
by Dr. Chris Ballard
Perhaps the outstanding irony in the Ipomoean revolution debate, in light of the Canberra seminar’s privileging of archaeological evidence (Brookfield and White 1968, White with O’Connell1982:187), is that there is no direct archaeological evidence for sweet potato from any site in New Guinea. Charred tuber fragments from a house site at Kuk dated to Phase 6, identified during excavation by Golson’s local assistants as sweet potato, and previously published as the first direct evidence for sweet potato in New Guinea (Golson 1977a:627), have recently been identified by Jon Hather as yam (Golson pers.comm.). Neither is there any direct archaeological evidence for taro, the presumed staple tuber prior to the introduction of sweet potato. 1 All models for the introduction of sweet potato to the Highlands thus depend, to a greater or lesser degree, upon correspondence between different tuberous crops and specific field forms inferred from ethnographic analogues (Golson 1977a:627-628), evidence for increased rates of sedimentation interpreted as forest clearance for sweet potato gardens in catchment headwaters (Worsley and Oldfield 1988), increases in pollen from tree species associated ethnographically with fallow practices in gardens dominated by sweet potato (Haberle 1993), or evidence for major settlement expansion at higher altitudes than are considered possible with other, pre-Ipomoean staples (Bayliss-Smith 1985a).
Differences of opinion over the likely date for the introduction of sweet potato thus derive not from direct evidence but rather from assumptions about the scope for dramatic social change; anthropologists (yV atson, Feil and Modjeska amongst them) tend to assume that a late introduction is feasible, whilst archaeologists and palaeoecologists (Golson, Gorecki and Haberle to name a few) have expressed a preference for an earlier introduction. Arguments for the “orthodox” or “late” date assume that the introduction of sweet potato to New Guinea followed the date of its initial post-Columbian transfer from South America to Southeast Asia; either along the “Batata” line of diffusion to the Indonesian archipelago via Africa and Europe in the 15th and 16th centuries, or along the “Camote” line to the Philippines in the early 16th century (Yen 1974, 1982). A third line of diffusion from South America, the “Kumara” line, represents an independent and direct introduction of sweet potato to Eastern Polynesia from South America prior to European contact; given the break in the distribution of sweet potato as a staple in the intermediate regions of Western Polynesia and Island Melanesia, the Kumara line had not been widely regarded as a likely source of sweet potato for New Guinea. The lateness of the introduction of sweet potato to New Guinea, presumably at some point during the 16th or 17th centuries, was the spur behind Watson’s description of any changes in Highlands society that might be associated with sweet potato as “revolutionary”.
As early as 1974, however, Golson (in Powell et al. 1975) was declaring himself troubled by the lateness of the “orthodox” date. The initiation of Phase 5 at Kuk, dated to 370 ± 70 BP (K-2643), fell tantalizingly close to the presumed date of the introduction of sweet potato to the western Pacific. 2
Initially, Golson (1976d, but written in 1973; see Appendix A2) had resolved the problem this implied for the speed with which sweet potato must have been introduced and adopted in the wetland agricultural system at Kuk by dividing Phase 5 into two: an earlier Phase (now identified as Phase 5) for which intensive wetland cultivation of taro was proposed, followed by a Phase 6, after the fall of the distinctive Tibito volcanic ash at between 305-270 cal BP (Appendix A3). The apparent tightening of the drainage grid in Phase 6, Golson (1977a:628) suggested, represented a modification to suit the edaphic requirements of sweet potato. But Golson was already toying with a radical alternative: that there had been an “early” introduction of sweet potato to New Guinea by about 1200 BP. This, Golson argued, would account for the apparent revolution in land use at that time, evident in the abandonment of Phase 4 at Kuk swamp and in the relative decrease of forest pollen and increase in tree fallow species such as Casuarina, evident in pollen diagrams from the Wahgi valley (Powell1970).
Although Golson (1977b, 1982b, Golson and Gardner 1990) has subsequently withdrawn his claim for an early introduction of sweet potato, citing the lack of direct evidence, others have continued to express dissatisfaction with the conditions required for the orthodox date. Haberle (1993:311), citing claims by Worsley and Oldfield (1988) for an association between the dramatic increase in Casuarina pollen at 400 BP and the introduction of sweet potato, has argued that an earlier rise in the regional evidence for Casuarina pollen after 1200 BP could also be taken as proxy evidence for the presence of sweet potato during this period. Haberle (1994:34, Haberle, Hope and DeFretes 1990) also cites the negative evidence for a late introduction from the Baliem valley, where there is little evidence in a pollen record from the valley floor for dramatic change during the past 400 years, and takes this to indicate that the impact of sweet potato must have been registered at an earlier date in the valley. Gorecki (1986: 164) had previously aired an apparent compromise, in which three distinct waves of introduction of sweet potato to the Highlands are proposed: the first corresponding to the early date of 1200 BP, the second to the orthodox date of 300 BP and the third to direct European contact in the Highlands in the 1930s. As Hather (1992) observes, each of the major transformations in the pollen record and the archaeological record at Kuk can thus be referred to the effects of successively fleshier and more productive sweet potato cultivars.
New evidence from Polynesia has strengthened the possibility for a pre-Columbian introduction to New Guinea. Carbonized tubers dating to earlier than 520 ± 70 BP (628-517 cal BP) (Beta-32829) and positively identified as sweet potato have been recovered from Mangaia in Central Polynesia (Hather and Kirch 1991) 3 The claim has also been advanced that analysis of the terms for sweet potato throughout Papua New Guinea suggests the possibility of a Polynesian and thus potentially early introduction via southeast New Guinea (Scaglion and Soto 1991).
Direct archaeological evidence may eventually provide an answer to the question of the date for the introduction of sweet potato to New Guinea. In the interim, the grounds for asserting either an earlier or later introduction are themselves of some interest. My principal objection to the early model is that it simply enlarges the scope for an existing tendency to refer historical change to technological innovations or introductions. Similarly, debate within the framework of Gorecki’s compromise model over the extent of change at either 1200 BP or 300 BP introduces discussion of the relative qualities of new and old cultivars but no further advance on the issue of the social contexts for the deployment of successive cultivars. Another observation on the debate between early and late introductions is that revolutionary change appears to have re-entered archaeological explanation in Highlands New Guinea since the Canberra seminar where it was felt that ‘no major technological revolution … was required to permit the adoption of sweet potato’ (Brookfield and White 1968:50). Dramatic but unexplained changes (revolutions) in the archaeological record, as at 1200 BP, are now ascribed to the introduction of sweet potato. Through a still more tortuous logic, the absence of revolutionary change in a period is taken to signal the fact that sweet potato was not introduced during that period (as in the Baliem valley).
Revolutions, in fact, have long been in vogue as a form of explanation: ‘Susian’ (Morren 1977:313), ‘Canine’ (Kelly 1988:166) and ‘Colocasian’ (Bayliss-Smith and Golson 1992a) revolutions have all been proposed to account for historical changes in Highlands society. What each of these models apparently shares with the notion of an Ipomoean revolution is the assumption that social change revolves around the introduction or redeployment of crop staples or domesticated animals. While I certainly do not seek to deny the importance of these materials in the historical constitution of modem Highlands societies, the selection of tubers and animals to represent revolutions in the structure of society is an illuminating choice: an obvious one for archaeology, perhaps, but nevertheless an emphasis that continues to place social change in the background of archaeological explanation and enquiry. The scope for a more complex model of sweet potato introduction that foregrounds the decision-making of human agents in the processes of introduction, adoption and deployment is discussed amongst the conclusions of Chapter D2.
In light of the importance attached to the relationship between sweet potato and pigs, it should be mentioned that the antiquity of the latter in New Guinea is also subject to some debate. The claims for a solitary pig incisor at a date earlier than 9780 ± 350 (ANU-358) from the Kiowa site (S. Bulmer 1966) and a second pig incisor from the Yuku site dated to 10 350 BP (sic) (S. Bulmer 1982:188) need to be considered in light of the broader range of evidence for pig from other sites. 4 Nowhere in New Guinea have pig bones been retrieved in significant numbers until after about 4500 BP, though major increases in their number are documented at dates earlier than 800 BP (White 1984:4, White with O’Connell 1982:187-189, Table 6.5). Most recently, accelerator mass spectrometry dating of a selection of the small sample of “early” bone fragments from Highlands sites has further undermined claims for pig earlier than 4000 BP (D.Harris, cited Spriggs in press). Given the small overall numbers of pig bones recovered thus far from archaeological sites in New Guinea, an attempt to interpret fluctuating quantities of pig material, most of which derive from rock-shelter sites, in terms of changes in the structure of pig production practices could be hazardous. Despite frequent claims of archaeological evidence for an increase in pig numbers in the Highlands (J.Watson 1977:61; Feil1987:22), there are not, nor are there likely to be in the foreseeable future, sufficient pig remains from archaeological sites to determine much more than their gross presence or absence.
Appeals to the available archaeological evidence relevant to the Ipomoean revolution debate thus leave us with little to work on. Yet archaeological models for the impact of sweet potato have been developed, debated and, in some cases, discarded. Obviously, archaeologists have been able to incorporate a degree of informed speculation in their explanations for the impact of sweet potato, if only as a means of directing the limited resources for field research. The short history of archaeological research in the Papua New Guinea Highlands and the limited resources available for future work there do not allow the level of detailed reconstruction from archaeological evidence common, for example, in Europe. However, the social and technological continuities in the Highlands between the present and a recent but non-literate past accessible only through oral history and archaeology allow archaeologists working in the region to consider the relationship between archaeological problems and ethnographic contexts for human agency in great detail. There is scope here for a study that sketches the parameters of explanation for an archaeological problem such as the Ipomoean revolution, even where archaeological evidence is insufficiently available to support or contest the conclusions. Such a form of study requires that we reconsider conventional archaeological practice and methods.
Conventional practice in archaeologically new areas of the Papua New Guinea Highlands has commonly consisted of locating and excavating cave and rock-shelter sites and establishing chronological control over a region (White 1972, Mangi 1988c). Yet it must now be clear that the results of rock-shelter excavations in the Highlands tend to reflect primarily upon the use of rock-shelters and not upon the broader field of human activity, a conclusion strongly confirmed by Gorecki’s (1991) local ethno-archaeological study of rock-shelter uses. Chronologies of the presence or absence of specific items in technological inventories can be compiled (though the problems surrounding the antiquity of the pig in New Guinea should caution against optimism even in this regard), but our ability to reconstruct social processes from such evidence is very limited.
In his study of the material correlates of Big Man leaders, White (1984, 1985) expresses his doubts about the possibilities for success of a direct archaeological approach to questions of social hierarchy through, for example, the identification of larger houses or greater numbers of valuables. Instead, he proposes that if we wish to derive social history in the New Guinea highlands from archaeological sources we will have to eschew the simple equations between burials and ranking, settlement size and hierarchy, specialised organisation of production and social differentiation that are widespread in areas with more complex data sets. White (1984:9)
Though there must be some question about the analogical foundations for these equations even in areas so favoured, the implication for archaeology in the Highlands is that a different approach to ethnographic analogy as a means of archaeological access to processes of social change must be sought. The use of ethnographic analogies in Highlands archaeology has largely been restricted to the identification of direct technical analogues for wooden tools (Golson and Steensberg 1985), house, cooking and garden structures (Gorecki 1982) and stone tool manufacture and use (White 1967b, White and Modjeska 1978a, 1978b). These ethnoarchaeological studies have tended to focus on the immediate contexts for production and use of structures and techniques identified first in the archaeological record. Few studies have attempted to relate material processes to broader social features; the notable exceptions include Golson’s (1982b) modelling of the relationship between social change and swamp use at Kuk, and Burton’s (1984) study of axe production and the movement of brides in the Wahgi Valley.
In the previous chapter I proposed that an approach identified here as archaeological ethnography be employed to address issues of social change from an archaeological perspective. The form that an archaeological ethnography might take in a Highlands context can now briefly be sketched, with emphasis on issues of human agency, of historicity and of scale.
The importance of human decision-making has been demonstrated in the Highlands for some of the key elements of Watson’s postulated Ipomoean revolution. Studies by Robin Hide (1981) amongst Sinasina in Simbu Province and by David Boyd (1984, 1985) amongst Awa in the Eastern Highlands Province, have shown that the intensification of pig husbandry in these communities, whilst ecologically constrained in certain ways, has been the result of historical decisions made within particular groups, acting consciously within the specific circumstances of both their social and physical environments. Paul Wohlt (1978) has examined exchange strategies amongst high-altitude communities whose sweet potatoes are exposed to the hazard of frost, and Michael Bourke (1988) has argued for the significance of long-term planning in the production of sweet potato. John Burton’s (1984) study of axe quarrying in the Wahgi valley, in many respects a model for this thesis, has emphasised the importance of ‘human decision-making’ in the development of a limited number of quarrying centres such as the Tungei source within an essentially homogeneous geological landscape. By combining genealogical recall of the clan origins of Tungei brides with his observations on the role of quarried axes in brideprice transactions, Burton (1987) has been able to explore changes in the movement of brides within the Wahgi Valley from the tum of the century, illustrating the scope for archaeological ethnography in Highlands archaeology.
As Burton’s work suggests and as Ray Kelly (1988) has shown in his analysis of Etoro pig husbandry, all such decisions and strategies are made within the terms of culturally specific systems of value. If an understanding of intentionality is to be a significant component in explanations for social change in the past, the universes of cultural meaning within which those intentions are formed and expressed must also be considered. It cannot safely be assumed that the strategies deemed appropriate in dealing with pigs, sweet potato, and swamps, or indeed the very meanings of these objects, are similar across different societies. Given that archaeology generally addresses social change in the long term, the models of these worlds of meaning likely to be most appropriate will probably operate at a more inclusive scale than that with which most ethnographers are accustomed- something akin to Godelier’s “global social logic” or anthropology’s notion of a “worldview”.
One consequence of recognising the influence of different worlds of meaning is the scope this introduces for alternative notions of temporality and of the relationship perceived by agents between history and action in the present. An archaeological ethnography differs from conventional anthropological ethnography in the attention paid to the historical constitution of society and of individual subjectivity. The littleexplored possibility that indigenous conceptions of history, usually expressed in the Highlands in the form of oral history, might play an important role in the production of a genuinely historical ethnography is currently the subject of a study by Polly Wiessner and Akii Tumu in Enga Province. Their preliminary results suggest, for example, that the actual historical genesis of the Enga tee system of ceremonial exchange is accessible through oral history, and that such an approach has the potential to yield a far more complex understanding of Enga history and ethnography than descriptions of the tee written and conceived of in the “ethnographic present” (Wiessner and Tumu in prep.). The histories of specific groups offer both access to the complexity of the recent past, which is surely critical in the formation of modern societies, and an alternative perspective provided by a non-Western sense of history.
The final theme to be emphasised in the construction of an archaeological ethnography is the issue of scale. White’s solution to the problem of identifying changes in the structure of Wahgi valley society turns upon the scope present in the archaeological record for greater predictability at more inclusive levels of analysis… [B]y looking at a concatenation of variables from a number of sites we may be able to see the changes over time that mark the evolution of Wahgi systems as we know them. White (1985:59)
Regional trends, such as the postulated Ipomoean revolution, can only be identified as such through the compilation of sufficient evidence from a wide range of sites. The challenge this poses for an archaeological ethnography is the need to model the relationship between processes operating on local and regional scales – to show, for example, how reclamation of a swamp relates not just to the immediate circumstances of the local population and environment but also to broader patterns of regional change. The goals of an archaeological ethnography, then, are to explore the relationships that obtain between worlds of cultural meaning and their material expressions. The methods of archaeological ethnography include the employment of dual or multiple perspectives provided by different historicities, thus opening up for question the categories within which the past is conceived, and the extension of the temporal and spatial parameters of conventional ethnography in order to consider the relationships between phenomena observed on different scales.
(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. 15-23.)
- Phytolith and pollen evidence for a range of other domesticates, including Eumusa section banana (Wilson 1985), Amaranthus, Oenanthe iavanica, and Rorippa (Powell et al. 1975:47) has been identified at Kuk. Haberle (1993) has discussed the problems encountered in trying to locate or identify pollen from taro and sweet potato. [↩]
- Note, however, that calibration of the raw radiocarbon date of 370 ± 70 BP yields a result between 520-290 cal BP (100% Pr.), which might be taken either to alleviate or to heighten Golson’s anxiety (see Appendix C11). [↩]
- Two Ipomoea samples from Mangaia are bracketed by dates of 980 ± 70 BP (Beta-32826) and 490 ± 50 BP (Beta-32818), and 790 ± 80 BP (Beta 32828) and 520 ± 70 BP (Beta 32829) respectively; note that this does not, as Haberle (1993:311) reports it, amount to evidence for sweet potato at the 980 ± 70 BPdate. [↩]
- At Kiowa, the single incisor in Layer 12A is the only cranial material identified as pig beneath the three or four fragments in Layer 2 of the same site (S. Bulmer 1979: Table 1). At Yuku, a total of six or seven pig cranial fragments, representing between 0.2% and 0.9% of the faunal cranial assemblage for each layer, have been identified from three layers immediately beneath a layer of human burials; this contrasts suspiciously with the much higher relative density of pig cranial material in the two layers above the burial1ayer, where pig crania represent between 12.0 and 28.6% of the faunal cranial assemblage (S. Bulmer 1979: Table 2). Post-cranial material was recovered from both the Kiowa and Yuku sites, but details have not yet been reported. [↩]