk." For many years,
not been explored. Are they important? anthropologists discussed fieldwork exper?
There is a strange complicity of silence at iences among themselves. In fact, the gossip
work here. The silence indicates that they concerning a particular anthropologist's field
probably are important. One wonders what experiences, her control of the language, etc.
is being covered up by this silence. The were important components of that person's
repeated refrain that looking at the construc? reputation. But such matters were not written
tion of power relations within the academic about "seriously." They remained in the
world is "navel gazing" while the micro corridors and faculty clubs. But what can
relations of, say, an East African missionary not be publicly discussed can be neither
community a hundred years ago is not, is analysed nor refuted. Those domains which
highly suspect. It is even more suspect when can not be analysed or rebutted, and yet are
such charges come from the deconstructive directly central to hierarchy and survival,
camp. Edward Said captures this well: "It is should not be regarded as innocent or irrel?
in the Age of Ronald Reagan that the poli? evant. Until very recently it was simply bad
tics of interpretation and the politics of taste to discuss the conditions which gave
culture are enacted. . . an aspect of the modern anthropology its own selfdefinition
present cultural moment, in which the social - fieldwork. Corridor talk became discourse
and historical setting of critical activity is a and we learned a good deal, none of it about
totality felt to be benign (free, apolitical, navels.
serious), uncharacterizable as a whole. . .and My wager is that looking at the conditions
somehow outside history" [29]. We know under which people are hired, tenured, pub?
from the work of Pierre Bourdieu that one lished, granted and feted within the American
of the most common tactics of an elite group academy in Said's Reagan's America would
is to refuse to discuss - to label as vulgar or repay our efforts. There is no doubt that one
uninteresting - issues that are uncomfortable of the major developments in the American
for them. We also know from Michel Foucault Academy in the last ten years is the explo?
that truth and its production is very much a sion of textual analysis, loosely grouped
thing of this world. The insistent refrain that under the banner of "deconstruction."
the academy is trivial and somehow outside How has this trend differed socially and
of the real world when scientific production politically from the other major movement of
is unquestionably central to the growth of the last decade, feminism? Why is it that
late capitalism leads me to underline Said's feminism and deconstruction have appeared
questioning of the rise of textualism in together to an extent in certain literary
recent years. Certainly some dimensions of departments [30] but almost not at all in
the nitty-gritty of academic production are anthropology? How are careers made now?
trivial but obviously all of them can't be. How are careers destroyed now? What are
Just as one could write an institutional and the boundaries of good and bad taste. Let's
micro-political history of the production of turn this corridor talk into discourse. What?
materialist discourse in American anthropol? ever else we know - whatever the relations
ogy, (which institutions were colonised and can be shown to be between late capitalism
how), it would be important to do the same and deconstruction or Marxist criticism - we
for less unified doctrines which have, none? certainly know that the material conditions
theless, gained important institutional strong? under which the textual movement has
holds in the recent decade. flourished must include the university, its
Another way of posing this problem is to micro-politics, its trends. How are these

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 12

micro-discourses made effective? and con? nial officials as well as social reformers (within
trolled? Why is it polite to say "yield" and one branch of French socialism) all concerned
snide to say "trendy"? Tropes are always with urban planning in the 1920s. By "stud?
part of a political project. Let us not forget ying up" I find myself in a more comfortable
that it is still impolite in many places to talk position than I would be were I "giving voice"
about institutional racism, sexism, and even to dominated or marginalized groups, or play?
colonialism and anthropology. Who enforces ing the role of universal intellectual, spokesman
these civilities and why? of the Truth. I have also chosen a group who,
Is it vulgar to ask: would longer, dispersive while unquestionably holding positions of
texts yield tenure? Is it bad taste to remark power and privilege, were nonetheless highly
that, for example, advocates of experimental critical of racism in the colonies and class
writing themselves produce texts which are oppression at home. But these men are no
resolutely academic and traditional in form. heroes; better than some, worse than others,
But these questions are posed in the corri? they invented and carried out some of the
dors all the time. They are real. This is some? initial programs of the welfare state. They
thing we all face in one form or another. seem to afford me an anthropological differ?
These questions are defined as small and ence; separate enough so as to prevent easy
petty; but those are the dimensions of power identification yet close enough to afford a
relations to which Nietzsche rightly exhorted charitable but critical understanding.
us to be scrupulously attentive. We know that Many of these reformers, in particular,
power relations on that level exist, affect us, Hubert Lyautey, governor general of Moroc?
influence our choice of themes, forms, con? co, were skillful and sophisticated writers in
tents, audiences. Less glorious than cham? a number of genres. Lyautey was extremely
pioning other causes, if more immediately lucid and adroit in his use of political analysis,
constraining, we owe them more attention. commissioned biographies, social science re?
Where do we go from here? I have no gen? ports, published letters, retrieval of Moroccan
eral prescriptions or proscriptions to offer. I archives, reconstitution of culture in general.
have emerged from this immersion in tex This raises a number of genealogical questions
tuality feeling inchoately anthropological. about the place of forms of social description
Anthropology for me has meant pursuing and political strategies. This poses, I hope, a
criticism of the barbarism of civilisation, means for questioning the place of writing
an openness to otherness, and a commit? in concrete historical and social terms. If in
ment to and great suspicion of Reason. These the last five years we have seen important
have been foregrounded in epistemological work showing us specific ways beyond the
and ethical issues for me. Given the con? transparency of language, I think it is now
straints of the contemporary historical time to take those advances and move back to
situation, what kind of subjects do we want the world.
to be? What kind of relations do we want to
NOTES
have with other subjects? How much can they
be forged? How? How does writing connect
1. I would like to thank James Clifford, Stephen Foster,
with these projects? What are the relations of
James Faubion, Michael Rogin, Marilyn Strathern and
ethics and politics at different conjunctures? the participants at the Santa Fe conference on "The
What is the place of reason in these activities? Making of Ethnographic Texts" for their constructive
responses to early versions of this paper.
Being temperamentally more comfortable in
2. A good example is: June Nash, We Eat the Mines and
an oppositional stance, I have chosen to study the Mines Eat Us (New York: Columbia University
a group of elite French administrators, colo Press, 1979).

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms
 13

3. Marcus and Cushman, "Ethnographies as Texts", An? Social Science as Moral Inquiry (New York: Columbia
nual Review of Anthropology, vol. 11 (1982); Clifford, University, 1983).
"On Ethnographic Authority", Representations, vol. 2 17. H. White, "The Value of Narrativity in the Representa?
(1983); S. Webster, "Dialogue and Fiction in Ethnog? tion of Reality," in WJ.T. Mitchell (ed.), On Narrative
raphy", Dialectical Anthropology, vol. 7 (1982). (Chicago: University of Chicago Press, 1980).
4. G. Canguilhem, "Du social au vital," Le Normal et le 18. Clifford, "Power and Dialogue in Ethnography: Marcel
pathologique, P.U.F. (1979), p. 181. Griaule's initiation," in G. Stocking (ed.), History of
5. Barthes, "Authors and writers" in S. Sontag (ed.), Anthropology Vol. I (Madison: University of Wisconsin
A Barthes Reader (New York: Hill and Wang, 1982), Press, 1983) p. 153.
p. 189. 19. Clifford, op. cit., 1983, p. 133.
6. Ibid., pp. 189-90. 20. Ibid., p. 135.
7. Ibid., p. 188. 21. Clifford, personal communication.
8. Ibid., pp. 186-87. 22. Clifford, op. cit., 1983, p. 134.
9. Ibid., pp. 191-92. 23. Ibid., p. 140.
10. K. Dwyer, Moroccan Dialogues: Anthropology in 24. S. Webster, "Realism and Reification in the Ethno?
Question (Baltimore: Johns Hopkins, 1982). graphic Genre," unpublished ms., p. 33.
11. e.g. P. Bordieu, La Distinction, critique sociale du juge 25. S. Fish, Is There a Text in this Class? The Authority
ment (Paris: Editions de Minuit, 1979). of Interpretive Communities (Cambridge: Harvard Uni?
12. Vincent Craponzano has given us a different and illumi? versity Press, 1982).
nating interpretation of Leenhardt. Compare his intro? 26. CL.R. James, Beyond a Boundary (New York: Pan?
duction to Leenhardt's in Do kamo (Chicago: Univer? theon Books, 1983 (orig. 1963)).
sity of Chicago Press, 1979). 27. B. Breytenbach, The Mouroir (Mouroir Mirror notes of
13. Ibid., pp. 123-24. a Novel), (New York: Faber and Faber, 1984).
14. G. Stocking, "The Ethnographer's Magic: Field work in 28. e.g., Bordieu, op. cit., 1979.
British Anthropology from Tylor to Malinowski," 29. E. Said, "Opponents, Audiences, Constituencies, and
History of Anthropology, vol. 1 (1983), pp. 70-120. Community," in W.S.T. Mitchell (ed.), The Politics of
15. D. Sperber, "Ethnographic interpretive et anthropologie Interpretation (Chicago: University of Chicago Press,
theorique," in Le Savoir des anthropologues (Paris: 1983).
Hermann, 1982). 30. E. Showaiter, "Critical Cross-Pressing," Raritan, Fall
16. R. Rorty, "Method and Morality," in Haan (ed.), (1983).

Dialectical Anthropology, 10 (1985) 1-13

Elsevier Science Publishers B.V., Amsterdam

This content downloaded from

129.180.1.217 on Sun, 15 Oct 2023 20:48:13 +00:00
All use subject to https://about.jstor.org/terms