Previous Next

Degei’s Descendants: Spirits, Place and People in Pre-Cession Fiji

1. Fijian Society: The Islands of Fiji (General)

But westward look, the Land is Bright

Fiji, a general geographical setting

Fiji and its neighbours

The islands of Fiji, Rotuma, Samoa and Tonga define an oceanic geographical quadrilateral in which Fiji has the largest landmass and the greatest diversity of climate and topography. Among these four regions, traditional Fijian society shows much the greatest variety in patterns of language and socio-political structure.

What is Fiji?

When investigations leading to this project began tentatively in 1951, Fiji (Viti in Fijian) was a British Crown Colony, after Ratu Seru Cakobau, Vunivalu of Bau, and twelve other high chiefs had ceded the ‘Islands of Fiji’ to Queen Victoria under the Deed of Cession dated 10th October 1874. The Colony was extended in 1881 when the chiefs of Rotuma ceded ‘Rotuma and its dependencies’ to Queen Victoria under the Deed of Cession dated 13th May 1881 and the British Government combined Rotuma with Fiji for administrative convenience. It was extended further in the 1960s when a small unowned reef to the southeast of the main group was annexed.

Ratu Seru Apenisa Cakobau had been installed in the position of Vunivalu of Bau and paramount of the matanitū of Kubuna (Bau) in the first half of the 19th century. In 1867, an allegedly pan-Fiji government with Cakobau as the head was established with the backing of some politically ambitious Europeans in the east. Cakobau was made King of Bau; and his self-adopted title of Tui Viti or King of Fiji was recognised by those who found it suited their political purposes to do so when the 1871 Government was established with Cakobau as the head.

On 10th October 1970 this overall area became the independent Dominion of Fiji, with a Governor-General who represented The Queen. Fiji ceased to be a member of the British Commonwealth on 7 October 1987, and became the Republic of Fiji with a President. It was temporarily readmitted to the Commonwealth in 1997 and became the Republic of the Fiji Islands with a President.

The term ‘Fiji’ today embraces a wider area than that which was originally covered by the term in 1874. The territory and society explored in my project are, however, restricted to those of ‘the whole of the group of islands in the South Pacific Ocean known as the Fijis’ lying within the area delineated in paragraph 1 of the 1874 Deed of Cession (Derrick 1950:252) and comprising the original Colony of Fiji.

The islands of Fiji

The Crown Colony as defined in the Deed of Cession has a total area of about 400,000 square km, of which the land area is about 11,288 square km. The largest island, Viti Levu, comprises 6418 square km; and to the north-east the second largest island, Vanua Levu, comprises 3419 square km. These two islands make up more than 85 per cent of the total land mass, the next largest, Taveuni, being only 272 square km. About 100 of the more than 300 islands in the Fiji group are inhabited. Many of the smaller islands are unsuitable for permanent habitation but are visited by the Fijian owners for fishing or to cut copra. Some islets, such as Narokorokoyawa in the Mamanuca group are of particular importance to Fijians because of their association with the Fijian spirit world, and are only visited for communication with the spirits.

To the northwest of Viti Levu lies the Yasawa Group, comprising seven main islands (including the outlier of Viwa), and many small islets, with a total landmass of 83 square km. To the south-east of Vanua Levu lies the Lau Group, an archipelago of twenty-nine main islands, some of which are closer to Tonga than to Viti Levu.

The island of Viti Levu

Excluding the continental remnants of New Guinea, the North and South Islands of New Zealand, and the Grande Terre of New Caledonia, and after the island of Hawai'i, Viti Levu is the largest island in Oceania. It is roughly oval in shape, about 144 km long from east to west and 104 km wide from north to south.

Some geological characteristics of Viti Levu

The structure of the island is mainly volcanic in origin, with volcanic flows and agglomerations near vents. The steep-sided, crested ridges are natural locations for defended sites. Outcrops of limestone occur, and caves and crevices provide a number of archaeological as well as natural sites that have come to be associated with the spirit world.

The dividing range

The main dividing range runs roughly north to south down the centre of the island and includes a number of peaks over 900 m in height. The highest, 1315 m, is Tomaniivi or Mount Victoria which lies towards the north end of the range. From Nadarivatu, to the east of Tomaniivi, the main range stretches towards the northeast, developing into the Nakauvadra range, the highest peak of which is Uluda, 861 m high. The terminal peak of the east of the range is Supani, 536 m high.

The dividing range lies across the path of the prevailing south-east trade winds; and the consequent patterns of orographic precipitation result in a basic climatic division of Viti Levu, with a wet zone in the south-east and a dry zone in the north-west. The annual rainfall in the wet zone may be as high as 5000 mm a year, and in the dry zone as low as 1375mm.

To the windward, wet side of the range, the rain forest is generally heavy and dense. To the leeward, dry side, in places where any forest remains, it is generally lighter and more open. Most of the leeward side is covered by open land with grass, ferns and reeds, with a few clumps of stunted casuarina trees in the gullies and along the banks of streams. Much of the land has been burned off by people searching for wild yams; without cover, the soil has become leached and will support little except tough bracken. Such land is referred to as talasiga or sun-scorched.

The broken highlands

On both sides of the main dividing ranges are tracts of broken highlands, mostly hemmed in by ranges of hills. On the seaward side of these hills, precipitous escarpments or steep slopes fall down to the lowlands facing the coast.

Forming a north-west spur to the west of the main central range, the Conua range (referred to as the Tualeita or Spirit Path) dominates the west-north-western side of Viti Levu. Starting some 15 km inland, the range runs westwards from its highest point at Mount Evans (Koroiyanitu or Peak of the Spirits, 997 m above sea level) nearly to the sea at Edronu. Its northern side rises abruptly from the southern piedmont lands of Vuda, and its southern side from the Sabeto valley. The Sabeto valley is overshadowed to the east by the Nausori highlands of which the highest point is Koroba (Mount Pickering, 1069 m).

The coastal lowlands

The coastal lowlands around the interior system of highlands mainly comprise rolling hills with a strip of flat land along the coast. This piedmont is generally 4 to 8 km wide. When, as in the Navua lowlands, the flat land is wider, it may be broken in places

The rivers

Viti Levu has an extensive series of river systems associated with the mountain ranges and the rainfall, some flowing south, others north and west.

Of those flowing south, the Rewa and the Sigatoka River systems are the most extensive. The former flows southwards for some 144 km (of which 64 km are navigable), and enters the sea at the south-east corner of Viti Levu, passing through an elaborate delta system with mangrove swamps. The latter flows southwards from near Nadarivatu for about 120 km through pockets of flat land hemmed in by hills, to the sea at the south-western corner of Viti Levu. The Navua River flows south for about 64 km through narrow gorges to flatlands and finally a delta of mangrove swamps into the sea, in the centre of the south coast of Viti Levu.

Of those flowing north, the Ba River goes through about 64 km into a mangrove-covered delta in the centre of the north coast of Viti Levu. The Penang River passing through my north-eastern focal area of research (see below), has three main tributaries, the Nakauvadra (developing into the Wailevu), the Vatudamu and the Dranayavutia (developed from the Naqorokawa). These rise from the northern slopes of the Nakauvadra range and its extension towards Supani peak. The Penang River passes through a wide valley and enters the sea amid the mangrove swamps to the west of the Rakiraki village complex.

Of those flowing west, the Sabeto and Nadi rivers pass through my western project area. The former rises in the slopes to the south-west of Mount Evans, runs parallel with the south side of the Conua range along the broad Sabeto valley, and empties into Nadi Bay through a maze of mangrove swamps. The latter has three principal tributaries, the Malakua, the Nawaka and the Namosi, which rise on the slopes of the Nausori highlands and Koroba peak. It flows through undulating countryside and mangrove swamp (now largely reclaimed) into Nadi Bay.

The Yasawa Group

Most of the Yasawa group of islands, known as Natu Yasawa, form an almost straight line in a north-north-east direction for a distance of more than 80 km from a point 40 km west-north-west of the north-west corner of Viti Levu. The main islands in this chain are, from north to south, Yasawa, Nacula, Matacawalevu, Yaqeta, Naviti and Waya, being mainly volcanic in origin. The exception is the limestone islet of Sawa-i-lau at the south end of Yasawa, with its system of caves and well known for its petroglyphs.

These main islands are relatively high. The summits range from 568 m and 497 m on Waya, to about 227 m on Yasawa. The islands are generally long and narrow, ranging from Yasawa which is about 17.5 km long and up to 1.6 km wide, to Matacawalevu which is 4.8 km long and between 1.2 km to 2.4 km wide. Waya is about 6.4 km long and 4.8 km wide, with high ridges parallel to the east and west coasts, and a transverse ridge between them to the centre.

The main ridges of the long, narrow islands have steep slopes to the western faces, drained by short watercourses flowing into steep valleys. The eastern, gentler slopes have larger streams which flow into lagoons or mangrove swamps.

The peaks and high ridges are wooded, especially on the western slopes; but the slopes are generally grassy with few trees, perhaps because of shifting cultivation and fires.

Barrier reefs occur between the islands and the open sea to the west, and isolated patches of coral are plentiful near to the islands, but reefs fringing the coasts are scanty.

The Yasawa group includes Viwa, the most westerly of the islands of Fiji, which lies about 32 km west-north-west from Waya. An isolated cay of coral sand, some 80 ha in extent, its highest point is only about 7 m above sea level. No natural source of fresh water exists on the island, although there is a fresh water spring in the sea in the southern bay.

Post-Cession Fijian society, land tenure and administration

Sir Arthur Gordon (later Lord Stanmore), the first substantive Governor of Fiji, was a son of former Prime Minister Lord Aberdeen and a ‘protégé of Mr Gladstone and Lord Selborne, who could write confidentially, confidently, to Secretaries of State and permanent officials at the Colonial Office’ (Scarr 1980:10). He took up his appointment in 1875, and his instructions, largely drafted by Gordon himself as Scarr has suggested in conversation, were that the Fijians should be governed ‘in accordance with native usage and customs’ (Carnarvon to Gordon, 4 March 1875, Fiji Archives), and that a system of land administration should be devised ‘with a view to disturbing as little as possible existing tenures’.

Gordon was faced with an evident dilemma. He had, first, to follow these instructions. Secondly, he had to take into account the demands by an increasing number of European settlers and traders for safety and security of tenure. Thirdly, he was faced with the need for an efficient and economic form of administration which would at the same time take account of the sort of colony he felt it appropriate to establish.

So, while developing Fiji along the lines of the classic pattern of a British colony with a Governor answerable to the British Colonial Office, a Colonial Secretary with overall executive responsibilities, and other Colonial officials such as European magistrates, Gordon proceeded to establish and develop a system of local government with jurisdiction over Fijians. This system, the Fijian Administration, was intended from the start to be based on what were understood to be customary Fijian forms of government and land administration.

Invaluable sources of material for my studies of pre-Colonial Fijian society are the detailed accounts and reports of a long-drawn-out series of official investigations into these matters. Also valuable is a study of the systems of Fijian administration and the administration of Fijian-owned land, adopted and developed by Gordon and his successors. These systems were based not only on the findings of these accounts and reports, but also on the widespread, continual and often contradictory discussions which took place before and after the submission of such reports.

Fijian administration

The establishment of a system of Fijian administration to accord with the instructions by the British Government would have required a detailed knowledge of the generally accepted principles of traditional Fijian society. Gordon may not have had such detailed knowledge but he had advisers such as J.B.Thurston who had first come to Fiji in 1863. Thurston had acquired a wide first-hand knowledge of Fijian society; he had held cabinet rank in the 1871 supposedly pan-Fiji Cakobau Government of the Kingdom of Fiji (see below) and played a major part in events leading up to Cession. He was Acting Colonial Secretary of the Interim Government established immediately after Cession, and later held the substantive posts of Auditor General, Colonial Secretary, Administrator, Lieutenant Governor and finally Governor until his death in 1897 (Scarr 1984:83). He was aware of the reality of the diversity of traditional Fijian society and also, when land legislation was considered, of the principles of Fijian land tenure.

Gordon followed the guidance of Sir Henry Maine that a local community's usages and institutions should be retained in the face of conflicting demands by outside settlers until ‘new social wants have taught new practices’, and had his own preconceptions of the importance of initially adhering to ‘native usage and customs’. An aristocrat, he recognised aristocracy when he saw it in the course of his encounters, both ceremonial and en famille, with the Fijian high chiefs such as Cakobau. Thurston, after years of experience in Fiji, had similar views of social change. In spite of much discussion, disagreement and misunderstanding in the course of consultations with Fijians (as well as old-timers), Gordon generally adopted the views of Thurston, and proceeded to determine the form of Fijian Administration. Generally he took advantage of Thurston's extensive knowledge about the nature of traditional pre-Colonial Fijian society. He also took into account the system of administration of the 1871 Kingdom of Fiji which had been divided into geographical areas known as yasana or provinces ruled by governors. These yasana had been established to correspond with traditional matanitū or major socio-political confederations, where they occurred.

So it was decided that the major administrative unit of the Colonial system of Fijian Administration should be the yasana or province, based on the yasana of the Cakobau Government. A yasana was divided into a number of tikina or districts (a term apparently invented for the purpose and presumably associated with the word tiki-na, a part of) based on traditional vanua or minor socio-political federations. Each tikina included a number of officially recognised koro or villages (koro was the eastern pre-Colonial term for a village, whereas the western word was usually rara).

After it had been decided that the yasana should be equated with the Cakobau Government yasana and as far as possible with the traditional matanitū, and the tikina with the vanua, Gordon and the Colonial government had to decide on titles for the official heads of provinces and districts (yasana and tikina). A problem arose because of the diversity of customary titles of heads of traditional matanitū and vanua, and the need for uniformity of official titles. The title of Roko, chosen as that of the administrative head of a province, originated from the traditional title of the spiritual paramount chief of certain major polities or matanitū, such as Roko Tui Bau of Bau, and Roko Tui Dreketi of Rewa. In the same way, the title of Buli which was selected as the official title for the head of a tikina, originated from the traditional title of the head of certain federations or vanua such as Buli Nadi of the vanua of Nadi in south western Vanua Levu. Such decisions at first caused a certain amount of confusion among those who had no understanding of the significance of these new titles, and especially among those in the west whose chiefly title was generally Momo.

The chiefs had ceded Fiji not to the British Government but to Queen Victoria, who thereby became recognised as the paramount chief of Fiji. Gordon considered that as Governor and The Queen's representative he had assumed, by acting on behalf of The Queen, the duties and responsibilities of the position of paramount chief of Fiji. So he personally was the only one appropriate to appoint heads of yasana. The person to be appointed was almost invariably the senior member of the principal polity of the yasana. Then, in the course of what Sir Arthur regarded to be an appropriate form of ceremony, he would hand the newly installed Roko a staff of office and charge him to look after his yasana. This procedure was a departure from ‘native usages and customs’, but was apparently accepted by the rather bemused or perhaps amused chiefs and people.

At the beginning the greater part of the Colony was divided into 12 yasana, and the Roko was the Governor's deputy in his own yasana. These yasana were in turn subdivided into a total of 86 tikina, each with a Buli responsible to the Roko. According to the 1881 census these tikina included about 1400 villages, each one of which had an officially recognised Turaga ni Koro or Village Headman answerable to the Buli. Most of the 115,000 Fijians recorded at the time of the census lived in a village with which they had traditional connections. Each administrative unit of the Fijian Administration therefore had an appointed Fijian official administratively responsible for the unit, and there was a chain of responsibility from the lowest official to the highest and eventually (later, through the District Commissioner as Deputy Secretary for Fijian Affairs) to the Governor.

Schematically, the Colonial system of Fijian Administration may be illustrated as follows:









Immediately following Cession, parts of the mountainous interior of Viti Levu were still politically unstable because local independent chiefs had not been formally consulted and did not recognise Cession. This mountainous interior, known as Na Colo, was placed directly under officials of the Central Administration until the Government considered that it had become sufficiently stable for it to come under the Fijian Administration. Part of this stabilisation process involved what is referred to as Gordon's Little War in 1876, in which coastal Fijians under the general direction of the Governor attacked and overcame the independent dissidents in the hills, many of whom happened to be their traditional enemies. Sir Arthur wanted to gain the confidence of the hill folk in the newly established Colonial Government. As a first step, he was determined to cause as few casualties as possible and therefore he kept fire-eaters like Colonel Pratt, a Royal Engineer, out of the campaign. The latter had for public works purposes been appointed the British Military Commander but he was also keen to demonstrate his military prowess in the field.

The final pacification of Na Colo resulted in, first, the alteration of the boundaries of some already established yasana and, secondly, the creation of two new yasana based on traditional relationships and administrative convenience, each under a Roko. Since then the number of provinces has remained at fourteen.

Some individual officials such as Thurston were aware of the diversity of pre-Colonial Fijian society and land tenure. Only later, however, did the Government generally accept that the principles of socio-politics and land tenure were not so uniform as was originally thought to be the case and that there was greater diversity than was at first realised in these aspects of ‘native usages and customs’.

As time went by, much more information (some correct, some misleading, some dubious) about traditional Fijian society became available to the Colonial administration as the result of the investigations and findings of Lands Claims Commissions appointed under the Lands Claims Ordinance of 1879 and Native Land Commissioners appointed under the Native Lands Ordinance of 1880. There was perhaps a growing realisation that local sensitivities and rivalries had not been given the recognition that was necessary if the new administration was to receive general acceptability. As a result, principally of having to meet the demands of local sentiment and local rivalries, and partly of having to extend the tikina system to the now pacified Na Colo, there had been a substantial increase in the number of tikina from 86 to 184 by 1945. Of these, there were no fewer than 109 in Viti Levu and even two in the associated offshore island of Beqa (part of the Viti Levu province of Rewa).

In course of time, it came to be realised that this increasingly widespread proliferation of Fijian Administration administrative units and officials, and the need to take into account local sensitivities and rivalry had to be balanced against what were seen to be the benefits of improved socio-political and economic conditions to be gained from a tighter and more efficient form of Fijian Administration. The Government accordingly reviewed the Fijian Administration. As a result, the Fijian Affairs Ordinance was enacted in 1945, whereby, inter alia, agreement was given, perhaps grudgingly, that the number of tikina should decrease. It is nowadays customary to refer to those original tikina as ‘Old Tikina’, in contrast to those recognised after the review, which were referred to as ‘New Tikina’.

When I began my investigations of the development of Fijian society in late prehistoric and proto-historic times, I studied the origins and development of Old tikina and yasana of the Colonial Fijian administration. I assumed that a study of the identity of tikina, being ideally based on vanua or pre-Colonial socio-political federations of yavusa or descent groups, would give clues to the final stages of the development of pre-Colonial Fijian society. The variations in the number of tikina and of their boundaries should provide evidence for the principles basic to the structure and development of such Fijian society and for the diversity within such principles acceptable to Fijian society. At a higher level, the same assumptions were applied to a study of Colonial yasana based immediately on those Cakobau Government yasana which in turn were generally based on matanitū or pre-Colonial major socio-political confederations. Similar studies to which I refer below have concentrated on Eastern Fiji (Na Tu i Cake), and so I focused my investigations on the following three western areas with which I had been officially familiar since 1951.

The three focal areas (see Maps I to VIII)

The north-east focal area is about 32 km by 8 km on the coastal piedmont to the north of the Nakauvadra range and the east of the Supani Peak, together with several off-shore islands including Malake, Nanuya i Ra and Nanuya i Cake. It is based on the present tikina of Rakiraki (in the yasana of Ra), which comprised before 1945 the Old tikina of Navolau, Rakiraki and Raviravi. Three peripheral villages in the project area which after 1945 were included in the present tikina of Saivou, had previously been included in the Old tikina of Naroko. Naroko now forms part of the present tikina of Saivou.

The western area is about 32 km by 10 km on the coastal lowlands to the west of the Nausori highlands, and is divided into two by the Conua range from Koroiyanitu peak to Edronu near the coast. It includes the island of Waya Sewa at the south end of the Yasawa group, because this island is socio-politically and administratively part of Vuda tikina on the mainland. It is part of the present yasana of Ba, and is based on the three present tikina of Nawaka, Nadi and part of Vuda. Before 1945 Nawaka comprised the Old tikina of Rukuruku, Nawaka and Vaturu; Nadi comprised the Old tikina of Sikituru and of Nadi which by then had absorbed a previous tikina known as Buduka; and Vuda included Sabeto, Vuda and Vitogo (of which I concerned myself mainly with the western part).

The western insular area is the main Yasawa group, also in the present yasana of Ba, which is based on the two present tikina of Naviti and Yasawa. Before 1945 Naviti comprised the two Old tikina of Waya (which included the islands of Waya and Viwa), and Naviti (based on the island of Naviti). Yasawa comprised the two Old tikina of Nacula which included the islands of Yaq(w)eta and Matacawalevu, and Yasawa (based on the island of Yasawa).

All three areas are included in that western part of Fiji known as the Yasayasa vakaRa.

Fijian land tenure

So far in this discussion, the term ‘vanua’ has referred to a minor socio-political federation. Vanua is also a term for the social concept of traditional Fijian society in which the elements of people, spirits and places run through and unite the society. Because of this inextricable association between people, their associated spirits and their lands, the governance of the Fijians ‘in accordance with native usages and customs’ required not only the provision of a system of administration of the people in accordance with such instructions. It also needed an understanding of the traditional system for the administration of land held by Fijians under customary tenure, and the recognition of such a system in Colonial native land legislation.

Customary land tenure

Investigations relevant to native land tenure were carried out by the Lands Claims Commission (LCC) and the Native Lands Commission (NLC) (see references under Native Lands Commission). Records of discussions between traditional landowners and the LCC, and of investigations by successive NLCs, as well as reports of discussions held by the Council of Chiefs formed the basis of the official Colonial understanding of the building blocks of traditional Fijian society. These records and reports included evidence, the critical study of which provided a useful basis for previous studies, especially those by Peter France (1969). They also formed the basis for my own investigations of the structure and development of pre-Colonial Fijian society and the principles of native land tenure.

The Deed of Cession made provision for categories of land not to be regarded as native land. These were land ‘deemed necessary for public purposes,’ (Crown Land Schedule A); land vacant at the time of Cession, (Crown Land B); and ‘lands shown to the satisfaction of the LCC, (see below) to be alienated so as to have become bona fide the property of Europeans or other foreigners’. Before areas of native land could be officially recognised and the officially accepted principles of native land tenure could be determined and secured by appropriate legislation, it had first to be decided which lands were ‘shown to be alienated’.

Under the provisions of the Deed of Cession, some 12,500 ha were set aside as Crown Land ‘deemed necessary for public purposes’. After investigations, the LCC set up under the provisions of the Lands Claims Ordinance of 1879 were satisfied that of 1335 claims made by ‘Europeans and other foreigners’, 517 claims to freehold title over more than 166,500 ha should be upheld and approved by the Governor in Council. The remaining 83 per cent of the land (about 1,625,000 ha) was regarded as held under ‘native customary tenure’, and was designated as native land (Roth 1953:88; Burns 1960:105–107).

The NLC, established under the Native Lands Ordinance 1882, investigated the principles of traditional Fijian land ownership and of Fijian society. Earlier efforts on the part of Governor Gordon to understand such matters were hindered partly by his inability to speak Fijian and his dependence on interpreters, and partly by the diverse views and misunderstandings of the Fijians whose traditional rights were being investigated. Nevertheless, the Council of Chiefs, after a most confused discussion in 1878, was reported by David Wilkinson, the interpreter, as having agreed that land ownership was communally based on a descent group generally known as a mataqali. Gordon accepted this report and proceeded with the enactment of the Native Lands Ordinance of 1882. The preamble to the Ordinance said that careful inquiry had revealed that such lands were held mostly by mataqali or family communities as the proprietary unit according to native custom. Section 2 of the Ordinance provided that ‘The tenure of lands belonging to the native Fijians as derived from their ancestors and evidenced by tradition and use shall be the legal tenure thereof’.

Gordon's views prevailed until Sir Everard im Thurn was appointed Governor in 1904. With the evidence of several investigations and Commissioners before him but with no previous knowledge of Fiji or previous experience as a Governor, im Thurn argued that registration of Fijian land and land owners by mataqali was not based on the principles of traditional ownership. He believed that the Fijians, a dying race, should not monopolise most of the land, and that his duties lay rather to the white settlers. In 1905, he made native land alienable and, by 1908, over 40,470 ha of fertile Fijian land had been sold. After considerable debate in the House of Lords, initiated by Stanmore in 1907, a decision by the House was relayed by order from the Colonial Office, and this practice was stopped by the incoming Governor Sir Henry May who initiated further investigations by the NLC into the increasingly uncertain nature of the principles of customary social units and traditional land tenure (see France 1969 for details).

A spate of activity followed on the part of the NLC, and in 1912 Commissioner Maxwell tabled before the Legislative Council a report on Fijian social structure. He gave distinctive names to the various social units, and set out definitively their relationships to each other. The findings became the generally accepted model of Fijian society, and were set out in Council Paper No. 27 of 1914. These findings, doubtless influenced in the first place by Ratu Sukuna (an influential authority on such matters in spite of his relative youth), were later endorsed by the NLC under the Chairmanship of Ratu Lala Sukuna in Council Paper No. 94 of 19271 (see references under NLC).

Fijian society: official colonial model

The following diagram represents schematically the official Colonial model of the polities of traditional Fijian society, based on that set out by Maxwell in Council Paper No. 27 of 1914, and endorsed in Council Paper No. 94 of 1927.

(a) Socio-political Constructs





(b) Descent Groups


Major descent group


Intermediate descent group


Minor descent group

This diagram indicates the socio-political constructs and descent groups that comprise the two main categories of polities forming the building blocks of Fijian society.

The polities in the model

Discussion of official accounts of the principles on which the model was based will be clearer if I first consider the descent groups (the yavusa, the mataqali and the itokatoka), and then the socio-political constructs (the matanitū and the vanua). The model did not include reference to the vūvale (household or family unit), one or more of which may have comprised an itokatoka.2

The descent groups

Maxwell (see above) explained that: ‘A yavusa consists of the direct agnate descendants of a single kalou vu or ancestor god, whose sons became the founders of component mataqali. As the population increased, the sons of the founders of the mataqali founded the various tokatoka which constituted the lowest order of subdivision exercising rights to land’.

Some current myths of origin recorded by me in Ra and in Vuda related that sometimes the original ancestor of a yavusa was a spirit who came directly or indirectly from the Nakauvadra Range (see Gifford 1951a:167), the home of Degei, or were descended from such a spirit. In other myths, the original ancestor was descended, directly or indirectly, from other mythical heroes who remained at Vuda having arrived there in the first canoe, the Kaunitoni.3

Each yavusa had a name, usually derived from the name of some natural feature near to the yavutū or of the yavutū itself. This term may be preceded by a word for ‘people of’ or ‘inhabitants of’, such as Kai, Noi or Lewe i. Thus a yavusa whose yavutū was known as Nasaumatua became known as the Kai Nasaumatua or Noi Nasaumatua or Lewe i Nasaumatua (Council Paper No. 94 of 1927). The head of a yavusa is recognised in the model as the senior member of the senior mataqali, and may or may not have a traditional title. For instance, the head of the Tububere, the second senior yavusa in the vanua of Vuda, has the title of iTaukei Sawaieke, meaning ‘the owner of the chiefly yavu called Sawaieke’. Among the Rakiraki heads of yavusa, the head of Navuavua was the Tu Navua, and of Natiliva was Lei Natiliva. Because of this diversity of traditional titles, the Colonial Government created the official title of Turaga ni Qali for a person holding the position of senior member of a yavusa. This title is non-traditional but presumably was derived from the word qali, a polity subject to the authority of another polity.

Under the NLC model, as the descendants of an original ancestor increased in numbers the resulting yavusa evolved into a number of mataqali, membership of which was based on common agnatic descent from one of the sons of the original ancestor.

Each mataqali had a distinctive name, and was ranked hierarchically within the yavusa, the senior one comprising the descendants of the eldest son of the original ancestor. Every mataqali had specific hereditary functions within the overall structure of the yavusa, and was categorised accordingly. In Council Paper No. 94 of 1927, Ratu Sukuna, leading expert on traditional Fijian society and Chairman of the NLC, explained that the functions of the various mataqali in a model yavusa were in order of seniority as follows:

(i) Turaga; including the person who was acknowledged as the traditional leader of the yavusa, as well as others of chiefly status at a yavusa-wide level;

(ii) Sauturaga, being the immediate henchmen and executive officers of the chief;

(iii) Matanivanua, being the chiefs’ heralds and masters of ceremonies;

(iv) Bete, being the priests into whose bodies the spirit of the original ancestor was supposed to enter from time to time and to issue advice to the chief and the yavusa;

(v) Bati, being the warriors.

This model was duly accepted by the Colonial Government and the British Government. Further the mataqali, again on the advice of Ratu Sukuna, was accepted as being the main communal land-owning unit suitable for the purposes of the official registration of recognised native land and of the names of landowners. The Colonial Government also accepted that the person holding the position of senior member of a mataqali should have the official title of Turaga ni Mataqali. The territory of a yavusa may then be regarded as the totality of the land owned by the component mataqali.

As the descendants of the sons of the original ancestor of a yavusa increased in numbers, the resulting mataqali evolved according to the model into a series of itokatoka based on common agnatic descent from one of the sons of the sons of the original ancestor. Although the official term itokatoka4 was recognised as traditional in some eastern parts of Fiji, the equivalent term was different in other parts.5

This tripartite division of categories of descent groups represents the Colonial model of Fijian society as a hierarchy based on recursive birth-order. The following diagram illustrates schematically the official Colonial model of those polities of Fijian society claimed as descent groups:








No recognised title

The model accords with the underlying principle of unity running through the formal structure of Fijian society, at any rate at the level of the descent groups. It also determines the ranking not only of groups and sub-groups but also of individuals. The titles show that the question of ‘who is the turaga or head or chief?’ has no absolute answer. A person may be head of his vūvale, or household, which is included in an itokatoka of which he is not necessarily the head. Similarly a person may be the head not only of his vūvale but also of his itokatoka which is included in a mataqali of which he is not necessarily the head. Finally a person may be head of his vūvale, itokatoka, and mataqali, but not necessarily of his yavusa. In each grouping of which he is head, he could be regarded as having the status of turaga or chief. Further, all members of the senior mataqali are regarded by virtue of such membership as having the status of Turaga or Marama (if female) and may take a traditional honorific pre-nomen such as Ratu (m), Adi (f), Ro (m or f), Bulou (f) or Lo(f).

The person regarded as the head of the yavusa is turaga over all members of the groupings subsumed within the yavusa. As such he has traditionally recognised socio-political responsibilities and privileges in relation to the yavusa as well as to land associated with the yavusa. Such responsibilities and privileges were recognised by the NLC, the person was accorded the official title of Turaga ni Qali and was allocated a small percentage of any rent money received from lessees of land associated with the yavusa.

Descent groups: the reality

Successive NLCs continued their investigations for many years. They recorded detailed traditional accounts of the origins and development of descent groups as explained under oath and sometimes unwillingly by Fijian landowners; and determined at least to their own satisfaction the identity of social units holding land at the times of the investigations. The boundaries of the lands held by each unit were surveyed and plotted, and the NLC registered the names of the units, their relative seniority and their associated land. These units, their relative seniority and their associated land then became the official basis of Colonial Fijian society, land ownership, social ranking and leadership.

A woman was registered as a member of the descent group of her father (or mother, if illegitimate), and remained a member after marriage. Women often married for political reasons out of their descent group in order to create alliances or seal relationships between polities, especially vanua or matanitū. Much of the power of pre-Colonial federations resulted from such marriages, and chiefs practised a system of polygamy developed for such socio-political reasons. However, no instance is known to me of polyandry in the case of a female chief. Marriages were generally virilocal.

Descent was customarily patrilineal, except in certain parts of Vanua Levu where it was matrilineal, and this pre-Colonial principle was adopted by the Colonial Government for the purposes of registration. There was, however, a custom, referred to by Roth (1953:72), known as ilakovi (the nominalisation of a form of lakova, meaning ‘to go to’), whereby a person could be transferred from the father's to the mother's descent grouping or vice versa, but to no other grouping. Such a transfer severed all interest in the land of the previous grouping.

The rigid situation embodied in legislation has resulted in departures from accepted practice in pre-Colonial times. In reality, a particular yavusa may have split over leadership problems or some insult or disagreement; and some of the component mataqali may have remained loyal to their parent yavusa, while others transferred their loyalty to another yavusa.

By establishing a standard form of Fijian administration and land tenure throughout the Colony, the Colonial Government did not provide in the legislation for certain customary practices prevailing in pre-Colonial times. For instance, pre-Cession Fijian society accepted the practice of including in a yavusa a mataqali which, typically following a dispute, wished to transfer its collective allegiance from its own yavusa of origin to another yavusa. Further, before the system imposed by the Administration, the practice prevailed of establishing a new and separate yavusa comprising one or more mataqali which might have decided to leave their yavusa of origin, following some internal split, perhaps over problems of leadership or as the result of some insult or breach of protocol.

The Colonial Government deemed that succession to leadership in traditional Fijian society should be based on the principle of veitarataravi vakaveitacini (seniority of descent), and this may have been the ideal in pre-Colonial times. In practice, if there was a disagreement about the leadership in pre-Colonial times, the matter was decided by formal discussion or, if unresolvable, by the war club. Dominant status and leadership mainly depended on achievement rather than on birth, reflecting individual ambition, internal rivalry and success in war. The possibility of usurpation was ever present, especially because of rivalry between members of the senior group. In the case of Bau, the Vunivalu or war chief usurped the position of the Roko Tui Bau, or spiritual chief, having attacked and burned the chiefly village, killed many of the inhabitants and driven away the paramount chief, the Roko Tui Bau. The incessant warring in the Nadi area resulted in the constant reshuffling of the loyalty of groups who left the district, sought refuge among more powerful neighbours or submitted to superior conquerors. ‘Stranger kings’ (Sahlins 1981) from other polities may have achieved chiefly status in a polity in the course of war either through conquest or by way of gratitude for assistance in war, or simply in recognition of the stranger's particular qualities of strength or leadership. In one case, according to tradition, when the people could not agree on who should be successor to the title of Kwa Levu of Nadroga, a shipwrecked Tongan chanced to be washed up on the beach and was found by women who were fishing. Wakanimolikula was installed as Kwa Levu, so the story goes, on the basis of his good looks and personality.

Such divergent pre-Colonial practices could not prevail under the system of Fijian administration, leadership and land legislation established under the Colonial Government. Such a system had to satisfy the instructions of the British Government that the Fijians should be governed in accordance with ‘native usages and customs’ and the claims of the Fijians. It also had to meet the needs of the increasingly vociferous members of other races who had settled in Fiji. They had been critical, though barely heeded, about what they considered to be, first, official disregard for their own safety and, secondly, the heavily Fijian-favoured land tenure situation, not only before Cession but increasingly so since then.

The general principles of the resulting officially-recognised systems of Fijian society, administration, land tenure and communal ownership may have been in general accord with the ideals of ‘native usages and customs’. These officially recognised systems based on the need for unification and simplification, could not, however, take fully into account the fact that traditional systems were, in practice, subject to widespread and significant diversity.

The model: the socio-political constructs

The vanua

The NLC determined that frequently throughout Fiji several yavusa had combined temporarily or more permanently to comprise a vanua or federation. For instance, in times of war, a number of yavusa might form a military alliance for mutual protection and military assistance, not only in the face of threats of assault by neighbours but also when Western Fiji was threatened by the Vunivalu of Bau, Ratu Seru Cakobau, or when Fiji was threatened by expansionist Tongans. Yavusa involved would then agree among themselves to vakarorogo or heed the paramount authority of the head of the strongest of the participating yavusa and accept his leadership. Such a resulting vanua might continue until the easing of the situation which had brought about the federation. Participating yavusa might then revert to an independent status (tu vakai koya). Alternatively, the participating yavusa might see some mutual advantage in forming a more permanent form of federation in times of peace (sautu) for the purpose of ensuring continuing access to goods and assistance or for the gaining of prestige through association with a yavusa with a forceful and powerful leader. An able and ambitious head of yavusa might seek to demonstrate his power and forward his ambitions by extending his sphere of influence through some form of alliance or through conquest.

Some yavusa combined to form a vanua on the basis of kinship relationships, either mythical, genealogical or matrimonial. Other factors leading to the formation of vanua included geographical convenience based on territorial proximity, or a mutual need for access to resources both natural and human. The NLC claimed to have found that such federations were of general occurrence and a feature of pre-Colonial Fijian society throughout Fiji. The territory of a vanua may be regarded as the totality of the territory of the various component yavusa. The membership of a vanua may be regarded as the totality of the membership of the component yavusa.

The Government decided that such vanua were appropriate forms of socio-political constructs in pre-Colonial Fijian society to serve as the basis for tikina or subdivisions of provinces in the Colonial Fijian Administration.

The matanitū

The NLC also found that in some parts of Fiji, several vanua might sometimes combine voluntarily from time to time under a powerful chief for purposes of mutual convenience or for protection in times of war. Vanua might also have been united forcefully, when an ambitious, able and powerful chief wanted to extend his area of authority and also to have access to natural resources or to military assistance. Such combinations of vanua might develop by consolidating to form relatively stable matanitū or confederations under a paramount chief, as in such cases as Verata (under the Ratu mai Verata), Bau (under the Vunivalu) and Rewa (the Roko Tui Dreketi). The traditional title varied but was recognised by the Government and retained for ceremonial purposes. Since the Government did not consider that this category of chief had traditional privileges over land sufficiently strong to justify a share of rent money, there was no need to create an official title. The NLC used the term matanitū to refer to such confederations of vanua. The Government accepted that such matanitū were recognised elements in pre-Colonial Fijian society and had formed the basis of yasana in the 1871 Cakobau Government for the Kingdom of Fiji. Accordingly it was agreed that matanitū should become the basis of what were termed yasana or provinces in the Colonial system of Fijian administration.

The socio-political constructs and administrative areas: general

The following diagram shows the equation of Fijian socio-political constructs and officially recognised titles of heads, with Fijian Administration areas and official titles of administrators in charge:






no official title







The Vanua and the Tikina

Vanua, being federations of yavusa, were recognised by the NLC as a widespread form of polity in pre-Colonial Fijian society. Names and boundaries of the tikina of the Colonial Fijian administration were ideally equated with those of the pre-Colonial vanua. These were duly prescribed under Colonial legislation and so given legal status for the purposes of the Fijian Administration. As far as land legislation was concerned, the traditional privileges of the paramount of the vanua were recognised. For instance, when land owned by a mataqali of a yavusa forming part of a vanua was leased, a specific percentage of the rent money was assigned to the paramount chief in recognition of his or her traditional privileges and responsibilities as paramount. In my time, I have known female paramounts in Vuda, Nadroga, Ba and Rewa. The occupant of such a paramount position would have continued after Cession to hold the traditional title associated with such a position. Such a title varied considerably over different parts of Fiji. For instance, the head of the vanua of Nawaka was the Momo i Nawaka; and the head of Natokea, Rakiraki, was the Ra'u (two long vowels) ni Na'okea.

The paramount of a vanua also holds the official title of Turaga i Taukei as selected by the Colonial Government, in the context of land legislation. The origin of this Government title is obscure, though it appears in the west of Fiji. For instance, the iTaukei Nakelo was head of part of Sabutoyatoya, the leading yavusa in Vuda. Nakelo was the name of a chiefly yavu. The holder of this title of Momo (Turaga) iTaukei, or owner of, Nakelo, had important ceremonial and advisory/executive duties in respect of or on behalf of, the Momo Levu i Vuda, or paramount chief of Vuda. The head of the Naua vanua (composed of a single independent yavusa) was before formal installation titled iTaukei Sawaieke or Owner of Sawaieke, the chiefly yavu of Naua. After installation he was given the title of Momo i Naua or paramount of Naua. Sawaieke is also the name of the chiefly yavu of of the yavusa of Tububere, Vuda, and the title of the head of the yavusa is iTaukei Sawaieke. In neither case could I find any agreed explanation of the origin of the name. Sawaieke does, however, also occur as the name of an area in the Lomaiviti island of Gau, and may reflect connections by marriage or that Gau was a place prisoners were sent by Cakobau in the aftermath of wars involving Naua and Vuda referred to later.

The matanitū and the province

In accordance with the model, traditional matanitū where they occurred formed the basis of Colonial yasana. For instance, at one time probably the most powerful matanitū in eastern Viti Levu was Verata, whose paramount held the title of Ratu mai Verata (see Appendix A). However, in 1829 Verata's powers had been finally eclipsed by the matanitū of Bau of which the leading yavusa was Kubuna. The paramount chief was the Vunivalu or war chief who had earlier deposed the spiritual paramount, the Roko Tui Bau. Another powerful matanitū was that of neighbouring and rival Rewa based on the yavusa of Burebasaga of which the spiritual paramount was the Roko Tui Dreketi and the second highest chief was the Vunivalu or war chief. A third major matanitū, that of Cakaudrove (commonly referred to nowadays as the Tovata), was led by the AiSokula polity of which the paramount was the Tui Cakau. These duly became the yasana of Tailevu, Rewa and Cakaudrove respectively.

It is interesting to explore how the Colonial Government established provinces in those areas where there was no recognised matanitū. For instance, there were vanua in Nadi, Nawaka, Vuda, the Yasawa Group and Rakiraki, but no recognised confederation of a higher order, in spite of kinship and marital ties between leading chiefly families. However, the provinces of Ba and Ra were created on the basis of such ties as the Government could identify from advice and investigations but also on contiguity of large geographical areas and administrative convenience. On the island of Kadavu, fifteen vanua were recorded, though only six were recognised in a pan-Kadavu ceremonial context. There has been no confederation of these vanua, and the ultra-independence of Kadavu polities is recognised in the well-known saying ‘Manu dui tagi’, roughly equivalent to ‘Every man is cock of his own dung-heap’. The province of Kadavu corresponded to the geographical bounds of the island and its associated offshore islets.

Two others were based eventually on what, at the time of Cession and for a while afterwards, had been areas of continuing political instability in the mountainous interior (Na Colo) of Viti Levu, where people claimed not to have been consulted about, or to have agreed to Cession (see above).

The constructs and reality

In exploring pre-Cession customary Fijian society, it is fortunate that the Colonial form of Fijian administration established immediately after Cession was intended to be based on what the Colonial government understood to be such customary society. Further, NLCs were set up to investigate land-owning units, land boundaries and the justification for these landowning units to own the lands which they claimed. A tremendous wealth of information was recorded and has been used by France (1969), Macnaught (1971, 1982) and Scarr (1980), and it is these records which form one of the main sources of information which I used in preparing for this project. It is these records which I have since used to check the accuracy and reasonableness of the oral accounts of the current Fijian understanding and explanation of the origins and development of their polities. It can, however, reasonably be pointed out that first, the NLC accounts and reports were biased, being based on evidence given, albeit under oath, by landowners who may have misunderstood the interrogations or more likely preferred to give only such evidence as would justify their claims to traditional rights over their lands. Secondly, such accounts were biased towards giving support for the NLC views that the principles of Fijian society and land tenure were (or should be) the same throughout the Colony. Thirdly, the accounts may have been incomplete or inaccurate, though given in good faith. Fourthly, Ratu Sir Lala Sukuna did pretty well as he pleased in the activities of the NLC, in order to ensure the acceptance by the Colonial Government of those principles of traditional society and land tenure which would be most advantageous to the Fijian people when Colonial legislation was being drafted. Indeed it may have to be admitted that neither the NLC accounts nor my oral accounts will ever be accepted as being historically accurate. The results of archaeological field surveys and linguistic investigations, however, should provide data from these other disciplines which throw light on the question of the accuracy or reasonableness of the recorded and oral accounts.

Records of discussions of the Council of Chiefs at meetings following Cession and during the 1880s determined the contemporary structure of the descent groups (the yavusa, the mataqali and the itokatoka or its equivalent) and the socio-political constructs (the vanua and the matanitū). These were then recognised officially and prevailed throughout Fiji. This official structure did not, however, by any means always accord with the customary principles of Fijian society found to prevail in many parts of Fiji. Some of the terms and groups of the model may not have applied in the west before they were imposed by the Government. For instance, there were apparently special western groups remembered as kete and lewe.6 Further, I quite often heard in some areas in the west that certain socio-political groups had been subdivided by the NLCs and re-created into separate groups in order to solve local disagreements. Some of this may well be regarded as parochial grumbling, but its nature, basis and significance should be taken into account.

This chapter, then, emphasises that a detailed study of pre-Colonial traditional Fijian society, especially in the west, surely begins with an understanding and appreciation of how the Colonial Government worked to form an effective and economic form of Government. The Government strove to combine what was understood to constitute ‘native usages and customs’, together with the needs and demands of the significant expatriate and part-European population. A close study of the records of the LCC, the successive NLCs and early meetings of the Council of Chiefs did indeed provide useful information to stimulate my own studies. This complemented information I obtained from oral traditions, archaeological surveys and linguistic discussions in the course of my own investigations. It was self-evident that in using oral traditions as a basis for my investigations, the problems were, first, to assess the accuracy and reasonableness of those oral accounts I recorded as to how Fijians currently understand and explain pre-Colonial Fijian society; and secondly, to attempt to elucidate how and why such accounts may differ from time to time and from person to person. The next chapter gives a general introduction to my project.

1 Ratu, later Sir, Lala Sukuna (of aristocratic lineage, Foreign Legionnaire and holder of the Medaille Militaire, barrister), was from early on a highly respected member of the Administrative Service of Fiji and a leading adviser to the Government on Fijian matters. Later, he was Chairman of the Native Lands Commission, Secretary of Fijian Affairs and Speaker of the Legislative Council.

2 Unless there is some special reason not to do so, I use Bauan terms in this thesis to refer to features of social organisation that are generally pan-Fijian in substance but not necessarily in name.

3 Sometimes sundry relatives accompanied an original ancestor, and current accounts usually include details of his wanderings. The place where he finally settled down and, if unmarried, took a woman usually of a neighbouring group and founded a family, was known as the yavu tu. On his death, his agnatic descendants who formed the yavusa would treat him as their kalou vu or ancestral spirit. The shrine or spirit house associated with a kalou vu was known as the bure kalou, or beto/bito in the west. Kaunitoni is the name of the vessel as used in current versions of the legend. The antiquity of the legend been queried in toto by Peter France (1966) who cautioned that it is but a missionary/anthropologist invention. Paul Geraghty (1977) has also queried the authenticity of the name on linguistic grounds. He pointed out that the western communalect equivalent for kau (tree) is kai. I might point out that the name Nakauvadra (currently translated as screwpine tree) contains the same word kau, whereas the Rakiraki word for tree is also kai. Even in 1952, I never heard the speakers of the broadest form of communalect use the words Nakaivadra or Kainitoni, and the use of such names was vigorously denied as not just inappropriate but incorrect and unauthentic. It is possible that the archaic pronunciation was kau in proto-western speech but changed in present everyday speech to kai in the west, though it survived in proper names. Similarly we find anuya in names of islands, from the old Fijian *(a)nuya), though ‘island’ is now ‘yanuyanu’.

4 itokatoka is used in eastern Viti Levu and in Standard or ‘Bauan’ Fijian. Although the model refers to ‘tokatoka’, the present generally accepted form of ‘itokatoka’ is used in the body of this thesis. Different terms for this sub-division, a pan-Fiji socio-political feature, are used in other parts of Fiji. For instance, in Lau, the term used is bati ni lovo; kausivi occurs in the Yasawas, kete in Waya, ma’anibure in Ra, beto in Vuda and Nadi, bito in Nadroga and part of Nawaka.

5 Whatever the name for it, this minor social grouping was recognised as a reality in traditional Fijian society. Under the model, each itokatoka had a distinctive name, and was ranked hierarchically within the mataqali. The senior itokatoka represented the descendants of the eldest son of a son of the original ancestor. The senior member had no traditional or official title.

6 An informant told the NLC at the turn of the last century that in Nadi kete was equated with mataqali and lewe with itokatoka. Their actual meaning is now obscure, though the terms appear in present western polity names such as Ketenitukani and Leweiwavuwavu.

Previous Next