You are on page 1of 7


Heterarchy and the Analysis of

Complex Societies: Comments
Elizabeth M. Brumfiel
Albion College

A good concept enables us to organize research find- butors to this volume, heterarchy includes a number of
ings and ask interesting questions about them. It gives us different structural forms:
the ability to predict, "This goes with that," and to ex-
plain, "Now I see why this goes with that." Judged by • an array of independent, homogeneous elements
these criteria, heterarchy is a remarkably good and useful (see Ehrenreich on the distribution of ironwork-
concept. ing among sites in southern England);
Joyce White, in her American Anthropological Asso- • the membership of elements in many different
ciation presentation, reported a wonderful "ah-ha" experi- unranked interaction systems with participation in
ence in applying heterarchy to Southeast Asian data on each system determined by the needs of each
burial and craft production. The concept of heterarchy element (see Rogers on tribal interaction in the
made it possible for her to interpret evidence that had Yadkin Valley, Levy on the relations of icono-
previously eluded understanding. Other contributors also graphic motifs in Bronze Age Denmark);
comment on the usefulness of the heterarchy concept, • the membership of elements in many different
Levy observes that when viewed from the perspective of systems of ranking where the same element
heterarchy, Danish Bronze Age data on settlement, icono- occupies a different rank in the different systems
graphy, and gender "come together rather neatly." Rogers (see White on individual status in Southeast Asia,
notes that thinking in terms of heterarchy enabled her to Potter and King on settlement functions among
question her initial assumptions regarding tribal structure. the Classic Maya, Small on economic, political,
Thus, the concept of heterarchy provides new per- and status hierarchies in ancient Greece and else-
spectives on the nuts-and-bolts foundations of archeology: where);
settlement pattern data, resource procurement, artifact • the existence of two or more functionally discrete
type distributions, design elements, and burial lots. It also but unranked systems that interact as equals (see
stimulates the critical review of such basic concepts as Potter and King on production sites among the
craft specialization, the functions of central places, the Classic Maya, Zagarell on tribal relations in the
structures of tribes and chiefdoms, and the definition of Nilgiri Hills);
social complexity. Clearly, heterarchy is an important • the existence of two or more discrete hierarchies
concept. that interact as equals (see Levy on gender rela-
tions in Bronze Age Denmark, Wailes on
church-state relations in Medieval Ireland).
Why has heterarchy been interpreted in so many
But what is heterarchy? Contributors cite Crumley's different ways? Partly because the concept of hierarchy
(1979:144) definition that heterarchy is an organizational includes a number of implicit assumptions: that ranking is
structure in which "each element possesses the potential present, that ranking is permanent, and that the ranking of
of being unranked (relative to other elements) or ranked elements according to different criteria will coincide. As
in a number of different ways...." As used by the contri- it turns out, none of these assumptions is necessarily true,
126 Elizabeth M. Brumfiel

and the authors in this volume have found it necessary to Intensive lithic production occurred in small settlements
challenge one assumption or another, or all three, with located literally on top of the resource; the output of these
various models of heterarchy. Let us consider what villages was distributed primarily to other small sites.
changes the concept of heterarchy suggests for existing Only the production of fine ceramics and the distribution
models of prehistoric economies, ideologies, and politics. of eccentric flints were centered at large sites, suggesting
elite control and decision-making.
Socioeconomics and Heterarchy Levy (Chapter 5) also doubts the importance of eco-
nomic control in the chiefdoms of Bronze Age Denmark.
Numerous proposals have been made linking econom- According to Levy, virtually all Late Bronze Age settle-
ic complexity with the existence of social hierarchies ments yield some evidence of metal-working; there is no
(Brumfiel and Earle 1987). One large body of literature indication of attached specialists. The absence of field
argues that the operation of complex economies requires boundaries, fortifications, or centralized storage argues
hierarchies of coordination and control (e.g., Engels 1972 against elite control of the economy. So does the absence
[1884]; Polanyi 1944:48-49; Sanders and Price 1968; of settlement hierarchies defined either by site size or
Wright 1969; Flannery 1972). A second body of literature function. On this basis, Levy argues that Bronze Age
argues that full-time specialization will not occur without chiefs did not exercise economic control and did not
an elite to generate a reliable basis of support for craft perform important economic functions. Rather, their
specialists (Hicks 1987; Brumfiel 1987). A third body of power was based upon their control of non-economic
literature proposes that political hierarchy cannot be main- rituals and esoteric knowledge.
tained except when it rests upon some form of economic Do these cases force us to reconsider the presumed
control (Earle 1987a, 1991; Clark and Parry 1990; Pere- linkage of economic complexity with social hierarchies?
grine 1991). This linkage of economic complexity with Yes and no. Thailand does seem to present a case where
political hierarchy has encouraged archeologists to assume widespread craft specialization was sustained in the ab-
that economic and political central places coincide, result- sence of elite control and decision-making, but craft spe-
ing in a single regional settlement hierarchy. cialization remained a seasonal, household-based activity
But the contributors to this volume challenge both the of rather low scale and intensity (using Costin's [1991]
necessary connection between economic specialization and terminology). Neither full-time craft specialization nor
political hierarchy and the universal existence of a single political hierarchy emerged. In Bronze Age Denmark,
regional settlement hierarchy serving economic, political, evidence for the chiefly control of material resources may
and religious functions. yet be found. Cattle are a mobile resource whose owner-
White (Chapter 9) reviews the evidence for the orga- ship may be difficult to establish archeologically. Pro-
nization of early bronzeworking in Thailand. In northern duction locales for the largest and most elaborate bronze
Thailand, the evidence suggests the intermittent exploita- artifacts must surely exist and may eventually be found.
tion of ore sources by many different groups, metalsmelt- In the absence of field boundaries, the earthen mounds of
ing at the sources, and bronze-casting in the villages with Bronze Age Denmark might have served as symbols of
a degree of community specialization in particular bronze territorial control (Earle 1987a). Among the Maya, politi-
items. In central Thailand, the evidence suggests more cal elites did maintain control over the production and/or
restricted access to the production process, with metal- distribution of wealth items, such as fine ceramics and
working more or less monopolized by villages located eccentric flints. Extensive specialization in utilitarian
near the ore sources. But in neither case does the evi- goods was present without elite control, but, as in Thai-
dence suggest elite control over production or production land, this production may have been seasonal and house-
for a narrowly defined elite class. There is, in fact, no hold-based, dependent upon the household's subsistence
evidence for elite organization of production or exchange. agriculture.
Potter and King (Chapter 3) present similar data for However, these discussions establish that certain
the production of utilitarian ceramics and lithics in the aspects of the economy have not been investigated thor-
Classic Maya lowlands. Ceramic production occurred in oughly enough and that some widely held assumptions
a series of small regional villages; neither production nor about the organization of pre-industrial economies are
distribution was localized around large regional centers. probably wrong.
Heterarchy and the Analysis of Complex Societies: Comments 127

The possibility of specialized production at locales the way of critical understanding. For example, in discus-
outside regional centers has been recognized for some sions of specialization, heterarchy is used to refer both to
time. For example, Costin (1991:13-15) discusses the specialists who are not under the control of political lead-
concentration of craft specialists and the factors that deter- ers and to specialization that is geographically dispersed.
mine whether production will be dispersed or nucleated. Costin (1991) argues convincingly that these are two
First on her list of variables is environmental diversity, as separate dimensions of specialization (which she calls
suggested by Potter and King for the Maya; second is "context" and "concentration") that are determined by
transportation cost, clearly an important factor in New different sets of variables. It would be best to maintain
World societies where draft animals were absent and this terminological and analytical distinction.
navigable waterways were limited. Given these environ- Second, more attention needs to be paid to the struc-
mental factors, archaeologists should have expected dis- tural contexts in which heterarchy (in its various forms)
persed craft production in the New World and should develops. Ehrenreich's (Chapter 4) analysis of metalwork-
have carefully considered its implications for economic ing in Bronze and Iron Age Britain is an example of the
organization, political organization, and settlement pat- kind of work that should be pursued in the future.
terns. Potter and King are absolutely correct in arguing Ehrenreich carefully documents change in metalwork-
that this type of specialization has been neglected, largely ing from a higher degree of craft specialization during the
because of assumptions about the importance of hierarchy. Middle Bronze Age (when it was hierarchically organized)
Heterarchical models are necessary for an accurate under- to lower levels in the Late Bronze Age and Iron Age. He
standing of the Maya economy. attributes this trend to the introduction of new technol-
Levy's review of the settlement data for Bronze Age ogies that simplified metalworking, making it more ac-
Denmark clearly establishes that the central places and cessible to individuals with minimum training. However,
three-tiered settlement hierarchies that information theo- these new technologies yielded a softer, lower-quality
rists have postulated for chiefdoms are not met with in the product. Therefore, Ehrenreich asks why these inferior
Danish Bronze Age. They are also not present in Hawaii products became popular. He observes that the new tech-
(Timothy Earle, personal communication, 1993). Thus, nologies were accepted in the context of climatic stress
while chiefdoms may have two or three levels of decision- and social instability. Following Crumley (1990), Ehren-
making as suggested by Johnson (1973), archeologists reich suggests that heterarchical organization was pre-
should not expect to find these expressed in settlement ferred in Iron Age Britain because heterarchy is more
pattern hierarchies. The implications of this for modeling flexible than hierarchy and more suited to unstable condi-
chiefly power and for identifying its existence in the tions. This is an intriguing general hypothesis that can
archeological record need some careful consideration. guide further research.
White observes that Southeast Asia sites in the first
and second millennia BC provide no evidence of elite Ritual and Heterarchy
organization of production or exchange, but they do ex-
hibit social differentiation, specialized production, greater Levy (Chapter 5) suggests that control of ritual cere-
elaboration of craft production, and increasing variation monies and esoteric knowledge supplies a basis for differ-
in wealth over time. She asks if this is not social com- ential status and influence in societies such as Bronze Age
plexity in the absence of hierarchy. She argues that the Denmark where the economy assumes a heterarchical
Southeast Asian data strengthen Crumley's (1987) case form and elite economic control is lacking. Wailes (Chap-
against of the general equation of social complexity with ter 6) argues that religious and political functionaries in
hierarchy, This is a fundamental issue to which I will medieval Ireland were ordered in parallel hierarchies of
return in the conclusion to these comments. equal status and power. These two cases raise important
In the future, the study of heterarchy in socioeconom- questions. Can ideological control operate independently
ics will benefit from research along two lines. First, a of political and/or economic control? Can ideology serve
more explicit terminology needs to be introduced. As as a basis for hierarchy when the political and the eco-
observed above, the term heterarchy refers to a number nomic organization is heterarchical? There are strong
of different structures. While this terminological impreci- arguments on both sides of these questions,
sion is probably useful for heuristic purposes, it gets in
128 Elizabeth M. Brumfiel

On the one hand, structural Marxists have argued Societal Control and Heterarchy
vigorously that control of "the imaginary means of pro-
duction" was a primary source of power to leaders during In this volume, the concept of heterarchy is applied
the early stages of the development of social inequality to what we have been calling egalitarian, ranked, and
(Friedman 1975:171-178; Godelier 1978). Mann (1986:1- stratified societies. This suggests that heterarchy is proba-
33) affirms the relative autonomy of power based on bly not any single type of social structure; rather, it is a
ideology, economics, military organization, and politics. principle of social organization, like kinship, that is re-
In contrast, Demarrais et al. (1996) propose that ideo- worked and assumes different roles depending upon its
logical control is necessarily linked to economic control. structural context. We probably should not use heterarchy
They argue that ideology can be monopolized only to replace the tribes-chiefdoms-states terminology with
through the control of its material expression in objects which we are familiar; instead, we should use heterarchy
and ritual action. Control over the material expression of to look at these constructs differently.
ideology requires, in turn, the control of at least some Rogers (Chapter 2) offers new insights into the con-
forms of production. cept of the tribe. Focusing on 29 Late Woodland villages
This issue is not easily resolved on the basis of along the Yadkin River, Rogers finds that no hard-and-
Levy's and Wailes' examples. As discussed above, elite fast boundary separates these villages into the discrete
control of ritual ceremonies and esoteric knowledge may political entities named in the ethnohistoric record. In-
have been linked to some form of elite control of the stead, the absence of fall-off frequencies for stone types
economy during the Danish Bronze Age. The Church was by distance from source suggests direct access to lithic
certainly endowed with substantial wealth in medieval supplies, perhaps facilitated by rules of exogamy and
Ireland (but lawyers and poets were not—did all three dispersed lineages that insured universal access to lithic
groups exercise comparable power?). We await further sources. In addition, the absence of discontinuities in
archeological assessments of ritual control in the absence associations of ceramic attributes suggest the free flow of
of elite control of the economy. In addition, future re- information, pots, and/or people (a situation also de-
search should explore the conditions under which the scribed by McPherron 1967).
differentiation of political and religious power emerges Such a lack of boundaries is found in other egalitarian
and the consequences of parallel hierarchies for subse- societies, such as those described in Hodder's (1982:73)
quent political and ideological development, ethnoarcheological study of Baringo, Kenya; M[i]n the
Levy (Chapter 5) and Rogers (Chapter 2) provide a Baringo area there is a continual tension between bound-
really striking method for measuring the degree of ideo- ary maintenance and boundary disruption; the one exists
logical control in prehistoric societies. Levy for metal in relation to the other." This unbounded social sphere is
objects and Rogers for ceramic vessels describe situations maintained, Rogers suggests, by the mobilization of sta-
where the association of symbolic motifs on decorated tuses and allegiances based on a number of different
objects do not follow any regular rules. Motifs are com- principles. According to a review by Flanagan (1989), the
bined and recombined in various ways, and as Levy ob- ability to mobilize statuses on a number of different prin-
serves, these multiple combinations provide great opportu- ciples is now understood by ethnographers to be the key
nities for social manipulation. Such situations can be to maintaining egalitarian relations. Flanagan (1989:259)
equated with the hierarchical organization of social dis- notes that, instead of being the simple homogeneous soci-
course, which would indicate very little elite control over eties that we once envisioned, egalitarian societies are
ideology. Such cases would contrast with situations where characterized by "not the absence of rules but the sheer
strict rules govern the associations of symbolic motifs, complexity of equality-maintenance rules they must imple-
which in theory would occur when elites more strictly ment." This is borne out by Rogers' work. If heterarchy
controlled ideology. Earle (1987b) describes instances of distinguishes a particular social type, then societies where
ideological control in Hawaii and the Inca empire; the egalitarian relations are maintained through the existence
study of iconography in these two contexts would provide of a number of overlapping and discrepant social princi-
an excellent test for the general validity of design pattern- ples are the most suitable candidate.
ing as a gauge of ideological control. White (Chapter 9) also presents a case where individ-
ual identity was a composite of various achieved and
Heterarchy and the Analysis of Complex Societies: Comments 129

ascribed statuses based on a number of different princi- ing economic resources and functions among various
ples. White uses burial data (i.e., differences in grave subject groups. This allocation enabled the state to main-
contents and treatment) to argue for a multifaceted, multi- tain power by balancing various power-holding groups
lateral system of status and social differentiation, possibly rather than attempting to control each of them directly
based on lineage, marriage, prowess, and wealth. As from above.
White's analysis continues, it will be interesting to see if Small (Chapter 7) argues that the Greek city-states
the various dimensions of social status can be teased apart present a situation where political, economic, and status
and identified as dimensions of variability in the burial hierarchies were not integrated. In Greece, politics, eco-
program. A promising beginning is already evident in nomics, and status were represented by separate, largely
White's identification of individuals whose social persona autonomous institutions. However, the aristocracy, an
include the role of craft specialist. economic elite who owned agricultural estates, slaves, and
Although ranking and stratification were absent in the trading ships, tolerated autonomous political and religious
groups studied by White, the differentiation of social institutions because it served their interests. An autono-
status through competition is suggested by the elaboration mous, democratic, state permitted aristocrats to find allies
of ceramics and metalworking. White resists calling these among non-elite citizens. These citizens provided the
societies chiefdoms, and her argument is convincing. But coercive force necessary to preserve slavery, which was
she goes on to question the utility of popular archeological the basis of aristocratic wealth. Thus, the political struc-
models of chiefdoms arguing that in Southeast Asia the ture was hierarchical but in the service of a highly strati-
"correlates are not correlating." I disagree. The data fied economic system. A parallel might be drawn between
suggest that these societies are not chiefdoms, and many the democratic governments of Greece that served the
of the archeological correlates of chiefdoms (e.g., eco- interests of aristocratic slave-owners and the interest
nomic control, settlement hierarchies, and distinctive group democracy of the United States that serves the
classes of burials) are absent in the archeological record. interests of the hierarchically structured capitalist system
In other words, the "correlates" do correlate pretty well; (see Foley and Yambert 1989).
none of them is present.
In this volume, heterarchy is often opposed to hierar-
chy so that the two appear to be mutually exclusive: the CONCLUSIONS: HETERARCHY
more heterarchy, the less hierarchy, and vice versa. It AND SOCIAL COMPLEXITY
seems as if heterarchy might be used as a synonym for
egalitarian. But Wailes (Chapter 6), Zagarell (Chapter 8), And so we return to the question posed by White: is
and Small (Chapter 7), who discuss hierarchical organi- heterarchy a form of social complexity? Certainly, heter-
zation within states, demonstrate how heterarchy cannot archy creates complexity in the lives of individuals. When
only co-exist with hierarchies of control but actually production is non-specialized or carried out by part-time
strengthen inequality and dominance. specialists, the lives of individuals are far more complex
The sacred and secular hierarchies described by than they are in complex economies. As White and Rogers
Wailes were created, historical records tell us, to accom- point out, the multiple activities, exchanges, and schedules
modate the cadet lineages of royal families. A similar of individuals in hierarchically organized economies
division of sacred and secular power occurred in response result in more varied and challenging lives for the individ-
to elite competition in Tonga (Gailey 1987:69-71). Thus, uals involved. The maintenance of hierarchical social
the creation of one form of heterarchy, parallel hierar- relationships also involves great complexity for individual
chies, has served as a means of containing tensions among actors. Rogers observes that in the Yadkin Valley, "au-
elites that might otherwise disrupt the system of elite tonomous individuals sustained access [to resources] by
control. Thus, heterarchy can be instituted for the purpose maintaining level social relations, forging alliances, inter-
of preserving hierarchy. marrying, and practicing high mobility." With these sce-
Zagarell arrives at a similar conclusion. He believes narios in mind, Rogers characterizes the change from a
that the hierarchical relations among tribes in the con- hierarchical tribal system to a hierarchical chiefdom as
temporary Nilgiri Hills originated in the administrative a "collapse" involving the homogenization and streamlin-
policies of premodern states. These states ruled by allocat- ing of organization.
130 Elizabeth M. Brumfiel

However, in the literature on social evolution, com- 1987 A dialectical critique of hierarchy. Pp. 35-56 in Power
Relations and State Formation. T.C. Patterson and C.W.
plexity is most often defined as a property of systems
Gailey, eds. Washington, D.C.: Archeology Section, Amer-
rather than individuals. For example, Flannery (1972:409) ican Anthropological Association.
states that the complexity of a system "can be measured 1990 A Critique of Cultural Evolutionist Approaches to Ranked
Society with Particular Reference to Celtic Polities. Paper
in terms of its segregation (the amount of internal differ-
presented at the Annual Meeting, Society for American
entiation and specialization of subsystems) and centraliza- Archaeology, Las Vegas.
tion (the degree of linkage between the various subsys- DeMarrais, Elizabeth, Luis Jaime Castillo, and Timothy Earle
1996 Ideology, Materialization, and Power Strategies. To appear
tems and the highest order controls)...." The challenge in Current Anthropology 37(1).
posed by heterarchy is, can segregation occur in the ab- Earle, Timothy K.
sence of centralized control? We return again to White's 1987a Chiefdoms in archaeological and ethnohistorical perspec-
tive. Annual Review of Anthropology 16:279-308.
observation that, in Southeast Asia, social differentiation, 1987b Specialization and the production of wealth: Hawaiian
specialized production, craft-production elaboration, and chiefdoms and the Inca empire. Pp. 64-75 in Specialization,
wealth variation clearly developed over time while cen- Exchange and Complex Societies. E.M. Brumfiel and T.K.
Earle, eds. Cambridge: Cambridge University Press.
tralized control was absent. 1991 The evolution of chiefdoms. Pp. 1-15 in Chiefdoms: Pow-
The coupling of differentiation and hierarchy is so er, Economy, and Ideology. T. Earle, ed. Cambridge.
Cambridge University Press.
firm in our minds that it take tremendous intellectual
Engels, Frederick
effort to even imagine what differentiation without hierar- 1972 The Origins of the Family, Private Property and the State.
chy could be. In this effort, archeological data can help E. Read, ed; R. Vernon, transl. New York: Pathfinder
Press [orig. 1884].
guide our thinking; the data will resist the appearance of Flanagan, James G.
order until we have developed the proper heterarchical 1989 Hierarchy in simple egalitarian societies. Annual Review of
model to fit the case. To a large extent, the possibility of Anthropology 18:245-266.
Flannery, Kent V.
complexity without hierarchy is an analytical and empiri- 1972 The cultural evolution of civilizations. Annual Review of
cal question that will require substantial model-building Ecology and Systematics 3:399-426.
and testing in the future. But archeologists who explore Foley, Michael W., and Karl Yambert
1989 Anthropology and theories of the state. Pp. 39-67 in State,
the heterarchical organization of past societies that were Capital, and Rural Society. B.S. Orlove. M.W. Foley, and
not linked to states, empires, or capitalist world systems T.F. Love, eds. Boulder: Westview Press.
Friedman, Jonathan
may be in an excellent position to give advice to social
1975 Tribes, states, and transformations. Pp. 161-202 in Marxist
activists who seek to institute a more egalitarian society in Analyses and Social Anthropology. M. Bloch, ed. Associa-
our own complex contemporary world. tion of Social Anthropology Studies, no. 2. New York:
Gailey, Christine Ward
1987 Kinship to Kingship: Gender Hierarchy and State Forma-
REFERENCES tion in the Tongan Islands. Austin: University of Texas
Bnimfiel, Elizabeth M. Godelier, Maurice
1987 Elite and utilitarian crafts in the Aztec state. Pp. 102-118 1978 Infrastructure, societies, and history. Current Anthropology
in Specialization, Exchange and Complex Societies. E.M. 19:763-771.
Brumfiel and T.K. Earle, ed. Cambridge: Cambridge Uni- Hicks, Frederic
versity Press. 1987 First steps toward a market-integrated economy in Aztec
Brumfiel, Elizabeth M., and Timothy K. Earle Mexico. Pp. 91-107 in Early State Dynamics. H.J.M.
1987 Introduction. Pp. 1 -9 in Specialization. Exchange and Com- Claessen and P. van de Velde. eds. Leiden: E.J. Brill.
plex Societies. E.M. Brumfiel and T.K. Earle. eds. Cam- Hodder, Ian
bridge: Cambridge University Press. 1982 Symbols in Action: Ethnoarchaeological Studies of Material
Clark, John E.. and William J. Parry Culture. Cambridge: Cambridge University Press.
1990 Craft specialization and cultural complexity. Research in Johnson, Gregory Alan
Economic Anthropology 12:289-346. 1973 Local Exchange and Early State Development in Southwest-
Costin, Cathy Lynne em Iran. Anthropological Papers. 51. Ann Arbor: Universi-
1991 Craft specialization: issues in defining, documenting, and ty of Michigan Museum of Anthropology.
explaining the organization of production. Pp. 1-54 in McPherron, Alan
Advances in Archaeological Method and Theory, vol. 3. 1967 The Juntunen Site and the Late Woodland Prehistory of the
M. Schiffer, ed. New York: Academic Press. Upper Great Lakes Area. Anthropological Papers. 30. Ann
Crumley, Carole L. Arbor: University of Michigan Museum of Anthropology.
1979 Three locational models: an epistemological assessment for Mann, Michael
anthropology and archaeology. Pp. 141-173 in Advances in 1986 The Sources of Social Power, vol. 1. Cambridge: Cam-
Archaeological Method and Theory, vol. 2. M. Schiffer. bridge University Press.
ed. New York: Academic Press.
Heterarchy and the Analysis of Complex Societies: Comments 131

Peregrine, Peter Sanders, William T., and Barbara J. Price

1991 Some political aspects of craft specialization. World Ar- 1968 Mesoamerica: The Evolution of a Civilization. New York:
chaeology 23:1-11. Random House.
Polanyi, Karl Wright, Henry T.
1944 The Great Transformation. New York: Rinehart. 1969 The Administration of Rural Production in an Early Meso-
potamian Town. Anthropological Papers, 38. Ann Arbor:
University of Michigan Museum of Anthropology.