Degei’s Descendants: Spirits, Place and People in Pre-Cession Fiji
2. Overview of Project
The first aim of this research is to determine from present-day oral accounts how Fijians, especially in western areas from 1951 to date, understand and explain, first, the origins, characteristics, development and interactions of the social and political divisions of late pre-Colonial Fijian society; and secondly, the general principles of traditional land tenure. The second aim is to assess the reasoning and accuracy of such understandings by taking into account anthropological, archaeological, historical and linguistic evidence based on my own investigations and on the records of others, especially the Native Lands Commission (NLC). The constant question with which my research was concerned was how and why oral accounts differed. Investigations of the understandings of these aspects of traditional Fijian society, including leadership, during the last 200 years or so of pre-Colonial Fijian society focused on the two areas of Rakiraki on the north-east and Vuda/Nadi/Nawaka on the west of the main island of Viti Levu, and on a third area comprising the western archipelago of the Yasawa group.
Scope of the project
The research explores:
- • secular and spiritual factors relating to the unity, identification, leadership and the dynamics of fusion and fission of polities, including their social and political relationships both internal or external to the polities;
- • what factors, internal or external, might have led to the development of polities of different degrees of complexity and stability at different times and in different parts of Fiji; and
- • certain aspects of Fijian cosmology, especially those features of the spirit world and the relationship between the realm of the supernatural and the realm of people which form the ideological basis for polities.
Areas of research
The three field areas on which this project focuses are:
- • the present district of Rakiraki in the province of Ra, on the north-east corner of Viti Levu;
- • the present adjacent districts of Vuda/Nadi/Nawaka in the province of Ba, on the west side of Viti Levu; and
- • the islands of the western archipelago known as Natu Yasawa or Yasawa Group, which comprise the present districts of Naviti and Yasawa.
These three areas were chosen because, first, they had attracted relatively little attention from prehistorians, though a modest amount of excavation work had been undertaken in each of them (see below); and secondly, field work based on the piedmont of the mountainous centre of Viti Levu and on the more easily accessible offshore islands was within the physical capability of my research assistant, my wife Tamaris, and myself.
Areas on Viti Levu
The two areas of Rakiraki and Vuda/Nadi/Nawaka have the following features:
- • Each is a piedmont area about 35 km by 5 km bounded by the sea on one side and by mountains on the other.
- • Each area is closely associated with sites connected with widely-known origin myths of the present inhabitants of Vuda and Rakiraki. These are sites in Vuda of the mythical first landing and the first settlements of the original ancestors who are said to have come by vessel from the west. A site overlooking Rakiraki is that of the mythical centre for a number of those who came on from Vuda and settled on the mountain range of Nakauvadra to the south of Rakiraki.
- • Each area has polities the social structures of which are generally different in complexity from those of the powerful and well-developed Eastern polities such as those based on Verata, Bau, Rewa and Cakaudrove (Appendix A).
- • Vuda/Nadi/Nawaka include a number of independent individual yavusa (descent groups) or small, independent politico-territorial federations, whereas in Rakiraki one descent group at present claims general paramountcy over a number of the polities. Some polities, however, including the Nakauvadra-hallowed federation of Navatu strongly maintain their independence.
- • Neither area has been the subject of detailed accounts by visitors, settlers or missionaries to the extent that the Eastern areas have been, although some Europeans had settled in both areas in pre-Colonial times.
- • Each area has one site excavated in depth by Professor E. Gifford (1951b) in the late 1940s. These are Korovatu in the Vuda area, and Navatu in the Rakiraki area. Navatu has more recently been re-excavated by Geoffrey Clark (Clark and Anderson 2001).
- • No detailed linguistic studies have been published in respect of either area, though a descriptive grammar of the Navatu communalect, in the Rakiraki area, has been completed (Parke n.d.- a).
The Yasawa Group
The islands of the Yasawa Group, including the western outlier of Viwa have the following features:
- • The Yasawa group is a chain of relatively small volcanic islands stretching in a north-north-easterly direction for a distance of 80 km from the south side of Waya to the north tip of Yasawa. This group, which has a total land area of 135 square km, lies between Bligh Waters to the east and the Ethel Reefs to the west. The greatest distance between islands is 5 km, between Waya and Naviti.
- • The longest island, Yasawa, is 17.5 km long; and the shortest of the inhabited islands is 5 km long. The greatest width in the chain is 5 km. Each island has a mountainous central ridge, with steeper slopes to leeward (west) and flatter land lying to windward (east). The highest peak (on Waya) is 568 m above sea level.
- • The sand cay of Viwa lies about 32 km north-west of Waya, and is about 80 ha in extent.
- • Each island including Viwa has a number of small, independent polities, comprised often of aggregations of small groups coming from Viti Levu as refugees or adventurers.
- • Perhaps reflecting this immigrant element of the population, the myths of origin are diffuse. Indeed, some islands such as Waya claim an autochthonous guardian spirit.
- • The islanders of Waya claim some early association with Samoans; and most of the islands except Viwa have had some association with the major eastern polities of Bau and Bua. The northern islands, especially Yasawa, were subject at one stage to Buan and Tongan influence. Vakawaletabua, Tui Bua, had family connections with Yasawa and his mother was Tongan. Some Europeans had also settled in various parts of the group in pre-Colonial times.
- • None of these islands has been described in any detail by early visitors or missionaries. However, a number of references to them had been made by early explorers and surveyors such as Wilkes in 1840, or by people concerned with politico-religious troubles involving Bua, Bau and Tonga, or by those concerned with the investigation of murders of Europeans and with retributive action.
- • Some excavation work has been undertaken by Terry Hunt on the island of Waya (personal communication).
- • Linguistic research has been undertaken on Yasawa by Raven-Hart (1953); and on Waya and Viwa by Andrew Pawley and Timoci Sayaba (1971 and the Wayan dictionary in press). Triffitt (2000) also recorded notes on Yasawa communalects.
Nature of data used
I now briefly describe the sources of data used in the project, and I will comment on the value of oral accounts and on the procedures used in assessing such accounts.
Sources of data
The various sources of data include:
- • oral accounts of the origins, structure, leadership and dynamics of polities from mythological times, generally based on the major descent group known nowadays as the yavusa. These accounts of 123 yavusa were written down by me and were based on personal interviews conducted in the early 1950s and during the 1990s;
- • written accounts into the circumstances under which land had been sold in pre-Colonial times, and who had approved such sales. These accounts were recorded by the Land Claims Commission (LCC) after Cession in the course of their enquiries into claims by Europeans to freehold title in respect of land allegedly sold to them by Fijian owners;
- • written accounts of the origins and histories of units of landowners and of boundaries of land, recorded by the NLC in the late 19th century and at the beginning of the 20th century, in the course of their enquiries into land ownership;
- • a few written accounts of polities recorded by leaders of these polities for posterity for their own people and for the Colonial Government’s education;
- • contemporary accounts of experiences in the research areas by European missionaries, visitors and settlers during the 19th century;
- • accounts of visits by me accompanied by representatives of the polities concerned, to over 200 archaeological and mythological/natural sites associated with the people and spirits of the polities in the research areas; together with some survey work, and limited excavation carried out by me in the Nadi area;
- • a basic 250-word glossary for eleven communalects recognised in the research areas, as recorded by me and taking into account the earliest form of the words as remembered; together with a supplementary 250-word glossary of one typical communalect each from the areas of Rakiraki and Nawaka. I have completed a descriptive grammar of the Navatu (Rakiraki) communalect (Parke n.d.- a), and George Milner and I are combining our resources to prepare a sketch of the Nawaka communalect.1 Andrew Pawley (Waya) and Raven-Hart (Yasawa) have recorded communalects for the north and south of the Yasawa group (for references, see above);
- • About 300 pages of texts of sere (songs) and meke (chants), including meke ni yaqona and meke ni tū yaqona, being chants used when preparing or serving yaqona in the course of ceremonies.
On the appropriateness of using oral accounts
One of the prime sources of data for this research is the corpus of oral accounts of the yavusa recorded in the three research areas. These accounts essentially record how current members of yavusa:
- • understand the origins, structure, development and interactions of their yavusa in the pre-Colonial period;
- • explain and evaluate their pre-Colonial inter- and intra-polity relationships; and
- • appreciate what influences (both internal and external) brought about such relationships at the yavusa and vanua levels and, so far as they apply, in a pan-Fiji context.
A decision on the appropriateness of using oral traditions as a primary source for my research first necessitates a general consideration of the pros and cons of the use of such accounts. On the one hand, current oral accounts, as recorded, may be regarded as being subject to memory loss about incidents experienced, or as being subject to faulty transmission from earlier times. Further, bias to promote one’s own good or the good of one’s polity is a constant factor when assessing the appropriateness of using oral accounts. Such accounts may, thus, be criticised as simply records of folklore which have no substantial backing and provide no basis for serious discussion, analysis or explanation of pre-Colonial Fijian society. In short, it may be claimed that ‘palaeoethnography’ or folklore should have no place in a prehistorian’s attempt to explore such a society.
Oral accounts may also be criticised as being chronologically inaccurate. No attempt has, however, been made to determine the absolute chronology and dating of events recorded in these accounts, although if the absolute date for an event can be determined from other sources, it is recorded. I record the relative chronology of what the narrators considered to be those events and features significant in the development of a yavusa and its relationship with other yavusa. Absolute dating is not, however, considered crucial for the purposes of this research.
On the other hand, this research suggests that these claims should be regarded as reflecting a narrow-minded, somewhat purist and old-fashioned point of view. My questions are rather ‘What is the significance of these oral traditions?’ and ‘For what explanations or interpretations can they be useful?’ A distinction is seen between myths of origin and oral histories. The former are regarded by narrators as their explanations of their spiritual and mythical origins. As such, they may be of interest to the ethnologist, but they have a much more significant aspect from the point of view of this work. They may be regarded as one of the symbols of unity and identification of a yavusa and one of the traditional grounds for bonding between yavusa.
The oral history of just a single yavusa in an area may be criticised, first as liable to bias towards the triumphs of that yavusa's development in relation to its neighbours; and secondly as atypical of yavusa generally in an area. It should, however, be emphasised first of all that the oral accounts record not only triumphs but also disasters; and secondly that the accounts are not simply those of a few selected samples of the yavusa. To avoid criticism of inadequate or atypical sampling, accounts of all the 123 yavusa in my three field areas were recorded as fully and carefully as the time available for the recording and the level of understanding of the narrators allowed.
An oral history may, indeed, be subject to faulty memory or to faulty transmission of handed-down accounts, or to deliberate alteration over a period of time, or to disagreement between the narrator and the rest of the yavusa. The version of the account recorded would be that given by the person who was regarded by the members of the yavusa as the most reliable ‘guardian of the lore’, of whom there was usually one such person. Such a person was usually, but not necessarily, the oldest or one of the most-high ranking members of the yavusa available, or the bete or priest, or the mata ni vanua or official spokesman and master of ceremony. Attempts were made to record in the presence of the head of the yavusa or nominee and as many of the members of the yavusa who wished to attend. The investigations involved, at this stage, simply the recording of how current members of a yavusa understood its origins and development, not an assessment of the reliability and accuracy of the accounts nor an external reconstruction of the history of pre-Cession society.
In recording such current accounts, it was sometimes clear that:
- • once evidence became available from other sources, the chosen narrator who was the oldest or the most high-ranking or the recognised ‘guardian of the lore’ may not necessarily have given the most reasonable account; or
- • two different oral accounts had been given by different contemporary narrators, perhaps reflecting bias, current internal conflicts or self-aggrandisement. In such a case, both the ‘authorised’ version and the other version would be recorded. In particular, it was appreciated that bias to promote one’s own group is a constant factor over time. Attempts are made to assess and evaluate which account was the most reasonable one. The oral accounts of individual yavusa should not however, be regarded in isolation nor as being of little interest save to the folklorist or to the members of those yavusa.
Questions about the reliability and reasonableness of the oral accounts of current understanding of what took place in pre-Colonial times and, ultimately, about the value and use of such accounts are crucial for this research. These are considered from the point of view of a ‘palaeo-ethnographer’, whilst paying due homage to the archaeologist, anthropologist, historian and linguist. As stated earlier, the constant question before me was how and why accounts from different narrators or evidence from other sources of information may have varied.
Evaluating the oral accounts
My project therefore aims, first, to evaluate the individual accounts recorded in the three research areas by setting them in a multi-disciplinary environment, and secondly, to place them within a pan-Fijian context.
In assessing the reasonableness and consistency of individual oral accounts of the current understandings of the origins, structure and dynamics of polities in pre-Colonial times, the research takes into account the totality of these accounts as well as other sources of evidence from archaeology, linguistics, and accounts written by early Commissions, visitors and early settlers, as follows:
- • individual yavusa accounts were compared and contrasted for consistency when more than one account referred to the same event;
- • written accounts by early European missionaries, visitors and settlers, records of enquiries by the LCC of the early 1880s, and by the NLC during the early part of the 20th century, archaeological and natural sites, linguistic evidence, and texts of meke (chants) were explored for any information relevant for comparative and integrative purposes;
- • these other sources of information were considered together with the oral accounts, in order to test to what extent they could be reasonably, reliably and significantly integrated and correlated;
- • the polities and associated sites, spirits and communalects in the three research areas were then considered in the general context of Fijian polities, sites, spirits and communalects, especially those in Eastern Viti Levu and Eastern Fiji.
Field surveys in the three areas and elsewhere indicate that there is an extensive amount of archaeological evidence available for the period covered by this project (Gifford 1951b, Frost 1979, Best 1984, Palmer 1969, Clark 2000). Even if archaeological evidence is not at present particularly intensive or extensive qualitatively or quantitatively, the potential is such that eventually archaeological research will provide a reliable and extensive source of evidence to be compared satisfactorily with that of the current oral accounts and linguistics. The evidence of archaeology, linguistics or early writers or Commissions alone is, however, inadequate for the aims of this project. A coordinated multi-disciplinary approach is the most satisfactory methodology to adopt in order to assess the significance of Fijians' current understandings and explanations of the origins, structure and dynamics of pre-Colonial Fijian polities in a pan-Fijian context.
Vanua 1 and vanua 2
In Chapter 1, the main polities or building blocks of traditional Fijian society were categorised broadly as descent groups termed yavusa, and socio-political constructs termed vanua or matanitū. Yavusa and vanua (federations of yavusa) were categories of polities recognised in many parts of Fiji, though perhaps less so in the west in pre-Colonial times. The matanitū or confederation of vanua was a category of the most complex form of polity which developed in the east and was more generally recognised at the time of Cession. After Cession, these terms were officially recognised as those for polities in traditional Fijian society for all parts of the Colony. At this stage, I should draw attention to the polysemous nature of the term vanua.
The term ‘vanua’ is used not only to refer to a certain category of Fijian polity of traditional Fijian society, with the meaning of a federation of yavusa. ‘Vanua’ was also used in pre-Colonial times, and is still used, as a complex and comprehensive term which formed, and still forms, the ideological basis of identity in traditional Fijian society. This second meaning of vanua had and still has social, spiritual and physical dimensions inextricably interrelated. The integration of these three dimensions within this meaning of vanua is the ideological basis of the nature and cosmology of pre-Colonial traditional Fijian society generally. Pawley and Sayaba in their dictionary (in press) distinguish in detail, as did Capell (1941) but more superficially, between vanua as ‘land, district, region, territory, country’ and vanua as ‘community, the people living in the community’. To distinguish between the semantics of the term vanua, I refer to the overall concept of Fijian society as vanua 1, and to the polity or federation of yavusa as vanua 2.2
Post-Cession Fijian society
The forms of post-Cession Fijian society and the systems of land tenure officially recognised by the British Government for the purposes of administration and legislation were intended, as explained earlier, to be ‘in accordance with native usages and customs’, following Colonial Office instructions to this effect. So, in setting the scene for this project, a brief account of those principles of traditional Fijian polities and land tenure which were adopted and the background to their adoption will be useful.
Post-Cession officially recognised traditional Fijian society is based on polities, as follows:
- • a series of registered descent groups known as yavusa, with sub-groups known as mataqali which in turn are divided into itokatoka (sometimes referred to simply as tokatoka);
- • a number of socio-political constructs known as vanua (that is vanua 2) or federations of yavusa, and matanitū or confederations of vanua; and
- • traditional leadership based on seniority of descent in the male line.
It was also officially determined that generally traditional land-tenure should be based on the precept that the land-owning unit is the mataqali.
Such a basis for Fijian society was adopted by the Colonial Government, following long-drawn-out investigations and often fiercely argued findings by the NLC, established under the Native Lands Ordinance 1880. After consultation with the Council of Chiefs and in the face of considerable debate, the British Colonial Government (Council Paper No. 94 of 1927) eventually endorsed these findings as authoritative and acceptable for official purposes of government. France (1969) has drawn attention to the controversial nature of the Commission's findings and their stormy passage over a period of thirty years, discussed in more detail later on.
Post-Cession society appears as a rather static form of society. It is, however, evident from the arguments that took place especially in meetings of the Council of Chiefs, that pre-Colonial traditional Fijian society in reality enjoyed a much more flexible and fluid socio-political way of life.
Post-Cession Fijian land tenure
Under the circumstances previously explained, Sir Arthur Gordon received instructions from the Secretary of State for the Colonies to devise a system of land administration ‘with a view to disturbing as little as possible existing tenures’ (Carnarvon to Gordon, 4 March 1875, Fiji Archives). Such instructions were in accordance with Gordon’s somewhat fraternal philosophy in dealing with the relationships between the rights of local inhabitants at a certain level of social advancement whilst taking into account the needs and some of the demands of expatriate settlers.
Gordon had arrived in June 1875, and the next year he asked the newly created Council of Chiefs to outline the traditional rights to land so that legislation could be framed to provide for a system of land registration which would embody such rights. This involved the recording of boundaries of recognised blocks of land and the recognised basis of the ownership of such blocks. Considerable argument followed about the terminology for polities. Further, some said that the land-owning unit was the itokatoka, others said the mataqali, others simply did not, purposely or otherwise, seem to understand the question. The record of the Council indicates that, at one stage, the Council came to agree that the mataqali was the landowning unit throughout Fiji. It is not, however, clear whether this surprising unanimity of opinion resulted from frustration or pressure from European ‘experts’; or whether, as was apparently the view of Ratu Sukuna, the Council record had been, perhaps deliberately, over-simplified by the recorder.
On the basis of this agreement, the Native Lands Ordinance was enacted in 1880, providing for the appointment of a Commissioner to investigate the system. Neither the councils set up in respect of each Fijian province nor the overriding Council of Chiefs nor the various successive Commissioners could however, agree on the basis of such a system.
Years of disagreement, changes of Commissioner, amendments of the Ordinance in 1892, opposition by Fijians and Europeans, and frustration were to follow. Eventually G.W. Maxwell, a Fiji Government administrative officer, was appointed Commissioner. He produced an analysis of the Fijian social system which served as the basis for all subsequent investigations of the NLC, and which was duly accepted by the Legislative Council (Council Paper No. 27 of 1914). The Secretary of State saw the necessity for the Commission to ‘secure some practical result at the earliest possible time’ and thought it best for land ownership on a mataqali basis to be settled, at any rate in the first instance. Maxwell objected that this was against the principles of traditional land administration, which were much more fluid, and pointed out that Fijians would not or could not mould their society into the form required by the Commission. Opposition mounted too among the Europeans who saw the NLC as a mechanism for frustrating the ambitions of planters to acquire land on terms of tenure which they understood and which were to their advantage. The climax came when the Europeans in Legislative Council criticised the NLC as a waste of time and public funds, ‘for which this colony gets no benefit, and never will’ (Fiji Legislative Council Debates, 1917:176). It was finally agreed in 1927, doubtless with the agreement and probably on the initiative of Ratu Sukuna, that for the sake of expediency the mataqali should be accepted as the land owning unit. The NLC would now cease to try to discover the varied traditional systems of land tenure in Fiji, and would confine its activities to recording mataqali boundaries.
Ratu Sukuna accepted the pragmatic approach of adopting a simple and homogeneous basis for land tenure even if such basis was not strictly in accordance with the principles of traditional land administration. He surely realised that the adoption of such a position was more important from the point of view of the traditional landowners than a fruitless, long-drawn-out continuation of hitherto unsuccessful attempts to determine the elusive and bewilderingly varied local principles of land tenure. Otherwise, ultra-conscientious Commissioners acting as amateur anthropologists would have continued to pursue the matter without practical solution, yea, to the very end of the rainbow. As a practical administrator with a great love for his own people, he accepted that it would be best for Fijians as traditional landowners if the question of land ownership was settled once and for all. Otherwise, it could well have been taken out of the hands of the NLC, and the administration of Fijian-owned land could have been undertaken directly by the Colonial Government. Indeed, following the investigations by O.H.K. Spate into the economic problems and prospects of the Fijian people (Spate 1959, Fiji Legislative Council Paper No.13 of 1959), the Council of Chiefs was forcefully opposed to the suggested land reforms. The Council was ‘of the unanimous opinion that the “Mataqali” should continue to be the landowning social unit… It is recommended that the present system of Fijian land tenure, ownership, administration, and reservation be rigidly maintained’ (Fiji Legislative Council Paper No. 29 of 1959:5).
So, somewhat paradoxically, the mataqali was accepted, and has been confirmed during the time of my investigations, as the social unit of ownership of Fijian land; even though the principles on which the NLC worked and continues to work may not wholly accord with the instructions of the Secretary of State for the Colonies.
1 (Editors’ note) Sadly George Milner (1918–2012) has now passed away as well and it appears that this linguistic work was never completed.
2 Vanua has cognates in other Austronesian languages, such as hanua (Rotuman), fanua (Samoan), honua (Tongan), benua (Malay) and whenua (Maori). Blust (1987) as well as Green and Pawley (1999) explore the etymology and polysemy of this term and its varieties. Single glosses given in bilingual dictionaries should not be regarded as accurate descriptions of their meanings but only as shorthand designed to fit the categories of the target language. I will use the simple expression ‘traditional society’ as referring to, but not accurately describing, the meaning of the complex term vanua 1 in Fijian. Some of the senses in various Austronesian languages do not involve all the dimensions and may be regarded as discrete. In the situation in Fiji, the two senses are doubtless related but the basic issue is whether they are discrete or not. I do not propose to pursue this argument here.