Volume 109 2000 > Volume 109, No. 1 > Changing identities: The ethnohistory of Roviana predatory head-hunting, by Shankar Aswani, p 39-70
                                                                                             Previous | Next   

- 39

Academics and other observers have put forward two general arguments to explain the intensification of predatory head-hunting in the New Georgia Group, Solomon Islands. The first argument proposes that, beginning in the 19th century, the progressive intervention of outsiders, including traders, whalers, missionaries and colonial government officials, in the affairs of indigenous New Georgians transformed the region's political economy. Authors emphasise the exchange of iron weapons and fire-arms for indigenous goods in propelling the intensification and expanded range of regional predatory head-hunting (Bennett 1987, Boutilier 1973, Hviding 1996, Jackson 1975 and 1978, McKinnon 1975, Schneider 1997, White 1991, Zelenietz 1979). A second argument stresses indigenous motivations including “native savagery”; the need to procure victims and slaves for offering and sacrifice in inauguration, burial and purification rituals (Coombe 1911, Goldie 1909,Hocart 1931,Knibbs 1929, Somerville 1897, Wilson 1932, Woodford 1890), and feelings of personal and social well-being purported to result from the act of beheading and the collection of human skulls (Bennett 1987, McKinnon 1975, Scheffler 1965).

Contemporary authors have developed such arguments relying on the chronicles of 19th and early 20th century observers (e.g., Shineberg 1971, Edge-Partington 1907, Guppy 1887, Hocart MSS, Mahaffy 1902, Somerville 1897). Their scholarly texts are rich in historical data and provide authoritative accounts of 19th century increasing predatory head-hunting and its commercial interlocking with Europeans, but they omit the narratives of New Georgians themselves and their views regarding the ritual and political significance of head-hunting. Such an indigenous view is undoubtedly essential to understand pre-European contact transformational processes and events, including changes in regional settlement patterns, political structures, technology and material culture, trading networks, regional military alliances, ritual and religious activities, and ideological domains. The favouring of static written historical accounts over diachronic indigenous narratives illustrates what Rosaldo (1980:37) calls “the false dichotomy between internal studies of oral traditions and external studies of written documents.”

- 40

This paper's first objective is to use ethnohistorical, historical and archaeological evidence to dispel the notion that organised large-scale interisland predatory head-hunting of Roviana, New Georgia, was primarily a 19th century phenomenon. There is no doubt that European agency, particularly commercial relations, had a profound impact on the nature and scale of localised warfare, as is indicated by historical and indigenous oral accounts. But the rosy picture espoused by some researchers stating that before any significant Western penetration indigenous military confrontations were small retaliatory incursions of an “interpersonal” and “interfamily” character is not supported by the evidence (Bennett 1987, Boutilier 1973, McKinnon 1975, Zelenietz 1979; cf. Jackson 1978, Spriggs 1997). It is shown that before the 19th century, the intensification of predatory head-hunting emerged from the cyclical political expansion and contraction of regional polities and that the indigenous ideological entanglement initially exacerbated head-hunting. This discussion centres on the period before European contact as other authors have extensively reviewed changes in head-hunting practices after 1830.

The second integral focus of this paper is to explain the role of trophy skulls as political symbols and their social, ritual and historical significance. Particularly, it addresses the assumption that enemy heads in New Georgia were perceived as a locus for the appropriation of a victim's “life-force,”, or mana, and supports Rivers' (1920) familiar rejection of the idea that heads were such repositories in Melanesia. Severed enemy heads are here construed as a medium to authenticate a chief's and his group's efficacious state and its ancestral endowment. Captured human skulls and those of captives destined to be sacrificed were a tangible way of counting and storing for display a group's success in war—a fruition that symbolically confirmed the precursory transaction of efficacy from the ancestors and deities to the living. Roviana chiefs and their kin, therefore, were able to secure their own regional ritual, social and political hegemony by constructing a quantifiable “currency of rank” out of persons' detached parts—a cultural idiom of political aggrandisement that was created long before European social and economic incursion. Key to a discussion on the nature and transformation of Roviana predatory head-hunting is understanding the gradual amalgamation of diverse inland and coastal Austronesian and non-Austronesian speaking populations during the 16th and 17th centuries, and the subsequent development of novel economic, ritual and political activities as these groups coalesced into the centralised Roviana coastal chiefly polities. 1

- 41
Figure 1: The Roviana and Vonavona Lagoon cultural area.

Indigenous oral narratives are coupled with other materials in this paper, but, without losing sight of an anthropological analysis, Roviana's own historicity is privileged, perhaps avoiding what George (1996a:59) calls “yet another act of pacification.” Insofar as Roviana's own historical representation is concerned, the accounts of Munda, Kalikoqu, Saikile and Vonavona informants are included to encompass a larger Roviana socio-cultural constituency (Fig. 1). Late 19th and early 20th century Western accounts have mostly privileged the military exploits and political influence of Chief Hinggava (Inqava) of Sisiata, Munda area (e.g., Edge-Partington 1907, Woodford 1890), going as far as calling him “Ingova, the king of Rubiana” (Somerville 1897:363). But what is absent from these accounts are commentaries on the political power and military exploits of other Hinggava contemporary warrior-chiefs, including Nona from Kalikoqu, Lepe from Kindu, and Bene from Haratana among others, or even those of earlier times such as Roviana-Kokorapa's Tae-Bangara and Qutu, and Saikile's Odikana and Mulongo, who were critical leaders in the development of a greater Roviana political confederation and its military regional hegemony.

Munda-centric “Roviana” accounts of head-hunting and oral history have been favoured largely because Western traders and missionaries settled the area first and opened a gateway for outside observers (e.g., Hocart in 1908). The breadth of contemporary Roviana oral history, however, also encompasses the inherited memories of Saikile, Kalikoqu and Vonavona - 42 elders, Austronesian and non-Austronesian speakers whose inland and coastal origins extend from Koqu Kalena in southeastern New Georgia to the northwestern tip of Parara Island. These original groups include Kazukuru-Roviana, Taghosaghe (Vinakiki tribes), Lio Zuzulongo, Hoeze, Hoava, Kusaghe, Kohinggo, Parara, Vuragare, Koloi, Viru, Nono (and others in Marovo), and Kalena Bay tribes among others (Fig. 1), as well as numerous filial links with groups throughout the Western Solomon Islands.

This cultural mosaic affords informants an opportunity to underscore their own political vantage point, potentially blurring an integral “historical” representation of the past. Yet Roviana narratives are consistent enough to produce a reliable body of evidence—a reliability obtained through comparative ethnohistorical and ethnographic research conducted across the New Georgia Region. 2 But in contrast to Rosaldo (1980), who worked with Ilongot members who had actually participated in a head-hunting raid, the accounts presented in this paper are based on the recollection and reconstruction of elders whose closest head-hunting practising kin are at least one or two generations removed.

From this discussion a crucial question follows: how can we manipulate, in Thomas's (1995a:286) words, “static constructs” to represent “some imaginary time that is at once pre-colonial yet accessible to our vision,” particularly with regard to the subject of head-hunting that is so profoundly obfuscated by a contemporary Christian discourse? But pre-colonial images of the past are historically germane, images that while contextually metamorphosed to incorporate alien ideologies, still convey the essence of past behaviours and events. Because head-hunting narratives are so much linked to the history of the Roviana chieftainship, head-hunting being in the past a principal means to attain political and spiritual authentication, they still survive intricately imbued with the contemporary political idioms of historical representation. Rosaldo (1980:31) suggests for the Ilongot that the selected rather than invented representations of the past survive, or rather, as Sahlins (1985) suggests, it is the consciousness and historicity of ruling elites in certain societies that endure through time. Similarly, Valerio Valeri (1990:191) maintains, countering the notion of “disinterested” history as real history, that the political objectives of elites, generally characterised as mythologising the past, are precisely what motivate “realistic” attempts at historical representation. Vansina (1985:199) concludes in his seminal work Oral Tradition as History that oral tradition “must bear the brunt of historical reconstruction” in regions where there are no or few written records.

The value of Roviana oral history is more convincingly demonstrated by its congruence with the regional archaeological record still being unearthed (Sheppard and Walter 1996, Sheppard et al. 1998). Miller (1980), for - 43 instance, has shown the usefulness of combining ethnohistorical and archaeological data to evaluate the settlement patterns of diverse Solomon Islands groups. The processes that spawned the intensification of head-hunting in Roviana are revealed, to some extent, by the archaeological record. Radiocarbon dating of the shrine sequence in Roviana indicates a major change in shrine form from mainland “stone platform” to coastal “coral cobble” and associated artefacts beginning in the 15th and 16th centuries A.D. Based on the dating of associated ovens (oputu),it is at this time that we see the first construction of shrines similar in form to those used historically, and in which we see large deposits of shell valuables, skulls and other artefacts previously missing (Aswani and Sheppard n.d., Sheppard et al. this volume). The appearance of these new sets of cultural artefacts corresponds remarkably with the socio-political changes described by the oral traditions as illustrated by the following regional model.


Most scholarly accounts have highlighted the historical and ethnographic aspects of New Georgia head-hunting's intensification and cessation, but have rarely considered the proximate causes for its prehistoric transformation. 3 The following model illustrates the autochthonous processes of social differentiation and political expansion that caused the alteration of ritualised head-hunting in the Roviana cultural region before European incursion. The temporal scale for this account is based on a genealogical chronology of between 13 and 15 generations. Vansina (1985:182-85) suggests that for most societies a human generation encompasses between 25 and 45 years, or approximately three generations per century, hence situating this discussion in the mid 16th to early 17th centuries. 4

The origin of head-hunting and its actual inception in the region cannot be pinpointed with any certainty. It is conceivable, however, that the maritime peoples who inhabited the lagoon barrier islands and coastal fringes of New Georgia before the inception of the Roviana “cultural complex” practised head-hunting for hundreds if not thousands of years. Roviana oral history suggests that New Georgia inland populations began their radiation to the coast around 15 generations ago and gradually fused with coastal groups to develop into a centralised polity. Synchronically, these peoples' system of social differentiation was transformed, or synergistically recast, into a system approximating a stratified society.

Fox (1995) has postulated that social differentiation in Austronesian groups emerges from two parallel processes, “lateral expansion” and “apical demotion with concomitant predatory expulsion”, each relying on different narratives for constructing a people's origin. This model is useful for - 44 analysing the ideational transformation of inland populations after their progressive mixing with coastal groups. Briefly, lateral expansion is a process where fragmentation of social groups occurs when antagonistic factions within a group divide, assuming land availability, and establish their own settlements. A group's origin is traced by remembering paths and places of earlier occupation. On the other hand, social differentiation via apical demotion occurs when the ruling class seizes political power by genealogically demoting subordinate groups away from apical figures. Class differentiation and status are rigidly traced to genealogy, particularly filiation to central apical founders. These processes of Austronesian social differentiation, however, are not monolithic and are prone to operate simultaneously.

After inland groups moved to the coast, powerful chiefly lineages emerged. These claimed descent from mateana ‘divine beings’ thus demoting the status of inland chiefly lines and places of earlier occupation, and elevating these newly sacramentalised coastal polities. Thereafter, genealogical association to mateana ancestors became a prerequisite for the attainment of chiefly power. The emergence of new ritual, social, economic and political activities, and venues for political empowerment typical of many Austronesian stratified societies, distinguish this pre-European contact period. The inception of new cultural media and channels for their distribution is particularly significant.

The oral accounts and archaeological evidence suggest that in the early pre-dispersal sequence, inland populations did not use shrines as repositories of dead ancestors' skulls or build elaborate burial receptacles (oru); nor did they participate in regional exchange networks, producing large amounts of shell valuables (such as bakiha), manufacture war canoes (tomoko) and ritual houses (paele and zelepade), or engage in large-scale ritualised head-hunting and the capture of slaves, except for the occasional raiding of inland settlements (Aswani and Sheppard n.d.). Following coastal resettlement, however, new ideologies were conceivably developed to accommodate inchoate sources and means to attain power. The belief in venerating one's ancestor skulls and objectifying those of one's enemies into quantified political symbols through their physical accumulation was engendered, or probably gained prominence, shortly preceding or following coastal resettlement, or several centuries before reported European contact.

Head-hunting evolved as the central form of ritualised warfare in unison with interisland trading, the former producing Roviana's regional military and political ascension, and the latter establishing its economic foundation and extending its collateral alliances across the region. Ritualised warfare in concert with other ritual and commercial activities became characteristic - 45 of the political economy and religious life of coastal chiefly polities. Chiefly success in war, however, required the monopolisation of the ritual activities and the cultural media, including genealogical knowledge and the pivotal ancestral shrines and their associated ritual artefacts, which were believed to direct all forms of supernatural power.

Predominantly manifest among these artefacts were a particular class of shell rings made of fossilised giant clam shell (Tridacna gigas) called bakiha. Consecrated bakiha emerged as divine signifiers and visible manifestations of chiefly authority. They embodied the actively manifest higher powers of the mateana ancestors, or what Weiner (1992:4) calls an object's “cosmological authentication”. By authenticating the authority of chiefs, these sacred objects legitimised their control over the flow of ceremonial and commodity exchange networks, and over ritualised warfare and the material means for its fruition (Aswani and Sheppard n.d.). Both these activities were dependent on voyaging technology, itself an “excludable resource” that had also been monopolised by chiefs and their closest kin (Aswani and Graves 1998:140).

These burgeoning “materialized ideologies” (Earle 1997:144) and their control also required chiefs to continuously validate their authority. The attainment of ancestral efficacy not only required the seizure of sacred objects, but also necessitated non-metaphysically derived human praxis. This demanded, among other things, chiefs to be, or command, successful warriors in raiding and attaining human heads and captives. A warrior's head-hunting prowess validated his previous appropriation of “blessings” from the ancestors and deities. It also substantiated the chief's accretion of mana and political control. The physical currency to show their ritual and political success was the accumulation of human skulls and captives. In brief, rank was to a certain extent hereditary, but it also necessitated, a chief or warrior to prove his entitlement through leadership and warfare.

The regional intensification and decline of head-hunting ensued from the cyclical centripetal and centrifugal political strategies of Roviana chiefly polities nurtured by the indigenous processes outlined above. As previously hinted, apical demotion through predatory expulsion (Fox 1995) neither resulted in the unrestricted control of a single chiefly line, nor in the extirpation of competing chiefly lineages. Parallel lines continued to genealogically coalesce with central lines and simultaneously to expand laterally to found new chiefly dynasties in the region. These events intensified the regional processes of competition within and between groups in greater-Roviana. Because political authority required, among other things, success in war and the securing of human skulls and captives for its validation, a growing antagonism between nascent Roviana polities was engendered. The - 46 competitive wrath of war, however, was generally directed towards neighbouring and regional groups rather than within Roviana itself. Social differentiation was attained through the chiefly manipulation of economic, military and ideological spheres of power (Earle 1997). Following is a more detailed analysis of the oral and early history.


The Roviana cultural region extends from Rarumana to Kalena Bay (Fig. 1) and covers a large section of southwestern New Georgia Island. Rain forests pierced by gardens and coconut plantations mantle New Georgia's mainland littoral fringe and the raised coral islands enclosing the Roviana and Vonavona lagoons. In Roviana, the Saikile and Kalikoqu chiefly districts are the largest, followed by the Nusa Roviana-Kokorapa and the Munda area districts of Dunde, Kekehe, Lodu Maho and Kindu. In Vonavona, there are several major hamlets and other smaller settlements in Kohinggo Island. Today, all these groups share cultural and linguistic affinities. This common ancestry derives from centuries of tribal intermarriage between inland and coastal populations. Kinship ties also extend inter-regionally to include bonds with other islands in the Western Solomons and in recent years with the Eastern Solomon Islands—a process of tribal encounters extending over thousands of years and ever replicating traditional patterns of population expansion and integration (Green 1991).

Before their radiation to the coastal margins and barrier islands of the Roviana Lagoon, several inland groups inhabited the mountainous interior of southwestern and central New Georgia. The Austronesia-speaking tribes of Taghosaghe, Lio Zuzulongo, Hoava, Hoeze and Kalena Bay lived around the central and eastern Roviana Lagoon, while the Kazukuru non-Austronesian and the Roviana Austronesian speakers lived at the western end (Fig. 1). The degree of interaction between these societies before dispersal to the coast is unclear. But oral traditions suggest that the intermarriage of Taghosaghe (Vinakiki), Roviana and Kazukuru groups led to the formation of the Kazukuru-Roviana polity 5 and to the eventual replacement of the Kazukuru language with the Roviana one. The initial mechanisms of Roviana political development are characterised by the integration and replacement of diverse cultural traditions. The following extract of a mythological tale before dispersal to the coast illustrates these processes.

Luturu-Bangara made a feast to celebrate his son's upcoming installation as chief of Bao. Luturu-Bangara was old and needed somebody to succeed him. He called his tribe (butubutu) and the tribe with which his sister Sogaduri - 47 resided, who were living at Hia Gore, to come together to install Ididu-Bangara. He called the people of Tirokiaba, Patu Kuti, Patu Kuna and Zorutu. Bao was a big settlement and there big feasts were celebrated. Before the feast, the chief told his people that when the Kazukuru dancers came nobody should laugh because their descent line (tuti-na) was one of lepers (tie popoqu). The Kazukuru came and danced; some had no hands, no eyes, and no ears. One woman from Moqala Qanaqana laughed, hihihihi! and zaaaa! The dancers sank into the earth and became stones, snakes and other things. But some Kazukuru people remained and returned to their inland settlements. Thereafter, Ididu-Bangara became the chief of Bao and began his move to the coast. 6

Interior groups began their progressive radiation towards the coast around the mid 16th to early 17th century. 7 The resettlement of the Kazukuru-Roviana tribes to Nusa Roviana Island was the principal population movement among these. The reasons for relocation are uncertain, but oral traditions suggest that Kazukuru-Roviana groups were galvanised by conflict with neighbouring Kusaghe, epidemics, and the desire to access the lagoon's resources, particularly the fossilised tridacna shells for the manufacture of shell valuables (Aswani and Sheppard n.d.), which prompted the abandonment of inland settlements. Oral traditions and allegories of this period do not specifically mention head-hunting as a significant ritual activity, but do suggest that some form of inland strife existed. The chief of Kalikoqu provides an early recollection of the consecration of kin skulls, and the consumption and hoarding of human body parts.

Luturu-Bangara's son Ididu-Bangara grew at Bao. Both lived with their families at Bao and had eighty warriors with them. The name of the warriors was si ka vesungavulu puta tuturu pa Bao (the 80 ‘crazed’ warriors from Bao) and they were man-eaters. The warriors stalked the interior at different places kidnapping people and carrying them back to Bao. The victims were cooked and eaten. That is the reason these warriors were called the 80 tuturu of Bao. One day they went to Moqala Qanaqana, but the people of that place “conjured their spirits” and cast a spell on the 80 warriors who died upon their return to Bao. In their memory, the large stones standing at Bao were raised as their memorials (vina-tigono). Ididu-Bangara carried their atlas bones (qanaqana) to the coast. Yet, the skulls remained at a shrine (hope) at Bao.

The New Georgia inland integration of non-Austronesian Kazukuru with Roviana and Taghosaghe Austronesian-speaking groups (Davis 1997, Lanyon-Orgill 1953) marks the initial transformation of social and material culture—a process that was further intensified when inland groups subsequently fused with coastal populations. Upon their arrival to Nusa - 48 Roviana, Kazukuru-Roviana migrants did not encounter a barren landscape as the island was apparently sparsely populated at the time. The eastern half of the island was visited and/or temporarily occupied by the Koloi tribe and the western half was claimed by the Vuragare 8 coastal peoples. 9 Roviana-Kazukuru dwellers rapidly took control of Nusa Roviana Island and formed or solidified existing alliances with coastal groups through intermarriage. 10 Tribal amalgamation extended their territorial authority to cover most of south New Georgia. Roviana oral accounts again characterise this period as one of integration and replacement, and hints at the cyclical intensification of localised warfare and raiding. The chief of Kalikoqu explains.

After Ididu-Bangara settled Nusa Roviana Island, the Koloi people called his people to help them caulk their war-canoes (tomoko). His son Peupeu-Bangara went to Koloi to help because they were relatives. Two Koloi boys followed the coast back to Peupeu's settlement and found old blind Ididu-Bangara near the shore grinding shell rings (bakiha). They took a branch and tickled his face. Ididu-Bangara thought that flies were bothering him, so he rubbed his face until it was covered with shell debris. During the evening, while people were still feasting at Kosianae, the boys returned and told of their prank. Peupeu-Bangara heard of the story and told his old man. Ididu-Bangara was angry and told his son to bring his basket (seki pagara). He said to his son, “You will take this bakiha and go to Kazukuru.” Peupeu-Bangara went to the mainland and presented the bakiha to his close relatives who accepted the compensatory payment. The war party (qeto minate) of Kazukuru began killing the Koloi at Zare Ibibu in Nusa Roviana and continued throughout the barrier islands of the Roviana Lagoon. Then, they crossed into the mainland and continued killing until they reached the Biribiri River where they saw a standing axe made of ebony (turu karamaho rige). This was the location where Ididu-Bangara had said that the killing should stop. Some Koloi survived because they went to Vuragare and hid from the warriors.

The emerging “Roviana” coastal polity began to consolidate its regional power as these groups continued to aggregate at Nusa Roviana Island. It is at this crucial historical point that Fox's (1995) model of Austronesian social differentiation becomes more relevant. Two generations after Ididu-Bangara's descent to the coast, his descendants Qorabele and Taua (or Tagua) were ostensibly transformed into mateana spirit beings. Qorabele died and flew to the sky, while Taua died and sank onto the earth to be resuscitated later. 11 Their children became the central chiefly ramage on Nusa Roviana Island and eventually the Saikile and Munda area enclaves. Schneider (1997:223) interestingly concludes from his research in Dunde that the “[s]ocial categories of banara [chief] and hiama [ritual priest] became - 49 separately recognised at the genealogical level of Hipobanara and Hiqebanara [Qorabele's siblings] respectively”. This period is marked by the progressive inception of distinct social hierarchies eventually approximating what Ehrenrich, Crumley and Levy (1995) call “heter-archical” societies, or internal hierarchies within each social segment (Earle 1997:1). The formalisation of three chiefly(bangara), positions in Kalikoqu during the 19th century, theatungu (sitting chief),adanga (working chief) and bangara varane (warrior chief), exemplify this process. Note that although power became centralised during Hipo-Bangara's time (c. A.D. 1700), parallel chiefly lines continued to maintain a measure of political autonomy.

The advent of the mateana cults epitomises the process of apical demotion of inland genealogies and places of origin, and the emergence of coastal ruling descent groups of divine origin. Hocart (1922) identifies mateana in Simbo as a class of natural events including shooting stars and meteors, and also as non-human spirits of varied disposition. In Roviana, mateana were believed to be mythical beings capable of flight and transmutation. A Saikile elder notes that “mateana chiefs were not people like you and me because these firstbangara were chosen by the ancestral spirits”. This supernatural genealogical association, however, required chiefs to control the means to bring into fruition the supernatural powers of mateana.

Control over the spiritual domain was actually consummated by the chieftainship's seizure of focal shrines, ancestor skulls and associated artefacts (such as sacred bakiha), together with their living custodian priests. Individual shrines were associated with particular powers that if properly propitiated endowed humans with supernatural abilities. Many Nusa Roviana hill-fort shrines associated with fighting powers are precisely those identified with mateana chiefly lines. It is no coincidence, then, that Roviana narratives, myths and allegories characterise this period as a time of burgeoning predatory head-hunting, trading and maritime technological innovations. The following local allegory, for example, attributes the origin of head-hunting to the necessity to fulfil cathartic rituals and to inaugurate newly manufactured war canoes.

Head-hunting became important due to the need to capture slaves, take sacrificial children (veala), and from enmity between tribes. Va-peza (inauguration) is an important concept because the tribes needed to prove the success of their war canoes. There is a story relating to this. The Kaleqe Mateana from Parara had a stone war canoe (tomoko). She asked her son to make the canoe fit for fishing. They built a canoe made from stone. The son went to inaugurate the canoe and asked dolphins to tow the canoe. They went far away until they arrived to Lauru. There he took no fish. When the Lauru people saw him they wanted to kill him, but he killed everyone. He - 50 said to himself, “Oh! I have been unlucky taking fish, but I have been able to kill many men.” So he cut off their heads. He took the heads because in burial custom they were taken and left in ancestral shrines. When the son returned the mother was angry because he had not taken fish. The boy responded that he had collected heads and that he had inaugurated the canoe. The woman was not happy, because she was worried of future enmity. The women sank the canoe in a place where today a stone resembles a canoe. The origin of head-hunting, then, is related to the concept of va-peza.

The pivotal mythological narrative of the period, however, is the story of Tiola. Most Roviana warfare narratives pertain to this spirit, tio meaning ‘man’ and la meaning ‘go’, or together meaning “man go to fight!” Today, Tiola stands as a broken stone dog statue at a major shrine complex at Botu in the Nusa Roviana hill-fort. Sheppard and Walter (1998:19) report that Tiola stands in the most highly defended area of the hill-fort which has large outer defensive terraces and deep ditches around the area. In the past, Tiola was considered an oracle capable of foreseeing the approach and direction of Roviana enemies and selecting the locality to conduct a raid. The Tiola origin story, while mythological, is replete with references to maritime technological innovations—the paele (canoe house) and tomoko (war canoe) being fundamental to the socio-economic and politico-religious foundation of Roviana society. A Saikile elder explains:

Before, people did not know how to make the tomoko. But when Tiola came to Nusa Roviana from Simbo he told the Vuragare people how to make the tomoko using certain materials. The high bow and stern of the canoe were designed to follow the tail of Tiola. People wondered where they could find the power to consecrate the tomoko. Tiola told them to make a canoe icon (nguzunguzu) and place it on the bow of the canoe. The dog sat down and folded its legs and said “like this”. Tiola could now follow them in war expeditions. The nguzunguzu conferred upon them the power of Tiola and prevented the Kesoko spirit from traversing the bow of the tomoko and jinxing the expedition. Wherever they went they exterminated the people. 12

The zenith of pre-European contact Roviana regional military, commercial, and political hegemony transpired during the rule of Tae-Bangara c.1750-1780 and continued through the rule of his son Qutu in the late 18th and early 19th centuries—a zenith that was only achieved again during Hinggava's time in the late 19th century. Yet, it was under the rule of these same men that existing parallel Roviana chiefly lines splintered to establish their own independent chiefly polities. New chiefly dynasties appeared across the landscape, including the Saikile polity in the mid to late 18th century and the Munda chiefly enclaves in the early 19th century. - 51 Furthermore, Nusa Roviana itself was formally fragmented into the chiefly districts of Kalikoqu, Kokorapa and Vuragare. By closely examining the genealogies of these splinter groups within Nusa Roviana itself, however, it is possible to infer that each sub-polity had all along maintained its own leaders and had mounted its own head-hunting expeditions (e.g., Vuragare). These sub-polities had converged with the central Kazukuru-Roviana chiefly line several times through the generations and, therefore, had maintained a subordinate position. Tae-Bangara, for instance, was Chief of Kazukuru-Roviana at Olobuki, Nusa Roviana. Yet, his mother was from Vuragare and Tae-Bangara himself had married a Vuragare woman among others.

Tae-Bangara is remembered as a ruthless chief who killed rival kinsmen and was quick to mount head-hunting raids against neighbouring groups. Such was the antagonism created by his rule that he was eventually murdered by warriors allegedly sent by his immediate kinsman, Odikana. Tae-Bangara's murder resulted in civil strife and only Odikana's move to Saikile to establish his own chiefly enclave prevented civil war. Qutu eventually succeed his father as chief of a now splintering Roviana polity. Once again, civil strife between Qutu and his half-brothers, Gove and Raro, prompted the latter to move to the Munda area and establish their own chiefly polities, hence further fragmenting Roviana. Nonetheless, all these groups continued to maintain strong socio-cultural links and to pursue the political interests of the greater-Roviana confederation that was witnessed by the first Europeans to visit the region.

Key to a discussion of territorial schisms is the Roviana kinship system of cognatic descent and cumulative filiation. In this system, an individual traces his or her descent bilaterally and can theoretically accumulate consanguine links through a genealogical association to various maternal and paternal ancestors (Aswani 1999). This fluidity allowed Roviana chiefs to redefine their kinship alliances to suit their shifting political aims. Even though most Nusa Roviana inhabitants shared the Kazukuru-Roviana apical ancestors, they did not all share the same pool of entitlement rights. For instance, Qutu's half-brothers, Gove and Raro, established their own chiefly enclaves in the Munda area and their descendants eventually moved to the Vonavona Lagoon because they simultaneously had entitlement rights to Kazukuru in the New Georgia mainland and Vuragare in the islands of Parara and Rendova. This intra-regional political mosaic with ever rearranging political configurations had a tremendous impact on the cyclical escalation and decline of ritualised predatory head-hunting. Regional competitive processes, therefore, were exacerbated by an increase in splintering chiefly enclaves within greater-Roviana.

- 52

Early encounters between Europeans and Solomon Islanders indicate the presence of head-hunting and cannibalism as early as 1568 when the Spanish explorer Alvaro de Mendaña y Neira visited Santa Isabel Island in the Central Solomons. That year, “the Spaniards met with one chief from the western end of Isabel returning with about 110 men in 15 canoes from an attack on a group said to be more than 20 leagues (120 km) from his territory. He offered the Spaniards part of the booty, the right arm and part of the shoulder of a boy…urging them to…eat…”(Landín Carrasco 1992, Spriggs 1997:227-28). Woodford (1909:506) suggests that the shape, carrying potential, size and speed of the Isabel canoes encountered by the Spaniards indicate that these were probably war canoes. The Spaniards' account is significant because it synchronises with the genealogical chronology and the radiocarbon dates, which as noted indicates a major change in Roviana shrine forms and associated artefacts beginning in the 15th and 16th centuries.

Later accounts marking the encounter of Europeans and Solomon Island war parties include Carteret 1767 in Malaita, Bougainville 1768 in Choiseul, Surville 1769 in Isabel and Shortland 1788 in Simbo among others (Haddon and Hornell 1975, Woodford 1909:506-7). In 1803, Mrs Kent a passenger onboard the H.M.S. Buffalo observed “what must have been a tomako [i.e.,tomoko], or head-hunting canoe: it carried fifty men, had high curving prows decorated with pearl shell and carvings of fish, birds and heads. In the centre of this canoe was an elevated seat for a ‘Chief’” (Jackson 1978:49). The high prow, particularly in the Western Solomons, is a distinctive characteristic of canoes primarily used in warfare. In 1769 Surville described Port Praslin, Isabel, canoes as follows: “The prow and stern are raised very high, apparently for the purpose of defending the warriors in them from arrows, by presenting either end to the enemy: and in general they are ornamented with pieces of mother-of-pearl…” (in Haddon and Hornell 1975:102). Ribbe (1903) suggests that in Kia, North Isabel, an area culturally related to Roviana, only war canoes had the distinctive high prows (cited in above). Roviana small canoes (mola) had raised prows, but among the large plank-built canoes only the tomoko war-canoe (Fig. 2) had the distinctive high prows and elaborate embellishments. Other canoes like the gopu, a canoe mainly used for trading expeditions, or qalo (occasionally used in warfare) did not have the raised prows and embellishments. These early historical accounts suggest that tomoko war canoes (or their regional equivalent) were being used in interisland warfare as early as the 16th century (cf. White 1979).

- 53
Figure 2: The tomoko war canoe (Aswani photograph).

The ethnohistorical, historical and archaeological evidence shows that McKinnon's (1975:290) conclusion that “large-scale raiding…was most likely a phenomenon that only came into being in the period following contact with Europeans” and Bennett's (1987:36) suggestion that “[e]xcessive head-hunting was an attempt to control he new forces unleashed… by the advent of iron” downplay the historical complexity and the time depth for indigenous forms of predatory head-hunting in the region. Woodford (1922:69), perhaps for the wrong reasons, rightly suggested that New Georgia large-scale head-hunting against Isabel and other islands “had beengoing on for at least three or four centuries” before colonial rule. Nonetheless, as was recognised by McKinnon and others, the entanglement of Roviana polities with Europeans contributed to the spiralling rise of regional trade and predatory head-hunting.


Janet Hoskins (1996a:2) defines head-hunting as “an organized, coherent form of violence in which the severed head is given a specific ritual meaning and the act of head taking is consecrated and commemorated in some form”. While this definition is appropriate for the New Georgia situation, it is necessary to discriminate between the act of severing a head in a fortuitous ambush, as often occurred among the Ilongot of the Philippines, and full mounted excursions, like those organised by the Iban of Borneo. In Roviana, as in the rest of the New Georgia Group, localised small-scale skirmishes in ambush attacks were often carried out by hired assassins as revenge killings for murder, adultery or serious customary infractions between kin or neighbouring groups, and not always involved beheading the victim. Roviana predatory head-hunting, therefore, is best defined, following Hoskins, as a - 54 highly organised form of ritualised warfare (including preparatory and post-raid activities), involving a sizeable force and requiring the voyaging technology for interisland travel (when not attacking inland settlements), mounted for the purpose of eliminating enemies and acquiring trophy skulls and slaves to fulfil ritual, social, economic and political demands. Rousseau (1990) distinguishes heuristically between Central Borneo Dayak “headhunting” as an act of headtaking, “warfare” as an act of political coercion, and “raiding” as an act to secure slaves and property (cited in Tsing 1996:197). Actually, Roviana predatory head-hunting, whether carried out as an act of war against neighbouring tribes or as a ritualised predatory act against distant groups, simultaneously fulfilled all these objectives.

The New Georgian rationale for head-hunting has often been ascribed to the necessity to acquire one's enemies' mana. Authors' conceptions have been nurtured by the idea that in the New Georgia religion the human head was considered as the “seat” or the “prime repository” of mana (Bennett 1987:35, Hviding 1996:89). This force was both “beyond the ordinary power of men” (Codrington 1891:118-19) and a non-metaphysically derived state of being efficacious (Keesing 1984). The former was a “‘supernatural’ or ‘nonordinary’ component of efficacy that is the necessary condition” for the latter, or the “‘natural’ or ‘ordinary’ components to take effect in certain areas of action” (Valeri 1985:97). From this standpoint, severing a human head imbues the captor with his victim's life force and collaterally provides his social group with spiritual and physical benefits—a notion tantamount to the classical anthropological interpretation of Southeast Asian head-hunting as a means of procuring a “life-fluid” or “soul-substance” and attendant social welfare (Freeman1979, Fürer-Haimendorf 1938, Kruyt 1906, Serpenti 1968; cf. Hoskins 1996a, Needham 1976).

Less attention, however, has been given to explaining the relation between the heads of one's kin and those of one's enemies as repositories of mana, and the ritual consecration of the former and the violent appropriation of the latter. The relationship between enemy skulls—their congenital mana being the postulated “causal nexus” (Needham 1976)—and the indigenous perceived feelings of well-being and ritual purification after their capture has yet to be established (see Dureau this volume). Bennett (1987:35), for instance, ambiguously notes that because the “specific location of mana was recognized by the practice…of preserving and venerating the skulls of the ancestors” in anticipation of positive interventions, taking the head of an enemy “did not conflict with general religious values”. Other authors have outright assumed the ritual equivalence of kin and enemy heads as the loci of mana and the mutual appropriation of “soul value” by keeping the former and taking the latter. For example, McKinnon (1975:301) states:

- 55

A vital force was thought to govern personal welfare…. One of the most direct ways of obtaining influence was to take the life force residing in another human being, centred in the head. It could therefore be obtained by taking the person's skull. Capturing heads was not only an opportunity to demonstrate bravery, the successful hunter also acquired…“soul value”. For all practical purposes reverence for the heads of ancestors had always been a part of custom. Skulls symbolised the lineage origins, its seed…. The predilection for capturing the heads of those considered to be alien appears to have been a logical extension of this practice.

In Roviana, however, the opposite was true. There was an unambiguous distinction between the ritual functions of deceased kin heads and those of slain enemies. In general terms, keeping the head of a Roviana chief or relative was an act of ritual consecration, a means to subsequently secure the power of the ancestors as vectored through the skull's physical presence. The skulls of enemies, however, stood as those of strangers and objectifiable others. Consumed vessels containing nothing that could be supernaturally taken (i.e.,mana) but, rather, something that could be supernaturally denied. By severing and taking heads and slaves, Roviana warriors abrogated the enemy's future benefits of securing the power of the enemies' own ancestors—the souls of many victims destined to become earthly prisoners seeking revenge against the living. 13 Roviana warriors' success in war also affirmed the ritual superiority of Roviana warring spirits and ancestors by showing that these were capable of “eating” the souls of their victims before their immolation. The skulls of kin and those of enemies, therefore, stood in an inverse relationship to each other. Keeping the skulls of kin amplified one's own ancestral power and the community's welfare, while taking those of enemies separated them from theirs or, as Josselin de Jong (1937:164) maintains regarding head-hunting in Indonesia, was a “supernatural means of redressing the balance between the groups concerned or even turning the scales to the advantage of one's own group” (Hoskins 1989:428).

This appropriation and inverted disowning opened for Roviana men a direct symbolic channel to transact with their ancestors and deities. Certain enemy skulls were kept as transient “inalienable possessions” (Weiner 1992) and were eventually canalised through a process of unidirectional exchange that Godelier (1999:13) calls Mauss's “fourth obligation”. Skulls were stored in ritual houses as heirlooms for display by chiefs and warriors,and the skulls of prominent victims were, at times, embellished and offered as prestations to Roviana ancestors and deities. 14 In a parallel fashion, designated captive children (veala) and slaves were offered to the ancestors and deities by being decapitated in purificatory and inaugural sacrifices. Such offerings, or a sort of pecuniary offering resembling the Malangan - 56 “gift to god systems” (Küchler 1988:626), not only “nurtured” the dead, but also provided powerful men a venue to authenticate their efficacious state. Accordingly, men capable of controlling the monopoly over enemy skulls and slaves, and the means to acquire them, opened for themselves, and by extension to their kin, a venue for their own sacramentalisation and concomitant social differentiation. The causal nexus linking the severed enemy heads and the indigenous perceived fertility and ritual purification, therefore, was not in the appropriation of the victim's “soul-substance”, but, rather, in the channels of “mana-ization” (Hviding 1996:91) that their taking unlocked. 15

For enemy heads to be transferred to the ancestors, or even to be disposed as worthless tallying vessels, they had to be first objectified and metamorphosed into detachable exchangeable objects. Persons were stripped of their personhood, conceptually “animalised” into fish or pigs, and converted into a ritual currency of rank through their decapitation and the subsequent quantification of their skulls (cf. McKinley 1976). Likewise, fated sacrificial victims were referred to as animals before their immolation, and their skulls eventually entered the chiefs' and warriors' head tally. Hoskins (1989:434) notes for the Sumbanese case, “the line between persons and things was crossed through detaching and hoarding parts of persons as trophy objects….” She also suggeststhat following pacification, unidentified heads could be bought with livestock and “treated with the casualness of commodities” (1989:435). But as her case study shows, commoditised trophy heads were polysemic enough to transcend their alienation and regain their personhood through purchase and subsequent burial by their kin. Objects in Roviana could similarly move between spheres of exchange, being “gifts”, “commodities”, or “inalienable possessions” depending on the social, economic and political context in which they were exchanged and/or transferred (Aswani and Sheppard n.d.). Although Roviana skulls were not bartered for other objects, as in the Sumba case, and did not have a set exchange value, the compensatory shell-valuable payments made to other participating chiefs for their success in war (i.e., their services), the remuneration made to partaking warriors by granting them access to an accruable currency (i.e., skull-trophies and slaves), and the trade of slaves (i.e., their sale) that were “really supposed to be dead” (Hocart 1931:306) for their ritual beheading are arguably tantamount to the commoditisation of skulls, at least at some point in their movement between spheres of exchange.

To rephrase Godelier's (1999:36) social exchange formula of “keeping-for-giving and giving-for-keeping” in a distinct social context, the ritual objective of Roviana head-hunting is best defined as taking-far-giving and - 57 giving-for-taking. Human heads were taken and captives abducted to ritually nurture Roviana ancestors and fighting spirits. In turn, nurturing these ancestors and spirits in purificatory rituals, for instance, guaranteed Roviana chiefs and warriors the ancestrally derived efficacy necessary for the successful taking of additional heads and capture of victims. Head-hunting, therefore, served Roviana chiefs and warriors as a vehicle to establish their own regional ritual, social and political legitimacy, or, as suggested by Robb (1997:138), it is through acts of violence, and the capacity to perform them, that the identity of categories of people is defined and reinforced.


Roviana head-hunting parties raided villages across the Western and Central Solomon Islands. War parties primarily coming from Rendova, Vella Lavella and Marovo Lagoon, in turn, sporadically attacked Roviana groups in Nusa Roviana, the Munda area and Saikile. 16 As Roviana polities were genealogically confluent, they maintained a confederated military alliance which also extended across to Simbo, Roviana's major head-hunting ally (though probably representing different factions at different times), to some Kolobangara groups, Ughele in Rendova, Kuboro in Choiseul and Kia in North Isabel among others. Warring expeditions were generally conducted under two circumstances: against neighbouring populations such as Lokuru (Rendova), Kusaghe (North New Georgia), several Marovo Lagoon groups (e.g., Vangunu and Gatokae) and Tetepare for revenge or enmity killings; and against distant islands, including villages in Choiseul, Isabel, and as far as the Russell Islands, Malaita and West Guadalcanal, for the ritualised collection of trophy heads and slaves.

Expeditions against neighbouring groups in the Western Solomons entailed two types of expeditions: (i) localised small-scale skirmishes in ambush attacks by hired assassins as revenge killings for murder, adultery or serious customary breaches between kin or neighbouring groups, 17 and (ii) larger war parties hired to kill powerful rival warriors or to annihilate entire groups resulting from intra-group enmity. In either case, the contracting party had to make a compensatory payment of shell-ring valuables (e.g., bakiha and poata) to the hired chief and warriors. Rarely did a killing ensue without the proper payment to the killer or war-party. Chiefs habitually kept shell-valuable coffers called nibaka, obtained in ceremonial and barter exchanges or locally manufactured, 18 to finance trading expeditions and, most importantly, to compensate chiefs and their warrior “mercenaries” for their successful killings and capturing of children intended for sacrifice (Aswani and Sheppard n.d.). Frequently, the granting of plots of land provided additional compensation for success in war.

The beheading of victims ensued when hired Roviana war parties attacked neighbouring non-Roviana groups or vice versa. But assassinations for murder - 58 or adultery between Roviana kin did not always entailed the victim's decapitation. The spirits of dead kin (boso lau) murdered by hired assassins or by the chief's warriors were considered extremely dangerous. To Prevent the angered spirit's revenge, the boso lau was buried upside down, its eyes covered with ash or lime, and a piece of its body (hair or ears) was burned as an offering in an ancestral shrine. Such an offering was believed to quell the angered spirit and activate one's own protective spirits who were “pleased” to “smell the cooking flesh” (humanga hibi).

Roviana conducted ritualised predatory forays against Choiseul and Isabel populations most intensively during the 19th century for the collection of trophy skulls and to capture slaves for sacrifices and for work. 19 Hocart (1931:303) reports that Simbo head-hunting forays against Isabel and occasionally Choiseul were chiefly conducted to procure heads for “vapenja”, or to inaugurate new canoes and new communal houses, and to commemorate the death of a chief and release a widow from confinement. To inaugurate (va-peza) a Roviana war canoe was a ritual means to please the ancestors and to ensure the canoe's future combative success. Failure to procure at least a victim's head in a canoe's maiden voyage would jinx (tamu garata) the canoe. The capture of children (veala) 20 for future toil and for sacrifice in purificatory rituals to propitiate a disharmonious natural order, and of very young women for service in ritualised prostitution (bibibolo and maqota), toil (nabulu), and adoption (pinausu) and eventual marriage was also an important aim. Proof of the intensity of slaving, particularly for captives taken from Isabel in the 19th century, is the large number of slave descendants presently living throughout the Western Solomon Islands (see McDougall this volume).

Yet oral history suggests that in pre-European contact times groups in the Western Solomons were also subject to Roviana and Simbo large-scale predatory head-hunting attacks. The burgeoning of regional intermarriages, often the corollary of abductions in raids, and trade networks prompted the eventual cessation of ritualised assaults against neighbouring groups, excluding revenge and enmity attacks that continued until the end of warfare in the early 20th century. This process inverts Schneider's (1997:121-24) causative chronology, which suggests that ritualised predatory attacks within New Georgia Island and the Western Solomons in general escalated during the late 19th century because of a shortage of skulls. The ability of Choiseul and Isabel populations to withstand New Georgia raids, he suggests, provoked a scarcity of the human prey necessary to fulfil ritual demands, thus re-directing the wrath of war towards regional groups. Most likely, however, attacks within the Western Solomons witnessed by Europeans resulted from an inflationary process. Trade with Europeans increased the supplies of both European and native goods, particularly iron weapons and shell valuables, allowing regional polities to further arm themselves and finance war parties to consummate long-standing regional political enmities.

- 59

Dureau (1994:72) maintains that Simbo “warfare” and “head-hunting” should not be conflated, as the “issue of the causes of local enmities is quite distinct from the causation of head-hunting”. But the line between an act of war against adjacent groups and ritualised predatory forays against distant populations was blurred in Roviana warfare. Notwithstanding distinct initial motivations for both warring acts as emphasised above—one caused by political dissension and the other by an indigenous metaphysical worldview—they both were profoundly intertwined processes that had unequivocally parallel outcomes. Both simultaneously served as a vehicle for political coercion within and beyond Roviana, they fulfilled the ritual necessity to procure human heads, and they provided slaves and property.

Head-hunting expeditions, conducted for one reason or another, were organised by chiefs only after ritual specialists (hiama) had consulted and properly propitiated the spiritual world. Failure to obtain the blessings of the ancestors and warring spirits would assure a raid's failure. Ritual specialists 21 had to ensure that all participating warriors were ritually unpolluted, to successfully prevent warriors from being harmed in battle. Such protection was acquired by mustering the tribe's ancestors' efficacious blessing (tinamanae) and the supernatural power of the warring spirits (liqomo-vovoso and vovoso-viriviri kana). 22 Ritual cleanliness also prevented the ire of higher order spirits, such as the tamasa kolo who could obliterate an entire expedition at sea if an offender was found amongst its crew.

Figure 3: A Roviana warrior (photo by Williamson, courtesy of the Royal Anthropological Institute of Great Britain and Ireland).
- 60

Particular warriors (Fig. 3) having their own genealogically intrinsic (kokolo) magical power (potana) avoided certain types of food and contact with women and children to maintain themselves ritually clean before a raid. Powers such as doma ‘stealth’, hurupete ‘ability to inflict fear’, malagigiri ‘capacity to weaken an adversary’, and mamahelo ‘lightness in movement’ were associated with certain kin-based groups. These magical powers were ignited when taking certain ginger plants (pasapasa). Warriors believed that during the heat of battle they could muster the powers of their warring spirits by the incantation of secret charms and battle cries like vovoso viriviri kana! The power of Roviana warring spirits and ancestors, they believed, possessed the head-hunter and gave him unmatched strength and fierceness.

Roviana ritual priests selected the targeted village through divination (sabusabukai and betubetue) and after conferring with local oracles, such as Tiola and the omnipotent liqomo talisman. Priests who invoked the ancestors at particular shrines (haro hope) by making them food offerings (vukivukihi) spiritually enthralled warriors. Young warriors, who had adorned their bodies with battle ornaments (e.g., dala head disks) and painted facial lines (busa sokovea), gathered near the zelepade ritual war-house and were incited into a ferocious state by chewing certain ginger plants and eating a customary pudding called qeqese. This pudding was mixed with ingredients such as a fearless lizard called leoputa and scraped tree bark from where women and children had previously defecated. By eating this mix, men obtained prowess and showed defiance. For some, particularly but not exclusively, those whose closest kin were varane warriors, it was also an initiation rite towards the attainment of the tie varane social distinction.

Hviding (1996:172) notes that the annual head-hunting peak in Marovo occurred during the calmest season in November and December. Roviana ritualised attacks followed a similar yearly cycle, but financed incursions against other Western Solomon groups were conducted throughout the year. Mounted expeditions generally consisted of two to five warring canoes each containing between 30 and 50 warriors (Fig. 2), although historical accounts speak of larger fleets towards the end of the 19th century. During a tour in the Roviana Lagoon, Woodford (1889:479) noticed that nearly every male inhabitant within several villages was away on a head-hunting raid to Isabel. White (1991:88) notes that Roviana chiefs used their links with the Kia people of North Isabel to expand their “cycle of trading and raiding” to the rest of the Island. 23 Roviana war parties would call in at Kia to rest, trade, and recruit local scouts (piko) for reconnaissance. The expedition, headed by a chief's most prominent warriors (nihana varane) and ushered by their Kia allies, rummaged along the coast in search for any signs of human - 61 habitation including lemuhe leaves (used in traditional ovens) and human excrement discharged by rivers. Once the unsuspecting village was located, the assault (rapata) was carried out at dawn or during heavy rain with the intent of the village's annihilation (eongo). A Saikile elder narrates on war tactics.

Once the tomoko had landed following the permission of the liqomo oracle, the warriors rapidly went to shore and started killing the unsuspecting victims while they were conducting their morning necessities. Some warriors were chosen to be in the front, usually a war leader called koimata. His rear guard was called totoba. The buko (a chief's spokesman) would assign who was koimata and who was totoba. There was also a warrior (varane) in charge of calling the withdrawal. The pule pae was at the back of the group during an attack and would shout if any of his men were wounded. But being wounded made them stronger, made them madder because their blood had been drawn (napo ehara). If the pule pae shouted and the adversary was not weakened, then the pule hite would shout again. These two men had the power of malagigiri, or to make people weak. Thereafter, they would proceed to kill the enemy. But if they wanted to spare the life of a woman or child they placed their shields (lave) on top of them. In Roviana it was taboo (hopena) to flee from the enemy and warriors had to receive wounds on their body-front. Scars in their backside made someone a coward (bubuhele) and unfit (maqoho). When varane faced death they wanted to “eat” the enemy's axe (gani karamaho) in their foreheads.

Hill fortresses (toa) were also stormed with the intent of killing prominent warriors and chiefs. Simbo informants indicate that a raid's major objective was to secure the head of the chief and prominent warriors. The scale of killing and capture of captives is difficult to guess, but historical accounts suggest that substantial numbers of people were killed and translocated. In 1844, Cheyne counted 93 heads taken in a Simbo raid and observed that the victims were of both sexes and all ages (Shineberg 1971:304). Woodford (1890) noted that in Roviana approximately 31 heads were taken in a single expedition. It is probable, however, as suggested by Dureau (1994), that in a typical raid between 10 and 20 heads were taken.

Bennett (1987:36) has argued that before the introduction of iron weapons the “number of casualties was limited by the comparability of the combatants' weapons….” Yet Roviana accounts indicate that, before and during the time of Tae-Bangara and Odikana, raids resulted in the near extermination of entire villages or groups as the Koloi example suggests. The depopulation of parts of Isabel and Choiseul Islands occurred with most intensity during the second half of the 19th century. But the progressive depopulation of Tetepare, parts of Kolobangara, and perhaps Ghizo and Mbava (Baga) Islands - 62 near Vella Lavella (both depopulated by the mid 1800s [Bennett 1987:36]) must have commenced earlier. Tetepare descendants' oral accounts confirm this idea by suggesting that the island was gradually depopulated between five and ten generations ago. 24 This regional context, however, did not mean that groups were in a permanent state of war, as Bennett also pointed out. Warfare in Roviana, as Macintyre (1983:20) has noted for the Tubetube of the Massim, “must have been counter-balanced by sequential appeasement or the formation of new alliances” through peacemaking and trade.

Roviana warriors attempted to kill all the inhabitants of a besieged village as a preventive measure against future retaliatory attacks and a supernatural means to abrogate the enemy's future benefits of securing the power of their own ancestors. Simultaneously this was a means of validating the ritual superiority of Roviana ancestors and warring deities. Most importantly, however, it was a medium for chiefs and warriors to respectively attain the distinction of bangara varane (brave chief) and tie varane (brave warrior). Roviana men validated the capture of ancestral efficacy and reaffirmed their political control by accumulating, ranking and enumerating skulls. The Roviana speak of ninae batu boso ‘tallying heads of the slain’ and hinena (or pinera) ‘spoils taken in battle’—concepts which support the notion that chiefs and warriors kept a tally of their killings. Knibbs (1929:102), for example, noted that Choiseul warriors kept “native cords…knotted at intervals to tally the count of killings in battle”. Isabel warriors similarly counted their victims by notching the handles of their shields (White 1979:113). 25

The indigenous quantification and accumulation of trophy skulls symbolise an ideology of ultimate alienation. This worldview, as argued above, was nurtured by a rationale that metaphorically converted persons into animals and detachable objects before their immolation. The Roviana animalisation of human prey, like the Huaulu assimilation of human victims into wild game (Valeri 1994), metamorphosed humans into “wild” fish and animal prey. During ritualised head-hunts against Isabel and Choiseul populations, Roviana warriors referred to the former as a red fish called heheoku (Lutjanus gibbus), a metaphor for the reddish skin colouration of Isabel populations, and to the latter as a black fish called valiri (Acanthurus nigricauda), a metaphor for the dark complexion of Choiseul people. Human prey was also depicted as slaughtered pigs. The association between fishing and hunting and head-hunting is manifest in their similar ritual context. Waite (1990:56), for example, suggests that Roviana bonito-fishing and war-canoe shrines were propitiated in a similar fashion before a trip's departure to assure safety in a fishing or head-hunting expedition. Barraud (1972) also identifies the powers of Simbo mateana as related to head-hunting and - 63 bonito fishing magic. This association is also evident in certain types of Roviana magic used in fishing and hunting and head-hunting, such as the power of tutulotanga, or the capacity to kill an animal or a human being with a mere scratch from the bearer's hand or weapon. This assimilation, however, did not entail the cannibalisation of humans, as there is little evidence to support the notion that war victims were eaten, save the piacular killings of veala children for their ritual consumption. However, oral traditions, e.g., the tale of the 80 warriors from Bao, hint that interior populations may have practised cannibalism before their dispersal to the coast and the development of the Roviana polity in the 16th and 17th centuries.

Mahaffy (1902:191 -92) noted that Western Solomon head-hunters “regard the yellow-skinned race, for who they have the greatest contempt and most insulting names, as their legitimate prey”. Wilson (1932:238) similarly observed that Western Solomons people “regard it as their mission to destroy” the people of Isabel and Choiseul. Such contempt for “other” Solomon Islanders prevented Roviana polities from ritually incorporating the skulls of their foes with those of their friends (cf. McKinley 1976), but made them treat the majority of captured skulls with irreverence. Once cut, the heads of victims were smoked and rinsed, and then carried back to the village “like a bundle of coconuts”. Following the return of a successful expedition with the blow of a conch shell (buki hogoto) and the warriors' welcome with dance and songs by an expectant village (peka aqa), the victims' skulls were placed in a tabernacle (patu kevuana) in the centre of a dancing ground (pavasa). A ritual specialist would spit on the skulls a mix of betel nut and ginger leaf, and recite a particular secret charm to ‘cool down’ (va-ibu-a) the revengeful spirits of the victims. In addition, an offering would be made at an ancestral shrine to celebrate the expedition's safe return (Waterhouse 1949:152).

Although Roviana warriors believed that their ancestors and warring spirits had already consumed the souls of their victims during their immolation, it was still necessary to exorcise the remnant spiritual force of the severed heads (batu boso). Just as people physically ate fish and pigs, Roviana ancestors and deities devoured the spirit of their victims, leaving behind empty receptacles to be accumulated by men. This consumption was depicted, for instance, by roping a dead victim's limb to the nguzunguzu icon in the canoe's prow to symbolically “feed” the Tiola warring spirit after a successful raid. Nurturing the dead and deities was also practised in Simbo where a ritual specialist would place a “slain enemy's ears into the fire” at inatungu shrines as an offering to the ancestors (to guarantee efficacy in communal undertakings) (Hocart 1931:314).

- 64
Figure 4: Trophy skulls placed on the rafters of a Roviana paele canoe house (photo by C. M. Woodford, courtesy of Solomon Islands National Museum).

Following their ritual inactivation, the skulls were placed in the canoe house (paele) (Fig. 4). The skulls of captured chiefs and warriors were occasionally translocated to the zelepade ritual war-house, while those of commoners were eventually left in interdicted sites or buried along walking tracks for their continued ritual desecration. Most skulls left in the canoe house were unidentifiable, but those of notable warriors and chiefs had a shell trinket (hinuili) attached to them or were caulked with parinarium nut paste, coloured and embellished (kibo) for their identification. The ultimate political impetus behind the accumulation of trophy heads of varied ranks was to elevate the status of the head-hunter and concomitantly that of the chief, characterised by a missionary's observation as “[h]e who has taken 200 heads is twice as great a man as he who has taken only 100” (Anon. 1888:563). The attainment of social prestige through head-hunting was manifested in religious ceremonies similar to the Simbo tundu (Hocart 1931), where a successful warrior would publicly assimilate humans into prey by counting the number of humans, pigs and turtles captured during his forays.

The human head as a trophy object, whether belonging to a raid or a sacrificial victim, was proof that a man had made a kill. Through their - 65 quantification, trophy skulls took the character of a commodity, or, more accurately stated, the rank and social prestige of chiefs and warriors were calculated in the currency of human heads and slaves. Furthermore, as discussed above, their taking opened for Roviana men a symbolic channel to directly transact with their ancestors and warring deities who in remuneration for their offerings bestowed Roviana polities with additional efficacy. In brief, the contextual mutability of persons and objects, or their polysemic permeation across different spheres of circulation, was the religious, economic and political impetus underlying social and political stratification and the development of large-scale predatory head-hunting and regionally extensive exchange networks in the region.

The initiation and transformation of Roviana predatory head-hunting cannot be fully understood without a study of Roviana's own diachronic historical representation. Particularly germane are narratives depicting the cultural transformation and amalgamation of inland and coastal Austronesian and non-Austronesian speaking populations during the 16th and 17th centuries and the efflorescence of novel economic, ritual and political activities as these groups congregated into centralised coastal chiefly polities. The historical and archaeological evidence supports this transformation. By focusing on indigenous oral history, this paper has dispelled the notion that Roviana large-scale organised interisland predatory head-hunting was primarily a 19th century phenomenon instigated by European agency. Rather, the asymmetric cyclical expansion and contraction of Roviana polities and their intrinsic ideological rationale spawned the inception and transformation of predatory head-hunting. In a more general way, I suggest that frequent violence amongst indigenous peoples was not necessarily a corollary of European contact.

The emergence of a new materialised ideology following coastal radiation established the foundation of the Roviana chiefly authority. Control of kin and enemy skulls, shell valuables and other sacred objects authenticated the chief's rule over the spiritual and material domains, and collaterally afforded chiefs control over the means of production required to organise raiding and trading expeditions. This appropriation reciprocally required chiefs to substantiate their authority by compelling them to engage in further head-hunting. Human trophy skulls were metamorphosed into a tangible ritual currency capable of travelling across more than one socio-economic sphere. While this transferability did not result in the direct exchange of trophy skulls for other commodities, the sale or gift of slaves for their ritual - 66 decapitation and the access given to warriors to tally trophy heads arguably resulted in the transient commoditisation of detached human body parts.

The accumulation of skulls demonstrated a chief's and his warriors' success in war and authenticated their previous acquisition of mana. Moreover, heads were severed and captives taken to ritually nurture Roviana fighting spirits and ancestors. Roviana chiefs and warriors believed that nurturing the dead and higher order spirits would guarantee their future success in war. Severed heads, therefore, did not directly invigorate Roviana warriors with their victims' mana but, rather, feeding the dead and warring spirits unlocked channels that sustained their political and ritual authority. Head-hunting ultimately served Roviana chiefs and warriors as a vehicle for their own political legitimacy.

In sum, Roviana political organisation and integration is distinguished by epochs of political expansion and contraction. Regional population movements and alliances resulted in the tribal fusion of several groups. Yet, the formation of a pan-Roviana political constituency did not eliminate political dissension within each composite group. Such differences afforded each aggregate polity the opportunity for schism and expansion through the reaffirmation of their own cultural identity and military prowess. This undertaking was nurtured by the outlined ideological rationale. Roviana head-hunting, then, ensued from a cyclical process, its intensification or decline synchronous with the ever rearranging political landscape and the social, historical and economic milieu of the times—a process that must have begun centuries or perhaps millennia before European incursion.


I am grateful to the people of Baraulu and Roviana and Vonavona Lagoons in general for allowing me to work with them for all these years. I also want to thank the National and Provincial Governments, and the Roviana Lagoon Area Council for their support. The Royal Society of New Zealand, the National Geographic Society and the University of Auckland, through the New Georgia Archaeological Survey project (NGAS) headed by Peter Sheppard (University of Auckland) and Richard Walter (University of Otago), funded this research. The National Science Foundation (SBR-9320498), Sea Grant (University of Hawaii R/MA1, and NA36RG0507), WWF-Pacific and ICLARM granted previous support. I wish to thank Ann Chowning, Christine Dureau and two anonymous reviewers for comments on earlier drafts of this paper and also Joan Lawrence for her fine illustrations.

- 67 Page of endnotes

- 68 Page of endnotes

- 69 Page of endnotes

- 70 Page of endnotes

1   Roviana political organisation in pre-European contact times is defined here as a chiefly system approximating a stratified society. The hereditary quality of the Roviana chieftainship expressed in the vernacular bangara (chief) and the existence of internal hierarchies within each social strata challenges the “big man” model (Sahlins 1963) in studies of Melanesian leadership (cf. Jackson 1978 for Roviana).
2   Data for this paper was collected during 27 months of fieldwork in the Roviana and Vonavona Lagoons. Fieldwork was also conducted in Rendova, Simbo, Kolobangara and Ranongga Islands. All work was carried out in the Roviana vernacular during 1992 and 1994-95 for doctoral research (Aswani 1997), and subsequently during 1998 and 1999 as part of the New Georgia Archaeological Survey (NGAS). The archaeological notes made in this paper result from six field seasons, headed by Peter Sheppard (University of Auckland) and Richard Walter (University of Otago), and conducted over the period 1996 to 1999 (Sheppard and Walter 1996, Sheppard et al. 1998).
3   By proximate causes I mean human intentionality as historical agency. Understanding the ultimate causes of warfare or even headhunting, in my view, requires examining the wider literature on human ecology and competition and conflict, e.g., the evolutionary ecology of competition and social complexity (see Aswani and Graves 1998, Boone 1992; cf. Hoskins 1996a:2-3). But this is beyond the scope of this paper.
4   Employing a conservative estimate of 25 years per generation still places this account in the 17th century, a period amply preceding any reported Western contact.
5   Numerous Munda informants claim that these two were different tribes remaining so until the late 19th century (Hall 1964). The majority of respondents in Kalikoqu and Saikile, however, maintain that Kazukuru and Roviana (via Taghosaghe) joined together before descending to the coast and again thereafter. While I concur with their account, it is probable that some Kazukuru speakers remained in the Munda area and retained some autonomy until their final absorption by Roviana lineages resettling the New Georgia mainland from Nusa Roviana Island in the early to mid 19th century.
6   Note that there are several versions of this tale (Hall 1964, Waterhouse 1931). Numerous discrepancies arise in indigenous ethnohistorical accounts as a result of different groups trying to establish their own social, economic and political hegemony. The goal here is to show general regional settlement and political processes rather than to identify particular historical events.
7   Miller (1980) cautioned researchers about literally interpreting Solomon Island oral narratives of inland groups moving to the coast because, he argues, they may be more symbolic than historical events. The New Georgia Archaeological Survey (NGAS), however, has mapped numerous inland settlements identified as ethnohistorically significant by informants, particularly the inland Bao shrine complex that is recognised as the cradle of the Roviana polity (see Sheppard et al. this volume).
8   Here the Vuragare polity is treated as an ancestor-based coastal group. But some Munda informants maintain that “Vuragare” are Kazukuru people who settled the sea-facing section of Nusa Roviana and should not be treated as a different ethnic group. It is possible, however, that the Vuragare tribe were earlier “Kazukuru” (or other inland) bush dwellers that settled the coastal fringe and intermarried with Parara, Kohinggo and Rendova groups to become a distinct tribal entity. Some Parara, Saikile and Kalikoqu elders support this idea(Aswani 1997,1999; cf. Schneider 1997:41). Even if “Vuragare” developed as a sub-polity after the Kazukuru-Roviana radiation, it is likely that the inland groups had interacted with the coastal Koloi, who supposedly spoke the Kozi dialect, and other coastal groups in Parara and Rendova Islands.
9   Some prominent informants maintain that there were some Taghosaghe settlements in Nusa Roviana at the time. Still others say that there were also Hoeze groups there.
10   It is reasonable to conclude that inland populations had all along interacted with coastal groups, but the oral history and the archaeological record suggest that this transformative period contrasts significantly with what preceded it (see Sheppard et al. this volume).
11   It is plausible that these stories have been syncretised with Christian narratives. Yet, the mateana stories recorded by Hocart (MSS.n.d.) early this century occurred shortly before the region-wide mass conversion to Christianity. Note that the Saikile chieftainship has a similar origin story that further distinguishes its descent line from those at Nusa Roviana.
12   Not all alikoqu elders endorse this story, rather they maintain that only the paele canoe house imicked the body of Tiola. Kalikoqu oral traditions suggest that the tomoko war anoe was built by the Koloi people who inhabited the barrier islands before the Kazukuru-Roviana coastal radiation and that it was not a technological innovation imported from Simbo.
13   Note that a victim's kin would replace the captured head by making a statue (beku) in the dead person's image (vina-tigono) in hope of propitiating the departed spirit. Magical plants associated with the victim, such as kuruvete and minila, were attached to the image to facilitate the transfer of their deceased kin's power. However, the powers vectored through the vina-tigono and associated magical plants were not as powerful as those canalised by the presence of the actual skull.
14   Saikile and Kalikoqu elders indicate that the skulls of commoners were eventually buried along walking tracks in order to be polluted as people walked over them. Also, this was done to prevent the revenge-seeking spirits from re-emerging. The skulls of enemy chiefs and brave warriors were kept as heirlooms until they were deposited at the base of a chiefly shrine upon a chief's death, or during other ritual offerings.
15   Nevertheless, some forms of magic could be directly transferred to the killer. For instance, the power of takarao, or a protective charm which guaranteed a person that if murdered his or her spirit (boso lau) would assail the killer and his kin, could be neutralised and transferred to the killer if, following a murder, the victim's rolled bark pendant (guzala) housing this power was taken.
16   Informants in Rendova, Ranongga and Kolobangara Islands indicate that, while their ancestors often engaged in internal feuding and in interisland predatory head-hunting, they did not raid as often as Roviana and Simbo did. The Roviana-Simbo military axis was presumably among the region's most powerful. Yet, alliances were extremely fluid because one's partners could be one's ally's enemies. Take, for instance, Lokuru's friendship with Simbo and simultaneous enmity with Roviana.
17   Revenge attacks could be averted if shell rings (bakiha) were paid to the offended party beforehand.
18   Roviana and Marovo were the predominant shell-valuable manufacturers in the region.
19   Lattas (1993:106), trying to counter the notion of their “slavement” and “alienation”, has argued that slaves were commonly incorporated into the New Georgia kinship system. While it is true that “slaves” often married into Roviana lines and were generally well treated, their lives were expendable and subject to sale or ultimate alienation by their clubbing or decapitation to fulfil sacrificial rituals—a situation of which adoptees had an “ever-present knowledge” (Coombe 1911:347). Furthermore, many of the “fed and reared” (Lattas 1993:107) captive children (veala) were spared from sacrifice because they had been ritually polluted (e.g., by menstruating women touching or placing them under a house). The incipient objectification and commoditisation process is revealed in the fact that powerful men kept a tally of captured children. Chiefs enumerated sacrificed veala by erecting a rock known as patu keresana for each victim at the chief's ancestral shrine (hope). Besides its cathartic purposes, chiefs gained prestige and ancestral efficacy from this activity (Aswani and Sheppard n.d.).
20   Note that this practice, while widespread throughout the Western Solomons and Isabel, was not as deeply entrenched as in Roviana. Lokuru (Rendova Island) informants suggest that children taken in their raids were seldomly sacrificed. This hints to a possible population deficit resulting from counter-raids against Lokuru and the concomitant need to replace casualties via the recruitment of human captives (see McDougall this volume on women captives).
21   There were two types of hiama in Roviana, the hiama va-sage-a batu and the hiama proper. The former were responsible for removing the skulls of the deceased and placing them in ancestral shrines. Often foreigners (mainly from Isabel), these men were assigned what were considered to be ritually hazardous tasks. The latter, were primarily from Roviana and were in charge of other ritual functions.
22   Waterhouse (1949:150) calls the liqomo-vovoso the Roviana “patron of war”. Some informants suggest that the viriviri kana was a desiccated body of an enemy who in life had shown great prowess in battle. The captured corpse was hung inside a paele canoe-house and used in pre-raiding rituals to secure the warrior's fighting power. Note that the liqomo and vovoso were also artefacts that vectored the power of these fighting spirits.
23   Trading expeditions followed a similar yearly cycle to raiding. Head-hunting and trading were undoubtedly closely intertwined. Head-hunting expeditions travelling to Isabel would often call at Kia to barter before raiding Bughotu in the south. Roviana traded, and at times raided, groups with which it had tenuous relationships, such as with Bilua in Vella Lavella (see Aswani and Sheppard n.d.). Macintyre (1983) has described a similar situation for the southern Massim.
24   Tetepare descendants living in Rendova also indicate other reasons for abandoning the island including rampant dysentery and raging interisland blood feuds.
25   Counting and ranking captured heads was also an important practice in Southeast Asia groups, such as the Huaulu of Indonesia (Valeri 1994).