Huli Archaeological Ethnography

by Dr. Chris Ballard

A fundamental point of departure for this thesis is the observation that much archaeological explanation is curiously ahistorical, insofar as it fails to address adequately the role of human agency in its accounts of the past. I take the principal purpose of archaeology to be the reconstruction of historical change in human society and in its interaction with the environment. Yet archaeology’s approach to the nature of social change and, in particular, to the structure of engagement between social and environmental change appears poorly theorized. My intention in this thesis is not to expound such theory directly – I have reservations about the claims for any theory that would seek to embrace the entire discipline – but rather to explore through a particular case study the closer texture of historical interactions between humans and their environment as a means of rethinking the pivotal role in this relationship of human agency.

Debate over the relationship between, and relative weight to be accorded to, social and ecological factors in archaeological explanation appears to have reached something of an impasse with opposed camps of opinion grouped around seemingly irreconcilable creeds. This difference of opinion reflects a broader division in archaeological thought between two loose coalitions of thought: the behaviourists, processualists, or materialists on one side and the humanists, post-processualists or “mentalists” on the other (from amongst a wide range of other invective labels). In as much as it is possible to generalise about the range of diverse views expressed within each of these groups, the respective positions they have taken on the question of the role for human agency in archaeological explanation frame the intermediate position that I want to adopt here, and are briefly sketched in the course of this chapter.

Behaviourist explanation is characterised by an avowedly scientific and empiricist approach to archaeological evidence in which explanation proceeds from the surer ground of ecological and technological observation. Cultural influences on this material record are engaged largely to address the residual ‘dissonances’ unaccounted for by strictly ecological and utilitarian explanations (Gould 1985:641, Gould and Watson 1982:367). In line with the chain of inference that this procedural sequence establishes, where technological or ecological explanations are not forthcoming, archaeologists are ‘forced back to a cultural one’ (Jones, cited in N. Thomas 1981:166). Behaviour and cognition are distinguished from one another and deliberately ‘disaggregated’, with the latter treated explicitly as a ‘black box’ (Gould 1985:642) to which archaeologists should seek no access.

The principal tenet underpinning this approach is the uniformitarian assumption that ‘processes in the past were not qualitatively different than those we observe today’ (Kent 1987:43, see also Gould 1980:x). Closely allied to this assumption is the presumption of a universal rationality which enables archaeologists to distinguish between decisions founded on sound economic and ecological rationales and those that are ecologically or economically irrational and hence a product of the specific cultural environment (e.g. Trigger 1982:35); conversely, when culture is ‘held constant’, the variability in behaviour which constitutes the real data of archaeology can more clearly be identified (Binford 1985:585) ¹ It is probably fair to suggest that the bulk of archaeological writing, and specifically the writing on archaeology of the New Guinea Highlands discussed in the following chapter, adopts this broad logic of explanation, unconsciously or otherwise.

The post-processualist critique of behavioural archaeology takes the black box of human cognition as the primary goal of archaeological enquiry. Although my sympathies lie firmly with this latter position and this thesis is itself an elaboration upon the critique of behavioural archaeology, I do not intend to rehearse in detail the central arguments of post-processual archaeology ² . It is sufficient here to note that the assumptions of uniformitarianism and of a universal and trans-temporal rationality are rejected in favour of an approach which assumes that subjectivity is culturally distinct in the present and that the past is qualitatively different from the present (J. Thomas 1991b:l6).

In its most extreme form, the post-processualist critique argues that, Social change can only be explained in terms of the social. Non-social factors, at best, set down parameters. They have no direct explanatory power. Tilley 1982:37 (emphasis in original).

While I am essentially in agreement with this position, it risks neglecting the role that non-social or environmental factors play in social change or the complexities of the relationship between change in the environment and change society. What this thesis attempts is a closer examination of this relationship through consideration of the role that perceptions of their environment play in the decisions of human agents. Bailey has attempted to find some middle ground between these two poles of opinion by relating social and ecological forms of explanation to different temporal scales in his theory of “time perspectivism” (1983, 1987) which he defines as: the belief that different time scales bring into focus different sorts of processes, requiring different concepts and different sorts of explanatory variables. Bailey 1987:7

The emerging pattern in archaeology which sees explanations for the distant past invoking ecological factors and those for the more recent past deriving from social factors may reflect the relative ease of such solutions but should not go unchallenged (J. Thomas 1988, Wobst 1990:338). A significant objection to time perspectivism, and one that Bailey himself acknowledges (1983:181, 1987:15), is that changes over longer periods of time are no more (or less) reducible to ecological causes than are changes over short periods (Ingold 1983:5). Alternatively put, while acknowledging the difficulty of conceiving of human agency compounded over millenia, it needs to be recognised that social factors are at play in change over long periods to the same extent that they are over short periods. The challenge for archaeology is to think of the scope for human agency even in the apparent absence of evidence and thus create the possibility of its identification (N. Thomas 1981:172). This possibility is not, as Murray (1994: 19) would have it, subject to varying circumstances of archaeological preservation, but rather a requirement that we look to forms of evidence and links in explanation other than those with which the discipline has traditionally equipped itself. This, in tum, requires that we reconsider the nature of archaeological evidence and its analogical interpretation.

Patrik (1985) has shown how different notions of archaeological evidence, constituted as the archaeological “record”, underwrite both the essentially positivist approach to knowledge of the processualists and the post-processual treatment of archaeological material as a text requiring interpretation. The terms employed by each approach expose contrasting sets of metaphysical implications and assumptions about the admissibility of different forms of analogical reasoning. Patrik raises the possibility that the two approaches reflect different stages in the process of archaeological interpretation: the processualist position emphasising the recovery of acts of deposition from the archaeological record and the post-processualists focusing instead upon the construction of the archaeological record as the material residue of worlds of meaning (Patrik 1985:55). Both approaches assume the presence of archaeological material as a record, though the interpretative constitution of evidence from different theoretical standpoints may yield radically different types of record (Wylie 1985:87).

We need, says Barrett (1987), to consider what it means to constitute material evidence as a record. How are the concerns of archaeological inquiry in the present and the evidence of meaningful practices in the past combined in the construction of the notion of a record? The crucial link in this process of construction is the role played by analogical inference. Archaeology, however particular positions within it might otherwise pretend, is unavoidably analogical (Wylie 1985:64). Attempts to prescribe a “scientific” model for archaeological reasoning do so at the expense of reducing analogical inferences to the status of universalist observation, flattening out the landscape of culturally informed human agency. Where analogues are admitted in archaeological explanation under the scientific model, they tend to be restricted to the most direct and mechanical forms, a process that treats ethnography as a quarry. Instead, as I argue below, an archaeological ethnography should be a forum for debate over the social contexts for material evidence, a means of understanding the role of human agency in the constitution of archaeological evidence.

Archaeology, as a form of history, must address human agency and, by extension, the intentional constitution of history. Human agency reflects intentionality, expressed as action (or practice) mediated by decision. Significantly, human action in the environment of the material world reflects particular cultural dispositions. Crops, technologies and landscapes, for example, are not universally perceived and thus acted upon in the same ways; their significance for historical agents cannot therefore be unproblematically understood across either cultures or time from the perspective of a single (and commonly Western) rationality ³ . This is not to adopt a position of extreme “emergentism” in which history is seen as entirely the product of an autonomous human agency, unmediated by its material environment. The material world influences human decisions, but it does so initially through the lens of human perception. Archaeological explanation, then, needs some understanding of environmental influences as they are understood and acted upon by human agents.

If a role for human agency in history is accepted, the extent to which agents are themselves temporally embedded and time-reflective must be appreciated, and an attempt made to comprehend the cultural constitution of that temporality. To understand the twin and inseparable influences of environmental and social factors upon the world of historical agents, we need therefore to consider the role of a “worldview” for those agents ⁴ Necessary though this perspective is, it remains but one component of the broader process of historical or archaeological explanation. While history requires an understanding of intentionality, it must also be recognised that human agents often have no sense of the broader social and environmental conditions of their actions (Shanks and Tilley 1987a:116); further, that their actions have unintentional effects, both in the sense that individual actions have unintended consequences and in the sense in which the broader significance of historical action lies beyond the comprehension of individual agents or groups ⁵ . Explanation must thus account for both intentional action and unintentional consequences, but must also seek to distinguish between them. Gardner (1989), referring to Weber’s injunction that interpretation be adequate at the levels of both cause and meaning, or structure and intention, notes that this entails that studies which address the actions of humans cannot ‘permanently dispense with a consideration of their intentional world’. Conflation of these two levels of explanation gives rise to teleological accounts of history (O’Grady 1986) which impute to historical actors intentions read backwards from their longer-term effects and legitimate an improperly temporalized vocabulary ⁶ . One means of drawing attention to the distinction between the levels of cause and meaning is to treat archaeology as a narrative discourse and consider the role of narrative structures in archaeological explanation.

The writing of archaeological history turns upon the notion of event. Yet events have no existential reality; they are not immanent in the past (Shanks and Tilley 1987b:135). Rather, events acquire valence and meaning through their position in a relative sequence of other events, their “employment” within a narrative context. Revolutions, for example, are defined as such in terms of the perceived transformation between previous and subsequent conditions (Landau 1987:112-113). However it might be available to us in the form of evidence, history is made known to us, is rendered meaningful, through the form of narrative ⁷ . While we may accept that the past “really” happened, there can be no absolute rendition or truth of that past but rather accounts of the past that bear the trace of their own historical contexts. Specific narratives, then, betray implicit assumptions or “prefigurative constructs”. The most pervasive of these in archaeological writing draws on what Jones (1990:49) identifies as ‘the progressionist legacy of the 18th century Enlightenment’. Variously labelled “progressivism”, “complexity”, “modernism” or “meliorism”, this tendency to represent historical change as a progression from simple to complex represents a temporal extension of the universalist assumptions of behavioural archaeology (Rowlands 1989, Tilley 1989:108, J. Thomas 1991a:l-2).

Relinquishing the possibility of access to an ultimate truth of the past is an uncomfortable move for many archaeologists. If there is scope for an infinite number of possible narratives to suit a finite past and yet no absolute measure of their relative value, how are we to judge amongst them? The criteria proposed by narrativists for such judgement centre around notions of the “usefulness” of narratives and their lack of “closure”. Narratives can be evaluated on the basis of their internal consistency, but those narratives with a capacity to address without undue harm the available evidence whilst simultaneously opening up the widest range of further avenues for explanation are the most useful. The least useful are those “totalizing” narratives that seek to encompass the past within a single framework of explanation to the exclusion of other possible forms (J. Thomas 1991a:178). It is important to recognize that my argument here is not a prescription for archaeological writing but rather a description of the way in which archaeological explanation appears in fact to proceed. Archaeological practice contains within it a significant wisdom about the nature of archaeological explanation, however much this wisdom might have eluded our attempts at a theoretical legislation of practice.

In the following chapter, I consider the debate over the Ipomoean revolution in Highland New Guinea, both as a case study of the role of narrative in archaeological explanation and as a context for my own study of the impact on local society and the environment of the historically recent adoption of sweet potato by Huli-speaking communities of the Southern Highlands. This focus on the recent past of the last three to four centuries affords the clearest access to the level of detail necessary to trace the processes of social and environmental change which I wish to address. Working with a single language community and paying particular attention to oral narratives about Huli history allow me to consider issues of the relationship between historical individuals and the social structures within which they acted, and which they acted upon, while also providing a sense of change over a period sufficient to inform our understanding of change on an expanded temporal scale ⁸ . A further advantage of this form of “archaeological ethnography” is the closeness of analogical interpretation it affords.

Ethnographic methods have rarely been used to full advantage by archaeologists, due in large part to misconceptions about the role of ethnographic analogy that stem from the use of ethnography in archaeology as a quarry for direct or mechanical analogues. In her review of archaeological uses of ethnography, Kent (1987) distinguishes between three different approaches: the early practice of
“anthropological archaeology”, whose goals were ‘cultural historical’ and whose methods were dependent largely upon direct analogies between present and past; “archaeologically oriented ethnography”, the provision of ‘potentially useful ethnographic material for analogs as aids in the identification of archaeological descriptions’; and “ethnoarchaeology” whose goals are ‘to formulate and test archaeologically oriented and/or derived methods, hypotheses, models, and theories with ethnographic data’. Ethnoarchaeology, as Kent views it, is specifically not concerned with cultural history or with the generation of analogues.

What all three approaches appear to share is Kent’s (1987:42-43) assertion that, ‘[a]t best, ethnographic analogies can only be limited identifications of cultural material’, that ‘analogy for identifications is valid but analogy for explanations or understanding is not’. This limited scope allowed to interpretation through analogy reflects the behaviourist framework within which much of ethnoarchaeology has been conducted, and is refuted in part by the work of Wylie (1985) referred to above. My objection to Kent’s vision of ethnography is that it hardly touches upon the real potential of ethnography for the interpretation of past social process. What I have in mind, and what this thesis attempts to illustrate, is a form of ethnography conducted by archaeologists on a temporal and spatial scale that is considerably more inclusive than that conventionally practised by anthropological ethnographers.

Most importantly, an archaeological ethnography is a means of confronting the radical difference of cultural otherness ⁹ ; a confrontation that is, in many ways, analogous to history’s engagement with temporal difference. As such, an archaeological ethnography forces reconsideration of covert universalisms and redirects attention to relationships between agents and their environments that are otherwise inconceivable. Contra Kent, culture history is a necessary component of such a study because of the important light it sheds on the historical contexts of ethnographic observations. Further to this, an archaeological ethnography that failed to concern itself with the historical constitution of its subjects would, once again, engage not “data” but living individuals and societies as a resource to be quarried. Finally, the attention paid to culture history, particularly where this is expressed through the medium of oral narratives, has the additional benefit to archaeologists of introducing alternative historicities, conceptions of history, that are at odds with Western assumptions about the nature of history and of historical explanation. A notion of Huli historicity is engaged in this thesis precisely for its value in exposing, and thus introducing new possibilities for, a conventional archaeological approach to history.

This thesis represents an attempt, through the medium of archaeological ethnography, to address the issue of archaeological explanation for social and environmental change in the recent past of Huli-speakers of the Haeapugua basin in the Southern Highlands of Papua New Guinea. The remainder of part A grounds the thesis within the relevant regional literature and in the context of a particular topic (the debate over the Ipomoean revolution), and sketches a route towards the Haeapugua basin and the specific goals of my project. Part B provides an ethnographic model for Huli society and its interactions with the environment that focuses attention on the meaning and historical significance for Huli people of their landscape. The history of a discrete portion of that landscape, the Haeapugua basin, is then described in Part C through the juxtaposition of oral historical with archaeological and palaeoecological evidence. Finally, in Part D, I seek to show how archaeological ethnography offers new perspectives on issues of interest to both anthropology and archaeology by providing a sense of the historical genesis of elements of modem Huli society and by identifying fresh areas of focus for archaeological research.

(Reprinted with permission of Dr. Chris Ballard. The Death of A Great Land: Ritual, History and Subsistence Revolution in Southern Highlands of Papua New Guinea. A thesis submitted for the degree of Doctor of Philosophy of The Australian National University, Canberra, January 1995. pp. 2-9.)

1. Gould elaborates upon Binford’s point: ‘Only by looking for and recognizing anomalies to general patterns of conformity to utilitarian expectations in human behavior can we reliably infer when and under what conditions symbolic and ideational factors make a difference in the ways people actually behave’ (1980:xi). [↩]
2. For details of the broad post-processualist critique see Hodder (1986), Shanks and Tilley (1987a, 1987b), J. Thomas (1991a). [↩]
3. Hodder (1986:13-14) offers a nice illustration of this point that has a particular resonance for the case study presented in this thesis: ‘Causes in the form of events, conditions and consequences (intended and unintended) in the world, cannot have social effects except via human perception and evaluation of them. Thus land erosion may be a cause with the effect that people abandon their village and disperse. But the effect of land erosion does not by itself determine any particular response because there are many ways of dealing with or avoiding or preventing land erosion. How land erosion or its effects are perceived, and how the possible responses are evaluated, depend on how land erosion is involved in individual social strategies within particular culture-historical contexts.’ [↩]
4. Some of the dangers of employing an encompassing “worldview” in this way are addressed in Chapters B1 andD2. [↩]
5. Binford is thus correct in his statement that archaeologists should study ‘an order of reality that was unknown to participants in ancient cultural systems’ (cited in Duke 1991:21) though it requires the proviso that the role of the “reality” which structured their actions, however inaccessible that might be, needs also to be addressed [↩]
6. Such as ‘incipient horticulture’ (Saggers and Gray 1987:119), ‘nascent agriculture’ (Feil1987:29) or . ‘protoagriculture’ (Sorenson 1976). [↩]
7. Useful accounts of narrative approaches to history are given by Spence (1982), Ricoeur (1984) and H.White (1987). [↩]
8. A strategy outlined and recommended by M. Johnson (1989:208-209); see also Irwin (1986:190) and Gosden (1989:47). [↩]
9. In spite of Kent’s attempted legislation, I use the term “archaeological ethnography” because it describes most succinctly the nature of my project. The history of the term “ethnoarchaeology” carries too heavy a weight of behaviourist usage for it to distinguish an alternative approach. [↩]