Sweet Potato and Highland Society

by Dr. Chris Ballard

The character of historically modern societies in the central valleys of the New Guinea Highlands owes a considerable debt to the conjunction of the imported South American sweet potato and the Southeast Asian pig ¹ . The first European miners, missionaries and colonial administrators to encounter societies of the Papua New Guinea Highlands between the 1920s and the 1960s made frequent comment on the obvious importance of sweet potato as the principal human staple and main source of fodder for pigs, which were both the most common domesticates and the most significant non-human form of valuable. The likely dates of introduction to New Guinea of sweet potato and pigs and the nature of their adoption by and consequent impact upon Highlands societies provide perhaps the most enduring problems for archaeologists working in the region; problems that also have a significant bearing on issues of concern to anthropologists and geographers, such as the causes of variation in agricultural production and social formation amongst Highlands societies.

One consequence of this convergence of interest has been the development of a wide-ranging inter-disciplinary debate over the historic constitution of Highlands societies, much of which addresses the role of the introduction and adoption of sweet potato. This debate over the possibility of a sweet potato or “Ipomoean” revolution, initiated by James Watson (1965a, 1965b), has provided a forum in which ideas about history and explanation have been brought to bear on an increasingly complex body of observations by archaeologists, ethnographers and geographers ² Although many of the positions and theories which I refer to under the rubric of the “Ipomoean revolution debate” have either opposed the very notion of a crop revolution or extended argument into regions or periods where the sweet potato is presumed absent, the issues initially raised by Watson have fundamentally shaped this wider debate.

This debate, which provides a critical context for my thesis, is presented in Appendix A1 within the framework of a chronological narrative. This allows for an understanding of the development of different positions within the debate and provides a perspective on the relationship between archaeological revelation and anthropological speculation. My account of the development of the debate is partial as there are two propositions that I want to pursue. The first is that the early phases of the debate were
crucially formative for archaeological thought and research in the region. The second is that anthropological speculations, however poor their apparent grounding in archaeological fact, have proved consistently to be more productive and insightful in terms of understanding historical processes of social change in the New Guinea Highlands than the models developed by archaeologists on the basis of their own evidence (Feil 1989:119). A brief synopsis of the major positions adopted in the debate is given here in order to set the context for a more focused discussion of the archaeological evidence for the introduction of sweet potato and pigs. This is followed by a review of the ways in which conventional archaeological evidence has been deployed in the explanation of social change and of the prospects for widening the net of archaeological enquiry to address other forms of evidence.

The Ipomoean Revolution Debate: A Synoptic Account

Building upon the common observations on the importance of sweet potato and the knowledge that the ultimate origins of this tuberous crop lay in tropical America, Watson (1965a, 1965b, 1967) proposed that the adoption of sweet potato in the Highlands, presumably after its introduction to the Indonesian archipelago in the 15th or 16th century and to the Philippines in the 16th century (see A2.3 below), must have had a revolutionary impact on local society. Archaeological research in the Highlands, which had begun only in 1959, had already demonstrated an antiquity of occupation slightly in excess of 10,000 BP, but shed little direct light on agricultural history (S. and R.Bulmer 1964). Watson’s modelling of possible scenarios for the impact of sweet potato adoption was thus speculative in terms of the archaeological evidence available. Drawing instead on the treatment of sweet potato in local oral traditions, ritual practices and agricultural technologies, Watson suggested that the dense settlement of the central Highlands valleys most probably reflected a post-Ipomoean increase in population. Sweet potato has a number of advantages over the postulated previous staple, taro (Colocasia esculenta), including a greater productivity on poorer soils, in drier environments and at altitudes above 2000 m (Clarke 1977), a potential which, Watson argued, Highlanders would quickly have recognised and exploited. A key element of Watson’s modelling, and one on which he later elaborated, was the link in the Highlands between sweet potato and the production of pigs, which he held to be significant both for their protein and for their value as the principal medium of exchange in the region.

Watson argued that the scope for population increase and the development of major networks of ceremonial exchange around the production and circulation of pigs, taken together, represented an ‘Ipomoean revolution’. Watson’s papers coincided with the discovery in 1966 of buried agricultural drainage systems in swamps in the Wahgi valley (Figure A4), for which a preliminary date of 2300 ± 120 BP (ANU-44) was soon available (Golson et al. 1967). Equipped with this new finding, a seminar convened in Canberra in 1967 sharply dismissed Watson’s preferred model of an Ipomoean revolution. In their review of the seminar’s conclusions, Harold Brookfield and Peter White (1968) expressed a preference for archaeological evidence over anthropological speculation, for ‘pragmatic’ or ‘technological’ accounts of history over Watson’s ‘sociological’ vision, and for ‘gradualism’ or ‘evolution’ rather than ‘revolution’ accounting for the rate of change. This reaction to Watson’s proposals was subsequently reflected in the ecological and technological emphases in research and explanation that have largely dominated Highlands archaeology since the 1960s.

Continued excavation of the Wahgi valley agricultural sites over six field seasons between 1972 and 1977 was conducted by an interdisciplinary team led by Jack Golson. This complex of sites, referred to collectively as “Kuk swamp” after the largest of the Wahgi valley sites at the government’s Kuk Tea Research Station, has been the focus of the most intensive archaeological study in the Highlands. Detailed results from Kuk have begun to emerge only recently (Bayliss-Smith and Golson 1992a), but the broad outlines of the chronology and significance of Kuk have been sketched in an extensive series of articles by Golson (listed in Appendix A2).

A deep stratigraphic sequence at Kuk swamp has revealed a succession of superimposed drainage systems, interlain with inwashed sediments and reworked volcanic tephras (Figure A5). this sequence, Golson has distinguished six principal phases of drainage and use of the Kuk swamp extending back over the last 9000 years, all but the two most recent phases being separated by long periods of disuse and abandonment of the drainage systems. The last phase, Phase 6, must have terminated by 1933, when the first European explorers documented the complete reversion to swamp of the Wahgi wetlands (Golson 1981b). An apparent general abandonment of the floor of the Wahgi valley by the 1920s, explained by Paul Gorecki (1979) in terms of an epidemic of malaria, deprived Golson’s team of any direct ethnographic analogies for wetland use in the Wahgi area. ³

Details of Golson’s explanations for wetland reclamation and abandonment at Kuk are described in Appendix Al. By 1977, Golson essentially accounted for wetland reclamation in terms of pressure on dryland resources, as a consequence of inexorable population increase, which forced communities into the labour-intensive strategy of wetland drainage. In explaining wetland abandonment Golson referred to the releasing effects of innovations in agricultural technology, which permitted higher yields and returns on labour from dryland gardens. Thus the end of Phase 3 marked the development of complete soil tillage (for which soil aggregates in the sediments infilling the Phase 3 drains served as evidence), the end of Phase 4 coincided with, and could be accounted for, by the introduction of tree-fallowing in dryland gardens (witnessed by an increase in Casuarina pollen) and the end of Phase 5 was the result of the innovation of raised bed gardening, assisting in the intensification of dryland agriculture.

A renewed interest amongst anthropologists in the issues surrounding Watson’s proposed Ipomoean revolution saw three influential pieces of writing in 1977 return to the role of pig production in the genesis of modern Highlands societies (Morren 1977, Watson 1977, Modjeska 1977). Each of these authors argued that the development of intensive regimes of pig production was crucial to the character of modern societies: George Morren suggesting that destruction of forest resources led to an increased emphasis on the production of domesticated pigs to supplement the loss of protein sources; Watson that the prestige associated with pigs could account for the historical diffusion of both pig breeding and sweet potato; and Nicholas Modjeska that pig production met social as well as dietary needs and that the history of Highlands society was thus linked to change in the nature of social demands on production.

Modjeska’s work, in particular, presented the first significant challenge to the assumptions of the Canberra seminar and to the prevailing logic of explanation in Highlands archaeology. In place of a vision of human agency bounded by calorific requirements and technological limits, Modjeska proposed that transformation in production regimes reflected the renegotiation of cultural concepts of value and of labour. In particular, Modjeska suggested that there may have been social motives at play in the commitment of labour to wetland drainage which turned upon the greater potential for production of drained wetland soils. Where phases of wetland reclamation had formerly been regarded as the enforced result of pressure on dryland resources, they could now be reconsidered as evidence for the development of social demands on production and for changes within, rather than external to, local society.

These changes, Modjeska argued, could best be identified in terms of transformation in a society’s relations of production: For a transformation of production systems to have taken place, new relations of production were required in addition to new materials. People had to choose to work harder in order to produce more. Modjeska (1977:87)

Through comparison of the production regimes of eight different Highlands societies, Modjeska proposed that the ratio of pigs to people in a society could be taken as a rough index of the intensity of production and, by extension, the complexity of social relations of production in that society. By ranking these societies in order of increasing intensity of production, Modjeska continued, some sense of the historical trajectory of change could be derived.

Modjeska’s comparative approach to the history of Highlands society has been taken up by a number of other anthropologists, including Daryl Feil, Maurice Godelier and Pierre Lemonnier. In his application of Modjeska’s model to a major review of the regional literature for the Papua New Guinea Highlands, Feil (1987) attempted to show how variation in all aspects of social formation can be related to the relative intensity of production. Arguing that productive intensity increases amongst Highlands societies along a cline that runs from the Eastern Highlands in the east to Enga Province in the west, Feil proceeded, like Modjeska, to suggest that this variation in modern “ethnographic” societies could be read as an analogue of the historic evolution of the most developed societies in the west. Although Fell’s commitment to his notion of east-west clinal variation tended to obscure more subtle forms of variation (see Appendix A1), he reiterated Watson’s important insight that the nature and the impact of the adoption of sweet potato were likely to have varied from community to community, depending upon the nature of existing technologies and relations of production.

Godelier (1982) and Lemonnier (1990) together have shifted the focus of attention from ethnographic variation in production regimes towards the nature of difference in principles of exchange, characterised as “global social logics”. Much as Fell has pursued Modjeska’s use of the ratio of pigs to people as an index of the general intensity of production in a given society, Godelier and Lemonnier have taken up another of Modjeska’s arguments in addressing the roles of leaders in different communities as a means of identifying the underlying logics of exchange. At either extreme of the range of structural transform-ations yielded by this approach are the “Big Man” and the “Great Man”, to which Lemonnier has more recently added the intermediate “Leader”; thus Great Men are found in societies where the logic of direct exchange of identical materials (marriage partners, for example) is dominant, while Big Men correspond to social logics in which the principle of indirect or non-equivalent exchange predominates. Though there must be reservations about details of their reconstructions of Highlands history, the significance of these more recent contributions of Modjeska, Feil, Godelier and Lemonnier to the Ipomoean revolution debate is the emphasis they have placed on the role that changes within society, such as change in principles of exchange or in demands on production, must have played in the development of modem Highlands society.

Modejska’s critique of the Canberra seminar’s influence on Highlands archaeology an immediate response the form of a series of papers by Golson, written after 1977, in which he has attempted to introduce a more active sense of human agency to his explanations for the Kuk swamp sequence by considering the possibility that swamp reclamation conferred a productive advantage upon those communities with access to swamps (Golson 1982b). This paper has also been the occasion for Golson’s most sustained consideration of the possible impact of sweet potato on Wahgi valley society, in which he speculates on the reconfiguration of local variations in inequality brought about by the introduction firstly of sweet potato and then, in the aftermath of colonial contact and in unprecedented quantity, of shell valuables. Although there has been little new field research at Kuk since 1981, excavations at Yeni swamp in the adjacent Jimi valley (Gorecki 1989) and in the Arona valley in the Eastern Highlands (Sullivan, Hughes and Golson 1987), and palaeoecological results from Norikori swamp and the Baliem valley in Irian Jaya (Haberle n.d., Haberle et al. 1990) have begun to provide a regional context for the Kuk swamp sequence.

Within this brief framework of the development of the Ipomoean revolution debate, we can now consider the specific evidence for, and models for, the introduction of sweet potato.

(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. The island of New Guinea is currently divided into two politically distinct halves: the Indonesian province of Irian Jaya in the west and the independent nation of Papua New Guinea in the east. My primary concern in this thesis is with the Highlands region and largely with the better-documented portion of the Papua New Guinea Highlands; more specifically, I address the history of the Huli speaking communities who occupy a cluster of basins and valleys between 1200 m and 2300 m in the Southern Highlands Province of Papua New Guinea. Definitions of the Highlands region, which extends along the central spine of the island of New Guinea, vary considerably (Hays 1993). For my purposes, a general and necessarily flexible distinction can be drawn between the Highlands, lying above approximately 1200 m above sea level, the Highlands fringe, between about 500 m and 1200 m, and the Lowlands, below 500 m. The term “Central Highlands” serves to distinguish between the central cordillera (abbreviated here as the “Highlands”) and other areas above 1200 m, such as the Finisterre and Saruwaged mountains of the Huon Peninsula (Figure A1). [↩]
2. The term “Ipomoean” was coined by Watson (1965a) in reference to Ipomoea . the scientific term for sweet potato. Watson’s formal definition for the term is cited in Appendix Al. [↩]
3. But see Gorecki (1982) for his use of observations on wetland reclamation at Kuk in the 1970s to interpret the archaeology of drainage systems. [↩]