Reconstructing Ethnicity

Author(s): Hal B. Levine

Source: The Journal of the Royal Anthropological Institute, Vol. 5, No. 2 (Jun., 1999), pp.
165-180
Published by: Royal Anthropological Institute of Great Britain and Ireland
Stable URL: http://www.jstor.org/stable/2660691 .
Accessed: 02/12/2013 22:58

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

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range 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 support@jstor.org.

Royal Anthropological Institute of Great Britain and Ireland is collaborating with JSTOR to digitize, preserve
and extend access to The Journal of the Royal Anthropological Institute.

http://www.jstor.org

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 RECONSTRUCTING ETHNICITY

HAL B. LEVINE
VictoriaUniversityof Wellington

Weakly theorized, lacking even an agreed definition, ethnicity has joined the list of
deconstructed concepts in anthropology. This article argues that an escape from the nihilistic
excesses of deconstructionism is possible if ethnicity is 're-located' from the narrow space it
inhabits in current theories (social, cultural or psychological) to the active interface between
the mind, society and culture. Two bodies of case material (from urban Papua New Guinea
and New Zealand) are presented here in an attempt to demonstrate that, firstly, ethnicity is
the product of an empirically available activity, classifying people according to their origins,
and secondly a great deal of cultural and symbolic content accretes to these classifications.
When ethnicity becomes subject to the elaborations of cultural or identity politics it often
develops into a focus of symbolic contestation. The symbolic and ideological elaborations of
ethnicity present analysts with phenomena analytically separate from ethnic classification
itself

In an academic version of fiddling while Rome burns, postmodern anthropolo-

gists deconstruct while other commentators use our concepts to analyse contem-
porary events (Ahmed & Shore 1995: 12).
This is especially so in regard to ethnicity. The news is full of ethnic cleansing
and genocide while the anthropologists stress that ethnicity is 'invented' and set
out to 'decentre' the notion (e.g. Hanson 1989). Another sign of ethnicity's
declining conceptual importance in our discipline is provided by Banks, in a
recent insightful textbook (1996). Banks notes that instead of confronting the
problems of 'the parent concept', anthropologists interested in things formerly
conceived as aspects of ethnicity are now writing about 'identity' and 'cultural
politics'. This is an unequivocal instance of what D'Andrade (1995: 4-5) calls
'agenda hopping'. When an area of research ceases to generate new ideas because
its practitioners now understand how complex the phenomenon under investi-
gation actually is, scholars begin to jump ship. They attack the old programme
and move on to something else. In a sense, ethnicity deserves what it gets. After
all, stronger and more central building blocks of anthropological theory such as
kinship, marriage, society and culture have been declared redundant,' so why not
ethnicity? It has obviously been in a sorry state, nomothetically, for a long time.
As Comaroff (1996: 164) says:
It is difficult not to be struck by the banality of theory in conceptual discussions of ethnicity
and nationalism ... What is remarkable ... is the sheer tenacity of this theoretical repertoire:
it has changed little in the past two decades, despite the fact that existing approaches have
repeatedly been discredited.

The goal of this article is to suggest a reconstruction of ethnicity, one that

avoids the equally deadly pitfalls of naive positivism and absolute relativism. The
main point I wish to elaborate below is that anthropologists, and sociologists,
have painted themselves into a corner by single-mindedly concentrating on the
J. Roy. anthrop.Inst. (N.S.) 5, 165-180

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
166 HALB. LEVINE

social and cultural dimensions of ethnicity. We certainly acknowledge that this

phenomenon, like everything else we find interesting about society and culture,
has a subjective dimension. However, because anthropologists neglect to incor-
porate cognition into their analyses, ethnicity remains an inadequately developed
concept.
The next section focuses on how excluding cognition from anthropological
studies of ethnicity has stunted the development of theory. This helps establish
the fact that there is a feasible alternative to stagnation, agenda-hopping and the
postmodern demolition of ethnicity. I later argue that, if we view ethnicity as a
particular method of categorizing people in the context of complex social situa-
tions, we may begin to see how it is produced in the conceptual space that Valera
et al. (1991: 179) call 'the interface between the mind, society and culture'.2

A missingdimensionof contemporatytheoty
What is banal and tenacious about ethnicity is the recurring argument about
whether ethnic identities are essentially primordial or situational.
The primordial approach situates ethnicity in the psyche, so deeply that society
and culture are bent to its will. Ethnic identities and hatreds naturally draw
people into persistent identities and antagonisms. Ethnic formation hardly
merits explanation once a resort to deep-seated passion is made. Most authors
attack primordialism by showing that ethnic formations change continually
despite ideological pronouncements about their essential nature and great
antiquity. It is, by now, well established that social and cultural traits and
identities respond to the contingencies of everyday life and entire categories of
people become submerged or transmute into something new. Banks (1996: 18-
24) says that only the Soviet ethnos theorists, particularly Bromley (1974),
continue to see ethnic groups as resilient enough to persist through long periods
of time and different types of social formation. Banks uses the essentialism of the
ethnos theorists as a foil for his conclusion that ethnicity, like other 'key
concepts', 'may not be out there in the world at all' (1996: 186). When Banks says
that these Soviet writers are the only ones that 'think that ethnicity really does
exist and really is a fundamental aspect of the human condition' (1996: 186) he
raises an important problem which is also relevant to that other hackneyed brand
of theory, situationalism. Where is ethnicity located?
Classic and influential examples of the situational approach (also called
constructivism) were produced byJ.C. Mitchell (1956) and A.L. Epstein (1958),
notable members of the Manchester School. They concentrated on the ways that
identities were expressed in micro-level social contexts in colonial towns in
Africa's Copper belt. This work contained a number of stimulating facets. The
sophisticated perspective these authors developed on the relationships between
situations and settings, micro-contexts and macro-structures, had the potential
to establish urban anthropology as a solid field of inquiry with real theoretical
problems. Furthermore, the development of new categories in towns, their inter-
relationships and influence on wider social fields, were important phenomena
that lent themselves well to exploration by ethnography, the method par excellence
of social and cultural anthropology. In particular,Mitchell and Epstein noted that
the classifications that constituted urban ethnicity served as a kind of 'cognitive
map' for urban Africans, reducing the bewildering environment of people to a

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 HAL B. LEVINE 167

few manageable categories. In essence they paved the way for the type of
theorizing advocated here, but, unfortunately, very little was built on these
foundations.
Fredrik Barth (1969) became far more influential. He developed an approach
to ethnicity that contained elements of both primordiality and situationalism.
Barth stressed the importance of boundaries rather than the cultural contents of
ethnic groups. Demarcating self from other, in-group from out-group, the
boundary component of ethnicity remained stable while society and culture
changed. Individuals, moving across boundaries in response to changing condi-
tions, could choose their ethnic identities from those available to them. Barth's
focus on the way an individual could adopt an identity (or choose a course of
action generally) was initially attractive as an antidote to the ills of the function-
alist preoccupation with the maintenance of groups. However, his ideas share the
limitations of other methodologically-individualist approaches.
To methodological individualists, social structures reflect, in summary
fashion, the strategic choices of individuals. The constraints people face, in terms
of identity choices or other courses of action, are not sufficiently analysed. Such
limits on social action appear in some of Barth's writing to be components of an
external environment, like mountains and trees, to be circumvented or ploughed
through by rational actors. Neither Barth nor the members of the Manchester
school provided a convincing account of how particular ethnic categories arise
and become salient in social action. Their points about maps, boundaries and
choice can contribute something to the study of ethnicity, but only after the fact,
when the categories in question have already become established. Abner Cohen,
who viewed ethnicity as an instrument for the articulation of informal organi-
zation in struggles for economic and political advantage, made this very point
when he heaped scorn on Epstein's, Mitchell's and Barth's flirtation with
psychology.
What it says is that people act as the members of ethnic categories because they identify
themselves, and are sometimes also identified by others, with these ethnic categories. How
do we know this? The actors say so, or so they act. Such statements and arguments will not
become more analytical if we attribute identification and categorisation to so-called cognition
and begin to construct 'cognitive maps' to 'explain' them (Cohen 1969: xii-xiii).

Although Cohen identified a crucial problem, he threw the baby away with the
bath water when he conflated cognition with psychological reductionism and
focused on informal political action. His instrumental approach is really nothing
more than situationalism with a focus on economic and political goals. These
goals may importantly influence the way that ethnicity develops, but again this is
only possible after the ethnic categories themselves have some salience in social
situations.
Cohen's critique proved influential, but social anthropologists did not all
eschew 'psychological' approaches to ethnicity. Epstein, for example, emphati-
cally went the other way. Using Erik Erikson's, Freud's and other psycho-
analytical theories, he mounted a counter-attack (1978) on Cohen's version of
instrumentalism. Epstein argued that instrumentalist approaches fail to explain
the nature of ethnic identity. However, Epstein did not adequately account for
this himself, nor did he explain why ethnicity should become an idiom of social
interaction in the first place (Banks 1996: 36-9). Much more recently, Jenkins

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
168 HALB. LEVINE

(1994) has attempted to 'rethink' ethnicity in a way that incorporates catego-

rization and other 'psychological' variables into the subject.
Although I, too, wish to establish that categorization processes are central to
ethnicity, my treatment of cognition and categorization diverges from Epstein's
and Jenkins's more conventionally psychological approach. The major problem
with their work is that they both locate important ethnic processes inside the
forces of personal development that produce an overall sense of the self
Cohen was right, at least in the limited sense that anthropologists should avoid
this kind of psychological explanation of social and cultural facts. Ethnicity may
influence psychological development and self-image, as Epstein andJenkins aver,
but considering mechanisms of personality and maturation side-track us from
explaining how particularidentities become available,to be incorporated into the
psyche or anything else. Jenkins's view of categorization shows some of the
limitations of this approach most clearly. Using the Macmillan Student
Encyclopaedia of Sociology'sdefinition of categorization, he says that categories are
externally defined while groups are defined internally. (So if someone in Port
Moresby calls another person a 'highlander' and that person thinks of herself as
a 'Chimbu', highlander is a category and Chimbu is a group.) When put this way,
the distinction raises more questions than it answers. How did the group identify
itself? Was it once a category? Can categories become groups and groups become
categories?
An important corpus of social psychological literature, unconsidered by
anthropologists writing about ethnicity, demonstrates that the relationship
between categories and groups is dynamic, something that produces social facts.
What is more, the processes of categorization engage centrally many of the long-
term concerns of anthropologists: history, class, culture, power and so forth. The
issue of how ethnic categories become salient components of social and cultural
action can be moved out of the shadows to the forefront of discussion in a
paradigm that stresses the interaction of mind, society and culture.

A cognitiveapproachto ethnicity
I propose the following simple and minimalistic definition: ethnicity is that
method of classifying people (both self and other) that uses origin (socially
constructed) as its primary reference. The use of the label 'ethnicity' for this
activity may be the anthropologist's, but the term refers to something ubiquitous
and empirically available.We know we are witnessing instances of ethnicity when
we observe people classifying people according to their origins. When the
categories in use refer to something other than origins (e.g. sexual orientation,
disability, etc.) they are not ethnic categories.3
Conceptualizing ethnicity as a particular method of classifying people has a
number of advantages. The most obvious is methodological. Ethnicity is
something we are already,as fieldworkers, preparedto observe. Charting its artic-
ulation with other aspects of our informants' social lives is a routine anthropo-
logical endeavour. A perhaps less obvious (and potentially more productive)
benefit to viewing ethnicity as a label for something people do, comes from the
attendant capacity to make a distinction, emphasized by Ellen (1993), between
classifying as an activity and the systems of classification that result from classi-
fying. The existence of a system of classification tells us nothing, a priori,about

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 HAL B. LEVINE 169

the activity that gave rise to it (Valeri 1995). Yet the ethnicity literature is full of
this (cf the objections made by Cohen, discussed above). Systems of ethnic
classification are simply taken as given, the point from which analysis flows. If,
however, ethnicity is re-conceived as a kind of classifying, analysts are naturally
drawn into fertile questions of dynamics, considering the relationships between
classifying and classifications in a more holistic fashion. How do ethnic
categories arise in social situations, become meaningful to the participants and
relate to each other? What factors enter into acts of classifying (e.g. class, culture,
religion, regionalism, etc.) to give the classifications content and meaning? Do
cognitive, non-cultural, aspects of classifying have any effect on the meaning of
categories, or is the content of ethnicity wholly accounted for by social and
cultural context?
The notion of an 'embodied mind' can help us to frame these issues. Valeraet
al. (1991: 139-40) suggest that we move away from viewing the individual as a
self-contained entity who engages an extrinsic, pre-given world. Rather, people
'enact a world ... that is inseparable from the structure embodied by the cognitive
system'. This enacting is recursive, exhibiting an 'operational closure' where
outputs become inputs into the system.
This embodied mind is intrinsically social because, as Toren (1994: 979)
reminds us, 'forms of human sociality ... are inherent in the way humans
conceptualize or bring forth their world by virtue of living it alongside others'.
We need to keep in mind, however, that such a mind is not entirely social and
cultural. People, in interaction, communicating with one another, actively
construct the experience of social and cultural groupings and identities. But the
cognitive aspects of classifying contribute something to ethnogenesis often
overlooked by anthropologists.4 The very act of classifying transforms percep-
tions of reality. Individuals become depersonalized, transmuted from unique
persons to exemplars of named groups. This group reification is typically accom-
panied by a dual accentuation: a magnification of the differences between groups
and an emphasis on homogeneity within a group.
There is an important aspect of rationality in the activity of classifying (in any
domain) emphasized by Hogg (1995). The cognitive system seems to maximize
meaning by choosing categories that best account for the similarities and differ-
ences relevant to a situation. During use, the differences between categories
become more and more accentuated. In all situations involving social classifi-
cation, out-groups are perceived as more uniform than in-groups, regardless of
the basis of categorization (Fiske & Taylor 1991). Stronger categories, like those
based on ethnicity and gender, are more resistant to disconformation, perceptual
change from subsequent interaction, than are weaker categories, such as 'day' v.
'night' people.
The implications for the study of ethnicity of these tendencies to reify groups
are profound. The simple act of labelling others, even if only 'to maximise
meaning in a specific context' and regardless of the existential reality of the classi-
fications, generates group conflict (Hogg & Abrams 1988: 20, 51). Social and
cultural facts arise that become progressively more solid and elaborate as they
continue to be used to describe identities and social situations. When, as
commonly happens in the case of ethnic classifications, this reification and
maximization of separateness is mapped onto differences in access to resources,
conflict and further ideological elaboration inevitably follow. In turn, conflict

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
170 HAL B. LEVINE

and ideology reinforce the tendency for ethnicity to become institutionalized, a

fundamental dimension of social and cultural life. Some specific examples can
help illustrate how these dynamic aspects of the act of classifying lie at the heart
of ethnicity.

appliedto casedata
A cognitiveapproach
Doing research in urban Papua New Guinea in the waning days of colonialism
(Levine 1976; Levine & Levine 1979), I encountered a kind of minimalistic
ethnicity, an ethnicity without established cultures, institutions or deep personal
significance, that seemed to lie somewhere between 'hollow categories' and the
more usual, clear-cut, meaningful ethnic groups discussed by other anthropolo-
gists. Looking back at this material, it seems to provide very clear examples of
how the embodied categorizing mind, discussed above, produces ethnicity. This
incomplete transformation of categorical labels into identities in Papua New
Guinea is discussed immediately below In the subsequent section it is compared
with a more consummated instance of similar developments in New Zealand,
which help us to separate the essential from the epiphenomenal in ethnicity.
Urban Papua New Guinea seemed permeated by ethnic hostility during my
fieldwork. Key informants, who came from the Eastern and Western Highlands
districts, felt 'coastals' dominated the country and worried that highlanders
would form a permanent underclass after independence. 'Coastals' in turn
seemed to fear and despise highlanders. Riots broke out in the coastal towns of
Port Moresby and Rabaul between highland workers and local tribespeople
during the fieldwork period, and two prominent Bougainvilleans were murdered
in the highland town of Goroka after their car hit a child. Politicians from the
affected coastal areas,citing these incidents, called for highlanders to be deported
from the coast and suggested that their constituencies should secede from Papua
New Guinea rather than continue experiencing the violence brought to them by
these outsiders.
Explosive as such incidents and accompanying rhetoric seemed, the Papua
New Guineans with whom I interacted did not explicitly set out to use or adopt
these ethnic categories, nor did they form ethnic groups in town. My informants
basically wanted ajob, some place to live and companionship. They had no other
motives than these when they conceptually carved up the social landscape of the
towns. They approached their own set of relatives, friends and acquaintances
from 'home' using a variety of ties (kinship, friendship, etc.) in their first efforts
to become established. The migrants' strategy for urban adaptation was so
pervasive and uniform that it came to be called the 'wantok system'. Wantok
(from 'one talk' or language) were people who came from the 'same area' and
were expected to help one another with food, lodging and employment. The
specific definition of wantok was highly contextual, although it was common to
use some colonially-defined administrative unit as a labelling device. For
example, to my Kepaka informants in Mt Hagen, a town close to their home,
neither enemy clansmen from their home area nor Melpa speakers from nearer
to town were considered wantok. But in the faraway capital city, these very
people, indeed all Western Highlanders ('Hageners' in Port Moresby), and
sometimes even highlanders in general, were considered to be at least potential
wantok.

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 HAL B. LEVINE 171

Even when individuals from a very specific home area used rural-group
idioms of relationship (kinship, affinity, school ties) to recruit or become
attached to wantok in towns, the networks of friends, neighbours and acquain-
tances that developed had no traditional group referent. Whiteman, for example,
noted how her informants in Port Moresby often congregated at someone's
house. 'The hosts might not know all their names or what, if any, blood or affinal
relationship they had to them. They only know that they were Chimbu, and that
they came to visit them with someone whom they did know' (1973: 93). Because
there is such a confusing profusion of 'traditional'group names in this geograph-
ically and linguistically variegated country, people often use terms that derive
from colonial maps such as 'Chimbu' (another highlands district) to talk about
themselves and others in many social situations.
This situation, where a classification is imposed on a set of people from
outside, is a very common occurrence. Ardener (1989) noted that such classifi-
cations establish what he called a 'taxonomic space' where external labels and
self-designations interact. Ardener and his students stressed the 'hollow' nature
of these designations, how they were 'constructed from without', far more than
how they came to be adopted or resisted by the people the category designated.
Chapman, for example, characterizes identifications like 'Celt' as 'balloons,
puffed up with the hot air of academic discourse and given substance in the
modern world of nationalisms and anti-nationalisms' (Banks 1996: 137). Other
students of Ardener also emphasize how ethnic categories 'tend to be both
defined and filled through the work of external agencies ... in order to define (by
opposition) state-level national identities' (Banks 1996: 142). This stress
highlights cultural inventions and constitutes a kind of deconstruction of ethnic
identity. The work is important because it reminds us that categories are neither
neutral or capricious. Their construction and deployment incorporates a whole
range of agendas: political, economic and nationalist. But the cognitive aspects of
the categorizations remain unexamined. This lacuna can be corrected by
focusing analysis on how seemingly 'hollow categories' and 'taxonomic spaces'
get filled in. My material from urban Papua New Guinea (Levine 1976; 1997)
shows that this filling-in happens in real-life situations where random events
become subjects of cognition, discourse and interaction in particular infrastruc-
tural spaces.
My prototypical example comes from conversations about an automobile
accident that occurred in Mt Hagen just prior to my first trip there in 1972. The
town was a small regional administrative centre with a population of 10,000.
Situated in a densely populated area of Melpa-speaking groups, it was inhabited
mostly by unskilled highlanders, small numbers of expatriates and educated
'coastal' Papua New Guineans. Although most of the highlanders came from the
Hagen sub-district, about 2,000 Enga speakers (from the adjacent Wabag sub-
district) also lived in the town.
Mt Hagen seemed very much the territory of the local Melpa, collectively
known as 'Hageners' after the sub-district. Peri-urban clanspeople regularly
visited the town to use the market and other facilities and visit kin. They seemed
to take a proprietary interest in the place, which was built on land alienated from
local clans. Some of these groups had recently been involved in a number of
battles over land, as population growth, expatriateplantations, cash cropping and
colonial use of land for development increased pressure for resources in the area.

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
172 HAL B. LEVINE

Self-government and imminent independence heightened 'tribal fighting'

throughout the Western Highlands. Members of rural kin groups attempted to
settle old scores and create new boundaries in the opportunity created by dimin-
ishing Australian control.
Some of my key informants in Mt Hagen came from Tambul patrol post, a
non-Melpa area, in the south-west corner of the Hagen sub-district of the
Western Highlands. Although they interacted frequently with people from local
Melpa clans, and became the friends and husbands of some, Tambul feared and
mistrusted Hageners to such an extent that they would not venture out on the
streets after dark. I had to drive my friends, Anton, Joseph and Loess, even the
shortest distances in this small town for evening interviews and on social
occasions.
During one of these get-togethers the automobile accident came up as a topic
of conversation. Four members of a local clan died when their utility vehicle
crashed with a truck driven by a man from Wabag sub-district. As the evening
wore on it became apparent that talking about the crash stimulated the articu-
lation of a rhetoric of ethnic grievance between two groups. My informants
talking about and classifying the parties to the incident, by no means academi-
cally inspired, helped 'puff up' the hollow colonial categories, 'Hagener' and
'Wabag',with social and cultural meaning. The more the crash was recounted in
terms of these group names, the more elaborate, reinforced and sedimented the
meaningfulness of Hagener and Wabag as labels for self and group identities
became. The accident served as an exemplar, a prototypical story of relationships,
stimulated by more general forces of urbanization and modernization, between
two very general categories of people.
It was difficult to understand how these general categories could be
meaningful to my informants. They seemed merely to cover over the cracks
between groups, which were not only distinct socially and culturally but often
fundamentally opposed and sometimes at war. I asked whether the terms
'Hagener' and 'Wabag' were really significant? Were these not just shorthand
labels for people competing for jobs and living space in the town from two
different parts of the Western Highlands district? Hakai (a Melpa speaker whom
I was soon to drive, hidden under blankets, through enemy territory to join in
preparations for battle against another local clan to avenge a killing) responded
like a true Hagener patriot. The Wabag, he said, could work and live where they
pleased but they had to be made to respect the property and lives of Hageners.
He said that Tambul people, because they lived in the Hagen sub-district, were
also Hageners. Anton, seemingly heedless of his own anti-Hagener xenophobia,
concurred.
A close examination of transcripts of this extended conversation (see Levine
1997: 57-61) reveals that these men conflated social and cultural traits with
administrative boundaries. They maintained, for example, that people spoke the
same language because they came from the same sub-district. They were all
Hageners. Of course they knew, from local gossip and radio reports of tribal
fighting in Wabag, that divisions and animosities parallel to their own existed
among the Enga. These counted for nothing, however, and my persistence in
pointing them out only created annoyance. Those people were 'just Enga, just
Wabag and that's all'. Later events seemed to prove the correctness of this view.
Further grisly incidents were interpreted as 'paybacks'for the crash, instances

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 HAL B. LEVINE 173

ofviolence between people labelled with these categories. Subsequent interviews

with an informant from an outlying area of Wabag, who initially felt such affairs
had nothing to do with him but who came to concede that he was a Wabag as
events proceeded, reinforced my conclusion that the classifying process itself
created social and cultural facts for residents of the town. This classification-
based ethnicity proved to be pervasive in urban centres, and in other sites of late-
colonial modernity in Papua New Guinea. Regional categories were always
employed when collections of people from different areaswere involved in inter-
action, at a segmentary level appropriate to the situation and place (Levine 1976;
1997; Levine & Levine 1979).5
The cognitive mechanisms and sequelae of ethnic categorization worked in
these situations to give ethnicity a salience that typically remains unexamined in
the literature. The use of administrative labels clearly shows how efficient
categorizations become important ingredients in the production of social and
cultural reality. Papua New Guineans could have used local group names to
identify the people they interacted with. These are, after all, their own designa-
tions, more accurate and elaborate than the classifications of Australian colonial
authorities. But they are, in fact, too elaborate to be generally meaningful. Papua
New Guineans typically ignored the complexity of village- and kinship-based
groups when describing identities to people with whom they were not closely
related.
'Class' categories also could have been used to make sense of people in situa-
tions like the incident in Mt Hagen where antagonisms over uneven devel-
opment, competition for housing, jobs, etc., were (at least to the anthropologist)
structural factors that contributed fundamentally to the realities of social differ-
entiation. These economic and political factors were intrinsic to the classifica-
tions themselves because the country was 'opened up' and developed through
the establishment of regional centres of colonial administrative control.
Differential development provided the content of the stereotypes (e.g. violent
and primitive highlanders) that attached to the categories. In the absence of an
equally efficient alternative, the readily available administrative categories
provided a system of segmentary labels that best accounted for the similarities
and differences between sets of people.
Once they began classifying people in this way, my informants proceeded to
maximize the clarity and salience of the categories by accentuating them in
conversation. As soon as the categories became attached to accounts of events,
discussions characterizing the parties became more focused and intensified. In
the Mt Hagen example differences between Hageners and Wabags were
magnified and cultural similarities of members of the same category (e.g.
Hageners all speak one language) were invented. However, Hageners and
Wabags did not create any ethnic institutions during my fieldwork. Indeed, all
this Hagener-Wabag categorical sharpening was occurring while tribal fighting
between local Melpa groups and also amongst the Enga in Wabag became
increasingly common. The problems people like my Tambul informants had
with Melpa-speaking Hageners, whose home villages were closer to the centre of
administration, also continued unabated. Since the intensification of animosity
between Hageners and Wabag was not matched by a diminution of conflict
within the categories, the ability of the regional designations to become foci of
enduring groups was continually being under-cut. They remained groups only

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
174 HAL B. LEVINE

in people's heads.
The fact that the ethnic categories can become extremely salient and wide-
spread in social interaction, while they remain devoid of the cultural markers and
permanent psychological significance stressed in much of the ethnicity literature,
suggests that culture and deep emotional attachments are not fundamental,
defining properties of ethnicity. Rather, ethnicity emerges from classifying an
activity located in Rosch's interface of mind, society and culture 'rather than in
one or even in all of them' (Valera et al. 1991: 179). Those other aspects of
ethnicity, so central to the academic literature, must then constitute elaborations
of a more basic phenomenon.

in New Zealand
ethniity:biculturalism
Elaborating
Biculturalism, an ideology calling for the implementation of a comprehensive
partnership in all areas of governmental policy and practice between Maori and
pakeha6 in New Zealand society, became a prominent cause in the late 1970s.
The Waitangi Tribunal, established by an Act of Parliament in 1975, developed
into a primary site of the production of New Zealand cultural politics in the
1980s. The Treaty of Waitangi, an 1840 agreement between the British Colonial
Office and Maori tribes, has long been seen as the cornerstone of Crown
sovereignty in New Zealand. The Treatygave Britain the right to govern and, in
turn, guaranteed chiefly control of tribal property. The subsequent history of
New Zealand has, nevertheless, seen the Maori reduced to the status of a 'fourth
world' population.
Like that of Australian Aborigines and Native Americans of the United States
and Canada, European settlement resulted in substantial social, cultural and
economic dislocation for New Zealand's indigenous population. Maori land and
other resources passed into the hands of settler governments. The people form a
minority of the nation's population,7 over-represented on all measures of social
distress: imprisonment, disease, unemployment, lack of educational qualifica-
tions, and so on. Bicultural ideology advocates that the solution to these
problems of inequality lies in 'honouring the spirit of partnership' said to be
contained in the Treaty.The Waitangi Tribunal hears claims by Maori individuals
and groups that their Treaty rights have been violated and makes recommenda-
tions to government about settling such grievances.
Tribunal cases, and commentary about them by politicians, academics,
journalists, etc., provide fascinating insights into the ideological development of
ethnic politics. Interpreting what the Treaty means (there are versions in Maori
and English that do not quite match up) and considering how a document from
1840 can be relevant to conditions existing today, has spawned an interpretativist
industry that almost rivals biblical hermeneutics (see Levine 1987; 1991b; 1997).
The aspect of all this most relevant to the concerns of this article is that a period
of nearly 160 years of interaction between settlers, the state and its indigenous
citizens is cast as the interplay of two originally hollow categories, Maori and
pakeha. Both names are as much products of colonialism as the Papua New
Guinean classifications examined above.
The indigenous people of New Zealand arrived from eastern Polynesia
approximately one thousand years ago. The kinship-oriented pre-contact
societies that they developed superficially resembled those of rural Papua New

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 HAL B. LEVINE 175

Guinea. The collective term 'Maori' (a word meaning ordinary,usual or normal,

Williams 1994) first came into use in 1820 as Europeans (pakeha) began to settle
(Ihi Communications n.d.: 6). Although iwi (tribe) and hapu (sub-tribe)
identities remain important, and can undercut Maori solidarity (Levine &
Henare 1994), the general designations have achieved a solidity and stability
noticeably absent from similar categories in Papua New Guinea. This solidity has
three main sources. Firstly, Maori iwi and hapuwere more centralized politically
than most New Guinean groups. Secondly, the autochthonous inhabitants of
New Zealand spoke a common language and shared many cultural traits. Finally,
and perhaps most importantly, the Maori experience of settler colonialism and
their status in New Zealand society provided a long-standing set of common
problems that stimulated ideology and organization, and filled in the category.
Maori, in contrast to Hageners and Wabag, developed enduring political
groups that have mounted increasingly successful campaigns to have their
grievances heard and addressed at all levels of society. Two sites of ideological
production, Tribunal cases and academic writing, show how a process of cultur-
alization of ethnicity elaborates categories in contemporary New Zealand.
In 1981 the Waitangi Tribunal received a submission from Aila Taylor of
Waitara, on behalf of Te Ati Awa, that the town's sewage system polluted
shellfish-collecting grounds used by the tribe. When a synthetic fuels plant
gained a right of discharge into the area the complainants felt the beds would be
completely destroyed. They initially phrased their claim in terms of material
deprivation, estimating the price of food lost to the tribe because of the pollution.
Te Ati Awa also maintained that the Treaty guaranteed the traditional fishery to
them. This strategy merely recapitulated a historical pattern that had brought no
redress to Maori people in the past. A large number of legal claims, where
members of a group sought compensation for actions contrary to the Treaty,had
previously failed in the courts. Legal precedent long established that since
sovereign states make treaties, the Treaty of Waitangi had no place in domestic
law.
The Tribunal re-interpreted Te Ati Awa's submission, changing it from a
narrow fiscal demand to a wider-ranging challenge to Crown sovereignty over
public resources. The reefs and seafood contained in them were taonga
(treasures) because they had cultural significance. They thus deserved protection
under the Treaty8regardless of legal ownership. This re-conceptualization was
accomplished by a procedural decision to move the hearings from the capital city
of Wellington to the tribal meeting grounds of the complainants. This change of
venue was, perhaps, as important to framing biculturalism as the re-construction
of the meaning of taonga.
In their meeting house, Te Ati Awa's etiquette of welcome, speech-making and
the like turned the proceedings into an effective Maori forum to present
grievances. The trappings of power and the repeated demonstrations of respect
which are usually directed towards the state in courtrooms went to the hosts
instead. Claimants were encouraged by the Tribunal to expand their evidence.
Freely given in a familiar setting with no cross-examination, it swamped the dry
technical accounts of the pakeha. Government officials and lawyers, used to
feeling at home and in full control of proceedings in court, had to hear Te Ati Awa
in an environment they found alien and unsettling.
The entire situation struck one lawyer as a farce where 'a decision that was not

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
176 HALB. LEVINE

a decision' (a recommendation) is made by 'a tribunal that is not a tribunal' (the

Waitangi Tribunal has no adjudicative power) 'about a treaty that is not a treaty'
(O'Keefe 1983: 137). This lack of legal power also led some Maori to doubt that
anything useful would come from the hearings. But the weaknesses of the
Tribunal as an adjudicative body, although it posed serious limitations on specific
claimants, became a source of strength for Maoridom as a whole, as the Tribunal
developed into a forum for the articulation of new interpretations of custom and
history that ultimately found their way into domestic law.
Although the Maori quest for improved status in New Zealand is far from
over, biculturalism has delivered concrete results. Maori groups now control half
of the nation's fishing resources, and a significant amount of Maori land
alienated by the Crown has been returned. The Treaty is taken seriously in law
and other forums that affect Maori interests.
While Maori organizations, businesses involved in the extraction of natural
resources, planners, the Tribunal, courts and other government bodies grappled
with the practicalities, a less anchored academic discourse also concerned itself
with the implications of biculturalism. Perhaps the best-known piece of writing
was an article by Alan Hanson (1989) in the AmericanAnthropologist.
Hanson combined the literature on 'cultural invention' developed by histo-
rians and anthropologists (Hobsbawm & Ranger 1983; Keesing & Tonkinson
1982) with the ideas of certain postmodernist thinkers and applied them to bi-
culturalism. The main point of the 'invention' literature is that rhetoric about
tradition validates contemporary interests by invoking images of continuity with
an inviolable past. Uncovering inventions raises issues about the nature of
culture, identity, objectivity and research ethics. Hanson sought to reconcile
these problems when discussing Maori culture with a new-age, postmodernist
version of cultural invention and biculturalism.
In the days when official policy envisaged assimilation or amalgamation,
representations of Maori culture stressed similarities with British New
Zealanders. Accounts of traditional religion, for example, re-cast elements of
belief systems, such as the Io cult, to resemble Christian monotheism. The
notion that the country was originally settled by the co-ordinated efforts of a
Great Fleet of canoes purportedly showed that the Maori also derived from a
group of settlers who crossed the ocean to live in a new and distant land. Hanson
maintained that Maori believe these reconstructed ideas about their culture and
origins. When they, and pakeha, 'open themselves to the emotional and mystical
impact of charisma and the non-rational ... these and other elements of the
current invention of Maori culture become objectively incorporated into that
culture by the very fact of people talking about them and practising them'
(Hanson 1989: 896).
The most interesting implication of the Maori case for anthropologists is that
it shows that anthropology itself invents culture. Hanson backs away from a
critical evaluation of such developments by emphasizing that all culture is
invented anyway. In Hanson's perspective, real traditions, transmitted faithfully
from the past, simply do not exist. He quotes a statement from Derrida to the
effect that we need a de-centred view where everything becomes discourse and
no particular discourse gets privileged treatment as truth. He argues that the
Maori of the 1760s played the same discursive games their descendants and
anthropological associates play now. 'It follows from this that the analytic task is

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 HALB. LEVINE 177

not to strip away the invented portions of culture as unauthentic, but to under-
stand the process by which they acquire authenticity' (Hanson 1989: 897-8).
Today, biculturalism promotes the idea that Maori culture provides a funda-
mental set of contrasts with pakeha culture. Maori-ness constitutes a counter-
balance to such negative pakeha traits as coldness, rationality,individualism, lack
of feeling for the land and pollution, because Maori culture promotes feeling,
sharing and communing with the environment. Hanson's mention of pakeha
culture and identity reminds us that biculturalism posits a 'pakeha-ness' with an
identity parallel to that of the Maori. But despite the fact that pakeha is part of the
same system of classification as Maori it has remained a hollow category, a mere
designation with little meaning to European New Zealanders (Pearson 1989).
This hollowness presents real problems to a biculturalist construction of reality
that reifies 'the pakeha'.
Spoonley recognizes that the viability of biculturalism demands the
construction of a cultural identity to make this category solid. He suggests that
non-Maori New Zealanders deliberately define themselves as Pakeha (note the
capital 'P'). This constitutes a political act that creates an 'imagined community'
of values, lifestyles and progressive orientations, 'a contemporary identity that
has been formed by interaction with iwi and a sympathy for their aspirations. It
is an identity informed by an understanding of both iwi histories and a self-aware
and self-critical appreciation of the ethnic history of Pakeha'. This 'fictive
ethnicity' 'affirms the centrality of biculturalism and the ambition of ... iwi. It
reflects a post-colonial position that privileges equity in terms of biculturalism'
(1995: 104-5).
Spoonley's invocation of Pakeha identity demonstrates clearly that the identity
politics of biculturalism focus on radically re-conceptualizing New Zealand
nationalism. The entire argument runs counter to the usual interpretation of
majorities as non-ethnic components of states. As Banks (1996: 158) notes,
'nationalisms ... actively seek both to enhance and reify the specifically ethnic
identities of deviant others within the nation state, and at the same time to efface
the idea of ethnic particularism within the national identity'. Biculturalism
builds upon ethnicity and provides a good example of how it becomes a
component of political ideologies with very wide ranging-goals.

Conclusions
Marcus Banks, at the end of his comprehensive and interesting review of the
literature (1996), argues that ethnicity is located not in the heart as the primor-
dialists maintain or in people's heads where situationalists find it, but in the
imaginations of analysts. He ends up adopting a postmodern perspective: the best
thing that we can do is de-centre ethnicity by deconstructing the concept,
examining its representations in the literature and exposing to critical assessment
the presuppositions that underlie these representations. The position argued in
this article is that ethnicity moves around in everyone'shead, not just the social
scientist's. It becomes shaped by consciousness and interaction, conceived here
as the interface between the mind, society and culture.9 Ethnicity remains an
ideal subject of empirical investigation because, at its core, ethnicity is an activity
that we all engage in frequently.
The main point of the case material from Papua New Guinea was that there

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
178 HAL B. LEVINE

are no completely objective differences between people that give rise to ethnicity.
Rather, the activities associated with classifying origin transform hollow
categories and create a sense of groups. These group classifications quickly
become foci of conflict, especially when origin is mapped onto differential access
to resources. Salience and elaboration flow out of categorization in social and
cultural contexts. It is after they become more solid categories that ethnic
identities and groups are further developed in ways that give the phenomenon its
great variety of expressions. The ethnicization of the schisms apparent during the
transition to independence in Papua New Guinea shows how social and
economic factors accrete to categories of origin. The development and elabora-
tions of biculturalism give us a view of a situation considerably more advanced,
where ethnic identities are consciously elaborated, manipulated and 'cultur-
alized' in the course of constructing a new nationalist ideology.'0 It is these later
elaborations that pose the greatest unresolved problems to a comprehensive
theory of ethnicity.
Once ethnicity develops into a cultural commodity, an object in the discourse
of various 'hegemonic projects' (Fairclough 1992), elaborations of its meaning
may come to overwhelm the apparent significance of the classificatory origins of
an identity. This is especially the case where public forums like the Waitangi
Tribunal emerge to shape grievances. Discourse about the Maori in New
Zealand has shifted far from contexts of classifying amongst individuals and
groups (talk 'on the ground') to the 'struggles to fix meaning' commonly called
cultural politics Jordan & Weedon 1995). These later battles are commonly
waged in terms of a hyper-reality,where classifying is buried under representa-
tions that appear more real than the things themselves.
Cultural and identity politics are rich in elaborate interpretations of ethnicity
but the ideological nature of this unfettered symbolic discourse makes it very
difficult to produce general statements about these representations.Jay (1991: 3),
speaking about sacrifice, notes that 'it is impossible to control the flood of
potential interpretations and it is very hard indeed to be really convincing about
having identified the true reasons for ritual' when one considers its diversity of
symbolic manifestations. The impossibility of separating valid from erroneous
interpretations led to the abandonment of efforts to produce general theories of
sacrifice in the nineteenth century. Something very similar seems to be
happening to ethnicity late in the twentieth century, with 'deconstruction',
'invention of tradition' and 'authenticity' becoming key words in the disen-
gagement from theory and empirical investigation. This unfortunate devel-
opment is both a consequence and a result of the ways that recent
anthropological discourse has framed ethnicity. No longer considered a
phenomenon in its own right, ethnicity has become an object of symbolization
itself both within and outside of anthropology; a label for any cause however and
wherever similarity and difference are manufactured. If we let out some of the
air and focus on how classifying origin is shaped by embodied minds, anthro-
pologists can begin to reconstruct the shattered concept of ethnicity.

NOTES
I am gratefulto Marlene Levine for her constructivecriticismsand carefulediting of earlier
draftsof this article.The commentsof theJournal'sHon. Editorand anonymousreviewershelped
me to clarifyfurtherthe analysispresentedhere.

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
 HALB. LEVINE 179

'See Needham (1971) and Brightman(1995) for examples.

2 Roschsuggeststhatthis is where all cultureknowledgeresidesbut leavesit to anthropologyto
'explorethis possibility'(Valeraetal. 1991: 178-9).
3 VWhether or not a categoryrefers to origin is a matter of empiricalresearch,not a priori
definition. For example, I discuss below how urban Papua New Guineans use regional
administrativecategoriesas shorthandsfor origin. Religioncan also be an ethnic marker,as when
it symbolizesoriginsforJews (Levine1997).'Race',at its core, is ethnicwhen 'racialcharacteristics'
areused to signalorigin, in the way that 'black'connotesAfrican'roots'in the United States.
4Hirschfeld (1988) is a notableexceptionto this generalization.He arguesthat childrenuse a
specificcognitivemechanismto classifyracialandethnicgroups.Althoughthe datapresentedhere
may be consistentwith the notion of an innatesocialcategorizingdevice, they cannotbe takenas
an indicatorof such a device in operation.Rather,we can note that categorization,whateverits
domain,has effects and these arediscussedabove.
These sourcesprovideother examplesfrom Mt Hagen, PortMoresbyand Rabaul.
6 Non-Maori people, usuallyreferringto those of Europeanorigin.
7 The 434,000Maorirepresent7.5 per cent. of the entireNew Zealandpopulation(Department
of Statistics1994:132),proportionallya much largerminoritythanAustralianAboriginesor Native
Americans.
8 The treaty of Waitangicontains a statementthat Maori may retain 'taonga katoa',all their
treasures.Defining taongato mean anythingof culturalvalue was one of the most interesting
developmentsof Tribunaldiscourse(see Levine1991b).
9 This position has implicationsfor other anthropologicalconcepts, e.g. culture. It supports
Sperber'sassertion(1985:73) that'thebelief thathumanmentalabilitiesmakeculturepossibleand
yet do not in any way determine its content and organisation'is naive. William James (see
McDermott1967:xxix) madeessentiallythe samepoint- thatconsciousnesswas not a merevessel
throughwhich ideaspass- backwhen anthropologistsfirstbeganto articulatetheirideasaboutthe
natureof culture.
10Ethnicitymay become 'deculturalized'or unelaboratedas well. Gans (1979) andAlba (1990)
note thatwhen the social-structuralsupportslikejob concentrationor neighbourhoodsdisappear,
saliencedeclines and ethnicitybecomes merely 'symbolic',somethingwith little content. I have
arguedthat this is happeningto Jews in New Zealand(Levine1997:73-118). With little structural
support, cultural content or political conflict, being Jewish in New Zealand can become a
categoricalidentity,not too dissimilarfrom being a Hagener.

REFERENCES
Ahmed, A. & C. Shore 1995. Is anthropologyrelevantto the contemporaryworld? Thefutureof
anthropology(eds) A. Ahmed& C. Shore.London:Athlone.
Alba,R. 1990.Ethnicidentity: thetransformationof WhiteAmerica. New Haven:YaleUniv. Press.
Ardener,E. 1989. Language,ethnicityand population.In Thevoiceofprophecy andotheressays(ed.)
M. Chapman.Oxford:Blackwell.
Banks,M. 1996.Ethnicity: anthropological London:Routledge.
constructions.
Barth,F. 1969.Ethnicgroupsandboundaries. London:Allen & Unwin.
Brightman,R. 1995. Forgetculture:replacement,transcendence,rereflexication.Cult.Anthrop.10,
509-46.
Bromley,Y 1974. Sovietethnology andanthropology today.The Hague:Mouton.
Cohen, A. 1969. Introduction.In Urbanethnicity (ed.) A. Cohen. London:Tavistock.
Comaroff;J.L. 1996.Ethnicity,nationalismand the politicsof differencein an age of revolution.In
ethnicpremisesin a worldof power(eds) E.N. Wilmsen & P McAllister.
Thepoliticsof diference:
Chicago:Univ. of ChicagoPress.
D'Andrade,R. 1995. Thedevelopment ofcognitiveanthropology.Cambridge:Univ. Press.
Department of Statistics 1994. The New Zealandoficial year-book,1994. Wellington:Dept of
Statistics.
Ellen, R. 1993. Theculturalrelations an analysisof Nuauluanimalcategoriesfrom
of classification: central
Seram.Cambridge:Univ. Press.
Epstein,A.L. 1958.Politicsin an urbanAfricancommunity. Manchester:Univ. Press.
Fairclough,N. 1992.Discourse andsocialchange.Cambridge:PolityPress.
Fiske,S.T. & S.E. Taylor1991. Socialcognition,2nd edn. New York:McGrawHill.
Gans, H. 1979. Symbolicethnicity.EthnicracialStud.2, 1-20.

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions
180 HALB. LEVINE

Hanson, A. 1989. The making of the Maori. Am. Anthrop.91, 890-902.

Hirschfeld, L. 1988. On acquiring social categories: cognitive development and anthropological
wisdom. Man (N.S.) 23, 611-38.
Hobsbawm, E. & T. Ranger (eds) 1983. The inventionof tradition.Cambridge: Univ. Press.
Hogg, M. 1995. Intra group processes, group structure and social identity. In Social groups and
identities(ed.) WP Robinson. Oxford: Butterworth-Heineman.
Hogg, M. & D. Abrams 1988. Social identifications. London: Routledge.
Ihi Communications n.d. Historical database. In Biculturaltraining.Auckland: Ihi Communications
and Consultancy.
Jay, N. 1991. Throughoutyourgenerationsforever. Chicago: Univ. of Chicago Press.
Jenkins, R. 1994. Rethinking ethnicity: identity, category, power. Ethnic racialStud. 17, 197-224.
Jordan G. & C. Weedon 1995. Culturalpolitics.London: Blackwell.
Keesing, R. & R. Tonkinson (eds) 1982. Reinventing traditional culture. Mankind 13(4) (special
issue).
Levine, H. 1976. The formation of ethnic units in urban Papua New Guinea. Thesis, SUNY Stony
Brook.
1987. The cultural politics of Maori fishing.J. Polynes.Soc. 96, 421-44.
1991a. Participation in the modern sector and width of ethnic interaction. In Man and a
haf: essaysin honourof Ralph Bulmer (ed.) A. Pawley. Auckland: The Polynesian Society.
1991b. Comment on Hanson's 'The making of the Maori'. Am. Anthrop.93, 444-6.
1997. Constructingcollectiveidentity.Frankfurt am Main: Peter Lang.
& M. Henare 1994. Maori self-determination. Pacf Viewpt.35, 195-210.
& M. Levine 1979. Urbanizationin PapuaNew Guinea. Cambridge: Univ. Press.
McDerrmott, JJ. 1967. The writingof WilliamJames.New York: Modern Library.
Mitchell, J.C. 1956. The Kalela dance.Manchester: Univ. Press.
Needham, R. 1971. Rethinkingkinshipand marriage.London: Tavistock.
O'Keefe, 1.1983. Waitangi Tribunal 'Decision'. New Zealand Law Review, 136-7.
Pearson, D. 1989. Pakeha ethnicity: concept or conundrum? Sites 18, 61-72.
Sperber, D. 1985. Explainingculture.Oxford: Blackwell.
Spoonley, P 1995. Constructing ourselves: the post-colonial politics of Pakeha. Injustice and identity:
Antipodeanpractices(eds) M. Wilson & A. Yeatman. Wellington: Bridget Williams.
Toren, C. 1994. Correspondence on childhood cognition and social institutions. Man (N.S.) 29,
976-81.
Valera, F., E. Thompson and E. Rosch 1991. The embodiedmind. Cambridge, MA: MIT Press.
Ethnos 60, 123-6.
Valeri, V 1995. Review of Ellen R., The culturalrelationsof classification.
Whiteman, J. 1973. Chimbufamily relationshipsin Port Moresby(New Guinea Res. Unit Bull. 52).
Canberra: A.N.U. Press.
Williams, A. 1994. A dictionaryof the Maori language.Wellington: G.P. Publications.

Reconstruire 1'ethnicite
Re'sume'
Faiblement developpee theoriquement, sans meme une definition consensuelle du terme,
l'ethnicite a rejoint la liste des concepts deconstruits en anthropologie. Cet article soutient qu'il
est possible d'echapper aux exces nihilistes du deconstructionisme en effectuant un 're-
placement' de l'ethnicit6 hors de l'espace etroit qu'elle occupe dans les theories courantes
(sociales, culturelles ou psychologiques), vers l'espace d'interaction active entre l'esprit, la
soci6te et la culture. Deux corps de mat6riel ethnographique (de Nouvelle Guinee
Papouasienne et de Nouvelle Z6lande), sont presentes dans la tentative de d6montrer qu'en
premier lieu, l'ethnicit6 est le produit d'une activit6 de nature empirique qui consiste a classer
les gens selon leurs origines, et qu'en second lieu, un contenu culturel et symbolique abondant
s'accrue a ces classifications. Quand l'ethnicit6 devient sujette aux elaborations de politiques
culturelles ou d'identite, elle est souvent d6velopp6e comme foyer de contestation symbolique.
Les elaborations symboliques et ideologiques de l'ethnicite offrent aux analystes des
ph6nomenes analytiquement distincts de la classification ethnique en tant que telle.

Dept ofAnthropology,VictoriaUniversity,PO Box 600, Wellington,New Zealand.

Hal.Levine@vuw.ac.nz

This content downloaded from 198.105.44.150 on Mon, 2 Dec 2013 22:58:19 PM

All use subject to JSTOR Terms and Conditions