Archaeology in Oceania, Vol. 51, Supplement 1 (2016): 61–72 DOI: 10.1002/arco.5104
The late prehistoric introduction of pottery to Rossel Island, Louisiade Archipelago, Papua New Guinea: Insights into local social organisation and regional exchange in the Massim BEN SHAW School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, Australia
ABSTRACT The introduction and exchange of pottery between Pacific Islands can provide insight into interaction and social organisation from both regional and local perspectives. In the Massim island region of far eastern Papua New Guinea, pottery is present in the archaeological record from 2800 to 2600 calBP. However, on Rossel Island, a relatively isolated landmass in the far east of the Louisiade Archipelago, archaeological excavation and AMS dating of several sites has determined that pottery on this island was a late prehistoric introduction, from 550–500 calBP. The introduction of pottery coincided with the establishment of increasingly complex exchange networks in the Massim, namely the Kula. It is argued in this paper that the desire for Kula participants to obtain high-quality shell necklaces (bagi), which are prominently manufactured on Rossel, led to the island becoming more actively involved in down-the-line regional exchange. Pottery is largely found on the western end of Rossel, where most bagi are manufactured. The uneven distribution of pottery across the island is further argued to indicate a socio-economic/political divide between the populations living on the western and eastern ends, which is supported by linguistic and anthropological evidence.
Keywords: exchange networks, Louisiade Archipelago, Massim, Papua New Guinea, pottery, Rossel Island
´ ´ RESUM E L’introduction et l’´echange de poterie entre les ˆıles du Pacifique peuvent renseigner sur les interactions et l’organisation sociale d’un point de vue a` la fois r´egional et local. Dans la r´egion insulaire de Massim situ´ee a` l’extrˆemit´e orientale de la Papouasie-Nouvelle-Guin´ee, la poterie est pr´esente dans le registre arch´eologique a` partir de 2800 a` 2600 calBP. Cependant, sur l’ˆıle de Rossel, relativement isol´ee a` l’extrˆemit´e orientale de l’archipel de la Louisiade, des fouilles arch´eologiques et des datations AMS de plusieurs sites ont d´etermin´e que la poterie avait e´ t´e introduite tardivement au cours de la pr´ehistoire de l’ˆıle, soit a` partir de 550–500 calBP. L’introduction de la poterie a co¨ıncid´e avec la mise en place de r´eseaux d’´echange de plus en plus complexes dans la r´egion de Massim, a` savoir la Kula. Il est soutenu dans cet article que l’int´erˆet pour les participants a` la Kula d’obtenir des colliers de coquillages de haute qualit´e (bagi), fabriqu´es principalement sur Rossel, a men´e a` l’implication grandissante des habitans de l’ˆıle dans des e´ changes r´egionaux de proche en proche. La poterie est surtout trouv´ee dans la moiti´e occidentale de Rossel, l`a o`u la plupart des bagi sont fabriqu´es. Il est sugg´er´e que cette r´epartition in´egale de la poterie a` travers l’ˆıle indique une fracture socio-´economique / politique entre les populations vivant des cˆot´es ouest et est de l’ˆıle, ce qui est soutenu par des preuves linguistiques et anthropologiques.
Mots Cles: r´eseaux d’´echange, archipel de la Louisiade, Massim, Papouasie-Nouvelle-Guin´ee, poterie, ˆIle Rossel Correspondence: Ben Shaw, School of Biological, Earth and Environmental Sciences, The University of New South Wales, Biological Sciences Building (D26), Randwick, NSW 2052, Australia. Email:
[email protected]
INTRODUCTION Missionaries, explorers and government officials in Papua New Guinea throughout the middle to late nineteenth century were intrigued by the finely made shell, stone, wood and pottery wares produced by cultural groups along the coast and on many of the islands (Lindt 1887; MacGillivray 1852; MacGregor 1892; Moresby 1876; C
2016 Oceania Publications
Thomson 1889). It was apparent to these early European observers that the objects were traded through what appeared to be well-established interaction networks, and in many instances were transported over several hundreds of kilometres by sailing canoe between groups who did not necessarily speak the same language. Continued interest in the organisation and operation of these exchange systems throughout the twentieth century led to the publication of
62 several seminal anthropological studies on New Guinea1 culture (Harding 1967; Malinowski 1915,1922; Seligmann 1910; Tueting 1935). These studies have since formed the foundation for archaeological investigations seeking to identify the origins of these maritime exchange networks. One region that has since become synonymous with complex exchange systems is the Massim (Young 1983). The Massim region encompasses the eastern tip of the New Guinea mainland and more than 600 islands, spread over an area in excess of 180000 km2 (Figure 1). At least since initial European contact, systems of specialised interdependence through exchange have been observed between many of the smaller islands as an adaptive cultural response to overcome constraints from impoverished ecological environments (Belshaw 1955; Macintyre & Young 1982). A large number of exchange routes had been established between the islands, with notable examples including the Kula network of the Trobriand Islands (Malinowski 1922) and the Kune network of Tubetube Island (Macintyre 1983b). The recent production and trade of pottery in the Massim has received a great deal of attention as a means of investigating the social intricacies of these inter-island exchange networks in a historical context (Lauer 1970; Leach & Leach 1983; Macintyre 1982; May & Tuckson 2000). Much less is known of the antecedent pottery industries in the region, from which the prehistoric configurations of these networks can be traced archaeologically. This paper therefore contributes to the understanding of prehistoric pottery trade in the southern Massim by presenting evidence for the relatively late introduction of pottery to Rossel Island, the easternmost island in the region, from 550–500 calBP. The implications for a late pottery introduction are discussed in terms of the incentives for involvement in exchange, the changing positon of Rossel in regional exchange and the impact on local socio-economic systems.
THE ARCHAEOLOGY OF COASTAL AND ISLAND EXCHANGE SYSTEMS IN PAPUA NEW GUINEA In the Port Moresby area on the south coast of Papua New Guinea, survey and excavation of several sites from the late 1960s onwards by Susan Bulmer (1969) and Jim Allen (1972, 1977) allowed, for the first time, the prehistoric configurations of ethnographically recorded coastal trading systems to be investigated. Bulmer (1971) presented the first archaeological synthesis of coastal exchange by combining excavated material culture evidence from a handful of known sites with data from historical accounts, climatic records, oral histories and linguistics. In this regard, Bulmer is a pioneer of archaeology on coastal exchange systems in Papua New Guinea. While she has never worked in the Massim, her work on the south coast has been influential in formulating hypotheses about long-distance connections between the two regions. Indeed, Bulmer (1975, 1999) argued for a connection between red slipped pottery and the
Late Prehistoric introduction of pottery to Rossel Is earlier spread of the Lapita Cultural Complex, and for the later appearance of “Massim” styles in south coast sites, which is now associated with the “ceramic hiccup” (Sutton et al. 2015; Vilgalys & Summerhayes in press). Archaeological research has since been undertaken along the length of the south coast, on parts of the north coast and on some of the Massim islands to investigate the immediate origins of the Hiri (Bulmer 1982; Frankel & Rhoads 1994), Mailu (Irwin 1978b), Siassi (Lilley 1988), Kula (Egloff 1978), Sepik/Koil Island (Terrell & Schechter 2011; G. Summerhayes, pers. comm. 2016) and Madang (D. Gaffney, pers. comm. 2016) exchange systems. These networks individually encompass relatively small geographical areas, but collectively they span much of the New Guinea coastline and many of the offshore islands. Archaeological evidence has indicated that many of these exchange networks have an antiquity spanning at least 2000 years, but with the mass production of pottery tradewares and the intensification of specialised “middlemen” traders having only occurred within the past 600–500 years (Irwin 1978a; Lilley 2007; Rhoads 1982). The processes that drove such prominent shifts in the socio-economic configurations of these communities differed locally and were often multi-faceted (e.g. increased population density, environmental decline, access to resources). Nonetheless, such a shift does appear to have been a widespread phenomenon relating to the formation of monopolies on coastal trade in late prehistory (Allen 1984; Bulmer 1982; Irwin 1978b).
THE PREHISTORY OF THE MASSIM REGION Colonisation and early settlement It is probable that the Massim region was colonised in the Late Pleistocene, when sea levels were considerably lower and many of the islands were joined to form a continuation of the New Guinea mainland, or larger island landmasses (Shaw 2016). Swadling (2016) has further suggested that regional exchange networks were established in the Massim during the mid-Holocene. Evidence for settlement of this antiquity comes from surface finds of distinctive robust-waisted stone blades, as well as mortars and pestles, on several islands, that have otherwise been found in Late Pleistocene – mid-Holocene aged sites in New Guinea and the Bismarck Archipelago. Settlement of this antiquity in the Massim has yet to be defined within a radiocarbon chronology. However, it does open up the possibility of a much greater time depth for interaction spheres in the region, which undoubtedly influenced the development of later exchange systems (see, e.g., Torrence & Swadling 2008). Late Holocene settlement and regional interaction The earliest excavated cultural deposits in the Massim are from the Kasasinabwana site on Wari Island (Figure 1), which yielded calcareous/hybrid tempered plainware pottery dating from 2800–2600 calBP probably to be associated with the Lapita cultural complex (Chynoweth C
2016 Oceania Publications
Archaeology in Oceania
63
Figure 1. A map of the Massim region, showing the location of Rossel Island and the extent of the major ethnographically known exchange routes. Inset: the location of the Massim in the Western Pacific.
2015; Negishi & Ono 2009). Red slipped pottery was also recovered from Wari, dating to 2300–2000/1600 calBP, which is consistent with the Early Papuan Pottery (EPP; 2000–1600 calBP) found along the south coast of New Guinea, (Allen 2010; Bulmer 1978). Depending on how widespread EPP pottery is in the Massim, it may signify cultural continuity with mainland New Guinea populations during the first 1000 years of pottery exchange (Lapita–EPP). Elsewhere in the Massim islands, excavated cultural deposits associated with pottery date to within the past 1500–1000 cal yr when increasingly localised interaction spheres were developing (Ambrose et al. 2012; Bickler 1998; Burenhult 2002; Egloff 1979). On current evidence, at least by 1000 cal yr ago the northern Massim islands were connected with mainland New Guinea in a network involving the distribution of pottery, Fergusson obsidian and shell valuables (Ambrose et al. 2012; Egloff 1979; C
2016 Oceania Publications
White et al. 2006). Within the past 500 cal yr, northern Massim exchange networks increasingly centred on the islands, to the greater exclusion of links with the mainland. The later focus on inter-island exchange has been argued to mark the formative development of the Kula exchange network as it is known ethnographically (Egloff 1978; Lauer 1971) (Figure 1). As such, it is within the context of intensified inter-island exchange elsewhere in the Massim region that the findings from Rossel Island are modelled.
ROSSEL ISLAND AND ITS PEOPLE Rossel is a moderate-sized volcanic island with a total land area of 290 km2 , and is situated 360 km off the eastern tip of the New Guinea mainland. The interior of the island is mountainous, with the highest peak elevated 838 m above sea level (Figure 2). Rossel is the easternmost island in the
64
Late Prehistoric introduction of pottery to Rossel Is
Figure 2. A topographic map of Rossel Island (80 m contours) with the surrounding reef marked (grey shading), showing the location of recorded and excavated archaeological sites, coded by the presence or absence of surface/subsurface pottery. Other places mentioned in the text are also shown.
Massim, and is somewhat isolated from the other islands, as it is contained within its own expansive reef system (1026 km2 ). It is separated from the Calvados/Sudest lagoonal system by a deep and swift sea channel 14 km in width, which can be treacherous to cross. The inhabitants on Rossel have been well documented as being culturally (Liep 2009), linguistically (Dunn et al. 2005; Levinson 2006a) and genetically (van Oven et al. 2014) unique from other island populations in the region. Taro, yam and sago are the staple crops on Rossel, with a range of edible plants available that are scarce or otherwise lacking on many of the smaller Massim islands. While steel pots are now predominantly used for cooking, hot stones (mumu in Tok pisin) rather than clay pots, which are widely used elsewhere in the region, have remained the traditional means of preparing food. Although the population of Rossel now exceeds 5000 (as of the 2011 census), the population from the late nineteenth century up until the 1950s, when modern healthcare facilities were established on the island, was around 1500 (Liep 2009; MacGregor 1889). The population in late prehistory probably did not exceed this, and indeed population density was probably never high on the island (Shaw 2015).
INCENTIVES FOR EXCHANGE IN THE MASSIM For those involved in exchange, the incentives were not only economic (subsistence based) in nature. Interaction often also involved religious, political and social motivations that were of equal or greater importance. The need to acquire economic resources not locally produced or
readily available year round has been widely argued as the underlying driver for many regional exchange systems in prehistory (Allen 1977; Ambrose 1978; Kirch 1991). In the Hiri of the central south Papuan coast, for example, pottery was exchanged with communities in the Gulf to the west in return for sago and canoe hulls (Oram 1982). Socially valuable items such as necklaces were also exchanged as part of the Hiri, but they were of secondary importance to an economic return. This is evident as trade with the Gulf did not occur when yields of locally produced food were large enough to feed the population. In contrast to the Hiri, long-distance exchange undertaken by the Mailu Islanders 270 km to the east was driven primarily by the desire to obtain shell armbands, pigs and other socially valuable objects. These items were either articulated within their own ceremonial traditions (e.g. bride price), or the Mailu acted as middleman traders and passed the valuables on to neighbouring cultural groups for a profit2 (Irwin 1978b). The Mailu were able to satisfy their economic needs through local trade with communities who gardened the fertile land on the adjacent mainland, with no need for long-distance subsistence-based exchange. Turning back to the Massim, the islands can be broadly divided into those that are large and resource rich, and those that are small, resource impoverished and drought prone. The patterns of exchange within and between the smaller islands, in part, reflect a need to sustain a stable basis for subsistence (Macintyre & Allen 1990). Prominent pottery tradeware industries were or still are mostly centred on small resource impoverished islands, namely Tubetube, Wari, Amphetts, Brooker and Panaeati, with the islands also strategically located allowing these communities to C
2016 Oceania Publications
Archaeology in Oceania manipulate trade routes. In this context, pottery was moved with other valued items in exchange for staple foods from the larger islands, where an agricultural surplus could be more readily maintained. Aside from subsistence exchange, there were complex and variable ceremonial and/or sociopolitical incentives that influenced individual involvement in exchange, as is well illustrated in anthropological accounts of Massim exchange (Damon 1983; Macintyre 1983a). It is therefore evident that while the primary incentives for exchange can vary between communities, they need not be mutually exclusive, as incentives can change over time in response to environmental and social influences.
THE POSITION OF ROSSEL IN ETHNOGRAPHIC EXCHANGE
65 Pottery Government reports of the late nineteenth and early twentieth centuries reported very few pottery vessels on Rossel Island compared to much denser distributions found elsewhere in the Massim (MacGregor 1889; Murray 1908). Elders on Rossel today attribute the general lack of pottery to the fact that it was never manufactured locally or widely used for cooking, and was only imported in small quantities within the past few centuries. Prior to European contact, pottery on Rossel was used as a prestige item during ceremonial events to prepare high-status sago pudding and to prominently display food offerings (Shaw 2015). In more recent times, pottery has become an increasingly utilitarian item as post-colonial influences have opened up trade routes following the abolishment of head hunting and warfare (Liep 1983b).
Within the past one and a half centuries since the first accounts of Rossel culture were published (MacGillivray 1852; MacGregor 1889; Thomson 1889), the island has been documented as being on the periphery of regional exchange networks. Certainly, the location of Rossel meant that inter-island trade was naturally limited to the islands to the west, with cultural contact not extending much beyond the neighbouring island of Sudest (Armstrong 1928; Murray 1912). (Figure 1). The communities on the western end of Rossel were especially advantaged by their closer proximity to Sudest, with those on the eastern end and in the interior more reliant on the redistribution of imported goods (Liep 2009: 82–3). Both Rossel and Sudest are large in size with rich ecologies and fertile soil. Therefore, there was no regular subsistence incentives for trade, with documented exchanges instead having involved functional and ceremonial items (Liep 1983b).
In 2011–2012, 79 sites were recorded on Rossel Island. Of these, ten were excavated, with four yielding pottery sherds in varying quantities (Figure 3 and Table 1). Three of the pottery-bearing sites (Ghakpo, Pambwa and Morpa) are located on the western end of Rossel, with the fourth (Wule Gha) situated on the small island of Wule, 2.6 km off the north-west coast of Rossel, within the confines of the main lagoon (Figure 2). Pambwa and Morpa have been mentioned in anthropological accounts concerning pottery trade links with Sudest, and were targeted in part for this reason (Lepowsky 1981; Liep 1983b, 2009). From a geomorphological perspective, all four sites were preferentially targeted as they were situated on large beachflats conducive to the preservation of prehistoric settlement.
Bagi shell necklaces The distinctive red shell necklaces (bagi; made from the Chama imbricata oyster) were the primary exported trade item produced on Rossel Island, along with woven baskets, sandalwood and some shell valuables (Liep 1983a). The Chama shell beds are particularly dense on the western end of Rossel in and around Yongga Bay (Figure 2), with divers having to reach depths in excess of 5–6 m to obtain the vibrantly coloured shells of higher value (Liep 1981). The manufacture of bagi necklaces was also most prominent in the Yongga Bay area, which had further influenced the dominant position of the western Rossel communities in external trade relationships. While bagi (known as soulava in the northern Massim) are still made on several other islands in the Louisiade Archipelago, the deeper colour of the shells and finer craftsmanship makes the Rossel necklaces of a recognisably higher quality. Furthermore, although Rossel bagi were incorporated into Kula exchange as a highly valued item, no direct trade was made with islands participating in the Kula. Instead, bagi necklaces were redistributed along with other Rossel exports out of Sudest and through the Calvados islands (Lepowsky 1983).
Pottery sites Ghakpo is a small sheltered beach on the south-west coast, adjacent to a large but slow-flowing river. Systematic excavation of four 1 m2 units spaced across 40 m of the beachfront yielded a low quantity of pottery only in the upper cultural layers (Table 1). Pambwa is an expansive beachflat 3 km west of Ghakpo. Two 1 m2 units were excavated, 125 m apart, with the first unit having pottery throughout, whereas the other had pottery only in the upper cultural layers. Morpa is an embayment on the north-west coast, and is flanked on three sides by steep-sided hills. Three closely spaced 1 m2 units were dug, which yielded pottery deposits throughout all cultural layers. Lastly, the Wule Gha site is a substantial, dense midden deposit along the base of a hill at the back of the only habitable beach on the islet, at the eastern end. Four closely spaced 1 m2 units were excavated, which yielded relatively dense concentrations of pottery throughout the deposit (Table 1). Pre-pottery occupation at Ghakpo in the lowest cultural layer, associated with shell tools/ornaments and shell refuse, dated from 900–570 calBP (1σ ranges), and at Pambwa from 650–550 calBP (Table 2). Overlying cultural
C
2016 Oceania Publications
DEFINING THE PRE-POTTERY/POTTERY TRANSITION ON ROSSEL ISLAND
66
Late Prehistoric introduction of pottery to Rossel Is
Figure 3. Excavated pottery from Rossel and Nimowa Islands: (a) Ghakpo, 420–150 calBP; (b) Wule Gha, 520–330 calBP; (c) Pambwa, 510–330 calBP; (d) Morpa,
