1

--
Ethnography and the

Historical Imagination

YSTIC WARRIORS GAINING GROUND IN MOZAMBIQUE WAR." The head­

M line was exotic enough to make the front page of the Chicago Tribune
one Sunday.l ,
"Call it one of the mysteries of Mrica," the report began. "In the battle­
ravaged regions of northern Mozambique, in remote straw hut villages
where the modern world has scarcely penetrated, supernatural spirits and
magic potions are suddenly winning a civil war that machine guns, mortars
and grenades could not." The account went on to describe an army of
several thousand men and boys, sporting red headbands and brandishing
spears. Named after their leader, Naparama-who is said to have been
resurrected from the dead-they display on their chests the scars of a
"vaccination" against bullets. Their terrain is the battle-scarred province
of Zambesia, where a civil war, with South Mrican support, has been
raging for some fifteen years. Now heavily armed rebels flee at the sight of
the Naparama, and government troops appear equally awed. Western
diplomats and analysts, the report recounts, "can only scratch their heads.
in amazement." The piece ends in a tone of arch authority: "Much of!
Naparama's effectiveness can be explained by the predominance of super.,j
stitious beliefs throughout Mozambique, a country where city markets~
always have stalls selling potions, amulets and monkey hands and ostrich\
feet to ward off evil spirits."
Faced with such evidence, anthropologists might be forgiven for doubt­
ing that they have made any impact at all on Western consciousness. It is
more than fifty years since Ev~§-£ritcbard .(19.37) shQwed.... i1Lthep~t
PEO!;<:l!.hat Zand~ 11laS!<:-:WjLal1__!li!air of practical reaso~2-.!hat "p.rimitive
mentality" is a fic!iqp.Q.f. the modern.Iiii.tid;moretE:a:ii1il'iy years of writing

3
 4 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination
5
in an effort to contextualize the curious. Yet we have not routed the reflex
ditional" communities are still frequently held, for instance, to rest upon
that makes "superstitious" most aptly qualify Mrican belief. No, the straw
sacred certainties; modern societies, instead, to look to history to account
;I huts and magic potions are as secure in this text as in any early nineteenth

for themselves or to assuage theirsense of alienation and loss (cr. Anderson

• century traveler's tale. There is even the whiff of a traffic in flesh (the 1983:40; Keyes, Kendall, and Hardacre n.d.). What is more, these stero1­
monkey hands; the ostrich feet). No matter that these wayward warriors n typic contra.sts.... are re~dily spatialized in the chasm between the West an
are in fact the victims of a thoroughly modern conflict, that they wear .1 the rest. TJ::y.as they mIght, the Naparama will never be more than primitiv
civilian clothes and file into combat singing Christian songs. IQ..th~p.opular h rebels, rattling their sabers, their. "cultural weapons," in the prehistory.of
i£!lilgination the)'.",\'!.!~fully fledged .~jgns of the primitive, alibis forao a~ African dawn. As Fields (1985) has noted, their "milleniary" kind are
evolutionIsm that puts them-and their fascinating forays-across,atl seldom attributed properly political motives, seldom credited with the
iireiil~~able gtllf from oursel",es.
... ry al. le g.edly consists. 11:1 the eve.n.t~
p
... These sensationalized savages, thrust across our threshold one snowy rational, pur . OS.iVe . aC. tions
in Wh.iCh histO
the Western eye frequently overlooks important similarities in the ways in
Sunday, served to focus our concerns about the place of anthropology in which societies everyWhere are made and remade. ~I!d, all too often, we
the contemporary world. For the "report" told less of the Mozambican
anthr.oPOlogists_~ay~ ex.,~~_erbat:~9. t.hi.'.~, For we have our own i. nv<:::st.m.ent in f
soldiers than of the culture that had conjured them up as its inverted self­ ereserving zones of "tradition," in stressing social repro~uction over
image. Despite the claim that meaning has lost its moorings in the late random.: change, cosmology over chaos (Asad 1973; Taussig 1987). Even
capitalist world, there was a banal predictability about this piece. It relied as we expose our ethnographic islands to the crosscurrents of history, we
on the old opposition between secular mundanity and spectral mystery, remain fainthearted. \~L.s}ill. seJ?arate lo(,':al communitie~_~J[Qm ,Zlobal
European modernism and Mrican primitivism. 2 What is more, the contrast systems, the thick description of particular cultures frof\l the.thin narrative
implied a telos, an all too familiar vision of History as an epic passage of world events.
from past to present. The rise of the West, our cosmology teUs us, is The bulletproof soldiers - remind us that lived realities defy easy dualisms, ..I
accompanied, paradoxically, by a Fall: The cost of rational advance has that worlds everywhere are complex fusions of what we like to call moder- .
been our eternal exile from the sacred garden, from its enchanted ways of nity and magicality, rationality and ritual, history and the here and now.
knowing and being. Only natural man, unreconstructed by the Midas In fact, our studies of the Southern Tswana have long proved to us that
touch of modernity, may bask in its beguiling certainties. none of these were opposed in the first place-except perhaps in the
The myth is as old as the hills. But it has had an enduring impact on colonizing imagination and in ideologies, like apartheid, that have sprung
post-Enlightenment thought in general and, in particular, on the social from it. If we allow that historical consciousness and representation may
sciences. Whether they be classical or critical, a celebration of modernity take very different forms from those of the West, people everywhere turn
or a denunciation of its iron cage, these "sciences" have, at least until out to have had history all along.
recently, shared the premise of disenchantment-of the movement of As it has become commonplace to point out, then, European colonizers
mankind from religious speculation to secular reflection, from theodicy to did not, in an act of heroism worthy of Carlyle (1842), bring Universal
theory, from culture to practical reason (Sahlins 1976a; n.d.). Anthropol­ History to people without it. Ironically, they brought histories in particular,
ogists, of course, have hardly ignored the effects on the discipline of the histories far less predictable than we have been inclined to think. ~or,
lingering legacy of evolutionism (Goody 1977; cf. Clifford 1988). None­ despite the claims of modernization theory, Marxist dependistas, or "modes
theless, it remains in our bones, so to speak, with profound implications ?f production" mod~ls, global for.ces played intQ~ocal forms and con~itions
for our notions of history and our theories of meaning. In unexpected ways, changing known structures into strange hybrids. Ow::
The mystic warriors underscored our own distrust of disenchantment, own· evidence shows that the incorporation of black South Mricans into a
our reluctance to see modernity-in stark contrast to tradition-as driving World economy did not simply erode difference or spawn rationalized,
a "harsh wedge between cosmology and history" (Anderson 1983:40). To homogeneous worlds. Money and commodities, literacy and Christendom
be sure, we have never given any analytic credence to this ideologically challenged local symbols, threatening to convert them into a universal
freighted opposition or to any of its aliases (simple:complex; ascrip­ Currency. But precisely because the cross, the book, and the coin were such
tive:achievement-driven; coUectivist:individualist; ritualist:rationalist; and saturated signs, they were variously and ingeniously redeployed to bear a
so on). For, dressed up as pseudohistory, such dualisms feed off one host of new meanings as non-Western peoples-Tswana prophets, Napa­
another, caricaturing the empirical realities they purport to reveal. "Tra- rama fighters, and others-fashioned their own visions of modernity (cf.
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 7
6 '­
Clifford 1988:5-6). Neither was (or is) this merely a feature of "transi·
exactly are we to do so? Contrary to some scholarly opinion, it is not so
tional" communities, of those marginal to bourgeois reason and the
easy to alienate ourselves from our own meaningful context, to make our
own existence strange. How do we do ethnographies of, and in, the
commodity economy.

In our essays, as we follow colonizers of different kinds from the contemporary world order? What, indeed, might be the substantive direc­
metropole to Africa and back, it becomes clear that the culture of capitalism tions of such a "neomodern" historical anthropology?
has always been shot through with its""own 'magicalities and forms of
enchantment, all of which repay analysis. Like the nineteenth-century
. evangelists who accused the London poor of strange and savage customs II
(see Chapter 10), Ma!x insisted on understanding commodities as 21:>i;.cts
or £rimitive wors~2...~.Jetish~fBeing social hieroglyphs rather than mere Both history and ethnography are concerned with societies other than the one

alIenating objects, they describe a world of densely woven power and in which we live. Whether this otherness is due to remoteness in time ... or to

remoteness in space, or even to cultural heterogeneity, is of secondary

meaning, enchanted by a "superstitious" belief in their capacity to be

importance compared to the basic similarity of perspective.... [I]n both cases

fruitful and multiply. Although these curious goods are more prevalent in we are dealing with systems of representations which differ for each member

"modern" societies, their spirit, as Marx himself recognized, infects the of the group and which, on the whole, differ from the representations of the

politics of value everywhere. If, as Chapter 5 demonstrates, we cast our investigator. The best ethnographic study will never make the reader a native.

gaze beyond the horizon where the so-called first and third worlds meet, ... All that the historian or ethnographer can do, and all that we can expect of

them, is to enlarge a specific experience to the dimensions of a more general

concepts like the commodity yield useful insights into the constitution of
one.

cultures usually regarded as noncapitalist. And so the dogma of disenchant· -Claude Levi·Slrauss (1963a:16-17)
ment is dislodged.
Save in the assertions of our own culture, in short, assertions that have
long justified the colonial impulse, there is no great gulf between "tradi­ These questions parse into two parts, two complementary motifs that start

tion" and "modernity"-or "postmodernity," for that matter. Nor, as out separately and, like a classical pas de deux, merge slowly, step by step.

others before us have said, is much to be gained from typological contrasts The first pertains to ethnography, the second to history.

between worlds of gesellschaft and gemeinschaft, or between economies As we have noted, the current status of ethnography in the human

governed by use· and exchange-value. But we are less concerned here to sciences is something of a paradox. 0Il the one hand, its authority has

reiterate this point than to make a methodological observation. If such I been, and is being, seriously challenged from both within anthropology
distinctions do not hold uR, it tollo~s that the modes of discovery and outside; on the other, it is being widely appropriated as a liberalizing '..
<l.~sOti:lffea with them-ethnog~aphyfor "traditional" communities, history method in fields other than our own-among them, cultural and legal
"~--'" . / 0 studies, sociology, social history, and political science. 3 Are these disciplines
forthe ~inodern"" world, past and present-also cannot be sharply drawn.
• - ..- - - . . . . . . . . . . . . . . •

We require ethnography to know ourselves, just as we need history to know suffering a critical lag? Or, more realistically, is a simultaneous sense of
non~Western others. For ethnography serves atlonce to make the familiar hope and despair intrinsic to ethnography? Does its relativism bequeath it
stran~e and thestrange f~.!li,~r, all th~ better,to.U1;lderstal1(:~ thepl both. It an enduring sense of its own limitation, its own irony? There does seem
is, as It were, the canon-fodder of a cntical anthropology. to be plenty of evidence for Aijmer's (1988 :424) recent claim that ethnog­
In respect of our own society, this is especially crucial. For it is arguable raphy "always has been ... linked with epistemological problems." To
that many of the concepts on which we rely to describe modern life­ wit, its founding fathers, having taken to the field to subvert Western
statistical models, rational choice and game theory, even logocentric event universalisms with non-Western particularities, now stand accused of hav­
histories, case studies, and biographical narratives-are instruments of ing served the cause of imperialism. And generations of journeyman.
what Bourdieu (1977 :97f), in a different context, calls the "synoptic anthropologists since have struggled with the contradictions of a mode of
illusion." They are our own rationalizing cosmology posing as science, our inquiry that appears, by turns, uniquely revelatory and irredeemably eth­
culture parading as historical causality. All this, as many now recognize, nocentric.
for two things simultaneously: that we regard our own world as a The ambivalence is palpable also in critiques of anthropology, which

problem, a proper site for ethnographic inquiry, and that, to make good accuse it both of fetishizing cultural difference (Asad 1973; Fabian 1983;

this intention, we develop a genuinely historicized anthropology. But how Said 1989) and-because of its relentlessly bourgeois bias-of effacing

 EthtWgraphy and the Historical Imagination 9

TH EORY, ETHNOGRAPHY, HISTORIOGRAPHY
8
In this sense, the "problem" of anthropological knowledge is only a
difference altogether (Taussig 1987). In a recent review, for example, more tangible instance of something common to all modernist epistemol­
Sangren (1988:406) acknowledges that ethnography does "to some degree, ogies, as philosophers of science have 10jg realized (Kuhn 1962; Lakatos
make an object of the 'other.''' Nonetheless, he goes on to assert, it was and Musgrave 1968; Figlio 1976). For ethnography personifies, in its
"dialogic long before the term became popular." Similar arguments, one methods .•md its models, the inescapable dialectic of fact and value. Yet
might add, are to be heard in other scholarly fields that rely on participant ~~t:' of its practitioners persist in asserting the usefulness-indeed, the
observation: Surveying the growing literature in cultural studies, for creative potential-of such "imperfect" knowledge. They tend both to
instance, Graeme Turner (1990:178) remarks that "the democratic impulse recognize the impossibility of the true and the absolute and also to suspend
and the inevitable effect of ethnographic practice in the academy contradict disbelief. Notwithstanding the realist idiom of their craft, they widely
each other." accept that-like all other forms of understanding-ethnography is histor­
But why this enduring ambivalence? Is ethnography, as many of its critics
ically contingent and culturally configured.? They have even, at times,
have implied, singularly precarious in its naive empiricism, its philosophical
found the contradiction invigorating.
unrcflectiveness, its interpretive hubris? Methodologically speaking, it does
Still, living with insecurity is more tolerable to some than to others.
have strangely anachronistic echoes, harking back to the classical credo that
Those presently concerned with the question of authority fault (unenlight­
"seeing is believing." In this it is reminiscent of the early biological
ened) ethnographers for pretending to be good, old-fashioned realists.
sciences, where clinical observation, the penetrating human gaze, was
Thus Clifford (1988:43) notes that even if our accounts "successfully
frankly celebrated (Foucault 1975; Levi-Strauss 1976:35; Pratt 1985);
dramatize the intersubjective, give-and-take of fieldwork ... they remain
recall, here, that biology was the model chosen, in the golden age of social
representations of dialogue." As if the impossibility of describing the
anthropology, for a "natural science of society" (Radcliffe-Brown 1957). encounter in all its fullness, without any mediation, condemns us to lesser
The discipline, however, never really developed an armory of objectifying truths. Likewise, Marcus (1986: 190) counterposes "realist ethnography"
4
instruments, standardizing strategies, and quantifying formulas. It has to a new "modernist" form that, because it "can never gain knowledge of
continued to be, as Evans-Pritchard (1950; 1961) insisted long ago, a the realities that statistics can," would "evoke the world without repre­
humanist art, in spite of its sometime scientific pretentions. And while it senting it."8Ifwe cannot have real representation:>.Jet us,~ave!!~­
has never been theoretically homogeneous, internal differences and disputes tationat.alV Yet surely this merely reinscribes naive realism as an (unat­
5
have seldom led to thoroughgoing revisions of its modus operandi. Indeed, tainable) ideal? Why? Why should anthropologists fret at the fact that our
the unsympathetic critic could claim that ethnography is a relic of the era accounts are refractory representations, that they cannot convey an undis­
of travel writing and exploration, of adventure and astonishment;6 that it torted sense of the "open-ended mystery" of social life as people experience
remains content to offer observations of human scale and fallibility; that it it? Why, instead, should ethnographers not give account of how such
still depends, disingenuously, on the facticity of first-hand experience. experiences are socially, culturally, and historically grounded or argue
Yet it might be argued that the greatest weakness of ethnography is also . about the character of the worlds they evoke, with the aim of fructifying
its major strength, its paradox a productive tension. For it refuses to put our own ways of seeing and being, of subverting our own sureties (cf. van
its trust in techniques that give more scientific methods their illusory
der Veer 1990:739~. Et~no&~ae~y, in any cas~~ .~9~?>L~pe~kf~
.objectivity: their commitment to standardized, a priori units of analysis, but afl~. Neither uhagltlatlvely nor empmcaIIy can It ever capture"
ffor example, or their reliance on a depersonalizing gaze that separates 'th:eTr reality. Unlikely as it may seem, this was brought home to us in a
,subject from object. To be sure, the term "participant observation"-an London School of Economics toilet in 1968. It turned out to be our first
oxymoron to believers in value-free science-connotes the inseparability foretaste ofdeconstruction; perhaps it was where postmodern anthropology
of knowledge from its knower. In anthropology, the observer is self­ aU began. On an unhinged stall door, an unknown artist-perhaps an
evidently his/her "own instrument of observation" (Levi-Strauss 1976:35). unhinged student-asked nobody in particular, "Is Raymond Firth real,
This is the whole point. Even if they wanted to, ethnographers could not, or just a figment of the Tikopean imagination?"9
pace the purifying idyll of ethnoscience, hope to remove every trace of the Ethnography, to extend the point, is not a vain attempt at literal
arbitrariness with which they read meaningful signs on a cultural landscape. translation, in which we take over the mantle of an-other's being, conceived
But it would surely be wrong to conclude that their method is especially of as somehow commensurate with our own. IUs a historic,!:Uy situated
vulnerable, more so than other efforts, to know human (or even nonhuman) mode of understanding historically situated contexts, each with its own,
worlds.
10 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 11

perhaQs, r,adic_allviiffu:rnt~kindsQf subjects and subjectivities. Qbjectsand besides, we have to situate_ the ITl within the systems of signs and relatioJ:is, ~.
<:>~jectives. Also, it has hitherto been an inescapably Western discourse. In q~~.£tan9..~eaning, that animate them. ?ur concern ultimately ~s with
it, to pick up our earlier comment, we tell of the unfamiliar-again, the the lIlterpliy of such systems-often relatIvely open systems-wIth the
paradox, the parody of doxa-to confront the limits of our own episte­ persons and events they spawn; a process that need privilege. neither the . ,
h
mology, our own visions of personhood, agency, and history. Such critiques (sovereign self nor stifling structures. Ethn.QS~~phy,wewould argtle, is an .;'
can never be full or final, of course, for they remain embedded in forms of \ exercise in diakctics rather than dialogics, although the latter is always
thought and practice not fully conscious or innocent of constraint. But
they provide one way, in our culture, of decoding those signs that disguise
pa~t of the former. In addition to talk, it entails observation of activity~nd
irlt$iac):iQ!lPoth formal and diffuse, of modes of control and constraint, of
silenfeas well as assertion and defiance. Along the way, ethnographers also
~
themselves as universal and natural, of engaging in unsettling exchanges
with those, including scholars, who live in different worlds. :.. re~d diverse sorts of texts: books, bodies, buildings, sometimes even cities
For all this, it is impossible ever to rid ourselves entirely of the ethno­ (Holston 1989; Comaroff and Comaroff 1991; see below). But they must
centrism that dogs our desire to know others, even though we vex ourselves always give texts contexts and assign values to the equations of power and
with the problem in ever more refined ways. Thus many anthropologists meaning they express. Nor are contexts just there. They, too, have to be
have been wary of ontologies that give precedence to individuals over constructed analytically in light ofour assumptions about the social world.
contexts. For these rest on manifestly Western .assumptions: among them, "The representation of larger, impersonal systems," in short, is not
that human beings can triumph over their contexts through sheer force of untenable in "the narrative space of ethnography" (Marcus 1986: 190).
that economy, culture, and society are the aggregate product of Apart from all else, such systems are implicated, whether or not we
individual action and intention. Yet, as we shall point out again below, it recognize them, in the sentences and scenes we grasp with our narrow­
has proven extremely difficult to cast the bourgeois subject out of the gauge gaze. But more than this: Ethnography surely extends beyond the
anthropological fold. Sjhe has returned in many guises, from Malinowski's range of the empirical eye; its inquisitive spirit calls upon us to ground
maximizing man to Geertz's maker oLmeaning. Ironically, sjhe appears subjective, culturally configured action in society and history-and vice
again in the writingsotthose who take ethnography to task for failing to versa-wherever the task may take us. That spirit is present, we shall see,
represent the "native's point of view." Sangren (1988:416) argues vigor­ in the work of historians who insist that the human imagination itself is
ously that this is a legacy of American cultural anthropology or, at least, perforce a "collective, social . . . phenomenon" (Le Goff 1988:5; our
the version ofit that would sever culture from society, experiencing subjects emphasis). In this sense, one can "do" ethnography in the archives, as
from the conditions that produce them. Under these conditions, culture Damton (1985:3) implies by the phrase "history in the ethnographic grain"
becomes the stuff of intersubjective fabrication: a web to be woven, a text (see p. 14). One can also "do" the anthropology of national or international
to be transcribed. And ethnography becomes "dialogical," not in Bakhtin's forces and formations: of colonialism, evangelism, liberation struggles,
thoroughly socialized sense, but in the na~~o~ers~~~of,~_.~~a<;UE'.d~c9I1- social movements, dispersed diasporas, regional "development," and the
tYx'tMtzei:texchangebetween anthropologist an~p}fQrqtant< IJ We would like. Such systems seem impersonal and unethnographic only to those who
~edtic~6fa1fthlopol6gical t-ese~r~tf to an exercise in "intersub­ would separate the "subjective" from the "objective" world, claiming the
jectivity," the communing of phenomenologically conceived actors through former for anthropology while leaving the latter to global theories (Marx­
talk alone. As Hindess (1972:24) remarks, the rendering down of social ism, world-systems, structuralism), under whose wing ethnography may
science to the terms of the experiencing subject is a product of modern find a precarious perch (e.g., Marcus 1986). In fact, systems appear
humanism, of a historically specific Western world view. JeJFeatethnog­ "impersonal," and holistic analyses stultifying, only when we exclude from
rap~y as an encounterbe!.Y'es~ a1! ob~erver and an other-COnversations them all room for human maneuver, for ambivalence and historical inde­
r
with' Ogotemmeli (Griaule 1965 or The Headman and I (Dumont 1978)­ terminacy-when we fail to acknowledge that meaning is always, to some
is to J:)1a~e anthropology ill!'o_a global, ~tlutoseQtric interview. Yet it is extent, arbitrary and diffuse, that social life everywhere rests on the
precisely this perspective that warrants the call for ethnography to be imperfect ability to reduce ambiguity and concentrate power.
"dialogical"-so that we may do justice to the role of "the native infor­ Ofcourse, like all forms ofinquiry, ethnography objectifi/~s as }t ascribes
mant," the singular subject, in the making of our texts. meaning-albeit perhaps less so than do those metho'Q6fdgi~s that eXplain
Gc::ne,ra.tioIts of anthropologists have said it in a wide variety of ways: 1u.. human behavior in terms of putatively universal Plotives. An exacting critic
order to construe the gestures of others, their words and winks and more from a neighboring discipline recently allowed that the work of anthropol­
 '"'­

THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 13

12
ogy, "which combine[s] a passion for detail with a humane aspiration, does are not realist transparencies. Then, finally, we will be on epistemological
not suffer in comparison with its ethnocentric competitors" (Fields turf that, if only provisionally, we comprehend and control.
1985:279). In this respect, ethnography seems no more intrinsically "ar­
rogant" than do other modes of social investigation (pace Turner 1990:178). * * *
Much of the difficulty has come from the fact that, for reasons deeply The second motif, recall, is the question of history. Or, more precisely, of
inscribed in the politics of knowledge, anthropologists have classically historical anthropology.
studied populations marginal to the centers of Western power-those who In the late 1970s and eady 1980s, it became common to temper the
were'unable, until recently, to answer back. In this, as we will argue, our anthropological turn toward history-as-panacea by posing the problem:
position is little different from that of often radical social historians "What history? Which anthropology?" We ourselves raised the issue,
concerned with society's nether regions, the lives of"little people" viewed arguing that
from the bottom up (Cohn 1987:39). Tile dangers of disclosure in such
any substantive relationship between disciplines is determined not by the
situations are real enough. Indeed, while all representations have effects,
intrinsic nature of those disciplines-if any such thing exists-but by prior
those imposed by academic brokers on communities without cultural theoretical considerations. It would seem obvious, for example, that historical
capital are more likely to have deleterious consequences. At the very least, analysis assumes different significance for structural functionalists than it
we have to confront the complexities of our relations to our subjects, texts, does for either Marxists or structuralists .... Hence to assert that anthropol­
and audiences-especially because the impact of our work is never fully ogy should be "more" historical, or history "more" anthropological, may be
foreseeapJe. This not only demands a serious regard, once again, for well-intentioned; but . . . the assertion remains vacuous without further
contexts~ our own as much as those we study. It also calls for a careful theoretical specification. [In our view] there ought to be no "relationship"
consideration of the real implications of what we do, a consideration that between history and anthropology, since there should be no division to begin
with. A theory of society which is not also a theory of history, or vice versa,
must go far beyond the now routine recognition that our writings are
is hardly a theory at all.ll
potential instruments of "othering."
~ut ethnography also has positive political possibilities. We ought not But there was more to the matter than this. For the space of intersection
be !Q~ 9,9-ick,for instanc~e:, to disr,egarq ~lte(~a!len~<:~:lt cuJwraJ relativism between the two disciplines was (inevitably, it now seems) pervaded by a
poses"ro bou~geois consciousness. Why else the speCial opprobrium heaped particular Geist-a politics of perspective, so to speak. Clearly, the kind of
upoil usb),' shrill absolutists, essayists of the closed mind like Alan Bloom? ' histories that were to find a sympathetic ear among anthropologists were
The fate of the Naparama may tell us that we are less influential than we unlikely to be the Chronicles of Courts and Kings. Neither were they liable
often suppose. Nevertheless, our work does reverberate in and beyond the to be event-full political narratives, however fascinating, of embassies
academy, although its legitimacy and impact vary with the way in which among empires, strife between states, or trade between chieftains; nor
we choose to phrase our questions. An important moment of choice is now latter-day quantitative accounts of past worlds that, by appeal to synchronic
upon us. It:W~Jalce ollrtask to be an exer~isei9. intersubj(:ctive translation, sociology, sought to write "general histories" in "numbers and anonym­
;(/ ip speaking for others and.thekpoint ()f:v:~~j.l:l.!"".·~~. .Q . rls.,~il.~j;:all:se. us no ity."12
~'i end of difficulties, moral and philosophical. And if We see iuo lie in the Bound to be much more attractive, save in some structuralist and Marxist
. t6~~ ~n<!ly~is of social systems. or sv-lturaCst~~ture;;'statis~ically qr circles, were richly textured accounts of things similar to what we ourselves
l~gi~ii.c~ived, we..ev:~d~t~ issue of ~epi:eseI}~at~<?na(1d·e;x.:periS!!!~e study-analyzed, broadly speaking, in similar ways. If the description was
altoget!l~t if, after an older Europeantradition~ we seels.tounde.rst~ suitably thick, the subject matter obviously remote, so much the better.
the making of collective worlds-the dialectics, in space and time, of How could we not be appealed to by, say, Carlo Ginzburg'S (1983) tale of
societies and selves, persons and places, orders and events-then we open sixteenth-century witchcraft and agrarian cults in Europe, or his account
ourselves to conventions of criticism widely shan,:d by the ~npoSitJvlst (1980) of the cosmos of a contemporary miller. Such studies in Fhistoire
human scienc~en, too, we may traffic in analytiC constructions, not in des mentalites l3 are not just chronicles of the quotidian, of "little peopfe"
unverifiablesuDjectivities, and can acknowledge the effects of history upon and their ordinary practices; nor-like their even more everyday English
our discourses. Then we may focus on interpreting social phenomena, not equivalents-are they merely studies of "the experience of living men and
on the endless quest for textual means to exorcise the fact that our accounts women" (Thompson [1978a] 1979:21; cf. Thomas 1971; Hill 1972). As
14 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
15
Darnton (1985:3) notes, they also "[treat] our own civilization in the same
mode. This is not merely because the latter concerns itself with "faceless
way that anthropologists study alien cultures." Hobsbawm (1990:47-48)
masses," people who have left few documentary traces of Promethean
puts it in the words, more or less, of the English novelist L P. Hartley: For
careers. Subaltern historiography also challenges the very categorj~s_t.hro.ugh
these historians "the past is another country where things are done differ­
which (,;oloruaLpasts. have beenrnade. In so doing, it resonates with the
ently ... [in which even] the best interpreters still remain biased strangers."14
democratizing impUlse of our own craft, of which we have already spoken:
However, it is Raphael Samuel (1989:23) who probably comes closest to
the well-intentioned-some would say self-satisfied-view that ethnogra­
us in spirit. Arguing for the kind of history that might best be taught in
phy celebrates the. I1arratives, the: consciousness, and the cultural riches of
British schools, he says, wryly: "If one were not, like the historians of high
non-Western populations, especially those threatened with ethnocide.
politics, mesmerized by the glamour of power, one might suggest that
In anthropology, as we have noted, the liberal urge to speak for others
horses were more interesting to study than politicians and, at least for
has had its comeuppance. Social history may seem less vulnerable to
younger children, more appealing." Perhaps for anthropologists as welL
But the intention here is not to jest at the expense of politicians or
c~unterattack: Its subjects, often well dead and buried, can neither answer
b;ck nor be affected any longer by the politics ofknowlcdge. This, however,
historians. It is to make a profound methodological point. As Samuel
is-much too simple. Not: only do scholars work increasingly on history-in­
shows, the move from cavalry charge to hay wain and horse-gin, from
the-making (cf. Bundy 1987), but also anyone who writes of times past
sporting prints to the text of Black Beauty, lays bare the cultural texture
must recognize that there will be people who stand to suffer from the way
of an age. "Cherchez la vachel" says Evans-Pritchard of the world of the
in which social memory is fixed (c£ Ashforth 1991). In addition, there are
Nuer-advice offered on the same conviction: that, in the career ofeveryday
those, both revi~i?n!~t _:i.~(U~(:g~~I, who champiQn lhe-cau~e o(hi$.to;icaJ
goods, of valued things, we grasp the constitution of complex social fields. ~opuraticiris. -Thus Rosaldo (1986) contends that, for aI1 his efforts to
We ourselves follow this object lesson in Chapter 5, extending the concept I capture the life-world of the peasants of Montaillou from within, Le Roy
of commodity fetishism to explore how cattle give analytical access to a Ladurie (1979) derives his narrative primarily from the standpoint of a
changing Southern Tswana universe. In fact, in making his case against
contemporary inquisitor; RosaJdo, in fact, likens his perspective to that of
history as the biography of big men, Samuel voices the same concern as we
a colonial anthropologist. Spivak (1988) goes yet further: She questions
did about an anthropology "from the native's point of view": that it tends whether the subaltern can speak at all, even through the texts of a
to focus on individual intention and action at the expense of more complex radicalized history. It appears that in representing the point of view of
social processes. Take, for example, the Battle of Trafalgar, which looms
"natives," living or dead, cultural historians are on no firmer epistemolog­
large in standard British textbooks, not least because of Nelson's heroic
ical ground than are ethnographers-and no less embroiled in the politics
death. This event, claims Samuel, was far less important to the making of of the present (Croce [1921] 1959:46f).
an epoch than, say, the Married Women's Property Act of 1882. A product
This calls to mind Jacques Derrida's critique of Foucault's history of
ofdrawn-out social struggle, the act had critical consequences for marriage,
madness and, as significantly, Ginzburg's rejoinder. Both are instructive
family, and gender in late nineteenth century England-in other words,
for anthropologists-especially for those drawn by deconstruction, those
for the construction of modern British society tout court. Yet it barely rates
troubled by the tyranny of a totalizing social science. They are also salient
a footnote in any major work.
in light of our own analysis (Chapter 6) of the historical consciousness
Cultural historians like Le Roy Ladurie, Ginzburg, Darnton, and Samuel
borne, in apartheid South Africa, by an alleged "madman." It is impossible,
give us comfort in the face of less friendly interlocutors partly because they
~errida U2?8~3~f)il1pismis~i1)g F()1l911Jt~~Hi~t:orJ tf967},toiv.a­
reassure us that our methods ("suspiciously like literature" to the hard
social sciences [Darnton 1985:6]) are more rigorous and revealing than
l~e deI.1!e:m.i_'uav.ein ~'ther~str.ainedanqr~straiJ)jng langllage" ofWestern
r~~()!l. Yet this is the very language that constituted/olie in the first place­
they appear. But most fundamentally, they see virtue in-indeed, ma~e no the very means of its repression. It follows, therefore, that there is no point
apologies for-disinterrinJ;li and. disseminating the lives of insignificant in the discursive structure of Western rationalism from which an interro­
<c()tiiers."'"for many of them, far (rom an act of dominati()n or3P.Pi.OiIDa­ gation of abnormality may proceed. Derrida (1978:35-36) adds: "All our
fion, this is':the first stepin,a subversive' historical sociology, a histpry European languages, the language ofeverything that has participated, from
written agaiQst the hegemony of high bourgeoisies, the power of parlia­ near or far, in the adventure of Western reason [are implicated in the
Objectification of madness]. . . . NI?!ki~Mthin..thisJangy_ag~<\l]dBIL!?ne
ments, and the might of monarchies. Their work, moreover, bears more
alllo tho~e who spea.k it, C:(ll1e.s_(;.(l~.: ... [T]he revolution against reason
than passing similarity to colonial historiography in the so-called subaltern ng
16 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
17
can only be made within it ... [and] always has the limited scope of ... a
disturbance." For all his determination to write a history of insanity (1989:23) observation that ~.:.~is.toJ::y:from below' .... without somel,arger 'CD
"without repeating the aggression ofrationalism," then, Foucault is accused framework . . '~(.:~J!l£L'L£Yl:ill:_:§i!.<:. !l,n<f loses its subversive potential."
of self-delusion; the project, implies Derrida, was itself pure folly, madness. I'!!!p!.()perly. 'QIl1:eXtualiz.eq~ the stories of Qrdinary people.pa~~ stand in
And so his act of subversion disappears before the deconstructive eye. The. danger of remaining ju.sUJ1<lt;...~tQJ:ic:s. To become something..rnore, these
P!~<l!~<:!~~~~~.~~(! .politics of ethnography is obvious. Its analytic gaze, t<><;l,
Rartiif,".~'hidden histories" have to be situat.ed in the wider worlds of power
a2~r~,entrapped in Western reason, a party to the very relationship­
and meaning that gave theJl1life. But those worlds were also home to other
b_e~~<:,~I!..~l!Qit!.~U![199Qieq,.t.h<=..stu:veyor and the surveilled-on whic!:, dramatls pers.o..·na.e, other texts, other signifYing practices ..A .. nd h.ere .is t.he ',.~,,"\
colonizing power/kQowledge is based. second point: there is nobasiuo...assume that the histories of the represse5i, <b
"To Ginzburg (1980:xvii), however, Derrida's critique is both facile and U1..th.fI11S~Y~.~~Jl.QJ.!L'1§p.~,jjlIJs:c:"y~!p r~velatioQ~ as we show in Part Three, .
nihilistic. For, against all the forces of repression in the world, it allows the;;, giscours(;sof the dO,min,mt ,;tlso. yield vital insights into the contexts
little by way of legitimate reaction: inaction, ironic indifference, silence (cf. apd e!~<:~S.S.~~_9_t1:YhicI~tlle.Y.w~r<:part. The corollary: There is no great

Said 1978).15 Even worse, it misses the fact that "the only discourse that historiographic balance that may be restored, set to rights once and for all,

merely by replacing bourgeois chronicles with subaltern accounts-by

constitutes a radical alternative to the lies of constituted society is repre­

sented by [the] victims of social exclusion." Extraneousness, irrationality, "topping and tailing" cultures past (Porter 1989:3). History, Antonio

absurdity, rupture, contradiction in the face of dominant cultures, to take Gramsci reminds us, is made in the struggle among the diverse life worlds

the point further, are all mirrors of distortion, angles from which are that coexist in given times and places-between the "tendentious lan­

exposed the logic of oppressive signs and reigning hegemonies. Despite his guages" that, for Bakhtin (1981 :263; Holquist 1981 :xix), play against one

location within the discourses of Western reason, concludes Ginzburg, another and against the "totality" (posited, realized) that gives them

meaning. For historiography, as for ethnography, it is the relations between

I:§1l9uK~id.succeed in using the history of madness, the politics of sanitY,

fragments and fields that pose the greatest analytic challenge.

to unmask the c:p~rciveness of convention and (self-) discipline.

We should not draw false comfort from this. It is one thing to acknowl­ How) then, do we connect parts to "totalities"? How do we redeem the
edge the possibility that rupture, absurdity, or resistance may disclose­ fragments? How do we make intelligible the idiosyncratic acts, lives, and (. ,
representations of others? How do we locate them within "a historically
even disable-the world from which it emanates, but quite another to
determinate environment"? It is here that cultural history, for all its
ensure that it does. More immediately, though, there is relevance for us in
brilliant achievements, runs out of answers for us. Not that this should be
the methodological implications of Ginzburg's argument, in the kind of
a surprise. Just as we were turning to history for guidance, at the moment
history to which it is dedicated. The latter, by definition, must be grounded
in the singular. It can make no pretense of representativeness, ofdisintering When our early paradigmatic foundations were crumbling, many historians
a typical seventeenth-century European villager or nineteenth-century began to repay the compliment. Just as we were inclined to see history as
urban merchant. For all the cultural historian can ever "see" are the "good"-as if time might cure everything-they seemed to see ethnog­
raphy as a panacea. This should have warned us that they were in as much
dispersed fragments of an epoch-just as the ethnographer only "sees" theoretical trouble as we were ourselves.
fragments of a cultural field. However, the point of recovering these
fragments-be they individuals or events-is to "connect [them] to an In fact, much historiography still proceeds as if its empirical bases were
self-evident, as if "theory" were an affectation only of those of philosophicaJ
historically determinate environment and society" (1980:xxiv).16They may
bent (Thompson 1978b; cf. Johnson 1978). Collingwood (1935:15) might
come to us largely by chance and may in some measure be unintelligible.
have asserted, long ago, that the "points between which the historical
But to recognize and respect that unintelligibility, which we have perforce
imagination spins its web ... must be achieved by critical thinking." But
to do, "does not mean succumbing to a foolish fascination for the exotic
there has been relatively little effort to interrogate the constructs through
and incomprehensible." It is, rather, to undertake redeeming them. For
Which the silences and spaces between events are filled, through which
"redeemed [they are] thus liberated" (1980:xxvi). Liberated, that is, in the
disjOinted stories are cast into master narratives. In practice, of course, the
sense of being restored to a world of meaningful interconnections.
Way in which the "historical imagination" does its work is culturally crafted;
Ginzburg's insistence on the redemptive connection between fragments
so, too, is the fabrication of events, as we are reminded by old debates over
and totalities brings together two critical points about cultural history in
general and, in particular, its subaltern variants. The first echoes Samuel's ~'histoire evenementielle (see below). As this suggests, the cultural historian
IS no less prone than the cultural anthropologist to read with an ethnocen­
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 19
18
eye. In the absence of principled theory, ethnographers of the archive Let us begirito ;\nsweqh~_q,Yestionil1 the negative voice-by disposing,
and the field alike tend to become hermeneuts by default, finding in that ls;'of the kinds of historical anthropology that we seek specifically !p
interpretive anthropology a confirmation of their own phenomenological ayo.id. The method in ourmatice;'to invoke the memory of Edmund Leach
individualism. Of those who did turn to systematic approaches-especially (1961 :2), will reveal itself as we proceed: Some ground clearing is necessary
to some form of structuralism or materialism-many have been attracted, if we are to cut fresh pathways through old thickets.
Many years ago, Nadel (1942:72) drew the attention of anthropologists
in the wake of recent crises, by the less deterministic visions of a Gramsci
to the distinction, already well inscribed in social theory and philosophy,
or a Foucault, or to such "counternarratives" as feminism, psychoanalysis,
between "ideoIQgical" and "objective" history. The first recalls Malinow­
and subalternism. Tbey have drawn, il1_q~~er ~g~I.i~, on ~riip<:r.ea.si.Qgly
ski's (e.g., T948:92f) description of myth: It is the past as told b;Lpeople
glo~<lLJegacy of social thought, to..whiciLweanthropoiogists have.equal to account-authoritatively, authentically, audibly-for the contemporary
access. shape of their world. By contrast, "factu,al"chronicles, the work of dispas­
What, in sum, are the lessons to be taken from this excursion into
siomite observers, are scr~Pte(t ~~inaccordance" .•• with. universalcritera of
history? Clearly it is culturaLhistorians, more than any other social scien­
cQIlnexionanrl sequence." Nadel did not go on to point out that "ideolog­
tists, who validate our endeavor as ethnographers. This th.c;yggl:?y_~~~enjpg
ical" history, rarely exists (or ever existed) in the singular. He wrote, after
the possibili.!X~f.a sU~"y"~rsive.historical.aIltQr.bpqIQgy} 9nC.1hat.~es all, long before culture was seen to be a fluid, often contested, and only
primarily on lilll~.1l~Ql?1e".!lJ1(:tJhei.r W()r~ds. Lik..ecultural studies, with partially integrated mosaic of narratives, images, and practices (see below);
which-at least, in Britain-it has had a rich conversation (see Turner
before we even perceived that, in a single African society, there may be
1990:68f; Johnson 1979),cultyraLhi.stor.~'L.hasbecn especiallya,~sp.t.at
alternative (gendered, generational, even stratified) histories and world­
revealing that all social fie~ds .a~<;,.9.2111.J.iQ§.....QL~ th3J"~l!\n.m:'~ maps.
often a matter of argument, a confrontation of signs andpractices,a,IQllS
The distinction between ideological and objective history may no longer
the J<l1l1!.~~ otPQ..\yf~i. that it is possible to recover fr0In fr!l$!ne_l!!§, go unquestioned in the musings of metahistorians. But it remains deeply
d~~QLq... :ln~L<:y~n from silences,the" rawrn.~~rT!1[~}ili·~&.{c.h:.tQ.,.wci.ts;..
st entrenched in Western popular discourse and, implicitly, in much historical
imaginatiYc sociQlogie~"(,fJhe pi~ii9·<,fi1i~·pr~cnt. But eventually we m.ll anthropology. How Qften are we not at pains to show that the chronicles
part company. Given the reluqal)ce. Qf b~!Qri.aIl,stq reflc.cLatt.ma.ners..o
f
of kings, conquerors, and colonizers-we follow Croce's ([1921] 1959:51)
the~~,r. ~e?~e.n,EY_~~t?~~~~_<:~iEi~~~.~211:l:tiQ!1s f()!..a,g"Jyt.ic,1?r:9bl~ms, usage here 17-are distortions, pure ideology in servitude to power, the
w.~~}19 pur:pwn way through thetnaze.,ofcommdruI11~ ,that lies,!IQ$ <:9r:ollary i:>eing that our version is more objective, more factual? The same
the road to a principled historical anthropology.
--,-----""'","'-."'.-,.. ..
<.~-"".-- ~-'-~' ",.' "," --' ,,~
is true of the past as perceived, from the bottom up, by the dispossessed
and the disenfranchised, the mute and the muzzled. How often do we not
explain away their failure to act in their own interest, or to act at all, by
seeking to show that they perforce misrecognize the "real" signs and
III structures that sustain their subordination? In so doing, it is all too easy
So, with all this in mind, toward what kind of histol!<;al anthropology do to cross an invisible boundary, the now familiar line that marks out the
w~e? And how, exacttY:doesethriography'fu'T~to it? It follows, from limits of authority, ethnographic and historical alike. For it is one thing to
the way in which we call the question, that we do not find a ready answer assume that no human actor can ever "know" his or her world in its
in the methods and models spawned by the recent rapprochement of totality; one thing to situate the natives' points of view-note, now, the
history and anthropology-or by its intellectual precursors, which go back plural-in their appropriate context. That, as we have suggested, is entirely
much further than we often realize (see, e.g., Cohn 1980, 1981; J. L. lc::.git.imate. But it is quite anotbcc thjng tQar,ro,gare to ourselves an exclusive,
Comaroff 1984; Rosaldo 1986). Nor, as our brief excursion into fhistoire e,mancipatory, suprahistorical purchase on reality. To dredge up the lexicon
des menta-lites indicates, is one to be found by surveying existing historiog­ of an age gone by, ~<?c.i;\lJ~.l!<?"'\l\lJ~<igc.A~\ls.:fL~~ gr I2ric~. And
raphies and choosing the most congenial candidate. Recall Thompson's there are no "universal criteria of connexion and sequence"; vide Joan
(1978b:324) admonitory metaphor-a little shopworn now, but still val­ Kelly's (1984) feminist critique of orthodox practices of periodization in
uable-that ideas, ways of knowing, are not like objects in a supermarket, European history. Uniyer~al historiography, as we should all be aware by
perishables casually bought or brushed aside, cast out or consumed. now, is itself a myth-wQ!:~~,.<l:.<:ol1cei!:. Indeed, the most striking thing

./
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical ImaginRtion 21
20
Most notable among them were studies of.-dj)westic groups (Fortes 1949;
about the very idea-the Western idea of universalism, that is-is how
Goody 1958) and villa,ges (Mitchell 1956b; Turner 1957) th~ts()u,gl1~ to
parochial it is. arrive at aggregate descriptions of social structures by illuminating.their
Any historical anthropology that sustains a fixed dichotomy between

cy:~l!ca,l dynainics. Do not misunderstand us: Some of these studies,

the ideological and the objective is bound to run into all the old problems
especially those of the Manchester School in Central Mrica, were based on
of brute empiricism-not to mention accusations of insensitivity to its

oWJ).~positioning and provis ionality.In short, it invites the justifiable

perceptive, blood-and-guts narratives of social struggles. But the latter
were removed from history, consigned to the uneventful register of "struc­
criticisms raised most recently by postmodernism but also by many before
tural time." However much human beings railed against the contradictions
on the long road from the early Marx to late phenomenology. If a distinc­
of their world, or fought with one another, their actions were always seen
tion between the ideological and the objective is to appear in historical
to reinforce the system in place, never to transform it. This reduction was
anthropology at all, we would argue, it is primarily as a cultural artifact, a

not purely the preserve of British functionalism: It was to reappear later,

distinction that itself is to be interrogated wherever it surfaces. Who does
more fashionably addressed, among Marxists concerned with the repro­
it empower and in what manner? Are there other forms of historical

duction of systems of domination (e.g., Meillassoux 1981).

consciousness in the same contexts? Are they expressed or suppressed? By
Altogether more recognizably "historical," if utterly undiachronic, was
what means? InjJJl!l l. ol!!.historical anthropology begins by eschewing the
~ the second use of the past. Reminiscent of Levi-Strauss's (1963a) statistical
very possibility of a realist, or ~n e!!g:ntialil!t, history. This is not to say models, of historiography founded on the "anonymity of numbers" (see
that there are no es~ences andrealities in the world. Quite the opposite. p. 14), its...point was. to verify,. in rates andincid~nces, descriptive accounts
But our objective, like the objective of rriany others, is to show as cogently of existing social systems. Evans-Pritchard «(1961] 1963:55) observed that
as possible how they are constructed: how realities become real, how
"a termlike 'structure' can only ,be meaningful when used as an historkal
essences become essential, how materialities materialize. "Symbolic real­
e~pres$~i~o_~<i~X~()t~. ~_~~!. . ()fr~~a!i()l!s_ ~~()~~ to have endwed .ovec.a
ism," a figure of analytic speech used for rather different theoretical ends
considerable pel;iQd.~Qf..time." Echoes of Annates, by way of Braudel (e.g.,
by Brown and Lyman (1978:5), captures well the spirit of the matter. To
1980). Thus Barnes (1954:171) retraced 130 years of Ngoni history in
the degree that our analytic strategy may still count as objectivist, then, it
order to show that "the form of (their] social structure [had] rernaine!i ,the
is highly provisional and reflexive. Perhaps this is the hallmark of a same"; and so a long, tortured story of state formation, migration, and
neomodernist anthropology. colonial conquest is distilled into a two-dimensional, lifeless aggregate
If our historical anthropology is anti-empiricist, anti-objectivist, anti­
termed "[the Ngoni] political order."2°Less grand in scope, but similar in
essentialist-except in the amended sense in which we deploy these terms­ object and spirit, was the historical study of social institutions. Ifwe could
it is also anti-statistical and anti-aggregative. Let us explain what we mean show, for instance, that, succession among the Zulu had actually passed
with reference, once again, to our intellectual heritage. Recall the early from father to s~nior son a certain number of times, we had empirical
days of the controversy in Britain over the relationship between history justification for the claim that the "principle" of primogeniture obtains.
and anthropology. These were the days when Evans-Pritchard (1950, Similarly, if Highland Burmese men had married the daughters of their
1961:20), invoking Maitland (1936:249), lined us up with art and aes­ mothers' brothers in a given proportion of cases, we might be persuaded
thetics against science; when Leach (1961), Schapera (1962), and Smith to say that they "have" an asymmetrical alliance system. 21 Note the gram­
(1962) argued that, in spite of our claims to the contrary, we had been matically awkward tense shift from historical past to ethnographic present:
doing history all along-and neither could nor should do otherwise; when It recapitulates the methodologically uneasy move from data to generaliza­
the Association of Social Anthropologists of the Commonwealth finally tion, event to structure, history to form.
blessed the rapprochement with our "sister discipline" at its annual con­
Judiciously and imaginatively used, this kind of history may be sugges­
-m" ference (Lewis 1968).18 It is clear, with hindsight, that there were three
19 tive. More often than not, however, stat!~tica.L~~~t:<lI.1ce!!~E1I:.rticul;u::ly
,quite different forms of historiography being discussed. But nobody, when read across cultural registers, .!lre. misleadipg. Not OEJy,.9QJh~yjnrite
other than Leach (1954), seems to have said as much. us to reify institutions,".!hcl!L~ndowing a slipper~ abstraction with false
1" The first form was confined to analyses of repetitive processes of thel Concreteness, but_tl:ley.al_so__c:!.e~tfQY!lt_e.J::fut sig~~~ts toward causal expla­
short and medium term-analyses we would barely recognize as histori,a
nations. In the Southern Tswana chiefdoms of the nineteenth and early
at all now, although they were often cited as proof that anthropology rea,lly twentieth centuries, for example, senior sons typically inherited their
was concerned with time (as if this were the same thing; see Chapter 4).
 22 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
23
fathers' property and position. But they did not necessarily do so because
For Leach (1954.;4)., ..'~eyer¥-tealso(-":t.y.is a.process.i.luime": Internal
of the principle of primogeniture, as has often been said. 22 By virtue of the
chang;--=-either transformation within an existing order or the alteration
manner in which succession struggles were culturally constructed, men
of its structure (p. 5)-is perennial, ongoing, inevitable. What is more,
made themselves into senior sons in the course of these processes 0. L.
social reality never "forms a coherent whole." It is, by nature, fragmentary
ComaroffI978);the rules of rank might have provided the rhetorical terms
and inconsistent (p. 8). "System," therefore, is always a fiction, an "as if"
in which claims were argued, but they simply could not decide matters one
model of the world, for actor and analyst alike. But, Leach added, it is a
way or the other. As this implies, it was the logic of practice, not a set of
necessary analytic fiction, because it affords a means by which otherwise
ascriptive norms, that gave form to such struggles (Bourdieu 1977:19f).
invisible connections between social phenomena may be traced out and
Likewise, for reasons having to do with the politics of affinity here, the
explained. Many are familiar with his ethnographic case: In Highland
close and ambiguous kinship ties that often linked spouses before marriage
Burma, we are told, Kachin. groups were caught up in a dynamic pattern
were commonly (re)negotiated during their lives together (see Chapter 4).
of movement between two polar types, two idealized representations, of
Consequently, numerically based generalizations about Tswana succession
political order. One was the highly centralized, hierarchical, autocratic Shan
and marriage may do worse than tell us nothing. They may manufacture
state; the other, the decentralized, egalitarian, "democratic" gumlao polity.
misinformation.
Most communities, however, fell somewhere between, in so-calledgumsa
Indeed, given that accounts of this kind come in a highly persuasive
formations. But the latter were not static: They were constantly moving in
for~;' they have the capacity to render soft facts into hard fictions;23
th;direction of either the Shan or thegumlao "type." As they did so, the
\ ': statistical statement,24 as we said, is the mode of enchantmellt t:JlaJ,in our
internal inconsistencies (i.e., contradictions) ofthat "type" would manifest
culture, makes'truth "empirIcal." But, most distressing of all, thes~mei:h­
themselves, encouraging a countermovement-itself impelled by the self­
odS--defkct-:uui"attenfion away from the problematicqualii:y of habii:ual
interested actions .of individuals who, appealing to diverse values, abetted
practices, hiding their historicity by mys,tifyingtheir meaningfuLconstruc-:'
the process of structural change by pursuing their own ends (p. 8). The
~" don and the bases .~:)JJh~ir <;.mpowerment. For normal sociology there maY
net effect over the long run, some 150 years, was a pattern of oscillating
beendur{ng appe~l in ignoring ~ultural a,rppi&yiry:•.in sacrificing p~lyphony equilibrium.
to the queSi-roiceiialrify~'ln'-feduCing ll).essy "nat};e~'~egones.to.me-a­
Political Systems of Highland Burma certainly has its shortcomings.
siu-abl¥.~'scieQ.titj.C" oge,s. ,.$0 be it. Tl.lal iSE...<?.t-!~~.9.l?ieSLQfour....histQJ'ical _
Leach has been taken to task for (I) relying on crass utilitarianism, a
al1~hropology.
universalist cliche, to account for hUII1anmoti~~, thereby separating
The third mode of historiography in contemporary British anthropology,
CUlture from society and reducingitto the "outer dress" ofsociaLa~tion;
C9 to which we find ourselves much closer in spirit, also provide~.auseful
c~SsOn-:-AnythingbiifstatrStlc'aror--mductlve, it was ba~ed on the (~)resor_ting" none..tEel~s_s,. t,o y.!lJg~!:j9.!;,~J.~sl}l,i1l9r.qeI"tQ.mioo.alizl!-the
behavior of homo economicus in the Kachin Hills;.~l)•. ~ibingDHmsa,
a~Qmth;lt all social orders exist in time; that all are i~herentlyunstable
gumlao, and Shan as ideal "types"-without .sub;ec.tWg!htUl.to...hi.s.tat:i.cal
and generically dynamic; that there are no prehistoric "anthropolog~cal
sOOeties~'''to recall Caocian's (1976) extraordinary term; and that, as
apalysis-apd,,!hen tr~atin~ the.!ll...~.~.J~<:~1!~L~eaW:j~s; (41J~i~ing to locate
Highland Burmese communities in. continental and glooalcoptext or
Dumont (1957:21) once put it, "history is the movement by which a
society reveals itself as what it is." Perhaps most representative of this Within line.aL processes Qf!h~JQQ& ,ruIl;i!.!ld, .ti.12~J1.y.. (§1_r.edu~iE.B.hi§t2ry
to a repetitive pattern of (bipolar) social equilibriuQ),
position was Leach's remarkable Political Systems ofHighland Burma (1954),
We are not concerned here with evaluating these criticisms. Whether or
a study somet.im. es sa.id to ~a.v.e. anticipated by many years (ILthe move in not they are justified (see Fuller and Parry 1989:12-13), each stands as a
ant!!topQ1QGY toward prac.t1ce theory (Fuller and Parry 1989:I3j;(:2) 1m:.
general admonition, something that any historical anthropology would
call, to situate local systems in the wider PQli~ical and social wo~lg..$_9f
Want to avoid; hence they are to be added to our negative checklist. But
whIch they are part (Ortner 1984:142), and (3) the recognition that aU
that, too, is only part of the story. There are also three constructive lessons,
hllman communities are shaped by an interplay between internal forms
or rather challenges, to be drawn from this worthy effort to give expression Cr. .',
and external conditions (Leach 1954: 212). It also resonated with Bakhtin's
to the assertion-often made, rarely made good on-that societies are
(1981:270) insistence that the holism of (linguistic) systems is posited, "processes in time."
not given, and that it is acted upon to ensure intelligibility in the face of
(; The first concerns the t1uid, fragmentary character of social reality and
fragmentary realities (see following discussion).
the question of order. Leach would have scorned any postmodern sugges­
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination 25
24
resolutely prehistorical. So do some notable attempts to resituate them
don that, because the world is experienced as ambiguous and incoherent,
within World History (e.g., Wolf 1982), most of which merely show that
it must therefore lack all systematicity; that, because social life seems
they have been enmeshed in global connections for longer than previously
episodic and inconsistent, it can have no regularity; that, because we do
thought; not that they were inherently, internally dynamic-if in their own
not see its invisible forms, society is formless; that nothing lies behind its
particul~r ways-all ~long. .. .
broken, multifaceted surfaces. The very idea would probably have struck
This 1S not to denIgrate the lllSlghts that have come from looking anew
him as a lamentable failure of the analytic imagination. Political Systems of
at worlds other than our own through the eyes of, say, feminist anthropol­
Highland Burma, remember, set but to disinter the dynamic structure
underlying a diverse (dis?)array ofsocial arrangements and representat~ons,
ogy (see, e.g., Collier and Yanagisako 1987) or, for that matter, Meillas­
soux's Marxism, Wolf's world system, Sahlins's structuralism, and Bour­
values and. events; to show that, if our models are supple enough, t.hSY
dieu's embodied practice. ~_hf!'pe learned much from them. But we have
should make sense of even the most chaotic and shifting social environ­
not ended up with any -senerally a~cepted theories. Qrmodels of the
ment. Again, whatever the merits of Leach's account, the implication is
h~!9l'icity of non-Western societies. 26 Perhaps that, in itself, is no bad
clear. We require good grounds for claiming the nonexistenceo.(;LS¥Stem thing. And yet, without some way, however provisional, of grasping those
or a structure-the fact that we are unable to discern one at first blu~hjs
h~-rdly.pi?orth;t i~ 'ther~;
is.!1ot cou~t~;chal­
Here, then, is a preemptive
historicities-note, again, the shift to the plural, the recognition of differ­
ences-anthropology: wi1Lc.Qotinue.to cast "other cultures" in the timeless
lenge to the deconstructive impulse of the 1990s: A!>sence .and discoo~c­
shadows of its own dominant narratives. It_Y"ill alsQ leave intact the
don, incoherence and disorder, have actually to be dem.og~trated. They can
disabling. ()pposition between historiography and ethnography..
neither be presumed nor posited by negative induction. 25 l"3 'The last lesson to be taken from Political Systems of Highland Burma,
The second lesson of Political Systems of Highland Burma applies to ./ and from the arguments that followed in its wake, has to do witlL"units
/l~
the historical anthropology of the modern world order, in particular, to oLanalysis": the terms, that is, by which social science breathes life into
the currently fashionable concern with the encounter between international
data, thence to arrange them into expository narratives. Leach's analysis
and parochial systems, universal and local cultures. Good intentions not­ raises the difficult question of whether historical anthropology is forever
withstanding, it is impossible to restore histo,ry to peopies.allegedly witQ.Qut compelled to share the two fundamental tropes of Western historiography,
it by appealing WIiisrorlcal modds ofgl?balpr<:>.cesses, especially processes the individual and the event. Note that the master motif of the Kachin
ill Westem J;l91ki.cal ~q1l9my, ~hI!e _~ustainiI!g.ahistorical models of non­ past, as he tells it, was Qscillating equilibrium, a great epochal movement
Eur~£~an "social f9!1Uatigns"-whetherthese be described in the language realized in a cumulative series of incidents animated by (universal) human
of Levi-Strauss, Marx, or Max Weber. For, as has been said ad nauseam, motives and (rational) modes of action. In offering his methodological
"peripheral" populations do not acquire history only when they are im­
individualist account of structural drift, Leach falls back on a classical, and
pelled along its paths by the machinations of merchants, missionaries,
classicaUy ethnocentric, conception of social history. Of course, he is not
military men, manufacturers, or ministers of state. ~ntly J2!1t,_.~!nUy alone in finding it hard to escape the liberal modernism of his own
histori~.<!J.!l:n~~r()p()I~,i~.()my:...p.ossibkt9 tht;,. e~Jent lh~L.iUs.~apab.lsJJf European culture. Structuralism has long obsessed over the individual and
UJ:\!Illin.~t!!lg.Jhe.I~~408.-enons. histm:icityJ?La!l..~~JCi~J ~<?~!9~· the event,27 never quite laying down their ghosts once and for all. The
This may seem old hat. Coquery-Vidrovitch (1976:91) said fifteen years
situation of structural functionalism is similar: For all its ostensible concern
ago that "no one doubts any longer that precolonial societies had a history."
with the nomothetic, it came increasingly to rest, as we said earlier, on an
Still, it is one thing to recognize the undeniable, another to give account empirical scaffolding of life histories, case studies, social dramas of inter­
of it. Models of noncapitalist orders abound, yet few demonstrate their
personal conflict, and the like. Indeed, albeit often unobserved, biogra­
internal capacity for transformation (cf. Sahlins 1981), stressing rather the
phy-the optic that fuses individual and event into both a worldview and
(a priori) mechanics of their reproduction. Terminological niceties aside,
a narrative genre-lies at the methodological core of much ethnography
how much have we really advanced on our old conception of "traditional"
and history.
societies, "cold" cultures? Of local worlds trapped in repetitive cycles of But there is danger here. Biography is anything but innocent. 28 Its most
structural time (Gluckman 1965:285f)-until, to cite Meillassoux's
articulate textual vehicles in our own society are the private diary, the
(1972:101) startling revision of genesis, they suffer "historical accidents, journal, and the memoir, which find their way into much, often methodo­
usually due to contacts with foreign formations"? Even recent efforts to logically naive, historical writing; in the ethnographer's notebook it typi­
reconceptualize "precapitalist systems" (see, e.g., Guy 1987) treat them as
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 27
26
actions of human beings in the world? What decides whether, in the first
caUy appears in the guise of the life history, a singular dialogic contrivance
place, the bounded individual is even a salient unit of subjectivity? What
of observer and subject. Yet the diary and the life history are culturally
is it in any social context that constructs utilities and rationalities, "private"
specific, patently ideological modes of inscription. The former is strongly
associated with the rise, in the eighteenth century, of bourgeois person­ motives and collective consciousness, dominant worldviews and polyvalent
hood; Barker (1984), among others, traces its roots back to the Cartesian symbols, consensual signs and contested images? Precisely how are mean­
"I," an image of a self-conscious being freed from the webs of enchantment ingful atoms of human action and interaction contrived? The lessons we
and possessed of the capacity to gaze out at, and measure, the world. As a draw from Political Systems ofHighland Burma-and, more generally, from
medium of (self-) representation, more generally, life-histories bespeak a the early rapprochement of history and anthropology-converge in these
notion of the human career as an ordered progression of acts and events; questions. Indeed, considered in light of our dialogue with Phistoire des
of biography as history personified, history as biography aggregated; of the mentalites, they pose three challenges to any historical anthropology: (1)
"biographical illusion," Bourdieu (1987) caUs it, a modernist fantasy about to address the equations of structure and indeterminacy, of form and
society and selfhood according to which everyone is, potentially, in control incoherence, involved in tracking the movement of societies and peoples
of his or her destiny in a world made by the actions of autonomous through time; (2) to disinter the endogenous historicity of local worlds,
"agents." It is this fantasy that leads historians to seek social causes in both perceptual and practical, in order to understand better their place
individual action and social action in individual causes; to find order in within the world historical processes of which they are part; and (3) to
events by putting events in order. rupture the basic tropes of Western historiography-biography and event­
Inasmuch as it records such actions and events, then, the "life story" is by situating being and action, comparatively, within their diverse cultural
an instrument of bourgeois history-in-the-making, one strand in the process contexts.
whereby private thoughts and deeds are woven into the collective narratives At this point a shift in voice is appropriate. Having set the scene for our
of epochs and civilizations. Nor is it a passive, impartial instrument. historical anthropology in a critical key, we should say something of its
Gusdorf (1980:29), noting that autobiography is peculiar to the Western positive conceptual foundations. We began to lay these out in OfRevelation
sensibility of selfhood, argues that it "has been of good use in the and Revolution (1991). Here, consequently, we offer the briefest synopsis;
systematic conquest of the universe." Anthropologists, as we well know, in any case, as we stressed earlier, we wish to allow the essays to speak for
are alleged accessories in aU this: By translating the experience of others themselves.
into our own measures of being-in-time (cf. Fabian 1983), we are said to Clearly, the place to begin is with the idea of culture itself. Still the
have laid down the terms in which they may be represented-and, in both anthropological keyword par excellence, if anything it is enhanced, not
senses of the word, made into subjects. threatened, by recent developments in cultural studies and "cultural poet­ ......
Gusdorf mayor may not be correct about the "systematic conquest of
~"
ics" (Greenblatt 1990:3). For.reas.ous..detaikd..dsewhere (Comaroff and
the universe." Our more immediate worry, at this point, is that for the Comaroff 1991: 13f), w,~ take culturc....tQ...b~.th~~~~~Jllis:"'.~~E}~h~,.nSld of
most part social science persists in treating biography as a neutral, trans­ signs and.. P..~~~.tis:~~~id1.hJJJni!.I}. b.ri.ogs..,mnnm<::.l.aoo I£present them­
parent window into history. In so doing, it serves to perpetuate the selves anlio~l1.e.r§. L<l,mi.h~n"_thcil: sD"ie.ti~andhi.Sl9r.ies~ It is not merely
"biographical illusion": to regard persons and performances in the Pro­ a!L!l\:>.s!!.act ol5ktQf..signs,.,.ot .rclatiQQ~ (lIp.QI)..g. ~jgm, ~qLis. At just.the .~\]Jl1
methean mode, to find the motors of the past and present in rational c:>f habitualpractkes. Neither pure langue nor pure parole, it never consti­
individualism, and to pay little heed to the social and cultural forms that tutes a closed, entirely coherent system. Quite the contrary: Cultt:lr.~~IW\lYs
silently shape and constrain human action. It is a short step from this to a
vision of History and Society as the dramaturgy of intersecting lives: a
contains within it polyvalent, potentially contestable messages, image~,.<ll'!9 ~
,.
~ctions. It is, in short, a historically situated, historically unfolding ensem­
theater in which, as the narrative spotlight narrows ever more sharply on
bJ~_.<?i~~.stlifit:~s.-!n:<lS_tion:l sig!1 ifiers, at Q!!~e. mil~e.r!<lL'!lld~mbDlik.Soiial
actors and their scripts, text-a sad proxy for life-becomes all. And and aesthetic. Some of these, at any moment in time, will be woven into
context dissolves away into so many shadows. more or less tightly integrated, relatively explicit worldviews; others may
If historical anthropology is to avoid recapitulating the eccentricities
be heavily contested, the stuff of counterideologies and "subcultures"; yet
and ethnocentricities of the West, the individual and the event have
others may become more or less unfixed, relatively freefloating, and inde­
everywhere to be treated as problematic. Just how are they constituted,
terminate in their value and meaning.
culturally and historically? What determines, or renders indeterminate, the
 28 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
29
i It has been widely ar~:u:ecl in recent years that the concept ofCilltux,,i.n
i~~!f,_~s}Ilcapable ofgrasping the meaningful bases ofeconomy anci so~ sl!!:>ject of explication or a~gwnent (c£ Bourdieu 1977:94, 1~7!- This ~s

of inhabited history and imagined wOl;I9.~, "Power," we are told, has to be why its power seems to be mdependent of human agency, to he m what It

).
t added into the equation, since it determines why some signs are domil}iut, silences, what it puts beyond thelimits of the thinkable. It follows that it

. others not; why some practices seem to be consensual, others disputed­ is seldom contested openly. Indte.d,. the. moment that any set of values,

evellwhen they are backed by the technology of terror. Th~ generale?~~~ meanings, and material forms comes to be explicitly negotiable, its bege:m­

is well taken, albeit with a cautionary amendment: P~iijt~lf ..QQt ony is threatened; at that moment it becomes the subject.. Qfjdl:O~_QJ

counterideology.
above,nor outside of, culture and history, but it is directly implicatt.d.in
.thei!:5C:mStltution and determination. It qI.nnot, therefore, be "adde(:CJo As this)mplies, ideology describes "an articulated system of meanings,
th~~in s~ch a way as to solve the great conundrums of history and society. valu'Cs:-~nd beliefs of a kind that can be abstracted as [the] 'worldview' " of
This, we would argue, is where h~~.I!l0nyand )deology, the terrible any social grouping (Williams 1977:109). Carried in everyday practice
twins of much recent social theory, b~come salit:nt. Although we regret and self-conscious texts, in spontaneous images and popular styles, this
the often unspecific, devalued use of these terms, they do, if carefully worldview may be more or less internally systematic, more or less consistent
deployed, offer a cogent way of speaking about the force of meaning and in its outward forms. Still, as long as it exists, it provides an organizing
the meaning afforce-the inseparabi,lity;. th~aU~QL1?~_a.nd '!llmn:~ stheme , a . m~ter narrative, for. collective symbolic. produc:tion. Obviously,
:Fhey also serve to reframe the i5k~ ,9f @i.mrt; itself in such a way as to to invoke Marx and Engels (1970), the regnant ideology of any period or
embrace, at once, its systemic and indeterminate features: the fact. that...it place will be that of the dominant group, although th~ degr~eofjts
appears, on the one hand, as an orderly worldview .and,on tbs:pther,...as...a preeminence may vary a good deal; so, also, will the extent to which it js
h.~te!:()_dQ~1. even chaotic,repertoire ofpolyvalent images an.d rrac~ises.
eiIlpowered by the instrumental force of the s.tate.. Rut.nther.....su..Q_Qrdinate ~
P~~!d.hen, is an intrinsic quality of the social and the cultural~.in populations also hav~ ideologies. And, insofar as they try to assert the!l1- ;....
short, their determining capacity. S9metimes it appears as the (rclar.ive) s~b'.!s,.t<:>,i,!!!!j9~econtrol over the terms in which the w.orlqjs. Clr..dere.d,
abIlity "of' human beings to shape the lives of others by. exerting control th~y too \ViII call actively upon them-even if only to dash their symQols.
over the"prodtiction, circulation, and consumption of signs and objects, Here, then, is the basic difference between h~gem()ny . ~nd i<feology. .,
OY.eLilie. makiOg of both subjectivities and realities. Thisis .power in the ~~eI11()fl:r£.<:l.n.sJS.~~qf~.on.st!:uq~ an~ .::onven~io~a! pract.ices thathav~ come ~..I..;'!' : .. f
(l-gpntive mode. But it also immerses itself in the forms of everyday life, t£.£ermeate a polltl<:alcoIll!l1~l'l~ty; Ideology orIgmates m t~e ass(!~tlQ,l!~ of .'

forms that direct human perceptions and practices along conventional a particular social gr~1!p, HegemOny is beyond direct argument; ideology

pathways. Being "natural" and "ineffable," such forms seem to be beyond is.:'itore lik.eiYIo)e£e,.r_~~i,ved "as a ma~teJ; of injmical opil'l~on and interest'

human agency, notwithstanding the fact that the interests they serve may aru:t fie!1~e is more open to contestation. Hegemony, at its most effef:r;ive,

iU!tl!!eiigeology. invites argument.

be all too human. This kind of nf}nagentive power saturates such things as
aesthetics and ethics, built form and bodily representation, medical knowl­ Hegemony, .then, .is that part of a dominant. ideology that h<!S--.been
egge ..and material production. And its effects are internalized-in their natu?alized and, having contrived. a. tangible w.Qrld i.11..its jin~ge} goes.not
negative guise, as constraints; in their neutral guise, as conventions; in appear to be ideological at.all. Conversely, the ideologies of the subordinate
their positive guise, as values. may express hitherto voiceless experience, often sparked by contradictions
• 'I.
This distinction between modalities of power and agency, we suggest, that a prevailing cuI ture no longer hides. The manner in which a sectarian
underlies the differences, and the relationship, between ideology and he­ world view actually comes to naturalize structures of inequality-or, con­
gemony-which may fruitfully be regarded as the two empowered dimen­ versely, the'commonplace comes to be questioned-is always a historically
sions of any culture. specific issue. Typically, however, it involves the assertion of control over
Let us elaborate. We take hegemony to refer to that order of signs am! various modes of production, both symbolic and material-control that,
material practices, drawn from a specific cultural field, that come to. be as Foucault understood, must be sustained in such a way as to become
taken for granted as the natural, universal, and true shape of social beiug­ invisible. For it is only through repetition that things cease to be perceived
a!though its infusion into local worlds, always liable to challens..t;;~y: tbe Or remarked, that they become so habituated as no longer to be noticed.
logic of prevailing cultural fo!":'ms, is never automatic. It consists of ~N.ngs ~t the same time, however, no hegemony is ever total (Williams 1977:109);
that go without saying: things that, being axiomatic, are not normally the ~t constantly has to be made and, by the same token, may be unmade. That
IS why it has been described as a process rather than a thing, a process to
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 31
30
which all ruling regimes have to pay heed. TQe_C!!.()~e. su<:cc:~sfl!~r.~e,
poles, mentalities, and mass media of Europe and America. And this,
finally, brings us back to the question of method.
the more of its ideology will disappear into the domain of .hegemonic
practice; the less successful, the more. its unspoken conventions. wilLbe
opened to contest. This, self-evidently, is most likely to occur when the
g~__l:?~twe~n the world-as-represc::nted and the world-as-experienced l)e­ IV
comes both palpable and insupportable.
In Of Revelation and Revolution we take this analytic scheme further, How, then, do we do an ethnography of the historical imagination? How
using it to explore consciousness and representation, historical agency and
social practice, domination and resistance, global and local social orders,
do we contextualize the fragments of human worlds, redeeming them
without losing their fragile uniqueness and ambiguity? To repeat: for us
~
and the politics and culture of colonialism. Here we seek to make a more the answer lies in a historical anthropology that is dedicated to exploring
general point: that it is possible for anthropology to live easily with the
concept of culture and to defend it cogently against its critics. But this
the processes that make and transform particular worlds-processes that
reciprocally shape subjects and contexts, that allow certain things to be
\
requires that we treat culture as a shifting semantic field, a field of symbolic said and done. Over time, all social fields are swept by contrary waves of
production and material practice empowered in complex ways. unity and diversity: by forces that diffuse power and meaning and by
In sum, far from being reducible to a closed system ofsigns and relations, counterforces that concentrate and fix them. The premise of unification, of
the meaningful world is always fluid and ambiguous, a partially integrated some limitation to the "chaos of variety" (Holquist in Bakhtin 1981 :xix),
mosaic of narratives, images, and signifying practices. Its forms-which is essential to collective life-and, hence, to the very idea of society and
are indivisibly semantic and material, social and symbolic-appear, para­ culture. But so is the inevitability of proliferation, polyphony, and plurality.
doxically, to be at the same time (and certainly over time) coherent yet Situating our fragments is thus a challenging task, for the systems to which
chaotic, authoritative yet arguable, highly systemic yet unpredictable, con­ we relate them are systems of a complex sort. Yet, we insist, they are
sensual yet internally contradictory. The paradox, of course, is illusory. In systems nonetheless. We should not deny them coherence merely because
its hegemonic dimensions, any culture does present itself as relatively they refuse to reduce readily to simple structures.
coherent, systemic, consensual, authoritative. After all, whatever forms are We are not alone in urging that anthropology shift its concentration
powered by the force of habit are naturalized and uncontested; they do away from simple structures and local systems, at least as traditionally
seem eternal and universal-at least for the continuing present, however defined. This shift, however, has practical consequences. Above all, it
long that turns out to be. But alongside them there are always countervail­ deprives us of our conventional, all-too-easy means of bounding analytic
ing forces: dialects that diverge, styles that do not conform, alternative fields, forcing us to enter rarified realms of floating texts and macro­
moralities and world-maps. Sometimes these are implicated in open power structures, where the connective tissues-the processes and pathways of
struggles, sometimes they erupt in parody, sometimes they express them­ face-to-face sodalities-seem to dissolve into thin air. In the past, our
selves in mundane activity of indeterminate intention and consequence. strategy for studying "complex" situations was either to turn to the
Whatever. But the conclusion is clear: With a sufficiently supple view of sociology of networks and symbolic interaction-to a methodological
culture, we may begin to understand why social life everywhere appears individualism, that is, without a generic theory of society and culture-or
to find enclaves within the alienating world of modernity. We looked for
dualistic, simultaneously ordered and disorderly.
In the great confrontation between modernist and postmodern perspec­ "SUbcultures," informal economies, and marginal minorities, for ritual and
tives on the world, each of which emphasizes one side of the dualism, we resistance to capitalism; all neady circumscribed phenomena, for us still
are asked to make a choice. To do so is to be misled, however. The thick with meaning. Until very recently, we have felt ill equipped to broach,
is everywhere dualistic-this being one of those realities for which in their own ethnographic right, such things as electronic media, "high"
ought to have respect. Note that we say everywhere. If a culture, the discourses of science, or the semantics of commodities. At
anthropology is to work creatively at the frontiers of ethnography and the best, these have been regarded as forces eroding traditional orders or as
historical imagination, it must be founded on a conception of culture and" "Significant causes of our modern difficulties" (McCracken 1988:xi).29
society that takes us beyond our traditional stamping grounds-one that And so we have remained largely in the countryside, on ethnic islands and
travels easily to a newer generation of field sites, among them the metro- CUlturally distinct archipelagoes.
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY . Ethnography and the Historical Imagination 33
32
We are the first to acknowledge that it is not easy to forge units of extension of global capitalism-have their feet on the ground. Being rooted
analysis in unbounded social fields. But it would be false to assume that in the meaningful practices of people great and small, they are, in short,
an ethnography of the nation-state, of empire, or of a diaspora presents suitable cases for anthropological treatment. Indeed, whether or not we
problems unprecedented in earlier studies of, say, domestic production, choose to write about them directly, they must always be present in our
possession rites, or lineage relations. That assumption appears true only as accounts (cf. Davis 1990:32).
long as we pretend that such "local" phenomena are visible in the round The methodological implications of all this are best explored by way of
and are separable for heuristic purposes from anything beyond their a specific instance. Several of the essays that follow address the anthropology
immediate environs; as long as we sustain the primitivist fiction that of empire, in particular the nineteenth-century encounter between British
traditional orders are natural and self-perpetuating-and radically different Nonconformist missionaries and peoples of the South Mrican interior. The
from the unruly, unbounded, even unnatural worlds of "modernity" or former were footsoldiers of colonialism, the humble agents of a global
"capitalism." But few, surely, would wish to condemn anthropology to movement. The latter, who would come to be known as "the" Tswana,
such pastoral archaism; what should define us is a unique analytic stance, inhabited a world with its own history, a history of great political com­
less our locus than our focus. Whether our topic be headhunting in the munities built and broken. But the African past would become subservient ­
Amazon or headshrinking in America (or is it vice versa?), voodoo exorcism to the European present, made into the timeless sign of the "traditional"
in the Caribbean or voodoo economics on Capitol Hill, we should approach periphery. In order to grasp this process, we had first to characterize each
it from the same perspective: as meaningful practice, produced in the party as a complex collectivity, each endowed with its own historicity. And .­
interplay of subject and object, of the contingent and the contextual. then we had to retrace the (often barely visible) minutiae of their interac­
It is precisely here that anthropology has shown a failure of imagination, tions. For it is in the gradual articulation of such alien worlds that local
however, and here that we return to our opening theme. Many of us and universal realities come to define each other-and that markers like
continue to be hampered, in conceiving open systems, by the dualisms of "ethnicity" and "culture," "regionalism" and "nationalism," take on their
an enduring evolutionism. We are still prompted to deal in a priori meaning.
contrasts-between stasis and change, gifts and commodities, theodicy Elsewhere (199l:35ff) we have discussed the general problem of re­
and theory, and so on-that assume the meaning and telos of social covering the histories of peoples like the Tswana from evangelical and
phenomena. The Naparama and their kind remain primitive rebels, not official records, a topic now receiving long overdue attention (Amin 1984;
Promethean heroes or universal soldiers. And this impedes us as we try to Guha 1983). Here we are more specifically concerned with the question of
dissolve the great analytic divide between tradition and modernity, to how to do a historical anthropology of dominant, world-transforming
confront global issues in more inventive, less pejorative terms. Nor, as we processes (cf. Cohn 1987; Cooper and Stoler 1989). Clearly, colonial
suggested, is the problem resolved by upgrading mechanical models of evangelism must be understood both as a cultural project in itself and as
local systems, grafting them onto universalist theories of society and the metonym of a global movement; its participants certainly saw them­
history; or by literary critical methods that make ethnographic fragments selves as an integral part of the grand imperial design. This, then, is an
into exemplary texts without adequately relating them to the wider worlds appropriate site for an imaginative sociology, a context in which anthro­
pologists might recognize their kinship with cultural historians and embark
that produce them. on an ethnography of the archives. In our own work, the point of entry
.
'.(
Ethnography does not have to respect a binary world-map, let alone the
axes of typological difference. As a mode of observation, it need not be was obvious enough: We began with the conventional chronicles of the
tied either to face-to-face scenes or to a specific sort of social subject. True, Nonconformist missions. But, in trying to make sense of the churchmen's
we have classically set our sights on particular persons and palpable various writings, as well as the wealth of reported speech about them, we
processes, and this has determined our point of entry into any cultural soon learned not to rely on any preconstituted "documentary record."
field. But we are not, for that reason, limited to the writing of microso­ Rather, we had to pursue what Greenblatt (1990:14) terms the "textual
ciologies or histories. The phenomena we observe may be grounded in traces" of the period, traces found in newspapers and official publications
everyday human activity; yet such activity, even when rural or peripheral, as well as in novels, tracts, popular songs, even in drawings and children's
is always involved in the making of wider structures and social movements. games.
Nor ought we to confine ourselves to history's outstations. Even macro­ Instead of a clear-cut chain of events, or a discernible perspective, the
historical processes-the building of states, the making of revolutions, the colonial archives revealed a set of arguments. They were dialogic in
 34 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
35
Bakhtin's (1981:272f) sense; that is, they partook of diverse genres, of
cultural and historical heteroglossia that gave voice to complex patterns of The writings of the South African evangelists are especially interesting
social stratification. If the colonizers formed a single block, it was one in this regard. They differed a good deal in their intent and formality; The
ambiguities, agonies, and self-doubts aired in letters to kin were not
fractured by internal difference-and by diverging images ofempire locked
in "socio-ideological" struggle (p. 273). The latter expressed itself in exposed to morally vigilant mission overseers, for example; nor to philan­
disputes about such things as abolition, evangelism, and the way to rule thropists, who were more responsive to evocative accounts of savagery; nor
to the churchgoing masses, with their strong taste for Christian heroics.
and save savages. But, at root, it involved a contest over both the shape and"
meaning of "natural facts" and the major constituents of modern knowl­ Not only was similar material carefully contrived for diverse audiences,
revealing the range of purposes and constraints at work in the civilizing;
edge: its constructs of person, agency, and work, of Mrica and Europe,
quest, but the historical role of these writings varied likewise. Once
wildness and civilization. It was only by reconstructing this field of
addressed to the mission societies, correspondence was political property,
argument-and, going perhaps beyond Bakhtin, by redeeming its poli- .;
to be liberally edited and recycled for campaigns in parliament and the
tics-that we began to understand the cultural revolution entailed in both
public domain. Letters became pamphlets. And pamphlets became books,
the rise of European capitalism and the imperial gesture. Here, amidst all
eyewitness epics of "labors and scenes" beyond the frontiers of civilization.
the contradictions of the age, were forged the precepts and projects of a
Thus were layers of texts produced-indeed, an entire stratigraphy. By"
new hegemony, a new bourgeois modernism with universalist horizons
excavating the career of a particular document it is possible to follow the
and global ambitions. These, of course, included the Christian overseas
editor's pen as it refigured authorial statements, rationalizing them into
n11SS10n.
publishable forms that framed the doctrine of humane imperialism. And I
A historical ethnography, then, must begin by constructing its own
so an ethnography of this archive begins to disinter the processes by which'
archive. It cannot content itself with established canons of documentary
disparate, even divisive, discourses were fused into a consistent ideology,
evidence, because these are themselves part of the culture of global mod­
by which coherence was distilled out of the often chaotic, episodic stream
ernism-as much the subject as the means of inquiry. As anthropologists, of missionary experience (cf. Bakhtin 1981 :272f).
therefore, we must work both in and outside the official record, both with
We would insist, though, that a historical ethnography must always go
and beyond the guardians of memory in the societies we study (Cohn
beyond literary traces, beyond explicit narrative, exegesis, even argument.
1987:47f). In order to reconstruct the annals of a cultural imagination,
For the poetics of history lie also in mute meanings transacted through ~
moreover, we have to operate with a working theory not merely of the
goods and practices, through icons and images dispersed in the landscape
social world, but also of the role of inscriptions of various kinds in the
of the everyday (Comaroff and Comaroff 1987; Cohn 1987:49). Again,
making of ideology and argument. For only then can we situate individual
this is as true of world historical movements as it is of the most local
expressions and signifying practices within a wider field of representation.
processes. Just as the Reverend John Philip saw that any effort to re-form
Mter all, locating our fragments requires a sense of the way in which they
"the" Tswana supposed "[bringing about] a revolution in their habits" (see
ride the crosscurrents of division and unity at any moment; of how the
Chapter 10), so Corrigan and Sayer (1985) hold that the making of the
autonomous creative urge runs up against cultural constraint. Sahlins
modern British state was a cultural revolution borne in large measure by
(1990:47) notes that, although persons and collectivities "somehow deter­
I -,i mine" each other, they cannot, by that token, be reduced to one another. the humdrum rituals and routines that shaped the lives of subjects.
Certainly, the great empires of the past established themselves as much in
But our methods should tell us something of the way in which personal"",
a welter of domestic detail and small-scale civilities as by assertive political
acts become social facts. In the case of colonial evangelism, we had to
and economic means. Such are the tools that build hegemonies, that work
address the matter by locating a flood of rapportage from the imperial
thoroughgoing social transformations behind the back of a declarative,
frontier in the complex textual field wrought by the industrial revolution, heroic history.
the consequences of so-called print capitalism (Anderson 1983). But this
This implicit dimension-the study of symbolic practice-is a crucial
is just a specific instance of the general problem of reading social processes
Contribution of ethnography to history, since it brings a nuanced under­
from exemplary representations: If texts are to be more than literary topoi,
standing of the role of meaning and motivation to social processes. At its
scattered shards from which we presume worlds, they have to be anchored
best, anthropology has never been content to equate meaning merely with
in the processes of their production, in the orbits of connection and'"
explicit consciousness. In fact, the relationship of individual experience to
influence that give them life and force.
the collective, often unconscious logic of sociocultural categories and
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 37
36
designs has long been our stock-in-trade. We may have come to distrust churchmen were themselves contradictory products of a contradictory
bourgeois world. Although they wished to recreate, in Mrica, the British v
formal, overly coherent notions of culture, but we ought not to jettison
yeomanry of yore, their tools and tropes also carried the imprint of the
the subtle semantic models that so enhance our sensitivity to the power of
industrial marketplace and its commodity culture. And their actions played
signs in the world. 30 For, however open ended, systems of meaning have
a major role in processes of proletarianization, the likes of which they had,
determinations of their own. They do not just bend to the will of those
decried back home. This is where the relationship between the two dimen­
who wish to know and act upon them; to the contrary, they playa significant
sions of colonial evangelism-itself a highly specific encounter of the
part in shaping subjectivity. The "motivation" of social practice, in other
"local" and the "global"-took on its real complexity. At one level, it
words, always exists at two distinct, if related, levels: first, the (culturally
involved an odyssey, a highly purposive journey aimed at converting
--,configured) needs and desires of human beings; and, second, the pulse of
"savages" into pious peasants and citizens of Christendom. At another, it
collective forces that, empowered in complex ways, work through them.
participated in seeding a pervasive new order that would, along with other
This distinction informs the analysis of all historical processes, but its
colonizing forces, make Mricans into impoverished, subordinated subjects
significance has been underlined by the humanist turn in social science,
which has led to calls for a greater concern with agency. Its salience of empire. At times, these two levels reinforced one another, at times they
produced nightmarish disjunctions and discontinuities. We have argued
becomes particularly visible when we examine epochal movements like
elsewhere (n.d.[b]), in fact, that it was in the space between the liberal
European colonialism, in which purposive, "heroic" action was a central
world view of the mission and the racist world of settler society that modern
motif, even a driving impulse. Yet, from our perspective, that impulse is
black nationalist consciousness was to take root. In the longer run, as this
not enough to account for the determination of the processes involved­
or even to tell very much of the story. Witness, once again, the imperial suggests, the implications of evangelical imperialism were to be fixed by
the wider context in which it was embedded, just as they were to be
mission, an initiative moved by contradictory forces whose consequences
mediated by the responses of the Tswana themselves (see Part 3).
differed radically from the stated motives of those involved. Although they
And the general methodological point? There are several. The first is
were eminently effective in transforming local lives, the evangelists failed
that our current conceptual obsession with agency, subjectivity, and con­
precisely where they most hoped to succeed, namely, in implanting an
sciousness can be addressed only in ethnographic terms, and thereby rescued
orthodox Protestant peasantry on Mrican soil. from vapid theoreticism, under a pair of conditions: that (1) we treat as
Here, then, was a paradox of motivation, a paradox that ran to the heart
problematic the manner in which persons are formed and action determined
of the colonial encounter. While the mission spoke of itself and its inten­
and (2) we insist that individual action is never entirely reducible to social
tions in the language of Christian conversion, its practice proclaimed
forces, nor social forces to the sum of unique acts (above, p. 25)/Second,
something else. Motivated, silent and unseen, by the very situation of the
because it is multiply motivated, social history, as we have stres~ed, will
evangelists in the European scheme of things, this narrative told of the
always be both predictable yet subject to the innovative and the unforeseen.
reconstruction of a living culture by the infusion of alien signs and
Renee our historical ethnographies must be capable of capturing the
commodities into every domain of Tswana life. Methodologically, it com­
simultaneous unity and diversity of social processes, the incessant conver­
manded us to pursue the colonizing gesture beyond audible ideologies and
gence and divergence of prevailing forms of power and meaning. But they
visible institutions into the realm of such unspoken forms as bodies,
must do so without falling into the trap of typifying history in general­
• --I buildings, magic, and merchandise. And this, in turn, took us back to our
or histories in particular-as an expression of the radical contrast between
archives-to letters, lists, illustrations, and photographs-albeit now less
modern (or postmodern) worlds and their "traditional" antecedents. Or
for what they declared than for what they disclosed as maps of the mundane.
between commmoditized soci~Iies and natural economies.
It also prompted a cultural archaeology of the sites of earlier evangelical
It follows from all this,rthird, that our methodological concern is less
activity: for example, the windswept ruins of Tiger Kloof, a mission school
with events than with meaningful practices-which, perhaps, remains one
built for Southern Tswana early this century, where it was possible to
of the principle distinctions between historical anthropology and social
disinter, from the sediments of a dead community, aspects of colonial
history.31 Like most anthropologists, we are more preoccupied with ambIg­
pedagogy invisible in the written accounts. uous processes than with contained acts or isolable incidents that, in
The scattered signs retrieved in this quest all pointed to wider social
themselves, can be said to make a difference. To us, social life is continuous
transformations borne unwittingly by the missionaries. In many respects,
activity-activity that, because it is always a product of complex experience
these actually ran counter to their own desires and motives. For the
 38 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination
39
. and contradictory conditions, simultaneously reproduces and transforms home, ideals and dreams especially compelling to those at the margins of
the world. It will be evident, too, that we take meaning to be largely, if the rising bourgeoisie. And their reception by the often bemused Mricans
not entirely, implicit in practice; we do not see it to reside in abstract was determined, in large part, by the predicament of these peoples in a
schemes or in categories that endure or change in all-or-none fashion. From fraught, rapidly transforming political arena.
this perspective, history involves a sedimentation of micropractices into The incorporation of "the" Tswana into the colonial world, as we have
macroprocesses, a prosaic rather than a portentous affair in which events
pointed out, was a drawn-out process involving two dynamic social sys­
mark rather than make the flow of existence (d. Cohn 1987:45). This is tems, two historical orders, each with its own indeterminacies and internal
not to deny the importance of extraordinary human agency. Some acts do contradictions. The players in this theater of the ordinary changed one
have more consequence than others and, in certain contexts, actors can another by means of humble acts within the terrain they came to share­
become metonyms of history or, more accurately, of heroic-history-in-the­ although their behavior also moved, increasingly and in ways barely real­
making (Sahlins 1985:35f). But it is this metonymy, some would even say ized, to the beat of global imperatives. The pulse of these processes may
fetishism, that we have to explain. How is it that particular persons and be discerned, as we show in Chapter 9, in everyday struggles over such
events seem, in their own worlds, to embody and motivate processes whose things as agricultural technique, language and speech, the use ofland and
origins we, from our standpoint, ascribe to more dispersed causes? Heroes water, and modes of healing, each small thing summoning up a hinterland
are born not of gods, but of social forces. Their charisma camouflages of signs and practices. The plough, for instance, seemed an innocent
complex conditions of possibility, just as it personifies ambiguously au­ enough instrument. In this context, however, it bore within it the whole
thored action. culture of commodity production and turned out to have enormously
In the "newer social history," says Davis (1990:28), events serve less as complex social consequences. It is not that the historical encounter between
motors of change than to exemplify the mingling of the prescribed and the the evangelists and the Africans was un-event-ful. There were many notable
contingent 32 and/or to reveal the effect of cultural form upon social episodes: epic first meetings, dramatic demonstrations of "miraculous"
processes. This approach does not so much "nullify" the event-as did an technology, acrimonious public arguments. Clearly, these made a differ­
earlier structuralist history (d. Sahlins 1990:39; also above)-as resituate ence. But they did not make the difference. Nor, in themselves, did they
it in an unfolding sequence of action. It also democratizes human agency occasion moments of great rupture, cataclysms that led to the reconstruc­
by shifting attention from the subjectivity of big men to the force of tion of otherwise unchanging social systems. They were, rather, significant
communal projects and cultural practices. This entails a move, to cite Davis icons of-and elements in-an unfolding, multi leveled engagement be­
(1990:28) again, from such major episodes as wars and revolutions to tween Worlds.
processes of domination, representation, and resistance. Recall our earlier Insofar as global systems and epochal movements always root themselves
discussion: in particular, Samuel's plea for the significance of the Married
somewhere in the quotidian, then, they are accessible to historical ethnog­
Women's Property Act over the Battle of Trafalgar in shaping nineteenth­ raphy. In Africa, as elsewhere, the colonial "state" was both a political
century Britain. The former has been neglected because it was the product structure q,nd a condition of being; hence the former (its institutional order
of diffuse conflict and long-term collective action; it is not reducible, except of governance) might be interrogated through the latter (the routines and
~. in the most banal sense, to an event. Yet, although the impact of the Act­ habits oriented toward it). Similarly, the body politic and the body personal
t "I, ~ j
the changing nature of property, womanhood, and marriage-calls for a are everywhere intimately related-so much so that their connection has
processual perspective, its history remains rich with agency, some of it become almost a truism. Yet, we would suggest, the human body-or,
even "heroic." more precisely, its analytic use and abuse-provides a nice commentary on
Much the same may be said of the revolution that occurred when the the interpretive methodology of which we speak.
forces of European imperialism sought to insinuate themselves into the While the body has long been an important construct in Western social
non-European world, giving rise to the double context-the global stage ' thought (Durkheim 1947:115-116; Mauss 1973), it has recently gained
and the local mise en scene-in which all "Third World" ethnographies extraordinary prominence in the discourse of the human sciences. Perhaps
would later be done. Colonial history does not lack for heroes or events. because critical postmodernism has challenged fixed notions of power and
But neither is it reducible to a series of fortuitous encounters or fateful meaning, it has assumed a unique concreteness; it is, to be sure, something
actions. As pilgrims to the South Mrican "wilderness," the Nonconformist on which we can always lay hands. For that reason, it has been treated as
evangelists were moved by humanitarian ideals and imperial dreams at one of the only permanent points in a shifting world, especially by Foucault
 40 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
41
and his followers. The human body, in short, has been fetishized. And,
We might anticipate, then, that those who seek to forge empires, or to
like all fetishes, it is given credit for animating social life, yet it is strangely
remake existing worlds, will try to impress themselves upon the physiques
elusive-notably so in recent writing on the topic. Often no more than an
of their would-be subjects. States old and new have built their esprit de
alibi, a site, for equally elusive constructs like the "person," "the subject,"
corps by shaving, clothing, vaccinating, and counting their citizens, just as
and "social experience," it is named only to be dismissed; thus we
rising classes, ethnic groups, religious movements, and political associations
" perpetuate what Corrigan (1988:371) terms the "Great Erasure of the
tend to wear their self-awareness on their skin. For their part, conquerors
Body," long characteristic of Western scholarly discourse. (How different
and colonizers seem typically to feel a need to reverse prior corporeal signs:
this is from the frank sensuality with which some creative writers, from D.
often making bodies into realms of contest. The ancient English subdued
H. Lawrence to Toni Morri£On, have expressed their opposition to estab­
recalcitrant Scottish highlanders by cutting their hair and banning their
lished conventionsl) A notable instance of this absent presence occurs in
kilts (Brain 1979: 150); their descendants in Africa would attempt to force
Bryan Turner's The Body and Society (1984), which uses corporality as the Tswana converts into the dress of Christian decency.
ostensible focus for "explorations in social theory." In his journey through
Such tangible processes are eminently susceptible to the kind of ethno-;
issues of selfhood, sexuality, and social order, Turner seldom confronts
graphic scrutiny that may divulge the hidden hand of history. Take our
physicality at all (see T. Turner 1986). Displaced by the text, by a concern
colonial evangelists once again: While they talked of spiritual verities that
with representation severed from material being, the body actually loses
disparaged the flesh and condemned Mricans for their "carnal" ways, their
all social relevance.
actions displayed an intense interest in corporeal politics. The black body
None of this is new. Admittedly, in the great dialectic of the "social"
was seldom far from their thoughts or deeds, disrupting their rhetoric
and the "natural," a classical concern of social theory, the body has long
when least expected. As our encounter with Foucault, Derrida, and Ginz­
been seen as quintessential raw material for collective representation (see
burg would lead us to expect, these disruptions yield vital dues. A close
Chapter 3; Durkheim 1947:115f). Still, in striving to demonstrate the sui reading of the churchmen's diaries and records proves that body work­
generis quality of society and culture, scholars have repeatedly treated the the effort to retune the physical registers of dark persons through groom­
human physique as a tabula rasa, plastic material to be formed by arbitrary
ing, dress, and comportment-was a crucial mode of colonial production.
semantic categories (cf. van Gennep 1960; Douglas 1970; Bourdieu 1977).
This was one of the basic methods implicit in the mission, an unremarked
Poststructuralist and deconstructionist writers have perpetuated this form
means by which the Christians hoped to create a new moral empire. By
of idealism. Outside of discourse or the splintering subject or the floating deciphering the small print of letters, requisitions, and reports from the
sign there is, for them, no enduring object world. Rejecting aU traffic with
field, as well as the inventories of local merchants, we were able to trace
reality as brute "positivism"-as a matter of physical properties imposing the paths ofdiverse goods and practices converging on the Mrican anatomy.
themselves on passive subjects (T. Turner 1990:1O)-they are unreceptive Again, this is an instance of a universal process (Comaroff and Comaroff
to the idea that material facts have any role at aU in human experience.
1991:19f): No technique was too trivial, no mannerism too meaningless
'" Yet there is undeniable evidence that biological contingencies constrain to be drawn up into the sweep of history-in-the-making.

human perception and social practice, albeit in ways mediated by cultural

In their campaign to domesticate the black body, moreover, the colon­

forms (see Sahlins 1976b; Chapter 3). And this is the point: Precisely izers intervened in "native" cooking, hygiene, sexuality, and work. Wher­

because history is a synthesis of the heterogenous, we cannot ignore the eVer they could, they set about breaking the "communistic" interdepen­

role in it of such culturally mediated materialities. These, in turn, find dence of Mrican persons and productive processes, thereby to create a
their prime instance in the body, the physical object that also becomes a world of "free" individuals; free, that is, to consume and be consumed by
social subject (T. Turner 1990: 1). It is here, where physical facts meet European commodities. We discern this process most clearly in the ex­
social values, that collective modes of being emerge as dispositions or panding stock of objects (pots, fabrics, soap, tools, docks, locks, and so
motives. That is why movements of social reform, whatever they do at the on) that the whites saw as essential implements of modernity and progress.
level of collective institutions, tend also to work on the body as fans et These objects moved along the prosaic pathways that bore the traffic of
origo of the world (below, Chapter 3). Hegemony, at least in the cultural global capitalism and its culture to Southern Africa-and carried their
;sense we give it, has its natural habitat in the human frame. As a result, recipients toward material dependency. Abroad, as at home (see Chapter
that frame can never be a struggle-free zone, least of all when major 10), civilizing goods ushered in new orders of relations-relations both
historical shifts are under way. symbolic and substantial-that bound local consumers to an expanding
 42 THEORY. ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
world order. Such were the fragments of which novel totalities were being 43
constructed. The mundane practices to which they gave rise speak coher­ the capillaries of a full-blooded imperialism, capillaries that ran from the
ently to us from the ethnographic record. Together they weave compelling palace gates to the "mud huts" on the colonial frontier. In the late
narratives of a world historical movement and its many local variants, each nineteenth century, evangelical effort Was increasingly superseded by the
different in critical respects. work of civic-minded professionals (like the engineers and dOctors of the
Such symbolic processes, we stress, are not limited to colonizing mo­ Domestic Sanitation Movement; Adams 1991) and by the rise of state
ments. The making of what we term modernity in Europe can be read as schooling. Ultimately, any anthropology of the bourgeois revolution wiH
much in the evolution of table manners, sanitation, or the passport photo­ have to explore how homemade hegemonies played into such national (and
graph (Elias 1978; Jephson 1907; Fussell 1980) as in the development of nationalist) projects. But that is a topic for another place (Comaroff and
Comaroff n.d.[aJ: Chaps. 4, 5).
formal state institutions. Body work also had its parallels in the realm of
architecture and domestic space: Rybczynski (1986), for example, finds a More immediately, as we have remarked, the effort to colonize bodies
hidden history of the bourgeoisie in the rise of the modern European sense and buildings did not go unchallenged. In South Mrica, indigenous rulers
of "home." And the relentless social engineering of twentieth-century at first resisted the Nonconformists' gentle persuasion. They seemed alive
totalitarian states, whether they be in Eastern Europe or Southern Mrica, to the fact that the white men's designs on their people were anything but
is nowhere more clearly revealed than in their oppressively uniform public trivial. Later, many Tswana would rework those designs into provocative
housing. patterns, giving free reign to an independent, often subversive imagination.
We ourselves draw on these insights in Chapter 10. There we explore In London likewise, Cockney costermongers, poor street traders, ignored
the remarkable similarity between, on the one hand, efforts of colonizers / middle-class moralism and fashioned flamboyant life-styles of their own
to reshape the habits and habitations of nineteenth-century Mricans and, (Mayhew 1851,1). We should learn from them. For costers, chiefs, and
on the other, the apparently unrelated attempt back home to "improve" churchmen alike appear to have sensed that it is things like clothing that
the domestic lives of the urban underclass. Evangelists in Britain and make subjects-again, in both senses of the term. The Tswana "style
Bechuanaland expressed the same conviction: that "uncouth" populations wars," in which local leaders tried to fight off Western dress and architec­
could be tamed through the orderly deployment of windows and walls, ture, were as much the site ofcolonial politics as were formal confrontations
soap and sanitation, locks and lamps. How are we to interpret this coinci­ with government personnel or settler statesmen. In the fantastic fashions
dence? Was it a co-incidence? The answer, once more, lay in drawing that flourished on the frontier we catch a glimpse of the consciousness of
fragments together and situating them within a wider historical field, thus ordinary Africans, those who left little other imprint on the historical
to make sense of the embracing totality of which they were part. By tracing record. Here, along the line that divided the increasingly marked domains,
out the imaginative linkages among disparate texts and tropes, we were the ideological spaces, of "tradition" and "modernity" they made new
able to see that these seemingly independent instances of domestic reform identities by retooling old values, redeploying the very signs that the
were complementary sides of one process; indeed, that colonialism was as colonizers imprinted on the supple surfaces of their lives.
much a movement of re-formation within British society as it was a global Reading these poetic practices is by no means straightforward. Because'

gesture; that each site, the sickening England slum and the bestial African bodies and domestic space were vital terrains ofcolonization, the struggles

• ';j'l
bush, became a model of and for the "other"; that this whole process was that occurred around them exhibited all the complexities of the colonial

the political expression of a universalizing hegemony, a push to rebuild process itself-all the multiple motivations, the indeterminacies, and inter­

"savage life" on both continents to the specifications of bourgeois en­ nal Contradictions of complex historical conjunctures everywhere. Great

lightenment. social movements seem always to achieve both more and less than intended.
In both contexts the process would succeed, above all else, in standard­ For, even as would-be subject populations take issue with the manifest'
izing an aesthetics of class distinction; an architecture of othering for the messages and overtures that intrude upon them, they often internalize
metropole as well as the colony. Such discourse-and the philanthropic alien culturalforms along the way-without either knowing or meaning to
practice it empowered-stressed the morality of properly inhabited space: do so. That, as we show in Chapter 9, is why new hegemonies may take
In a world driven by property and propriety, the home was heavily invested root amidst ideological argument, why people may be deeply affected by<
with elemental values, framing middle-class images of personhood, pro­ the media that bear the messages they reject, why such processes are nevet'
duction, sexuality, and gender. Bodies, houses, and everyday routines bore redUcible to a simple calculus of accommodation or resistance. Thus even
those Tswana who most strenuously refused the dress of baptism, prefer­
 44 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Ethnography and the Historical Imagination
45
ring to pick and choose what they fancied from the miSSion, were pro­
foundly changed by the world of commodities admitted with these inno­ warriors" of Mozambique compel us to consider our wanton ambivalences,
cent objects. and so to reflect upon the way in which we ourselves reflect on others.
By the end of the nineteenth century, black identities in South Africa Such reflections persuade us that the conundrum of similarity and
were being shaped less by either indigenous or mission intentions than by difference is only to be resolved by turning anthropology on itself, by
the gathering forces of the colonial state. Whatever were their local mean­ treating modernity (and postmodernity) as a problem in historical ethnog­
ings, bodies, dress, and "life-style" were made over into signs of gross raphy. For the malignancy of primitivism-and its most notable symptom,
difference; into the distinctions of race, gender, and culture by which exoticism-should disappear when we estrange our own culture, treating
Africans were being incorporated into the lowest reaches of a rising its signs and practices as we would theirs. This is not a call for rewriting
industrial society. There is a general point here, and a concluding one. Far all anthropology as "We the Nacirema" (Miner 1956) or for making all the
from being primordial, "ethnicity," "tribalism," and other forms of iden­ world into an imaginary village. The purpose of estrangement, rather, is to
tity reside in tangible practices-as, of course, does "modernity." They are remind ourselves that the West and the rest, long locked in historical
the social and ideological products of particular processes, of the very embrace, cannot but be interrogated together. This, then, is our challenge.
conjunctures that set the terms of, and relations between, "local" and It is to explain the great conjunctures, the processes and practices through
"global" worlds. Such phenomena, we have argued, are not to be treated which have been fashioned the significant social phenomena of our times,
as received categories or analytic objects conjured up as universals from both global and local.
our own folk sociology. They are both polymorphous and perverse. Our These are issues of broad concern within the discipline at present;
task is to establish how collective identities are constructed and take on historical anthropology, patently, is more than a Chicago-cult. In order to
their particular cultural content; how they are made real, essential, embod­ address them we have appealed to a neomodernist method that takes
ied qualities for those who live them; how they become the natural atoms seriously the message of critical postmodernism yet does not lose the
ofsocial existence. Only then will the diverse forms of the modern world­ possibility of social science; that takes to heart the lessons of cultural
indeed, the very terms of modernity itself-become the subjects of an Marxism, seeking a conception of culture that recognizes the reality of
ethnography of the historical imagination. power, yet does not reduce meaning to either utility or domination; that
builds on the techniques of cultural history, pursuing the dialectic of
fragment and totality without succumbing to brute empiricism; that, above
v all, proceeds, as it must, by grappling with the contradictions of its own
legacy, seeking to transcend them-if only provisionally and for the
moment.
And so we conclude our voyage into method. The journey began with the
Naparama-or, at least, their representation in the Western mass media. It
was they who confronted us with the paradoxes and ironies that propel this Notes
essay: that, for all our obsession with the effect of anthropology on the
• ';1 ~
F i "other," the discipline has had very limited impact on our own culture;
that, for all the efforts of generations of ethnographers, the radical oppo­ 1. Chicago Tribune, Sunday, 9 December 1990, Section 1, p. 1.
sition between prehistorical "tradition" and capitalist "modernity" sur­ 2. There is widespread evidence that this ideological opposition has continuing salience
vives in the discourses of our age, popular and professional alike. Indeed, in our Culture. Take just one example, a token of a very common type: In a review of the
sUccessful, well-intentioned film Dances With Wo/pes (1990; director, Kevin Costner), Dorris
in directing much of our attention to peoples on the other side of the great 19
( 91:17) notes that, even today, "Indians embody the concept of 'the other'-a foreign,
rift, do we not still foster a lurking primitivism? And, with it, all the myths exotic, even cartoonish panorama against which modern (that is, white) men can measure
of our own disenchantment? The Naparama, in short, are a powerful and test themselves, and eventually ... be dubbed as natural leaders."
metonym of our scholarly predicament, a mirror in which we see ourselves 3. Note, however, that the move has not been without criticism, even in apparently
divided. They reveal our tendency, as a caustic critic once put it, to see receptive fields; see, for example, Johnson (1983). We return to this issue later.
people as everywhere the same except where they are different-and as 4. There are exceptions to this, especially in modern American anthropology. But they
everywhere different except where they are the same. In sum, the "mystic ire restricted to such relatively marginal areas as mathematical anthropology, cultural ecology,
and highly specialized forms of network analysis and economic anthropology.
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Ethnography and the Historical Imagination 47
46
5. Once more, the exceptions prove the general rule here. Although approaches like change" to otherwise synchronic ethnographies. These, typically, were no more than narrative
ethnoscience and mathematical and cognitive antbropology have called for new methods and dumping grounds for everything that escaped the deadening vision of the descriptive present.
theories, they have made little lasting impact on the practices of the discipline as a whole. 20. A few years later Smith (1960) published the more ambitious, more theoretically
6. For an insightful exploration of the tropes that ethnographic writing shares with the sophisticated G01Jernment in Zazzau, which covered a span of 150 years. In contrast to Barnes
(1954), his object was less to disclose the logic of stasis than to arrive at the causes of change.
earlier genre of travel writing, see Pratt (1986).
7. The evidence for this is everywhere at hand, from Evans-Pritchard's curt reminder that
In historiographic terms, however, Smith's procedure remained aggregative in spirit. Events
his facts were selected in light of his theories (1940:261), through Leach's (1954:5f) insistence
and relations were distilled into a generalized account of a political system that persis ted, as
if in equilibrium, until ruptured (by one of the forces specified in a set of abstract "laws";
that ethnographic accounts of social systems, like native models, are merely "as if" construc­

tions of the world, to Geertz's (1973:29) allegorical suggestion that culrural analysis is a
Smith 1960:Chap. 8).
2!. Our irony will be dear to those familiar with the distinctly ahistorical debates during
matter of "turtles all the way down." the 1950s and 1960s over prescriptive marriage systems and, more generally, over alliance
8. Marcus (1986:190-191) adds that "experimental" ethnographers "perhaps do not

theory. For a sample of the issues involved, in the words of the protagonists, see Needham
even recognize the priority or privileged validity of such abstractly represented realities

(1962), Leach (1951), Homans and Schneider (1955), and Levi-Strauss (1969, especially the
statistics]." Yet, in sustaining the opposition between "evocation" and "representation," he

preface to the second edition).

himself perpetuates a straw man: a "realist ethnography," whose "holistic commitments"

22. The notion that Tswana inheritance and succession are governed by the ascriptive
defy the "open-ended mystery" of experience, and ostensibly the possibility of "alternative

principle of primogeniture goes back to missionary ethnographies, although it is often

explanation." attributed to the classic writings of Schapera (e.g., 1938). Social scientists have reiterated it,
9. Nonanthropologists might wish to know that Raymond Firth was a senior professor of usually without question, ever since. The most recent to do so are Crowder, Parson, and
social anthropology at the London School of Economics, then about to retire. A very Parsons (1990:12f), who take issue with our early work on the topic. This is not the place to
distinguished scholar, he did much of his ethnographic research on the island of Tikopia in rebut their argument-which is based partly on a misrepresentation of our analysis and partly
Polynesia. on a curiously ethnocentric, culturally barren interpretation of the historical record (see also
10. For an especially clear example, see Marcus (1986:191). J. L. Comaroff 1990:561, n.14 for brief comment). Indeed, theirs is the kind of account that
11. See, for example, J. L. Comaroff (1982:143f), from which the quote is drawn. makes it clear why history needs anthropology every bit as much as anthropology needs
12. The phrase is from Ginzburg'S (1980:xx-xxi) brief but acid comment on those history.
historians who, like Fran~ois Furet, have found panaceas for large-scale problems in demo­ 23. The same general point has been made in a number of discourses on the human
graphic sociology. sciences-most memorably, perhaps, in McCloskey'S (1985) study of the rhetoric of econom­
13. Damton (1985:3) points out that there is no standard English translation ofl'bistoire ics.
des mentalites, until recently a predominantly French historiographical movement (see, e.g.; 24. It should be clear that we do not use the term "statistical" here in its narrow, purely
Vovelle 1990). Darnton himself suggests that "it might simply be called cultural history." It numerical sense. We mean it [0 refer, generically, to any inference of prevailing pattern or
is difficult not to agree; as Ginzburg (19S0: xiv-xv, xxii-xxiv) indicates, the final object of a probability derived from past rates of occurrence.
history of mentalities is an account of a particular ("popular") culture. 25. Leach does not phrase the implications of his analysis in the terms that follow.
14. Hartley's original phrasing, in the first sentence of the prologue to The Go-Between However, they flow from his comments on the nature of social change and history (see, e.g.,
(1956), was "The past is a foreign country: they do things differently there." 1954:212,228ff).
15. There is an irony lurking here. As Derrida (1978:35) notes, Foucault (1967) himself 26. This has not been for want of trying, of course. Sahlins (1985), for one, has argued
speaks much of silence in his History, especially when he situates madness within the trap of cogently for a structuralist historical anthropology. But his exertions have not gone unchal­
Western reason (not to mention the repressive language of psychiatry). Indeed, concludes lenged.
Derrida, "[Foucault's] history of madness itself is .. the archeology of a silence." 27. See, most notably, Sahlins (1990). However, it is not only cultural structuralism that
16. The comment is made in the specific context of his discussion of Febvre's treatment continues to struggle with the individual and the event. Structural Marxism has had similar
of Rabelais. problems. Ikcall, for example, the debates surrounding Althusser's portrayal of history as "a
17. Croce ([1921] 1959:51) contrasts chronicle with history, treating them as different process without a subject" or, in anthropology, Hindess and Hirst's (1975:45f, 78) claim
"spiritual attitudes. History is living chronicle, chronicle is dead history; . . . history is that Meillassoux's (1964; 1972) accounts ofGuro political economy owe less to Marx than
principally an act of thought, chronicle an act of will." It is this last phrase-the image of an to methodological individualism.
act of will-that we seek to emphasize here.
28. This passage on biography and the diary is excerpted, in amended form, from J. L.
IS. We pass glancingly over this terrain in Chapter 4.
Comaroff (1990).
19. Some might say that there was also a fourth, exemplified by Cunnison (1959; cf. also 29. Ikcent writings suggest that such conventions might be shifting at last; for some
Barnes 1951), which explored "what [people] made of their history" in the course of their diverse examples, see Martin (1987); Lave (1988); McCracken (1988); Spitulnik (1991).
social lives. However, such studies tended to limit themselves to the role of ethnic his[Orical This has been a result, in part, of the influence of cultural studies, a relatively new discipline
consciousness in repetitive social processes; most were produced within the "custom and that has challenged us by applying some of our own concepts and methods to Western
conflict" approach of the Manchester School. As a result, their methodological bases were nO phenomena (see, e.g., Hall, Jefferson, and Roberts 1976; Willis 1977; Hebdige 1979,1988).
different from those we will discuss. In this respect, [00, we do not dignify as history an old But it has also been helped along by a more general erosion of the boundaries between the
practice of British structural functionalism: the appending of residual chapters on "social human sciences.
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
48
30. These models, many of them originating in linguistics, come from a variety of sources,

_.­
ranging from orthodox structuralism to Jakobsonian pragmatics and Bakhtinian dialogics.
Whereas the first implied a static conception of culture-a conception now heavily under
attack-both the second and third inform current concerns with the practical, political,
ambiguous, and transformative qualities of meaning.
31. See Comaroff and Comaroff (199l:34f). This difference is often not acknowledged
by social historians, who sometimes fault historical anthropologists for not writing "real"
2
histories; that is, detailed chronicles of events (see, e.g., Shillington 1987).
32. This is what Sahlins (1990:47), after Ricoeur, terms a "synthesis of the heteroge­

ncous. n
Of Totemism and Ethnicity

well-known in some quarters, about

T HERE IS A SOCRATIC PARABLE,
a teacher who gives his students two magnifying glasses and invites
them to look at the one through the other. When each has told of all
he has learned, the sage delivers his lesson in the form of a question, a
coup de grace:
"Of what have you told me," he asks, "the thing you have seen or the thing
through which you have seen it?"

The same conundrum lurks, usually unremarked, behind the study of

ethnicity. Is the latter an object of analysis, something to be explained? Or
is it an explanatory principle capable of illuminating significant aspects of
human existence? Does it really refer to "idols of the tribe" (Isaacs 1975),
or is it in fact an idol of the scribe (Mafeje 1971)? It certainly has been
treated in both ways, sometimes simultaneously. As a result, there is still a
notable lack of agreement on even the most fundamental of issues: \YhaL
is ethnicity? Is it a monothetic or a polythetic class of phenomena, 0i!e
~mg- or-nf.ftfy1 Has it the capacity to derermine-sociai-activity;or IS it a
product of .other forces and structures? Do its roots lie in so-called
primordial consciousness or in a reaction to particular historical circum­
stance? And how is it related to race, class, and nationalism? In addressing
I "j,~ . these questions, we shall use a wide-angle lens rather than a magnifying
glass, and shall focus it, somewhat eclectically, on various Mrican contexts.
In so doing, moreover, we seek deliberately to turn the sage's moral on its
head. For we are concerned to examine, at once, both an analytic object
and its conceptual subject: on one hand, those processes involving the rise
of ethnic consciousness in Mrica and elsewhere, and, on the other, the
theoretical terms by means of which ethnicity may itself be comprehended.
Contrary to the usual canons of scholarly enquiry, we proceed not by
situating our discussion within the relevant literature, but by stating five
propositions about the nature of ethnicity. These propositions, though, are

49
 50 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Of Totemism and Ethnicity
51
not presented in axiomatic form; rather, they are developed and exemplified
lenged, goes the argument, ethnic sensibilities either do not exist or remain
as cumulative, if yet tentative, steps in pursuit of an analytic position dormant. 3
capable of accounting for the genesis, persistence, and transformation of Both these theses, we believe, are simultaneously correct and incomplete.
ethnicity and ethnic consciousness. This is not to pretend that any of them, On the one hand, the precapitalist world, where primordial affiliations and
regarded in isolation, is necessarily new-although each flies in the face of loyalties are tacitly presumed to have had their origins, was never so
some, and sometimes most, received wisdom. Whatever their individual atomistic that communities did not have relations with others. And, in so
provenance, however, their theoretical significance lies more in the system­ far as this is true, it would be plainly absurd to pretend that their members
atic relations among them than in the substance of each in its own right. could have lacked common identities or a concern for sociocultural differ­
ences; consider, for example, the acute awareness that the diverse peoples
of the Luapula valley are known to have had for each others' "customs"
(Cunnison 1959:53-61). On the other hand, this form of awareness is
distinctly different from ethnic consciousness suigeneris (Skinner 1978: 193).
The first proposition is intended primarily as a point of departure, a very Let us elaborate. In as much as collective social identity always entails
general statement of orientation toward the conception of ethnicity: Con­ some form of communal self-definition, it is invariably founded on a
trary to the tendency, in the "JIVttberian tradition, to view it as a function of 'marked opposition between "ourselves" and "other/s"; identity, that is,
primordial ties, 1 ethnicity always has its genesis in specific historical forces, is a relation inscribed in culture. Patently, the social and material bounda­
forces which are simultaneously structural and cultural. The corollary of this -ries involved in any such relations-not to mention their content-are
proposition ought also to be underscored. If it is true that ethnicity is a historically wrought; they change in the course of economic and political
product of particular historical conditions (cf. Wallerstein [1972] processes [see below]. Still, whatever the substance of particular relations
1979:Chaps.10, 11), and not an ontological feature ofhuman organization, between groupings, the irreducible fact of identity implies the cultural
it follows that it cannot be treated as a truly "independent" explanatory structuring of the social universe. All this merely echoes the anthropolog­
principle, a "first cause" in and of itself (Moerman 1968:160f).2 This is ical truism, after Durkheim and Mauss (1963), that dassifu::ation,-the
not to deny either its reality or the fact that action is regularly conducted meaningful construction of the world, is a necessary condition of social
in its name. Nor is it to ignore that such action has concrete implications existence. But, we stress, it is the marking of relations-of identities in
opposition to one another-that is "primordial," not the substance of
for everyday relations [see below, p. 60]. It is, rather, to recognize that, in those identities.
order to understand ethnicity at all, we have not merely to reveal the
conditions of its genesis, but also to establish its place in the sociological This, it will be recalled, was the point that Bergson (1935:172-5) made
about the essence of totem ism, the point upon which Levi-Strauss (1963b)
chain of being. It is to these two problems that much of this essay is
was to build his thesis. The genius of Bergson's insight lay in the observa­
addressed; for it is only by resolving them that the first proposition-and
tion that it was not the intrinsic nature of totemic objects ("their animal­
the overall position to which it speaks-may finally be sustained.
ity") that gave them their significance. It was, rather, "their duality"; the
It is instructive to begin with the long-standing contention that the
fact that relations between these objects stood for relations between social
-t 'ii, ~ • roots of ethnicity lie in the original "fact" of human cultural difference
groups (1935:175). Totemism, then, is just one form-and, it is to be
and ascribed status group affiliations; after all, the tenacity of the "primor­
added, an historically specific form-of the universaL process of dassifica.­
dial" thesis is itself significant. This thesis rests, in large part, on the
compound notion (i) that culturally defined communities-or, in Weber­
~iC!n. It is one in which groupings define themselves as independent or
~nterdependent units within a common humanity; formulate collective
ian terms, "status groups" (stande )-everywhere entertain an intrinsic
identities in cpntrast to one another; and portray themselves and others, in
awareness of their own identity; (ii) that the traditional loyalties vested in
~mbolic terms, as similar yet different. Whether such relations are signified
this identity are the source of ethnic consciousness and affiliation; and (iii)
tn animate or inanimate objects makes little difference. These are the media
that the latter provide the basis for collective action and intergroup rela­ of totemic conscioumess, a particular species of conscience collective.
tions. The counter thesis, of course, suggests that expressions of ethnicity
Ethnic consciousness also entails the formulation of collective identities
do not arise in any community save as a reaction to threats against its and their symbolic embodiment in markers of contrast between social
integrity and self-determination. As long as that integrity remains unchal- groupings. For ethnicity, like totemism, exists above all else as a set of
 52 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Of Totemism and Ethnicity
53
relations. In this respect, they are formally similar. But they differ clearly
in their substance. However the former is defined-and, to be sure, it has ago, this perc~ption persists.. Indeed, he tells an anecdot~ that makes the
been defined in very many ways (see, for example, R. Cohen 1978)-it point with frIghtemng clanty. It concerns an elderly mformant who,
appears to have two generally recognized and closely related properties. though deeply religious, had developed a quixotic theory as to why the
bible should not be taken literally:
One refers to the subjective classification, by the members of a society, of
r the world into social entities according to cultural differences. The other [If you read it literally ...] "you Wouldn't be able to explain how the Black
j involves the stereotypic assignment of these groupings-often hierarchi­ man got here," he said. "They come from baboons. That's what evolution
cally-to niches within the social division of labor.4 Neither property, of has taught us, and the Bible doesn't say anything about evolution. God
course, is unique to ethnic consciousness; the first applies equally to created the White man in a day. That, the Bible teHs us. It took evolution to
totemism, the second to class. But it is in their fusion that the particular create the Black man."
character ofethnicity resides [below, p. 54]. Moreover, it is not coincidental
that these features, among all those associated with ethnic consciousness, For the subordinate, ethnic affiliation may originate in an attribution of
are the most commonly observed: for reasons which will become evident, collective identity to them on the part of others. On occasion, as we shaH
they reflect the manner in which the forces that yield an ethnically ordered see, the creation of such identities has little foundation in pre-existing
universe impress themselves upon human experience. sociological reality, in which circumstances it usually involves what has
been termed "the invention of tradition" (Hobsbawm and Ranger 1983).
But even where they have had a sodal identity contrived for them, sub­
ordinate groupings typically come to define their "ethnicity" as an emblem
"
There is a good deal more to ethnicity than this, however. Not only may

of common predicament and interest; through it, too, they may begin to
assert a shared commitment to an order of symbols and meanings and,
its character change over time-which is one reason why it is so resistant
sometimes, a moral code (Moerman 1968). This, moreover, is often
to easy definition (Hechter 1975:311 )-but the way in which it is experi­
expressed in the reciprocal negation of the humanity of those who dominate
enced and expressed may vary among social groupings according to their
them. Those same Kavango and Ovambo mineworkers, for instance, repay
positions in a prevailing structure of power relations. For dominant group­
their degradation by referring to the Europeans as "barbarians" (R. Gordon
ings-be they Mrikaners, self-styled chosen people of South Mrica (Adams
1978:216). Similarly, the Tswana vernacular for "whites," makgoa belongs
J

and Giliomee 1979; Coetzee 1978:249), or precolonial Alur, spreading

to a class of nouns (sing. prefix: leo; pI. prefix: ma-) reserved for non-human
their particular concept of chiefship over the hinterland of Uganda (South­
animate objects and human pests; it includes such terms as legodu (thief)
all 1956:181)-it takes on the assertive stamp of a protectionist ideology;
and letagwa (drunkard). The word makgoa itself originally denoted the
a legitimation of control over economy and society. Concomitantly, it
"white bush lice" associated with the hindquarters of large animals (1.
involves the negation of similar entitlements to others, often on putative
Comaroff 1985:137). In sum, then, ethni0Q~nt~~J.l.JoY!Iich always assumes
cultural or «dvilizational" grounds, and may call into doubt their shared
both an experiential and a practical saIifnce- forthose who bear it, entails
humanity. Thus Robert Gordon (1978:215f) notes that Kavango
and the complementary assertion of the collective self and the negation -of the !
• >~, ~ i
Ovambo mine workers in Namibia believe, correctly, that their white su­
collective other; it may 9l! into question shareq. hu~; and its sub- I
pervisors «think we are animals." And nineteenth century Tswana saW
Stance is likely to reflefuhe tensl~ns~mood~eaIn 'fclations of.~Ci1iality.l
Sarwa ("bushmen"), who peopled their underclass, as phologoloJ beasts
We shalrexptore~these substantIve features below. For now, It IS enough
the wild. Although these Sarwa allegedly controlled wondrous knowledge
to reiterate that, while ethnicity is quite different in its content from
of herbal substances, they were thought properly to live in the undomes­
tOtemism, there is, beneath this difference, a common denominator: both
ticated bush, and were only allowed into the towns of their masters by
are, ultimately, modes of social classification and consciousness, markers
night (Mackenzie 1871:368). Ironically, the Tswana, themselves being
?f identity and collective relations. Herein, in fact, lies the point of
subjected increasingly to the domination of Boer (Mrikaner) settlers, had
Interrogating them together. For it calls into question the roots of the
become known to the whites as skepselsJ "creatures" (cf. Marais 1939,
Contrast and, by extension, the particularity of ethnicity: why should the
quoted in Crapanzano 1985:31; also Livingstone 1857:37). According to
primordial fact of social classification, of the consciousness of identity and
Crapanzano (1985:40), who did fieldwork among Afrikaners some years
distinction, take on such diverse forms? The easy answer, of course, would
be an evolutionary one: that totemism is the precursor of ethnidty, the
 Of Totemism and Ethnicity 55
THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
54
collective identities by contrast to one another. At times, such relations are
latter being a product of the movement from "simple" to more "complex" enshrined in cycles of exchange of varying kinds; at others, they may entail
social systems. But this explanation will not do. Just as it is demonstrable raiding, hostility and warfare; often they embrace both. Moreover, it is
that ethnic consciousness existed in precapitalist Africa [see p. 57 below], possible, in fact usual, that such encounters yield short-term inequalities
so, as Linton's classic ethnographic vignette proves (1924; see Levi-Strauss among the parties to them. But, if each retains its integrity and is not
1963b:7-8), totemic consciousness occurs in industrial societies. The subordinated as a group in and of itself-which is most likely where the
solution, we suggest, is to be sought elsewhere; namely, in the historically units bear formal structural similarity-the overall symmetry of relations
specific social contexts in which totemism and ethnicity, respectively, arise is perpetuated. And, as long as it is, totemic identities and affiliations, in
and persist. For the signs and practices involved in each have their source the general terms in which we have defined them, will be sustained-with
in the very construction of economy and society. the added qualification that the particular content of these identities will
Drawing all this together, then, the second proposition may be stated depend both on the sociocultural orders of those who bear them and on
thus: ethnicity, far from being a unitary ((thing,» describes both a set of the exact nature of their engagement with others.
relations and a mode of consciousness; moreover, its meaning and practical Totemic consciousness, and the processes that give rise to it, occurred
salience varies for different social groupings according to their positions in the widely in precapitalist Africa. It was realized, for instance, in relations
social order. But, as a form of consciousness, it is one among many-totemism among Sotho-Tswana chiefdoms, and between so-called acephalous soci­
being another-each of which is produced as particular historical structures eties such as the Nuer and Dinka in the Sudan (Evans-Pritchard 1956:82­
impinge themselves on human experience and condition social action. 4); to be sure, the ethnographic literature abounds with examples. It is
This proposition also has important corollaries, among them, the final true that, historically, totemic relations have also arisen in circumstances
repudiation of the "primordial" thesis. For, as the comparison of totemism where a political community has incorporated either immigrant or con­
and ethnicity has revealed, this thesis flows from a confounding not merely quered populations. Significantly, though, this seems to happen only when
of twO modes of consciousness, but also of quite separate levels of analysis. such populations are assimilated into an existing social order, not as a
We reiterate that the marking of contrasting identities-of the opposition subordinate class, but as units of structure like those already there. In such
of self and other, we and they-is "primordial" in the same sense that situations, a form of totemic pluralism tends to emerge. The most striking
classification is a necessary condition of social existence. But the way in case of this is provided by the Tswana merafe ("nations"), which regularly 1;
which social classification is realized in specific forms of collective identity, absorbed groupings from outside. These incoming groupings, whose to-I
ethnicity no less than any other, is always a matter to be decided by the
5
terns differed from those of the host chiefdoms, were generally established I
material and cultural exigencies of history. as integral and independent wards and sections, the major politico-residen- \
tial divisions within Tswana polities; their internal organization, distribu- \
rions of authority and productive arrangements became an indistinguish­
III able element of the structure in place (Schapera 1938; J. L. Comaroffl
1973). Consequently, Tswana chiefdoms usually contained a large number
This leads directly to the third proposition, which addresses those forces of totemic groupings whose identities in contrast to each other were marked·
that produce totemism and ethnicity, and their associated modes of con­ in. ritual contexts (Schapera 1952).6 As this implies, totemic consciousness)
~ '4 '. j
sciousness. In its most general form, it may be put as follows: while totemistn mIght also surface in the wake of conquest and in complex states, depend­
emerges with the establishment of symmetrical relations between structurally ing on the manner in which social units are ordered within the polity.
. similar social groupings-groupings which mayor may not come to be inte­ Where they are incorporated into symmetrical relations with other like
grated into one political community-ethnicity has its origins in the asymmetric units-as they were, say, in the "snowball state" of nineteenth century
incorporation of structurally dissimilar groupings into a single political econ-·· Ngoni (Barnes 1954)-the historical conditions for the production of
totemic consciousness may be expected to take their due course. 7
\ omy. More specifically, totemic consciousness arises with the interaction of The emergence of ethnic groups and the awakening of ethnic consciousi
I social units that retain-or appear from within to retain-control over the ness are, by contrast, the product of historical processes which structurd
• means of their own production and reproduction. It is, in short, a function relations of inequality between discrete social entities. They are, in otherl
of processes in which autonomous groupings enter into relations of Words, the social and cultural correlates of a specific mode of articulation
alence or complementary interdependence and, in so doing, fashion
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY Of Totemism and Ethnicity 57
56
1842:244; Burchell 1824,2:383)-so Tswana held an equally skewed
between groupings, in which one extends its dominance over another by
image of the culture of their nineteenth-century Sarwa serfs. This is yet
some form of coercion, violent or otherwise;8 situates the latter as a
more likely to occur when underdasses are composed of people of diverse
bounded unit in a dependent and unique position within an inclusive
origins: in such circumstances, the substance of their identities, as contrived
division of labor; and, by removing from it final control over the means of
from both within and outside, is inevitably a bricolage fashioned in the
production and/or reproduction, regulates the terms upon which value
very historical processes which underwrite their subordination.
may be extracted from it. By virtue of so doing, the dominant grouping
Ethnic consciousness, as we stated earlier, existed in precapitalist Mrica.
constitutes both itself and the subordinate population as classes; whatever
It was not purely the product of colonialism or urbanization, as has
the prior sociological character of these aggregations, they are, in the
sometimes been suggested, although both did (and do) contain the requi­
process, actualized as groups an sich. site conditions and have generally been accompanied by manifestations of
The creation of structured inequality of this kind, as we have already
ethnicity. Perhaps the most celebrated instance in the precolonial context
noted, demands meaningful signification; there can, after all, be no social
lay in Tutsi-Hutu relations in Rwanda and Burundi (Maquet 1961), the
division of labor without its representation in culture. And, in a stratified,
history of which conformed closely to the above characterization. But there
segmentary social environment, this entails marking out the social world
are many other examples (see Wallerstein [1972] 1979), the conquest and
into identifiable classes. 9 Now these representations are not arbitrary: since
subordination of the Bairu by the Bahima in Ankole providing another
they apprehend and rationalize the unequal distribution of material, polit­ immediately familiar one (Oberg 1940). Less well-known is the case of
ical and social power by virtue of group membership, they must, by the Betsileo of Malagasy, of whom Kottak (1980:4-5) writes:
definition, ascribe such inequalities to the intrinsic nature of the groups
concerned. It is thus that the "ascriptive" character of each becomes its Betsileo have not always shared ... consciousness of themselves as a distinct
"ethnic" identity-even though the groups themselves might only have ethnic unit. Prior to their conquest by the Merina, there appear to have been
been established, at least in their contemporary mold, in the very process no Betsileo. Rather, there were several statelets and chiefdoms located in
different parts of what is now the Betsileo homeland. Their conquerors ...
of articulation which contrived that identity in the first place.
created the Betsileo province of the Merina state . . . and, in so doing,
It follows that the identity imputed to a social group from the outside provided a basis for Betsileo ethnic consciousness to develop through the
may be quite different from that same identity as subjectively experienced. present.
For the construction of the collective self-and, by extension, any account­
ing of its entitlements (for the dominant) or predicament (for the subordi­ Kottak goes on to note (1980:303, n. 1) that Bara ethnicity also originated
nate)-depends on its differentiation from the collective other. "Other­ in the nineteenth century under very similar conditions (see Huntington
ness," then, becomes a contrivance in the counter image of social selfhood, 1974). Ethnic consciousness, it seems, has increased throughout Malagasy
not an empirical description of any particular population. Indeed, the as a function of political consolidation.
Janus-faced nature of ethnic consciousness-the fact that it involves both With colonial penetration into Africa, indigenous populations were
the assertion of a collective self and the negation of collective other/s-is integrated-to differing degrees and with varying rapidity-into the po­
a cultural expression of the structuring of inequality; it emerges as group­ litical economies of Europe, thereby becoming part of an increasingly
ings come to signify and symbolize their experience of a world of asym­ ~obal division of labor. And, as we might expect, their asymmetric
",j,", metrical "we-them" relations. lncorporation into a wider order of productive and exchange relations
The extent to which ethnic identities are rooted in previous cultural yielded novel ethnic affiliations and groupings (Wolf 1982:380f). As
realities, moreover, is highly variable (Young 1976:34). Apart from all else, Greenberg (1980:14) observes:
cultures tend to be transformed as the groups who bear them interact with The Kikuyu, for example, whose coherence is now so important to under­
one another (Sahlins 1981). But even if they were not, it is rarely the case, standing Kenyatta and nationalism in Kenya, had no certain identity before
for reasons just given, that they are accurately represented in the identity the imposition of British rule and the alienation of land to the settlers;
assigned to others-this being true even where emergent ethnic affiliations distinctive groups like the Sikhs in India, Ibo in Nigeria, and Malays in
are created out of previously bounded, homogeneous populations. Thus, Malaysia were barely conscious of their "sameness" one hundred years ago.
just as the agencies of British colonialism had little appreciation of African
But there were differences between the forms of ethnicity which had
social orders and acted upon distorted impressions of them-witness the
preVailed in the precolonial epoch and those bred in the wake of the
common view that black South Africans had no religion (e.g. Moffat
 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY 0/ Totemi.m and Ethnicity 59
58
such processes varies, of course. But the overall structure of nesting,
expanding world system. As has been argued elsewhere (Long 1984; J. L. opposed identitie~-:-of "tri~e," "?ation". and "race," each a particular
Comaroff 1982), the leitmotif of modern "Third World" history lies in a refraction of ethmClty-mamfests Itself with remarkable frequency.
mutually determining interaction, a "dialectic of articulation," between
global forces and the diverse social orders of Mrica and elsewhere. This
encounter established new, multilevelled structures of inequality both
IV
within Africa itself and between Africa and Europe. And, in so doing, it
laid the basis for what might be termed "segmentary ethnicity" (cf. R.
The fourth proposition, which takes all this a step further, concerns the
Cohen 1978:387); that is, a nesting hierarchy of ethnic identities.
distinction between historical forces and structures, and the manner in
At the lowest levels, local groups, which often became bounded political which they are experienced in everyday life. We have argued that ~
units by bureaucratic fiat, sought and gained dominance over others within
which pr09_t1-c:e_s()rJsciousness:::-::--lQ1CJ!!i~~~1!~ . 5).r_~~y... ot~~E___)j~__i!l.J:he
colonial-and, later, postcolonial-states, thereby constructing internal construction and transformation of economy and society.. But the terms in
ethnic relations. At the highest levels, the conjuncrure of "Europeans" and
"Mricans" expressed itself in an encompassing ethnicity. This was re­
which they are unde~st~o-d hy'
-soclilactors-haYe fo -do with the way in
which the world is signified: human beings perceive and act upon their
flected, on the one hand, in the growth of pan-Mricanism and the concept contexts not as they are formally constituted, but as they are construed in
of an "African culture"; on the other, it saw the crystallization of a settler­ shared signs and symbols. With the emergence of class formations in which
colonial social order wrought largely out of a caricature of (aristocratic) positions in the division of labor are signified by the labels of ascribed
Victorian English society (Ranger 1983). And, between these polarities, status and cultural distinction, ethnicity becomes a dominant medium
there developed a variety of middle-order allegiances, attachments that through which the social order is to be interpreted and navigated.
crossed parochial boundaries but justified common political cause with In other words, where it becomes the basis of social classification and
reference to shared cultural affinities; these have been described by terms status relations, ethnicity, rather than the forces that generate it, takes on
such as "supertribalism." Of the three levels, significantly, the lowest has the ineffable appearance of determining the predicament of individuals and
conventionally come to be portrayed as "tribalism"; the middle, as "na­ groups. Mter all, viewed from within any such social context, ethnic
tionalism," sometimes with the qualifier "ethnic"; and the highest, as affiliation cannot but be represented as the "independent variable" that
"race." But all had their genesis alike in the processes that gave rise to shapes careers and biographies. For, at the experiential level, it does seem
ethnic consciousness in its more complex segmentary guise. They are to be ethnicity which orders social status, class membership, and so on­
interrelated products of an historically specific confrontation between the and not class or status that decides ethnic identities. In fact, Marx's image
populations of Mrica and the various agencies of colonial domination. of the camera obscura ([Marx and Engels] 1970:36), whatever its other
It is not difficult to point to cases of segmentary ethnicity, since the limitations (Lichtman 1975:49), is apposite here: the origins of ethnic
emergence of a hierarchy of identities, all culturally marked and politicallY groups_and ~()!1.s.ciousness may lie in the structurin.g ]}rinequai:ity:--But,
salient, has been a recurrent theme throughout modern Africa. For exam­ once objectified as a "principlc~' by which the division oflabo'r-fs-organized,
ple, the colonial history of Zimbabwe caned into being an intricate eth­ ethnicity assumes the autonomous character of a prime mover in the
nological map of "tribes" whose administration under indirect rule de­ unequal destinies of persons and populations. To wit, just as working class
manded that their "traditional" political constitution be recognized-; black Americans do not view their blackness as a function of their class
even, in places, where they had none. It also fostered the pre-eminence position, but their class position as a function of their blackness,lO so
tWO such entities, the Shona and Ndebele, each of which spawned the Underclass Butu in Rwanda or Kgalagadi in Botswana see their status as
of a (supratribal) nationalist movement. The alliance of these movementS; being ascribed by virtue of their ethnic affiliation and not vice versa.
within a "patriotic front," in turn, fought the war of liberation that led Under these conditions, then, ethnicity becomes an essential feature of
Zimbabwean independence; a war between adversaries-each with the "natural" order of things, the "given" character of the world with
beyond the national borders-riven along the racial line of fault regard to which people must conduct their lives. Nor is this confined to
divides southern Mrica into inclusive (transnational) pigmentary any particular sociological category. It applies as much to those for whom
ings. Similar processes, if not always with the same climax, have occur fthnic ideologies legitimize dominance as it does to those for whom ethnic
in East Mrica, most notably in Uganda (Mazrui 1978) and Kenya ( abels are signs of subordination. As this suggests, such labels are not
1978), and in West Africa. The exact hierarchy of groupings yielded
60 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Of Totemism and Ethnicity
61
merely terms in a system of classification, although they are certainly that
too. They also become a pragmatic basis for the formation of interest phenomenon to be distinguished from those that sustain it, but any such
groups and networks, social resources for pursuing individual and com­ phenomenon, once created, may acquire the capacity to affect the structures
munal utilities. Consequently, ethnic consciousness enters a dialectical that gave rise to it. So it is with ethnic identity. lVhere it becomes an
relationship with the structures that underlie it: once ethnicity impinges objectified (<principle)) in the collective consciousness of a society, ethnicity may
upon experience as an (apparently) independent principle of social classi­ be perpetuated by factors quite different from those that caused its emnlfence,
fication and organization, it provides a powerful motivation for collective and may have a direct and independent impact on the context in which it arose.
activity. And this, by turn, must perforce realize an everyday world We have already stressed that, in systems where "ascribed" cultural
dominated by ethnic groups and relations, thereby reproducing the very differences rationalize structures of inequality, ethnicity takes on a cogent
social order that gave rise to ethnic consciousness in the first place. existential reality. It is this process of reification, as we have seen, that
However, as we shall see, the matter is yet more complex. For social action gives it the appearance of being an autonomous factor in the ordering of
conducted in the name of ethnicity also reveals contradictions inherent in the social world. As a result, ethnic identities regularly assume, for those
systems of structured inequality and may transform those systems from who share them, a pervasive functionality in everyday social, economic,
within. and political life (cf. Patterson 1977:102f). What is more, this functionality
All this may sound much the same as saying that ethnicity is merely a may itself seem to sustain the practical relevance of ethnic affiliation for
form of false consciousness, a "phantasmic" mystification of class. But that both persons and groups. Behind it, however, there lies a subtle relationship
would miss the point. For the way in which relations in any system are between social experience and the exigencies of collective and individual
practice.
signified is an irreducible part of its reality. As Genovese (1971:32; see
Greenberg 1980:390) puts it, Since it is cultural indices that are perceived to underlie inequalities in
these systems, it follows, in the very nature of social experience, that such
... once an ideology arises it alters profoundly the material reality and in asymmetries would be eliminated (for groups) and upward mobility be
fact becomes a partially autonomous feature of that reality. possible (for persons) if the relevant cultural markers could be reversed or
Indeed, structures of inequality and the terms of their cui rural represen­ relinquished. After all, in order to have meaning, any sign must imply its
tation, whether they be subsumed in "ethnicity" or anything else, are complementary opposite. Thus, if the signs and principles which apparently
mutually constitutive. For reasons already given, the one cannot exist mandate relations of inequality were no longer to apply, the inequality
without the other: both alike are elements of the dialectic of structure and itself would be removed-or so it seems from the actors' perspective. In
practice which shapes concrete social and material relations-and, there­ other words, such systems represent themselves as potentially navigable,

fore, the very essence of economy and society. their internal lines of division more or less porous, to the extent that

The fourth proposition, in sum, argues that, while ethnicity is the product sociocultural differences may be negotiated. This, in turn, establishes the

of specific historical processes, it tends to take on the «naturafJ appearance 0/ appropriate terms in which social action is to be joined; and interests to be

an autonomous force, a «principle)) capable ofdetermining the course of social pursued, at the levels of both collective and individual enterprise. Let us

life. That is, by (i) configuring the particular manner in which a social examine each, for not only do they have contrasting dynamics, but a

complex connection obtains between them.

system is experienced from within, it (ii) motivates social practice and
rationalizes the pursuit of individual and collective utilities. This, in turn, At the level of collective action, the logic of common interest is, for the
(iii) realizes manifest relations, groups and identities; and, so, (iv) a dominant, plain enough [see above, p. 52]. It lies in the authoritative
dialectic is established between structure and practice that, in time, repro­ protection their exclusive cultural identity and, with it, their material
duces and/or transforms the character of the social order itself. POsition. In ideological terms, such protectionism entails a stress upon the
Contrasts between themselves and others; although the creation of alliances
With particular groupings, the internal segmentation of the underclass, and
v the admission of individuals into their ranks, may become strategic consid­
erations in the defense of privilege. For the subordinate, the issue is not so
- ~Iear cut. If and when any cohesive response to the common predicament
This brings us to the fifth and final proposition, which echoes a very general
IS perceived as possible, two options present themselves. Either they can
analytic injunction: not only are the conditions that produce an historical
engage in some form of concerted direct effort-usually, but not always,
62 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Of Totemism and Ethnicity
63
political-to remove entirely the structures of inequality. Or they can seek
to negate cultural differences by "proving" that these have ceased to be as is indicated by the entire history of capitalism, the promise of upward
relevant: vide the attempts of South Mrican and American blacks, earlier mobility, as just reward for properly industrious achievement, is ordinarily
seen to enjoy both practical and ideological virtue.
this century, to claim their civil rights by establishing that they had
"become" sufficiently like those who oppressed them (i.e. "civilized"). It Once upward mobility is seen as a possibility and energies are expended
is not surprising, in light of the manner in which such systems engrave to achieve it, ethnic groups must inevitably become internally stratified.
themselves upon human consciousness, that the second option tends to be For, in order to garner the assets with which to renegotiate their class
the one of first resort (cf. Greenberg 1980:7-8); nor is it difficult to position, individuals require first to consolidate their situations within the
understand why it regularly fails, and gives way sooner or later to overt underclass itself Thus Wallerstein ([1972] 1979), among others, has shown
political action. In the colonial context, with the emergence of segmentary that intra-ethnic relations provide the arena in which such persons might
ethnicity, parochial groups (i.e. "tribes"; [see above, p. 58]) find that such acquire the resources for upward movement; indeed, there are few other
exertions, while not removing higher order inequalities, bring relative contexts within which they might do so. II In addition, under colonialism,
success as they "rise" within the lower reaches of the (colonial) division of where segmentary ethnicity developed, lower order ethnic groups often
labor at the expense of similar groupings. assumed the character of fractions of the underclass; fractions more or less
However these historical movements work themselves out in particular exclusively associated-much as in the segmented labor markets of the
contexts, a fundamental contradiction inheres in them. For any activity industrialized world-with relatively stratified niches within the division
aimed at the reversal of "ascribed" inequalities may reinforce the primacy of labor. 12 This, in turn, opens the way for individuals to manipulate ethnic
ofethnicity as a principle of social differentiation: the very fact such activity identities so as to ascend within the nether echelons of the hierarchy: it
is conducted by and for groupings marked by their cultural identities charts a series of steps that may finally lead out of the underclass itself. The
confirms the perception that these identities do provide the only available regular effort to navigate this route, of course, is the very stuff of the
basis of collective self-definition and action. This is more dramatically the ethnography of urban Africa. It gives rise to what is usually glossed as
case where common action is seen to have been successful, but failure also "situational ethnicity" (R. Cohen 1978; A. Cohen 1974), the strategic
tends to affirm shared affiliations, and the attribution to them of social management of personal identity for social and material gain. Those who
predicament. The affirmation is clearly tautological; in analytic terms, traverse the route, and ascend far enough within the class structure, become
though, it will be evident that the reproduction of ethnicity, and the an identifiable fragment of the dominant class, an emergent local bour­
reinforcement of its pragmatic salience, is as much a function of efforts geoisie. Both as individuals and as a group, however, they face an unavoid­
directed at its erosion as it is of activities that assert its positive value. And able choice. For, on leaving the underclass, they must either seek to discard
their ethnic identities-which, after all, marks the predicament from which
as long as social practice continues to be pursued as if ethnicity did hold
the key to the structures of inequality, the protectionism of the dominant they contrived to escape-or sustain the contradiction of being a member
and the responses of the dominated alike serve to perpetuate an ethnically of a group whose primary class position is different from their own. In

ordered world. either case, their situation is paradoxical, since the two dimensions of their

For the same reasons that collective melioration is held by the subordi­ identity are at odds with each other-a contradiction most vividly embod­

ied in the classical "black white men" of Africa (cf. Fanon 1968).

nate to lie in the negation of ethnic differences, individual mobility is often

perceived as a matter of relinquishing those cultural attributes which appear Drawing together the collective and individual perspectives, a significant
to ascribe social status. For individuals, the adoption of new identities is pattern emerges. On one hand, as long as ethnic affiliations and identities
relatively easier than it is for groups; again, there is abundant documenta­ provide the terms of communal action, such action-whatever its imme­
tion from Mrica of the renegotiation of such identities, perhaps the most diate goals, and regardless of the successes or failures of any given group­
poignant case being the practice of "trying for white" by those classified ing-reinforces the experiential salience of ethnicity as a social principle.
as "colored" in South Mrica (Watson 1970). At this level, then, the lines In consequence, the seemingly ascribed character of ethnic identities is
repeatedly confirmed; so, also, is the conception of ethnic groups as
of ethnic division may indeed be breached if and when particular persons
accumulate the social and material wherewithal to do so [see below, p. 63]. bounded units (cf. Barth 1969), despite the reality that membership in
As noted before, too, there may be little resistance, on the part of the them is frequently the subject of social management. On the other hand,
dominant, to the rise of a limited proportion of the underclass. If anything, at the individual level, the achievement of upward mobility effects a
transformation in the relationship of ethnicity and class. For it leads both
 THEORY. ETHNOGRAPHY, HISTORIOGRAPHY Of Totemism and Ethnicity 65
64
to the internal differentiation of ethnic groups and, with the emergence of in Jamaica in the 1930s, Marcus Garvey's Universal Negro Improvement
bourgeois fractions, to the loss of a one-to-one correspondence between Association (UNIA) among them. These organizations tended to acquire a
ethnic affiliation and class membership. (Were that correspondence to be solidly middle class leadership (for whom political action often involved
sustained, self-evidently, the paradoxes involved in both collective and demands made in their collective class interest; 1978:208), and a large
individual action would either not surface, or would take a different form.) following among the poor, who had come to experience their suffering,
This does not obliterate the fact that, as a statistical frequency-and as an "in the natural order of things" (1978:145), as the heritage of color. Post
indigenous perception-ethnic groups continue to be predominantly as­ goes on to argue that such "generalized Black Nationalism" (1978:161)
sociated with particular class positions. Nonetheless, class and sociocultural ultimately "refracted and atomized potential collective class consciousness"
differences cease to be coterminous in any absolute or prescriptive sense.
13 (1978:187), especially when it either receded into Ras Tafarian millenia 1­
All this is nicely exemplified in ethnic political movements which arise ism, or sufficed itself with the crumbs of economic palliation. In other
among populations that were once subordinated en toto but, over time, words, by virtue of its dualistic quality, ethnicity itself became a factor in
have become diversified to the extent that some of their number have left the maturation of a colonial and post-colonial capitalist order characterized
the underclass. These movements are often joined, and led, by such by marked asymmetries. On the one hand, it continued to provide a cultural
upwardly mobile members-sometimes, we suggest, in the effort to resolve and organizational basis for a highly stratified division of labor; on the
the contradiction in their own lives-and are usually framed in the rhetoric other, it rationalized the possibility of upward mobility and an ethos of
of protest and resistance; of rebellion rather than revolution, to invoke achievement.
Gluckman's well-worn distinction. In some instances they have a single More generally, then, as ethnic groups become internally differentiated
sociocultural grouping at their core, as did Mau Mau in Kenya; in others, and lose their uniform correspondence to class-as they must inevitably
like the Mrican National Congress in South Africa and many nationalist do in the long run-they mature into status groups in the classical
political parties throughout the continent, they cut across local "tribal" Weberian sense of the term. Now there is a double theoretical irony in
divisions. In the latter case, too, they may reinforce a shared symbolic order this. First, in as much as ethnic affiliations are realized and solidify into
where it existed only tenuously before, and weld a relatively heterogeneous status groups by virtue of such historical processes, they have precisely the
social entity into a bounded status group that acts for sich. In so doing, opposite trajectory to that theorized by Weber. In the Weberian tradition,
ethnic political movements underscore, at once, the continuities and dis­ affinities based on status, being primordial, ought to come before those
continuities in the articulation of ethnicity and class. In one respect, they based on class; and they should only give way to the latter with the growth
express the strong (statistical) association between the two; such action is, of an increasingly rationalized industrial economy (Hechter 1975:313;
patently, directed against a prevailing distribution of social, economic, and Greenberg 1980:6). They certainly should not emerge with the genesis of
political power, and its mass support comes from those who most palpably class structures, let alone enter a complex dialectical relationship with
suffer inequality. Yet they also breach the lines of class division and, as them. But, second, it is not only Weber who is turned on his head by the
such, give the appearance of being manifestations of status group solidar­ rise and persistence of ethnic groupings. In classical Marxian terms too,
ities. Moreover, the dualistic character of these movements-as statements ethnicity should not appear with the emergence of class differences. Quite
of underclass resistance, and as assertions of status group unity-may have the converse: the processes that produce class societies ought to submerge
implications that are not envisaged by those who engage in them. "traditional" ideologies and, eventually, cause them to wither away as class
Thus, under colonialism in Africa, ethnic struggles certainly had an consciousness is brought to the fore. However, far from disappearing, or
impact on the surface contours of the social world, and they did occasionally remaining a mere epiphenomenon of "real" antinomies, ethnic identity
lead to significant liberal reforms. But, insofar as they refracted class aSSUmes an important role in the dynamics of many historical systems­
antagonisms in the cause of status group interests, they did not finally Sustaining yet masking, reinforcing yet refracting, their dominant lines of
remove the structures of subordination. If anything, they masked those ~leavage. For the contradictions inherent in structures and signs of inequal­
structures in an ideology of individual achievement, and, by leaving intact tty not only give ethnicity the appearance of an objectified "force"; they
the correlation between ethnicity and class, served ultimately to consolidate also motivate and rationalize the very social practices which assert, with
relations of collective inequality. Nor was this restricted to Africa. As Post baroque circularity, that ethnicity truly is a pervasive force in the social
World.
(1978) reveals, much the same applied to black nationalist organizations
66 THEORY, ETHNOGRAPHY, HISTORIOGRAPHY
Of Totemism and Ethnicity
67
*' *' *' f collective identity; that is an irreducible social fact, whether or not "human nature," if it
If the five propositions bear scrutiny, it is no wonder, pace both Weberian ~xists at all, is as Gordon asserts. But why ethnic identity? For Gordon, the answer finally
comes down to the assertion that it cannot be willingly shed. Yet the anthropological evidence
and Marxian orthodoxies, that ethnicity is as ubiquitous and tenacious as
is unequivocal. Ethnic identity often can be set aside, and may be a highly situational attribute.
it is; and that, in spite of undergoing changes in its content, experiential Of course, the relative ease with which it is sloughed ofl'varies dramatically: for, say, Jews in
character and historical relevance, it refuses to vanish-notwithstanding Nazi Germany-who shared language, skin color, and (to a large extent) culture with their
the once commonplace sociological tendency to predict its imminent oppressors-it was well-nigh undeniable; for Central Africans on the copperbelt, although
demise. The major problem, in both theoretical and empirical terms, it divided by language, culture, and diverse skin pigmentation, it seems always to have been
negotiable. Quite clearly, ethnic mobility-like all social identity-is historically determined.
seems to us, is not to account for the genesis and persistence-or even the
But then, for many anthropologists and comparative sociologists, so too is "human nature."
transformation-of ethnic consciousness and affiliation. Once ethnicity is 6. Schapera (1952) refers to these groupings as "ethnic." By any current definition of
understood to exist as a set of relations, a product of specifiable historical ethnicity, however, this usage is inappropriate; clearly, it reflects the ambiguities surrounding
forces and processes rather than a primordial "given," those issues become the term in the 1950s (du Toit 1978:1£).
readily understandable. Much more vexing, in light of everything we have 7. A comparable argument could be made for the rise of the Zulu state in the early
nineteenth century. Here, it seems, most vanquished populations had social orders which did
argued, is the question of when and why ethnic ideologies break down and not differ greatly from that of their conquerors. When absorbed into the Zulu polity, they
class consciousness rises to replace it-if, indeed, this ever happens in such became integral units in its social and productive structure rather than a subordinate class
straightforward terms. But that is a problem which demands separate (cf. Suret-Canale 1969; Walter 1969; Guy 1980). For a complementary discussion on ethnic
treatment. The purpose of the present exercise has been more modest. It identities in precapitalist Africa, see Wallerstein ([1972] 1979).
has been to explore just some of the fundamental analytic issues surround­ 8. McGuire (1982: 168), following Spicer (1971 :797), offers the parallel observation that
ethnic groups, the bearers of "persistent cultural systems," are born out of "opposition"; that
ing the signs and practices ofinequality. And, if the propositions formulated is, out of the "efforts by those in control of the surrounding state apparatus to incorporate or
along the way remain tentative for now, they are offered in the belief, to assimilate" minority populations.
parody Levi-Strauss, that ethnicity is good to rethink. 9. Warren's (1978) revealing study of collective identity among Guatemalan Indians­
among whom subordination Within a class society is vividly portrayed in ethnic terms-is
built on this very premise; it shows, in considerable detail, how signs of inequality come to
be elaborated into a coherent set of cultural representations.
Notes
10. It is precisely this point that lies behind much of the debate over Wilson's claim

(1980) that the significance of race is declining in America. Wilson's critics have argued that

I. For discussion of this tendency and its Weber ian origins, see, among others, Hechter skin color remains the main factor shaping the lives of blacks; its salience, therefore, cannot

(1975:313f); Greenberg (1980:13f). have decreased. At the experiential level, this may be true. But, whatever other arguments may

2. The treatment of ethnicity as a "first cause" capable of producing autonomous effects be brought to bear on his case, Wilson, to his credit, does not deny the experiential importance

remains the norm throughout the human sciences: see, for example, Bronfenbrenner (1979:258) of race. His thesis, rather, is that the life chances of blacks are determined primarily by other

in respect of psychology, Cohen (1978) of anthropology, and Wilson (1980) of sociology. structural factors; namely, class differences. The Controversy, in sum, flows from the confusion

3. See, for instance, du Toit (1978:lOf); Fried (1967:15,170); Skinner (1978: 192f); and of social experience with the forces that give rise to it.
Cohen and Middleton (1970). As Wallerstein ([1972] 1979:184) puts it, "Ethnic conscious­ II. Greeley (1974:300) makes much the same point about ethnic relations in the United

ness is eternally latent everywhere. But it is only realized when groups feel either threatened States-a point classically illustrated, as William Wilson has reminded us, by E. Franklin

with a loss of previously acquired privilege or conversely feel that it is an opportune moment Frazier's Black Bourgeosie (1957).
politically to overcome long-standing denial of privilege." 12. For an especially clear example, see Mazrui's account (1978) of colonial Uganda,
4. These properties appear with equal regularity in Webcrian and Marxist discourses on where the cleavage between "Nilotes" and "Bantu" was unambiguously reflected in the
ethnicity-albeit with different analytic weight. Rex (1970:48), in fact, sees them as the Contemporary division of labor. Mazrui might easily have extended his analysis to take in the
defining conditions for what sociologists treat under the rubric of "race relations." Indian
of labor.and white settler communities, which had equally well-defined niches in that division
5. Another variant of the "primordial" argument is the "human nature" thesis. For
example, Milton Gordon (1978:73) holds that, since ethnicity cannot be shed by mobility, it . 13. Sec Wilson (1984), whose penetrating analysis of race and social policy in America
"becomes incorporated into the self." Man acting in the name of his ethnic group, then, is 1lluminates and gives comparative support to this aspect of our discussion.
man defending himself; and, given that "human nature" is aggressively narcissistic, ethnic
consciousness and conflict is endemic. Wilson (1981 :113) has countered by pointing out that
this "human" predisposition is always mediated by social and cultural factors-the implica­
tion being that it may be a necessary, but can never be a sufficient condition for ethnic
antagonism. We would go still further, however. It seems that Gordon has confused precisely
the two levels of which we have spoken. Certainly, "I" is fused into "we" in the construction