Putting Hierarchy in Its Place

Author(s): Arjun Appadurai

Source: Cultural Anthropology, Vol. 3, No. 1, Place and Voice in Anthropological Theory
(Feb., 1988), pp. 36-49
Published by: Wiley on behalf of the American Anthropological Association
Stable URL: http://www.jstor.org/stable/656307
Accessed: 29-08-2016 17:16 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted
digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about
JSTOR, please contact [email protected].

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

American Anthropological Association, Wiley are collaborating with JSTOR to digitize, preserve and
extend access to Cultural Anthropology

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

Putting Hierarchy in Its Place

Arjun Appadurai
Department of Anthropology
University of Pennsylvania

In the essay that follows, I shall be concerned with the genealogy of an idea.
But before I put forward this genealogy, I need to make two preliminary arguments. The first involves the anthropological construction of natives. The second
involves a defense of one kind of intellectual history.
The Place of the Native

On the face of it, an exploration of the idea of the "native" in anthropological discourse may not appear to have much to do with the genealogy of the idea
of hierarchy. But I wish to argue that hierarchy is one of an anthology of images
in and through which anthropologists have frozen the contribution of specific cultures to our understanding of the human condition. Such metonymic freezing has
its roots in a deeper assumption of anthropological thought regarding the boundedness of cultural units and the confinement of the varieties of human conscious-

ness within these boundaries. The idea of the "native" is the principal expression
of this assumption, and thus the genealogy of hierarchy needs to be seen as one
local instance of the dynamics of the construction of natives.
Although the term native has a respectable antiquity in Western thought and
has often been used in positive and self-referential ways, it has gradually become
the technical preserve of anthropologists. Although some other words taken from
the vocabulary of missionaries, explorers, and colonial administrators have been

expunged from anthropological usage, the term native has retained its currency,
serving as a respectable substitute for terms like primitive, about which we now
feel some embarrassment. Yet the term native, whether we speak of "native cat-

egories," or "native belief-systems" or "native agriculture," conceals certain

ambiguities. We sense this ambiguity, for example, in the restricted use of the
adjective nativistic, which is typically used not only for one sort of revivalism,
but for revivalism among certain kinds of population.
Who is a "native" (henceforth without quotation marks) in the anthropological usage? The quick answer to this question is that the native is a person who is
born in (and thus belongs to) the place the anthropologist is observing or writing
about. This sense of the word native is fairly narrowly, and neutrally, tied to its

Latin etymology. But do we use the term native uniformly to refer to people who
are born in certain places and, thus, belong to them? We do not. We have tended
36

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

HIERARCHY 37

to use the word native for persons and groups who belong to tho

world that were, and are, distant from the metropolitan West. Th

in part, tied to the vagaries of our ideologies of authenticity ov

centuries. Proper natives are somehow assumed to represent their

history, without distortion or residue. We exempt ourselves from th
to authenticity because we are too enamored of the complexities
the diversities of our societies, and the ambiguities of our collect

When we find authenticity close to home, we are more likely to l

native, the former being a term that suggests authenticity without

derogatory. The anthropologist thus rarely thinks of himself as a

place, even when he knows that he is from somewhere. So what d

be a native of some place, if it means something more, or other, t

that place?
What it means is that natives are not only persons who are from certain

places, and belong to those places, but they are also those who are somehow incarcerated, or confined, in those places.' What we need to examine is this attribution or assumption of incarceration, of imprisonment, or confinement. Why are

some people seen as confined to, and by, their places?

Probably the simplest aspect of the common sense of anthropology to which
this image corresponds is the sense of physical immobility. Natives are in one
place, a place to which explorers, administrators, missionaries, and eventually
anthropologists, come. These outsiders, these observers, are regarded as quintessentially mobile; they are the movers, the seers, the knowers. The natives are
immobilized by their belonging to a place. Of course, when observers arrive, natives are capable of moving to another place. But this is not really motion; it is
usually flight, escape, to another equally confining place.
The slightly more subtle assumption behind the attribution of immobility is
not so much physical as ecological. Natives are those who are somehow confined
to places by their connection to what the place permits. Thus all the language of
niches, of foraging, of material skill, of slowly evolved technologies, is actually
also a language of incarceration. In this instance confinement is not simply a function of the mysterious, even metaphysical attachment of native to physical places,
but a function of their adaptations to their environments.
Of course, anthropologists have long known that motion is part of the normal
round for many groups, ranging from Bushmen and Australian aborigines, to
Central Asian nomads and Southeast Asian swidden agriculturalists. Yet most of
these groups, because their movements are confined within small areas and appear
to be driven by fairly clear-cut environmental constraints, are generally treated as

natives tied not so much to a place as to a pattern of places. This is still not quite
motion of the free, arbitrary, adventurous sort associated with metropolitan behavior. It is still incarceration, even if over a larger spatial terrain.
But the critical part of the attribution of nativeness to groups in remote parts
of the world is a sense that their incarceration has a moral and intellectual dimen-

sion. They are confined by what they know, feel, and believe. They are prisoners
of their "mode of thought." This is, of course, an old and deep theme in the

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

38 CULTURAL ANTHROPOLOGY

history of anthropological thought, and its most powerful exam

in Evans-Pritchard's picture of the Azande, trapped in their mo
by a way of thinking that admits of no fuzzy boundaries and i

internal consistency. Although Evans-Pritchard is generally car

gerate the differences between European and Azande mentality,
gests that the Azande are especially confined by their mode of t

Above all, we have to be careful to avoid in the absence of native d

ing a dogma which we would formulate were we to act as Azand

elaborate and consistent representation of witchcraft that will accoun

workings, nor of nature which expounds its conformity to sequenc

interrelations. The Zande actualizes these beliefs rather than intell

and their tenets are expressed in socially controlled behavior rather th

Hence the difficulty of discussing the subject of witchcraft with

ideas are imprisoned in action and cannot be cited to explain an

[Evans-Pritchard 1937:82-83; emphasis mine]

Of course, this idea of certain others, as confined by their w

in itself appears to have nothing to do with the image of the n
who belongs to a place. The link between the confinement of
idea of place is that the way of thought that confines natives i
bounded, somehow tied to the circumstantiality of place. The li
tellectual and spatial confinement, as assumptions that underpi
native, are two. The first is the notion that cultures are "who

taken up in the section of this essay on Dumont. The second is t

ded in studies of ecology, technology, and material culture ove
the intellectual operations of natives are somehow tied to their
situations. They are seen, in Levi-Strauss's evocative terms, as
concrete. When we ask where this concreteness typically inheres

in specifics of flora, fauna, topology, settlement patterns, and the

it is the concreteness of place. Thus, the confinement of native

reflects in an important way their attachment to particular pla
the concrete can thus be written as the poetry of confinement.2

But anthropologists have always known that natives are not

cerated. The American anthropological tradition, at least as far
most recently in the voices of Sidney Mintz (1985) and Eric W
always seen cultural traits as shared and transmitted over large
capable of change, and as creating shifting mosaics of technolo
The French tradition, at least in that part of it with roots in Herd
more recently in Mauss, Benveniste and Dumezil, has always s
least of the Indo-European "linguaculture" (Attinasi and Friedr
many geographically scattered places. Even in British anthropo
been minority voices, like those of Lord Raglan and A. M. Hoca
that the morphology of social systems and ideologies is not con
territorially anchored groupings. It is now increasingly clear
stances where anthropologists believed they were observing and

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

HIERARCHY 39

tine or historically deep systems, they were in fact viewing product

transregional interactions. Diffusionism, whatever its defects and i
guise, has at least the virtue of allowing everyone the possibility of e
world larger than their current locale.
It is even more evident that in today's complex, highly interconne
dia-dominated world, there are fewer and fewer native cultures left. T

pressed by the international market for the objects once iconic of the
which are now tokens in the drive for authenticity in metropolitan

cultures. They are pushed by the forces of development and nat

throughout the world and are attracted by the possibilities of migration

in new places. Natives, as anthropologists like to imagine them, are

rapidly disappearing. This much many will concede.
But were there ever natives, in the sense in which I have argued
must be understood? Most groups that anthropologists have studied h
way been affected by the knowledge of other worlds, worlds about w
may have learned through migration, trade, conquest, or indigenous
As we drop our own anthropological blinders, and as we sharpen ou
torical tools, we are discovering that the pristine Punan of the interio
were probably a specialized adaptation of the larger Dayak communit
a specialized function in the world trade in Borneo forest products
1986); that the San of Southern Africa have been involved in a comple
with other groups for a very long time (Schrire 1980); that groups in
have been trading goods across very long distances for a long time,
reflects complex regional relations of supply and demand (Hughes 197
rican "tribes" have been reconstituting and deconstructing essential
principles at their "internal frontiers" for a very long time (Kopytof
Even where contact with large-scale external forces has been, till
minimal, as with some Inuit populations, some populations in lowlan
America, and many Australian aboriginal groups, these groups have c
very complex "internal" mosaics of trade, marriage, conquest, and lin
change, which suggests that no one grouping among them was ever t
cerated in a specific place and confined by a specific mode of thoug
example, Myers 1986). Although assiduous anthropologists might alw
cover some borderline examples, my general case is that natives, peop
to and by the places to which they belong, groups unsullied by con
larger world, have probably never existed.
Natives, thus, are creatures of the anthropological imagination. In
ogic age, this may not seem like a very bold assertion, but it ramifie
directions. If anthropologists have always possessed a large amount of
tion that has militated against the idea of the native, how have they s
holding on to it? How have places turned into prisons containing nat
The answer lies in the ways that places have been married to idea
ages, and here I resume an argument initiated elsewhere (Appadurai 1
thropology has, more than many disciplinary discourses, operated thr
bum or anthology of images (changing over time, to be sure) whereb

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

40 CULTURAL ANTHROPOLOGY

ture of a group is seen as quintessential to the group and as especially tr

group in contrast with other groups. Hierarchy in India has this qualit
discourse of anthropology, hierarchy is what is most true of India and
of India than of any other place.
In the subsequent sections of this essay I shall show that ideas that

metonymic prisons for particular places (such that the natives of that

inextricably confined by them) themselves have a spatial history, in th

discourse of anthropology. Ideas and images not only travel from plac

but they periodically come into compelling configurations, configurati

once formed, resist modification or critique. By looking at Dumont's

tualization of hierarchy in India, I shall explore the archaeology of hie

an image that confines the natives of India. In the last part of the artic
I shall propose a theory about the circumstances under which such res
figurations tend to occur in the history of anthropological discourse.

The Genealogy of Hierarchy

The recent wave of reflexivity among anthropologists, especially th

ticing in the United States, has already created a backlash, founded on

ervations, including temperamental and stylistic ones. But one of the r

the backlash has been the suspicion that the self-scrutiny of ethnogra

fieldworkers might be a prologue to the extinction of the object of ou

Faced with the disappearance of natives as they imagined them, some
ogists run the risk of substituting reflexivity for fieldwork. I belong to
of anthropologists who wish neither to erase the object in an orgy of self
nor to fetishize fieldwork (without carefully rethinking what fieldwor

mean and be in a changing world), in the way that Victorian educators

cold baths and sport as character-building devices for the public-schoo
why engage in any sort of genealogy?
All genealogies are selective, as any good historian of ideas would
nize. They are selective, that is, not through sloppiness or prejudice (th
could always creep in), but because every genealogy is a choice among
infinite set of genealogies that make up the problem of influence and
intellectual history. Every idea ramifies indefinitely backward in time, an
critical historical juncture, key ideas ramify indefinitely into their own h
contemporary contexts. Nontrivial ideas, especially, never have a finite
nealogies. Thus any particular genealogy must derive its authority from
it seeks to subserve. The genealogy I have constructed in the case of

conception of hierarchy is one such genealogy, which subserves my intere

spatial history of anthropological ideas. Thus my genealogy, like any

nealogy, is an argument in the guise of a discovery.
There is another way to characterize my position. The sort of genea
interested in has something in common with Foucault's sense of the p

calls "archaeology," a practice which, when successful, uncovers not

netic chain, but an epistemological field and its discursive formation

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

HIERARCHY 41

cursive formation with which I am concerned, at its largest level, is

of anthropology over the last century, and within it the subdiscours

and about India. This sort of genealogizing is intended to occupy the

between the atemporal stance of certain kinds of contemporary cri

cially those affected by deconstruction) and the exclusivist and gen
tions of most standard approaches to the history of ideas.
In Dumont's (1970) conception of hierarchy as the key to caste
India, we see the convergence of three distinct trajectories in West
These separate trajectories, which come together in recent anthrop
tice, are threefold. First there is the urge to essentialize, which cha
Orientalist forebears of anthropology. This essentialism, which has
genealogy going back to Plato, became for some Orientalists the pr

for characterizing the "other." As Ronald Inden has recently a

1986a), this led to a substantialized view of caste (reified as India's

stitution) and an idealized view of Hinduism, regarded as the religio

of caste. The second tendency involves exoticizing, by making dif

tween "self" and other the sole criteria for comparison. This tenden
cize has been discussed extensively in recent critiques of the histor
pology and of ethnographic writing (Boon 1982; Clifford and Mar
bian 1983) and has its roots in the "Age of Discovery" as well as in

century "Age of Nationalism," especially in Germany. The third t

volves totalizing, that is, making specific features of a society's tho

tice not only its essence but also its totality. Such totalizing probably
in the German romanticism of the early 19th century and comes t

variations of the idea of the Geist (spirit) of an age or a people. C

Hegel's holism, its most important result was the subsequent Mar
ment to the idea of totality (Jay 1984), but it also underlies Dumont
of the "whole," discussed below. In this sense, the dialogue betwee
istic and the materialistic descendants of Hegel is hardly over. In an
and in history, particularly in France, it is to be seen in Mauss's id
as a total social phenomenon and in the Annales school's conceptio
totale.

Hierarchy, in Dumont's argument, becomes the essence of caste,

its exoticism, and the form of its totality. There have been many c
Dumont's ideas about hierarchy. I shall be concerned here to decon

archy by unpacking its constituents in Dumont's scheme and by tracin

of the genealogy of these constituents that moves us out of India

places in the ongoing journey of anthropological theory. As we sh
genealogy is in part a topographic history of certain episodes and ce
the history of anthropological thought in the last century. Since my
concerned largely with the extra-Indian implications of Dumont's
briefly place them in their Indian context.

Arriving on the scene in the late 1960s, when American cultura

ism, British structural-functionalism, and French structuralism h

rather dull standoff in regard to the study of caste, Homo Hierarch

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

42 CULTURAL ANTHROPOLOGY

vanizing effect. It was widely (and vigorously) reviewed, and it gene

merous symposia, an army of exegetes, acolytes, and opponents. For a
decades it has dominated French structuralist studies of rural India, f

intellectual charter of the influential journal Contributions to Indian

and generated much empirical and theoretical activity both in Engla
United States. Dumont's ideas have been subject to careful and sympat
icism by a host of scholars who have pursued his French intellectual
conception of ideology, his model of renunciation and purity, and the

ideas with Indian facts (see, for example, Berreman 1971; Das 197

1976; Marriott 1969; Parry 1980; Srinivas 1984; Yalman 1969).

While difficulties have been seen with almost every important aspe
mont's methodology and claims, most scholars working on the caste sy

South Asia (even the most obdurately empiricist critics of Dumont) will g

Dumont's idea of hierarchy captures the distance between the value-as

of India and post-Enlightenment Europe like no previous characterizat
There are thus two trajectories within which Homo Hierarchicus fi
the trajectory that has to do with the history of Western values. Dumont
as the mid-1970s, had shown his concern with the dynamics of individ
egalitarianism in the West. This latter concern has intensified since the p
of Homo Hierarchicus, and Dumont's latest collection of essays (Dumon
makes it clear that the argument about hierarchy in India was an episode
term exercise in the archaeology of modem Western ideology.
Yet, since Homo Hierarchicus also made a bold and sweeping struct
argument about the ideology of the caste system, it demands assessmen
tique in its areal context as well. This it has amply received. What is n
for is an effort to bring these two trajectories into a unified critical d
discussion in which areal and theoretical issues are not invidiously sep
ranked. This essay is a preliminary contribution to this sort of unified
Pauline Kolenda (1976) has shown the ambiguities in, and polysemy
mont's use of the word hierarchy in Homo Hierarchicus, and has prov
uable basis for extending our understanding of the roots of his idea of
Dumont owes a very large part of his understanding of caste society t
Bougie, about whom I shall have more to say shortly. Bougle (1971) arg
the caste system was a product of the unique configuration of three relat
erties of the castes: separation, hierarchy, and interdependence. Dum
vance is to find a principle linking and underlying all three and developin
sweeping and abstract conception of hierarchy than Bougle's.
The ingredients of this conception of hierarchy, each of which has a
genealogy, are (1) a particular conception of the whole; (2) a particular
of the parts; (3) a particular conception of the opposition of pure and im
(4) a particular commitment to the idea of the profoundly religious bas
society. I shall consider each of these in turn, starting with the idea of th
Dumont's idea of the whole is consciously derived from Hegel, to w
attributes the view that the hierarchy between castes is a matter of the r
a whole. Hegel's Philosophy of History (1902), his most important co

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

HIERARCHY 43

to German Orientalism, is the main link between Dumont and the

back to Plato, in which a conception of the social or collective wh

mary source of values and norms.3 India, in this Hegelian view,
showcase for rank and stratification (which is a commonplace of f
of India from the beginnings of the Christian era) and becomes in
museum of that form of social holism that has been lost to the West. Less con-

scious, but equally decisive for Dumont's idea of the whole, is the conception of
the Ann6es Sociologiques, in which certain archaic social forms, especially gift
and sacrifice, are seen as total social phenomena. Although I shall have more to
say on the topographic genealogy of Mauss's ideas, it is worth noting that they
are the product of a particular French philological tradition that seeks to link the
Indo-European world with the world of the primitive. Its topos is the spatiotemporal landscape of the vanished Indo-European heartland and the scattered islands
of early ethnography. In Dumont's conceptualization of hierarchy, Hegelian holism and Maussian totalizing come together, and a decisive break is made with the
earlier Western obsession with Indian stratification. The subordination of parts to
the whole is at the heart of Dumont's understanding of the ideological basis of the
system of castes. This whole ("the system of castes") is taken by Dumont to be
complete, more important than its parts, stable, and ideologically self-sustaining.
Dumont's idea of the whole represents one variant of the wider anthropological
commitment to holism, a commitment that has elsewhere been opened to critical
examination.4

So much for Dumont's conception of the "whole." What about his conception of the parts? Here the plot gets thicker. Dumont's understanding of castes as
parts of a very particular type of hierarchical whole comes from two sources, both
of which he acknowledges. The first is Evans-Pritchard, whose classic study of
the segmentary nature of Nuer society influenced Dumont greatly (Dumont
1970:41-42). As Srinivas has recently emphasized, the topographic roots of the
segmentary nature of Indian castes comes from Evans-Pritchard's analysis of the
Nuer data, a special sort of African case (Srinivas 1984). In turn, Evans-Pritchard's view has complex, though obscure, roots. One aspect of the Nuer model
doubtless goes back to Robertson Smith's classic work on Semitic religion, which
contains a particular English Orientalist picture of Arabian society (Beidelman
1968; Dresch, this volume). On the other hand, the general roots of the classic
British social anthropology of African political systems surely goes back to the
19th-century Anglo-Saxon tradition in studies of ancient law. Especially central
here is the work of Henry Maine, who is a critical theorist of kinship as a basis
for jural order.5 Since Maine also worked on Indian law and society, in comparison with ancient Rome, we have here a wonderful circle. From the ancient village
republics of India, via ancient Rome and comparative law, through African political systems and Nuer segments, back to Indian castes.
But the other source of Dumont's conception of the castes as "parts" is Bougle's image of the "repulsion" of the castes toward each other, a fascinating Gallic precursor of Evans-Pritchard's conception of the fissive tendencies of Nuer
segments (Bougle 1971:22; Evans-Pritchard 1940:148).

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

44 CULTURAL ANTHROPOLOGY

It is not easy to trace the roots of Bougle's emphasis on the "r

the parts for each other, except as a synthetic insight based on the e
and Indology available to him at the turn of the century. Since Bo

(unlike Mauss) an Indologist, and since he was mainly concerned with

of egalitarian values in the West, we can only guess that the areal int

colleagues in the Annees Sociologique group had some effects on hi

specific influence we shall note shortly.
Before we come to Dumont's critical contribution-the oppositio
and impure as the axiom of the entire caste system-we need to ask
larger view on which it is based, namely that in India, religion is th
shaper of ideology and values. Although this is something of a comm
has been noted by centuries of observers from the West, Dumont pl
emphasis and interpretation upon the religious basis of Indian society
for this orientation again is Hegel. But more proximate are Bouge1
already, and A. M. Hocart. As to Bougle, he attributed the hierarch
conception of castes to the utter predominance of the priesthood in
one hand, this predominance was attributed to a weak state organizat
we have a parallel to the link that leads from Henry Maine to Evans

On the other hand, Bougle attributes it to the centrality of the sacrifi

India. This, in turn, Bouge1 derives from Hubert and Mauss's class
sacrifice and-you guessed it-Robertson Smith on the Semitic religi
rifice. So we are back in the shadow of Arabia.

But the other crucial source of Dumont's ideas about the religious basis of
Indian society is the work of the English anthropologist-administrator, A. M. Ho-

cart. Although Dumont makes many criticisms of Hocart's work on caste, he is

explicit in acknowledging his debt to him on the centrality of religion to caste.
What is interesting about Hocart's own anthropological career is that it began in
the South Pacific, where he conducted anthropological researches in Fiji, Tonga,
and Samoa. He was also Headmaster for some time of a native school at Lau

(Fiji), and he wrote a learned monograph on the Lau Islands. It was this expe

ence of the centrality of chieftainship and castelike specialization that was on

mind when, after World War I, he was appointed Archaeological Commission

in Ceylon, where he further developed his ideas on caste and kingship. In fact
his entire model of Indian society-centered on the ritual of kingship-is base
on his apperception of Ceylon, where the ritual of royalty remained a macro re

ity. When he finally wrote his comparative study of caste in the 1930s, it reflect

an understanding of Indian caste that echoed a Ceylonese redaction of his und

standing of rank, chieftainship, and religious order in the South Pacific, especiall

in Fiji.
An interesting variant on this genealogy can be seen in Dumont's understanding of the contrast between the pure and the impure. Dumont acknowledges

the important but mistaken ideas of scholars like H. N. C. Stevenson (1954)

(whose work on status evaluation in the caste system may have been influenced
by his own earlier work on the Chin-Kachin group in Burma, whom Leach subsequently immortalized). But he must have also been greatly influenced by Ho-

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

HIERARCHY 45

cart, whose comparative work on caste (Hocart 1950) contains an

though tacit, emphasis on the problem of ritual separation and the p

Although Dumont does not explicitly attribute this part of his think

there is a very interesting section in Homo Hierarchicus where h

ideas about food prohibitions in India are owed to an unpublishe
and expiation taught by Mauss at the College de France, where Mau
his ideas from Hocart's work on Tonga (Dumont 1970:140). Thus,
matter of food prohibitions, which exemplify the contrast betwee
pure, which is in turn the cultural pivot of Dumont's ideas on hier
pographic genealogy leads back to Hocart on Tonga.
Let me conclude with a review of my findings. Dumont's conc
erarchy leads from India in at least four major topological directi
regard to its conception of the parts; ancient Arabia, for its concept
segmentation and solidarity; ancient Rome, for its conception of ju
absence of a powerful state; and the South Pacific (via Ceylon), for
of the power of taboo and the ritual implications of specialization.
there are two other discourses that mediate this one, discourses
lies outside the scope of this essay. One is the metropolitan discou
pology, conducted at places like Oxford, the College de France, a
sites of colonial administration. The other is the grander discourse

whose strengths and weaknesses are still with us in the anthropol

India (Inden 1986a, 1986b).

Hierarchy in Place

It remains now to ask, more generally, about the circumstance

certain anthropological images-such as hierarchy-become heg

confined to, certain places. This question is inescapably both histo

parative.
From the comparative point of view, ideas or images that become metonyms
for places in anthropological discourse appear to share certain properties. First,
for the nonspecialist they provide a shorthand for summarizing the cultural complexity that has already been constituted by existing ethnography. By extension,
they provide a handy guide for navigating through new (or newly discovered)
ethnographies without getting lost in the minutiae of the locality. Although some
ethnographies become classics because they are compelling works of literature,
most routine ethnographies profit from those summarizing metonyms that provide
a point of orientation for the nonspecialist reader. Of course, this does nothing to
increase the likelihood that the nonspecialist is likely to pick up the situational
diversities of these local worlds.

Second, from the point of view of the specialists who work on a place, certain ideas or images are likely to become hegemonic because they capture something important about the place that transcends intraregional variations and that
is, at the same time, problematic, because it is subject to ethnographic or methodological question. Thus, hierarchy is (at least in some of its Dumontian mean-

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

46 CULTURAL ANTHROPOLOGY

ings) undeniably a striking feature of Indian society, but its exact

foundly debatable. For the specialist, images like hierarchy acquire
not because they ease the labors of traveling through the jungles of
ethnographies, but because they are compelling ideas around whic
debate, whether such debate is about method, about fact, about ass
about empirical variations.
Finally, neither of the above properties is quite sufficient to guar
particular idea (expressed as a term or a phrase) will become hegemo
to the construction of a place. It is also important that the image prov
link between internal realities (and specialist accounts of them) a
preoccupations (and their larger discursive contexts). The most resi
linking places and cultural themes, such as honor-and-shame in the
iterranean, hierarchy in India, ancestor-worship in China, compa
panic America, and the like, all capture internal realities in terms
discursive needs of general theory in the metropolis.
This hypothesis about the images of place in anthropology need
into a historical perspective as well. Such hegemonic ideas not onl
being in specific conjunctures but are also liable to being pushed ou
other such ideas. What accounts for such shifts is not easy to talk a
eral way, or in a brief space, since it involves the gradual accumula
changes in metropolitan theorizing; in local, ethnographically cen
and in the relationship between the human sciences in (and in regard

places.
Assuming that such topological stereotypes cost us more in terms of the richness of our understanding of places than they benefit us in rhetorical or comparative convenience, how are we to contest their dominance? Here three possibili-

ties present themselves. The first, exemplified in this essay, is to remain aware
that ideas that claim to represent the "essences" of particular places reflect the
temporary localization of ideas from many places. The second is to encourage the
production and appreciation of ethnographies that emphasize the diversity of
themes that can fruitfully be pursued in any place.
The third, and most difficult possibility, is to develop an approach to theory
in which places could be compared polythetically (Needham 1975). In such an
approach, there would be an assumption of family resemblances between places,
involving overlaps between not one but many characteristics of their ideologies.
This assumption would not require places to be encapsulated by single diacritics
(or essences) in order for them to be compared with other places, but would permit
several configurations of resemblance and contrast. Such a polythetic approach to
comparison would discourage us from thinking of places as inhabited by natives,
since multiple chains of family resemblance between places would blur any single
set of cultural boundaries between them. Without such consistent boundaries, the
confinement that lies at the heart of the idea of the native becomes impossible.

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

HIERARCHY 47

Notes

Acknowledgments. Earlier versions of this article were presented at the panel on "Place
and Voice in Anthropological Theory," at the 85th Annual Meeting of the American Anthropological Association, Philadelphia, December 1986, and at the Research Colloquium
of the Department of Sociology, University of Delhi, in January 1987. I am grateful to
colleagues present on each of these occasions for useful comments and suggestions. Comments by Paul Friedrich (on several drafts) and Paul Dresch (on an earlier draft) forced me
to clarify key points and eliminate certain errors.

'For a fascinating account of the ironies in the historical evolution of such terms as native,
inlander, indigenes, etc., in the context of Dutch colonialism in Southeast Asia, see Anderson (1983:112-128).

2Lest I be seen as excessively critical of the attention that anthropologists have paid to this

poetry of confinement, I should add that some of the ethnography that best combines description and theorizing capitalizes on the enmeshment of consciousness in culturally con-

stituted environments: Evans-Pritchard on Nuer time-reckoning (Evans-Pritchard 1940),

Irving Hallowell on Saulteaux measurement (Hallowell 1942), Steven Feld on Kaluli poetics (Feld 1982), and Fernandez on the imagery of African revitalization movements (Fernandez 1986).

3Hegel's own ideas about Indian religiosity were greatly influenced by the romantic Orientalist treatises of Herder and Schlegel (see Inden 1986a and Schwab 1984).
4When I published my own critique of anthropological holism, in the context of a critique

of Dumont's ideas (Appadurai 1986b), I had not had the opportunity to see Fernandez
(1986). In this essay, Fernandez is concerned with the mechanisms that create "the conviction of wholeness" in African revitalization movements. He is thus able to propose a
more optimistic solution to the problem of "cultural wholes" than I was. The time seems
ripe for a full-fledged debate about the many dimensions of the problem of cultural wholes

and the relationship between them.

'Evans-Pritchard seems to have been conscious of this debt, and has stated that one of
Maine's most important generalizations was that "kinship and not contiguity is the basis
of common political action in primitive societies" (Evans-Pritchard 1981:87). Of course,
Dumont was also influenced by Maine, but I believe that in this regard, the influence was
mediated by Evans-Pritchard.
References Cited

Anderson, Benedict

1983 Imagined Communities. London: Verso.

Appadurai, Arjun
1986a Theory in Anthropology: Center and Periphery. Comparative Studies in Society

and History 28(2):356-361.

1986b Is Homo Hierarchicus? American Ethnologist 13(4):745-761.
Attinasi, John, and Paul Friedrich

1987 Dialogic Breakthrough: Catalysis and Synthesis in Life-Changing Dialogue. In

Toward a Dialogic Anthropology. Bruce Mannheim and Dennis Tedlock, eds. Forthcoming.

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

48 CULTURAL ANTHROPOLOGY

Beidelman, T. O.
1968 Review of W. Robertson Smith, Kinship and Marriage in Early Ar

thropos 63/64:592-595.
Berreman, Gerald D.
1971 Review of Homo Hierarchicus. Man 4:515.

Boon, James

1982 Other Tribes, Other Scribes: Symbolic Anthropology in the Comparative Stu
of Cultures, Histories, Religions, and Texts. Cambridge: Cambridge Univers
Press.

Bouge1, C.
1971 Essays on the Caste System. D. F. Pocock, trans. Cambridge: Cambridge University Press.

Clifford, James, and George Marcus

1986 Writing Culture: The Poetics and Politics of Ethnography. Berkeley: University
of California Press.

Das, Veena

1977 Structure and Cognition: Aspects of Hindu Caste and Ritual. Delhi: Oxford University Press.

Dumont, Louis

1970 Homo Hierarchicus: The Caste System and Its Implications. Chicago: University
of Chicago Press.
1986[1983] Essays on Individualism. Chicago: University of Chicago Press.
Evans-Pritchard, E. E.
1937 Witchcraft, Oracles and Magic among the Azande. Oxford: Clarendon Press.
1940 The Nuer. Oxford: Clarendon Press.

1981 A History of Anthropological Thought. New York: Basic Books.

Fabian, Johannes

1983 Time and the Other: How Anthropology Makes its Object. New York: Columb
University Press.
Feld, Stephen
1982 Sound and Sentiment: Birds, Weeping, Poetics and Song in Kaluli Expression.

Philadelphia: University of Pennsylvania Press.

Fernandez, James
1986 The Argument of Images and the Experience of Returning to the Whole. In The
Anthropology of Experience. Victor W. Turner and Edward M. Bruner, eds. Pp. 159187. Urbana: University of Illinois Press.

Hallowell, Irving
1942 Some Psychological Aspects of Measurement among the Saulteaux. American
Anthropologist 44:62-67.
Hegel, Georg W. F.
1902[1837] Philosophy of History. J. Sibree, trans. New York: Collier.
Hocart, A. M.

1950 Caste: A Comparative Study. London: Methuen.

Hoffman, Carl
1986 The Punan: Hunters and Gatherers of Borneo. Ann Arbor: UMI Research Press.

Hughes, Ian
1977 New Guinea Stone Age Trade: The Geography and Ecology of Traffic in the
Interior. Canberra: Department of Prehistory, Research School of Pacific Studies, The

Australian National University.

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms

HIERARCHY 49

Inden, Ronald B.
1986a Orientalist Constructions of India. Moder Asian Studies 20(1): 1

1986b Tradition Against Itself. American Ethnologist 13:762-775.

Jay, Martin

1984 Marx and Totality: The Adventures of a Concept from Lukacs to Habermas.
Berkeley: University of California Press.
Kolenda, P.

1976 Seven Types of Hierarchy in Homo Hierarchicus. Journal of Asian Studies

35(4):581-596.
Kopytoff, Igor

1987 The African Frontier: The Reproduction of Traditional African Societies.

Bloomington: Indiana University Press.
Marriott, M.

1969 Review of Homo Hierarchicus. American Anthropologist 71:1166-1175.

Mintz, Sidney

1985 Sweetness and Power: The Place of Sugar in Moder History. New York: Viking.

Myers, F.
1986 Pintupi Country, Pintupi Self: Sentiment, Place and Politics Among Western Desert Aborigines. Washington, D.C.: Smithsonian Institution Press.

Needham, Rodney
1975 Polythetic Classification: Convergence and Consequences. Man 10:349-369.
Parry, Jonathan

1980 Ghosts, Greed and Sin: The Occupational Identity of Benaras Funeral Priests.
Man 15(1):18-111.
Schrire, Carmel

1980 An Enquiry into the Evolutionary Status and Apparent Identity of San HunterGatherers. Human Ecology 8(1):9-32.
Schwab, Raymond

1984 The Oriental Renaissance: Europe's Rediscovery of India and the East, 16801880. Gene Patterson-Black and Victor Reinking, trans. New York: Columbia University Press.

Srinivas, M. N.
1984 Some Reflections on the Nature of Caste Hierarchy. Contributions to Indian So-

ciology 18(2):151-167.
Stevenson, H. N. C.
1954 Status Evaluation in the Hindu Caste System. Journal of the Royal Anthropolog-

ical Institute 84(1-2):45-65.

Wolf, Eric

1982 Europe and the People Without History. Berkeley: University of California
Press.

Yalman, Nur
1969 Review of Homo Hierarchicus. Man 4:123-131.

This content downloaded from 143.106.1.138 on Mon, 29 Aug 2016 17:16:02 UTC
All use subject to http://about.jstor.org/terms