|
Article Excerpt
Abstract
Excavations at Mask Cave on the sacred islet of Pulu off Mabuyag in the central west of Zenadh Kes (Tortes Strait) reveal four occupational phases: Phase 1 (2900-3800 years ago), Phase 2 (2100-2600 years ago), Phase 3 (1500-1700 years ago) and Phase 4 (last 1500 years). Faunal remains indicate marine specialization (turtle and fish) during all phases. Petrographic analysis of sherds of finely made red-slipped pottery dating back 2400-2600 years reveals a unique fabric in terms of current understandings of Oceanic ceramic technologies. Mineral inclusions are consistent with local geology suggesting local manufacture and the existence of Indigenous Australia's first pottery tradition. Pre-ceramic Phase 1 is associated with demographic expansions across the western islands of Zenadh Kes by local populations of marine-based hunter-gatherers who were primarily Aboriginal language speakers. Phase 2 is associated with the immigration of Papuan maritime, horticultural and pottery-making peoples to the eastern and western islands of Zenadh Kes commencing 2600 years ago. Australian then Papuan settlement expansions across the western islands of Zenadh Kes explain why the local Western-Central Language has an Aboriginal base with a Papuan overlay. First colonization of the eastern islands by Papuans explains why the local Meriam Mir language is Papuan. Early red-slipped pottery in Zenadh Kes is linked to southern coastal Papuan pottery traditions that are reassessed to have a comparable 2600 year antiquity. Papuan settlement of the southern Papuan coast and Zenadh Kes was an extension of the post-Lapita settlement of the Pacific, an event memorialized in part by Torres Strait Islander oral tradition.
Keywords: Torres Strait, island colonization, maritime specialization, pottery, Papua New Guinea
**********
In the central western part of Zenadh Kes (1) is a small granite boulder-strewn islet called Pulu (Figure 1). The islet is located 300m off the west coast of the residential island of Mabuyag (2), home to the Goemulgal (people of Mabuyag). Pulu is a sacred place, revered by the Goemulgal and their neighours today as much as it was when Alfred Haddon and other members of the famous 'Cambridge Anthropological Expedition to Torres Straits' were escorted to the great kod (ceremonial complex) at the northern end of the islet by Mabuyag Elders in October 1898. Haddon was also taken up the ridge behind the kod to a rockshelter known as Awgadhalkula ('Augadalkula'). For as long as Goemulgaw oral history recalls, Awgadhalkula has been their most sacred site and the place where the sacred emblems of the culture hero Kuiam resided, keeping watch over the more than 80 skulls and jaws of decapitated warriors from enemy neighbours. Today many of these skulls rest patiently in boxes in the British Museum another world away.
[FIGURE 1 OMITTED]
Haddon had an intense interest in the history and migrations of the world's 'races' and this interest extended to Torres Strait Islanders, a people that for him were positioned curiously along the frontier of the Papuan 'race' to the north (New Guinea) and the Aboriginal 'race' to the south (mainland Australia) (Haddon 1909). But despite months of fieldwork in Zenadh Kes and visiting numerous islands and countless cultural sites, Haddon believed the archaeological potential of Zenadh Kes was minimal: 'Unfortunately, I was unable to discover anything concerning the archaeology of the Torres Straits Islands.... I consider it improbable that much will ever be found to illustrate the former condition of the people.... There never has been any pottery' (Haddon 1890: 303). Alternatively, the long-term cultural history of Torres Strait Islanders was to be accessed largely through reading beyond the metaphorical dimensions of their oral traditions and origins stories. But had Haddon continued walking along the ridge for 100m past Awgadhalkula on Pulu, he would have come across a cave that may have changed his view of the archaeological potential of Zenadh Kes.
On the 17 August 1998, 100 years after Haddon's team visited Pulu, one of us (IM) was accompanied to Pulu by Mabuyag Elders. Unlike Haddon's visit a century earlier, the cave past Awgadhalkula was visited (Figure 2). Exposed across the cave surface were found flaked stone artefacts and scattered fragments of marine shell and marine turtle bone. The most recent evidence of use was a broken bottle, a rusty piece of metal (possibly a large knife blade), and a matchbox tin, all probably dating to the late 19th/early 20th century. The most striking feature of the site was a large face or 'mask' painted with red ochre located on the western roof of the cave (Figure 3). The face has stylistic similarities to the low-relief carved faces on Gope boards of the Papuan Gulf (see Lewis 1973; Newton 1961).
[FIGURES 2-3 OMITTED]
On a subsequent visit to the site on 1 May 2000 by senior members of the Mabuyag community (Colin Bani, John Bani, John Mooka, Cygnet Repu, Terrence Whap and John Whop), IM (then University of Melbourne), BD (Monash University), Garrick Hitchcock (then Native Title Office, Thursday Island), and John Brayer (University of New Mexico), two small, thin, weathered pottery sherds (both with remnant red slip) were found amongst a lag deposit of stone artefacts associated with a c.20cm-deep erosion channel running through the middle of the site. The significance of the finds was recognized immediately. The sherds were the first examples of Indigenous pottery recorded for western Zenadh Kes--pottery being unknown ethnographically for Zenadh Kes (Gill 1874: 218; Haddon 1912: 122). These sherds appeared similar to pottery excavated two years earlier by Melissa Carter from the Murray Islands of eastern Zenadh Kes and pointed to the existence of an ancient and forgotten trans-Zenadh Kes pottery tradition. It was assumed that the pottery entered Zenadh Kes from ethnographically and archaeologically-known pottery producers of southeastern Papua New Guinea through a previously unknown western extension of the Papuan Gulf pottery trading system (e.g. Bickler 1997). Thus, both rock-art and pottery at Mask Cave pointed to possible ancient cultural connections between the central western islands of Zenadh Kes and Papua New Guinea. If Haddon and his Mabuyag escorts had been present with us that day they would surely have been impressed with what we found at Mask Cave, because they knew Torres Strait Islander oral histories point to New Guinea peoples to the east as the first migrants to many of the islands of Zenadh Kes. And after 100 years of archaeological scholarship across northern Australia, New Guinea and the Pacific, we also knew Mask Cave had the potential to yield fundamentally new insights into Goemulgaw history and perhaps into Torres Strait Islander origins. In this paper, we present results of our excavations at Mask Cave in October 2002 and October/November 2003. To set the scene for appreciating the significance of the site, we begin by outlining previous understandings of Torres Strait Islander origins, first from Torres Strait Islanders themselves (oral traditions), followed by the work of outside researchers (anthropologists and archaeologists).
Torres Strait Islander origins
Oral traditions
According to some Murray Islanders in the Eastern Group, the 'original settlers of Mer' (Murray Island) were Pop and Kod from the 'Fly River district of New Guinea' to the northeast of Zenadh Kes (Haddon 1908: 19; see also Lawrie 1970: 303). An alternative version of the story tells how after arriving on Met, Pop (a man) fashioned a 'mate' from white clay in the form of a female dugong with two legs (Haddon 1935: 103-4). Another tradition tells of the first settlers of the eastern islands of Mer and Erub as three women castaways from 'New Guinea' and men from New Guinea who had gone looking for them. On Mer, '[o]ther people [subsequently] came to the island', apparently also from New Guinea (Haddon 1935: 102; Laade 1968: 142).
Origin narratives have been published for the three islands of Dauan, Saibai and Boigu (Top Western Group). The first man of Dauan was Kogea who lived in a 'cave' where he continues to reside in the form of a large snake (see Laade 1971: 54-9; Lawrie 1970:139-42 and Wirz 1932: 292-93 for details). Kogea settled on Dauan after traveling across the sea from the east, possibly from Kiwai Island. The first man of Saibai was Melawal (see Haddon 1904: 27; Laade 1968: 142; 1971" 12-4; Lawrie 1970: 153-60). He lived underground in a baler shell which suggests that 'he came from the sea' according to one of Laade's informants (Laade 1971: 12). The second man, Budia, 'came from the west, nobody knows from where' (Laade 1971: 12). Another of Laade's informants suggested Budia came from the 'other side [of] New Guinea, might be Malayo country' (1971: 13). 'These two men have quite a mythical character' and lived on the western end of the island (Laade 1968: 142). The next immigrants, settling at Air village on the eastern end of Saibai, came from the Pahoturi River district of the adjacent Papuan coast (Laade 1968: 144). On Boigu, Laade (1968: 145-6; 1971: 84-6) was informed that while a man named Boigu was the first inhabitant of the island, the island was visited subsequently by a canoe crewed by 'Maori' (a likely generic term for 'Polynesians' according to Laade). Three of the crew stayed behind (a woman and two men) while the 'Maori chief went back again with the rest of the crew)'. The first children of Boigu were fathered by the two Maori men (cf. Ingui and Toby 1991: 2; Lawrie 1970: 207-8).
At a more general level, Laade (1968: 148) distilled the following oral tradition on Torres Strait Islander origins related to him by the Rev. Seriba Sagigi of Badu island in the Western Group:
"Pacific men come to Torres Strait Island and married New Guinea women. That's how these islands got population. The Pacific men traveled first up to New Guinea coast, Maori, (men from) Tonga, Fiji." Specifying the origin of the migrants does not mean much; Seriba simply gives some names familiar to him. Then he continued, (When some of the newcomers had obtained women they left New Guinea). "First people they dropped at Parema. Leave one man on Saibai (allusion to Melawal), two at Gebar, three at Mabuiag."
Rev. Sagigi's narrative continues, documenting subsequent settlement of various islands of the Central Group (Laade 1968: 149; see also Lawrence 1989: 103-5).
Landtman (1917: 366) recorded a story from coastal Papuans opposite Zenadh Kes relating how Mua and Murulag islands of the Western Group were settled by people from Daru island (located off the Papuan coast) who headed south fleeing Kiwai Papuan raiders. In a related sense, Lawrence (1989: 103) recorded a story from 'the olden days' on Daru of how Zenadh Kes was settled by people who 'ran away' from the adjacent southern Papuan coast.
Early anthropological and archaeological speculations
Prior to major archaeological excavations in Zenadh Kes during the 1980s, modeling of local culture history was based upon local oral traditions, racial categorizations, and broader linguistic and material culture affiliations. The first hypothetical, long-term culture history model for Zenadh Kes was presented by Alfred Haddon (1935: 275-78, 410-14; cf. Rivers 1926: 161). Using a racialised migrationist theoretical framework employed typically by late 19th century scholars to understand Aboriginal origins (see McNiven and Russell 2005: Chap 4), Haddon suggested the Western Group was populated by an 'ancient stock' that may have come across from northwestern Australia during a period of lower sea level (1935: 276, 410). After sea level rise and island formation, peoples from the Papuan mainland southwest of the Fly River mouth moved into the islands of western and eastern Zenadh Kes. Finally, peoples from the Fly River 'estuary' mixed with Islanders from western Zenadh Kes. Significantly, Haddon (1935: 278) speculated on Austronesian influences in the Strait:
The cultural traits that Austronesian-speaking peoples brought from Indonesia to New Guinea on their way to Melanesia have spread over a considerable part of that island. By secondary diffusions, whatever routes they took or whatever modifications they underwent, some of these traits reached the southern coasts of New Guinea and influenced the Torres Strait islanders.
The strong New Guinea elements in Haddon's migration model was clearly based on his (generally valid) observation that 'physically the islanders are Papuans and can be easily distinguished from Australians' (Haddon 1935:410). While Haddon saw his New Guinea origins model as consistent with the 'Papuan' language of the Eastern Group and the oral tradition of Pop and Kod (see above), the 'Australian' language of the Western Group indicated 'an ancient ethnic movement' that required 'further investigation' (Haddon 1935: 410).
The next hypothetical culture history model, advanced by Jack Golson (1972: 384-89), posited that after formation of Zenadh Kes some 6500 years ago, the western Strait became a 'marginal area of Aboriginal occupation' with 'residential and seasonally visited islands', thus 'ensuring the survival of the traditional populations of the area and the persistence of their hunter-gatherer way of life'. The descendants of the subsequent influx of Melanesian seafaring peoples with a 'horticultural economy' is found with Papuan language speakers of the Eastern Group, and the mixed Aboriginal-Papuan language of the Western Group (who subsequently expanded northwards and eastwards to settle the Top Western and Central Groups respectively). This Melanesian maritime settlement of Zenadh Kes was linked with the broader colonization of Melanesian islands by peoples using double-outrigger canoes. As these early maritime peoples were thought to use pigs, a date for this maritime expansion (including into Zenadh Kes) was put at 5000-6000 years ago based on the then apparent age of pig bones in the Papua New Guinea Highlands. A second Melanesian maritime migration event dated 2000-3000 years ago involving Austronesian speakers using single-outrigger canoes was thought to reach Zenadh Kes around 2000 years ago based on the earliest pottery dates for the adjacent southern Papuan coast (cf. Alien 1972; Vanderwal 1973). Austronesian influence in Zenadh Kes was demonstrated in part by Austronesian words for canoe outrigger floats amongst the non-Austronesian languages of Zenadh Kes, Cape York, and Fly River delta (Golson 1972: 392-94). David Moore (1979: 308-13) followed Golson, adding that in more recent times further Papuan migrants moved south into the Central Group (to speak a dialect of the Western-Central language-a.k.a. Kala Lagaw Ya) and Eastern Group.
Haddon, Golson and Moore relied heavily on the relative degrees of Australian and Papuan linguistic elements within the two major languages of Zenadh Kes. Detailed analysis of Zenadh Kes languages by Wurm (1972), building on the foundational research of Ray (1907), confirmed that Islanders of the Eastern Group spoke a 'typical Papuan language' (referred to as Meriam or Meriam Mir) that belongs to the Eastern Trans-Fly language family of the adjacent Papuan coast. In contrast, Islanders from the Western, Top Western and Central Groups spoke an 'Australian language apparently adopted by speakers of an originally Papuan language'. The 'Australian influence on Miriam [sic] is negligible', while the Papuan influence on Cape York Aboriginal languages is 'relatively superficial' (Wurm 1972: 349, 356). To account for these linguistic arrangements, Wurm (1972: 364) posited that when Papuan peoples moved southwards into Zenadh Kes within the last few thousand years, they found the eastern islands uninhabited, while movement into the western islands was more complex as 'on at least some of the[se] islands' were peoples who spoke an Australian language.
Recent archaeological understandings and limitations
David et al. (2004) provide the first long-term chronological framework for the human history of Zenadh Kes based on local excavation results. Badu 15 rockshelter on the island of Badu in the Western Group reveals a stone artefact sequence representing 'three phases of island use and occupation'. A lower phase dated 6000->8000 BP indicates 'the presence of resident populations on Badu at a time when Badu was separating from the mainland as postglacial sea levels rose' (David et al. 2004: 72). A middle phase with 'sporadic' artefacts dated between 3500 and 6000 BP is associated with 'low level use (likely fleeting visitation)' by people venturing 'from the Australian mainland' (David et al. 2004: 72, 75). An upper phase dated 300013500 BP with higher artefact concentrations and 'sustained cultural deposits and heightened sedimentation rates' signals a return of 'resident populations' linked to an 'influx of people from the north or northeast' of 'Papuo-Austronesian' origin (David et al. 2004: 74, 76). In broad terms, this three phase occupation model is consistent with occupational phasing hypothesised by Golson (1972) and Moore (1979).
The case for substantial Austronesian influences followed recent linguistic work revealing considerably more Austronesian influences in local languages than previously thought (Mitchell 1995). Such findings support previous hypotheses for Austronesian influences in Zenadh Kes advanced by Haddon (1935), Golson (1972) and Moore (1979). David et al. (2004: 76) note that 'colonisation of Torres Strait is unlikely to have taken place long after initial Austronesian seafaring began in southeastern Papua.' They suggest further that these early Austronesian links are revealed by the 'presence of red-slip ceramics in low numbers at various sites across Torres Strait' (2004: 76). However, Carter (2004: 340) questions this link, pointing out that earliest available dates for pottery in Zenadh Kes at 2000 BP (Eastern Group) matches the earliest dated ceramic assemblages from the Papuan Gulf which are linked to immigration of Austronesian speakers from island Melanesia (Allen 1972: 121; Vanderwal 1973: 233). Carter (2004: 340-41) concludes that it is more likely that 'the occupation of Torres Strait not only pre-dated the Austronesian expansion into southern Papua, but also occurred entirely independent of this event'.
Badu 15 revealed no faunal remains, thus precluding insights into subsistence activities. As such, it shed little light on the apparent widespread development of midden deposits (and associated maritime specialization) across Zenadh Kes that had independently been argued to start 2500 BP (Barham 2000). In addition, no radiocarbon dates are available at Badu 15 for the last 3000 years BP (the upper 11cm has not been dated). Thus, two issues remain unresolved in our new understandings of Zenadh Kes colonization. First, the historical relationship between 'colonisation' events of 3500 BP and the onset of major midden development and marine specialization at 2500 BP. Second, the mismatch between the suggestion of 3500 BP Austronesian influences into Zenadh Kes (David et al. 2004) and the more temporally restricted and well-accepted timeframe of 2000 BP for Austronesian expansions along the southern Papuan coast. It is in this connection that excavations at Mask Cave on Pulu islet in western Zenadh Kes shed fundamentally new insights into the history of Zenadh Kes and southern Papua New Guinea.
Mask Cave: excavations and stratigraphy
Mask Cave is an elevated granite boulder supported by two major granite bedrock 'pillars'. The fiat-roofed cave today has low average headroom of 1.5m with a parallel sediment and granite rock floor sloping downwards to the north (Figures 2, 4 and 5). Two trenches located 3m apart were excavated to sample deposits from different sections of the site (Figure 4). The west trench comprised four contiguous pits--Pits A, B and E (each 50 x 50cm in area) and Pit F (50 x 30cm in area)--excavated to a maximum depth of 112 cm below the surface. The east trench comprised two contiguous 70 x 70cm pits (Pits C and D) and was excavated to a maximum depth of 111 cm. In total, 2037kg (1572 litres) of deposit divided into 266 spits or Excavations Units (XUs) averaging 2.3cm in thickness were excavated (Figure 5). Sediment samples were taken for all XUs. Elevations, recorded for the beginning and end of each XU, totaled over 1000 for both trenches. A datum point was established on the highest point of an exposed granite boulder at the centre of the site (Figure 4). The northwest corner of Pit C is 156mm below the datum while the southeast corner of Pit A is 143mm below datum. As such, the surface of both trenches is on roughly the same horizontal plane. The consolidated nature of sediments made excavation physically difficult. All excavated sediments were dry sieved through 3mm mesh (measured on diagonal). For the purposes of this paper, investigation focuses on Pit A, the deepest, oldest and best-dated cultural sequence recovered from the site. Dating and preliminary analysis of cultural materials from the other squares appear consistent with Square A findings.
[FIGURES 4-5 OMITTED]
Both trenches revealed broadly similar stratigraphies, albeit with different erosional structures reflecting topographic and water runoff variations across the site. The resulting complex stratigraphy is summarized as follows. Excavated sediments were divided into three major lithostratigraphic units (SUs) (Figures 6 and 7). Basal sediments (SU3) rest on granite bedrock across much of both trenches. They comprise pale brown (10YR 6/3), slightly acidic (pH = 5.0-5.5), fine-grained to coarse-grained granitic sediments (subdivided into SU3a and SU3b respectively in Pits C and D) with few granite rock inclusions and tabular fragments of granite roof fall. Sediments were mostly highly consolidated and difficult to excavate. Sediments exhibit scattered fibrous roots and the occasional larger root (maximum diameter = 5mm). The upper surface of SU3 in Pits A, B, E and F is undulating, probably as a result of subsequent erosional events associated with water runoff through the northwest sections of the cave. This erosion is most obvious in the c.25cm-deep channel feature that runs through the lower sections of Pits E, A and F (see south section of Pits A and E, east section of Pit A, and north section of Pit F in Figure 6). The northeast side of this channel is decomposing granite bedrock that represents the now buried sections of the westem bedrock support 'pillar'. As such, this channel runs along the southwest edge of this 'pillar'. Prior to removal (erosion) of sediments with formation of this channel, SU3 in Pit A was at least 60cm thick. While SU3 in Pits C and D is culturally sterile, SU3 in Pits A, B, E and F contains a low number of flaked stone artefacts and small, dispersed fragments of bone and charcoal.
[FIGURES 6-7 OMITTED]
SU2 is similar in form and texture to SU3 and is located above most of SU3 in the west trench. It represents a brown (10YR 5/3), slightly acidic (pH = 5.5), coarser-grained granitic sediment, ranging from consolidated to highly consolidated (subdivided into SU2a and SU2b respectively), with scattered fibrous roots and the occasional larger root (maximum diameter = 6mm), and few granite rock inclusions and tabular fragments of granite roof fall. The surface of SU2 is largely flat and parallels the surface topography of the cave. The major exception is another major channel feature located in Pits E, A, and E Here SU2a sediments filled and buried the large channel feature in SU3 but subsequent erosion removed much of this fill to create a new, c.40cm-deep channel also running along the southwestern edge of the granite bedrock 'pillar'. As with SU3, SU2 contains a low number of flaked stone artefacts and small, dispersed fragments of bone and charcoal.
SU1 forms the major culture-beating deposit of the site. In the west trench SU1 sits above SU2 whereas in the east trench SU1 sits upon culturally-sterile SU3 sediments (SU2 not represented in the east trench). The base of the unit varies from 35cm below the surface (Pit A) to more than 90cm (base of Pit F), and consists of consolidated, granitic sediments that range in texture from gritty (SU1b) to coarse grained (SUle). Colour is dark grayish brown (10YR 4/2) grading to grayish brown (10YR 5/2) and brown (10YR 5/3) with depth. Sediment is slightly acidic (pH = 5.0-5.5) with neutral sediments (pH = 6.0-6.5) within 5cm of the surface. Numerous fibrous roots occur through the deposit, along with a low number of larger roots (maximum diameter = 36mm). Loose and often powdery sediments forming the surface 1-2cm of the deposit were designated SU1a. SU1 filled and buried the c.40-deep channel feature cut through SU2 sediments (Pits E, A and F) and continued to accumulate a further 40-60cm of deposit on top of SU2. SU1 contains scattered fragments of granite roof fall in marked contrast to SUs 2 and 3. Flaked stone artefacts and fragments of marine animal bone, charcoal, and ochre form the moderate to high-density cultural deposit. Significantly...
|
