Canoes of the Grand Ocean 9781407302898, 9781407332963

Citation preview

BAR S1802 2008

Canoes of the Grand Ocean Edited by

Anne Di Piazza and Erik Pearthree DI PIAZZA & PEARTHREE (Eds) CANOES OF THE GRAND OCEAN

B A R

BAR International Series 1802 2008

Canoes of the Grand Ocean

Edited by

Anne Di Piazza Erik Pearthree

BAR International Series 1802 2008

ISBN 9781407302898 paperback ISBN 9781407332963 e-format DOI https://doi.org/10.30861/9781407302898 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

ACKNOWLEDGEMENTS

Writing a book about canoes is, in its own way, to embark on an odyssey. And while looking for insights about canoes through archives, drawings, publications and conferences, we also made fruitful encounters, in particular with the many authors who participated in this volume. We received valuable and insightful comments from these colleagues or friends, and express our deepest reconnaissance to them. We also owe a special debt of gratitude to our own institute of affiliation: CREDO (the Centre for Research and Documentation on Oceania) who facilitated this work in Marseille, France. Such a project must be funded, and we are very grateful for the grant awarded by the committee (Action Concertée Systèmes Complexes en Sciences Humaines) of the CNRS (National Centre for Scientific Research). We would also like to acknowledge David Davison, our privileged contact with BAR in Oxford for his patience and advice. Many thanks too to Daniel Laurent, a friend and artist to whom we confided the difficult task of portraying first encounters. His collaboration and enthusiasm to listen to our stories about Pacific voyaging were a great stimulation.

i

TABLE OF CONTENTS INTRODUCTION CANOES OF THE “GRAND OCEAN” by Anne Di Piazza

1

SESSION I I.

II.

III.

IV.

EARLY EUROPEAN DESCRIPTIONS OF OCEANIC WATERCRAFT IBERIAN SOURCES AND CONTEXTS by Carlos Mondragón and Miguel Luque Talaván Early Sources for the History of Oceanic Navigation Of pig tusks and canoes At the Crossroads of the South West Pacific: Taumako and the Western Polynesian World Conclusion References

9

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS by Anne Salmond European voyaging to Tahiti Tahitian voyaging Captain Wallis in 1767 Bougainville and Ahutoru in 1768 Captain Cool and Tupa’ia 1769 The Spanish explorers and Puhoro 1774-5 Conclusion References

23

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS by Eric Rieth The documentation of Oceanic canoes Historical problematics for the study of traditional watercraft Conclusion References

47

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS by Richard Feinberg Anuta’s dual symbolic organization Quadratic and concentric structures The mapping process Anutan terms for marine features Naming reefs and marine features Visual inspection Conclusions References

69

9 10 13 19 20

23 24 26 27 28 34 35 36

47 63 67 67

71 73 74 76 79 79 80 82

SESSION II V.

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS by Adrian Horridge The constraints A warning The beginning The 5-part canoe Outrigger canoes in historical times

ii

85 85 85 86 88 93

VI.

The influence of other cultures on Pacific canoes The triangular sail in the Indian Ocean Colonial influences Conclusion References

93 100 100 103 103

DUGOUT AND SEWN PLANK CANOE CONSTRUCTION ON RAIVAVAE, AUSTRAL ISLANDS by Robert Veccella Canoes on Raivavae The earliest descriptions The carpenter The tools Measurements The woods used today Building a canoe today Sewing Storage and maintenance Conclusion References

107 108 108 112 112 113 113 113 117 118 119 120

SESSION III VII.

VIII.

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION by Chris Avis, Álvaro Montenegro and Andrew Weaver The colonization of Oceania Navigation and maritime technology Scope of work Climate of the study area Model methodology Experiment design Results Discussion Conclusions References SIMULATING POLYNESIAN DOUBLE-HULLED CANOE VOYAGING. COMBINING DIGITAL AND EXPERIMENTAL DATA TO PREPARE FOR A VOYAGE TO RAPA NUI (EASTER ISLAND) by Bradley M. Evans Settlement of Polynesia Previous investigations - computer simulations and experimental archaeology Overcoming limitations Discussion Conclusion References

iii

121 121 121 123 124 125 126 128 130 139 140 143

143 145 146 149 151 153

LIST OF FIGURES I.

1. Island Melanesia, with the location of Taumako and the Central Solomons 2. Two views of contemporary Taumako te puke 3. Rendering of a Taumako te puke in motion 4. Sketch of the Duff’s Groupe 5. Contemporary topographic chart of Taumako 6. Tongan kalia 7. Earliest European rendering of a tongiaki 8. Two double-hulled tongiaki 9. A contemporary rendering of a tongiaki from Tonga

11 14 15 15 16 18 18 21 21

II.

1. An engraving of a fleet of Tahitian war canoes after William Hodges 2. Cook’s chart after Tupa’ia ‘s lists of islands

25 34

III.

1. Tonga-Tabou, sea canoe known as a Vaca 2. Canoe or Tafahanga from Tonga-Tabou 3. Tonga-Tabou, Double canoe known as a Calié 4. Calié of Tonga Tabou 5. Outrigger raft from Tevai, Vanikoro 6. A canoe of the Grand Ocean seen at Valparaiso 7. Ancient canoe from Tahiti, an Ivaha from Cook’s time 8. Po-motou archipelago, double canoe measured on Tahiti in 1838 9. Calié, double canoe from Tonga-Tabou 10. Double canoe from the Tuamotu Islands 11. New Caledonia. Double canoe after drawings by M. Amourous. 12. New Caledonia. Canoe by M. Amourous 13. Outrigger canoe after M. Amourous’ drawings 14. Double canoe from New Caledonia 15. Caroline archipelago. Canoe from Duperey island onboard the Coquille. Caroline archipelago. Canoe from Rotuma

48 49 50 51 52 53 55 56 58 58 59 60 61 62 64

IV.

1. Map of central Pacific Ocean showing relationships of Anuta to other islands and island groups 2. Founders of Anuta’s commoners and chiefly ‘clans’ 3. Diagram illustrating the binary ritual structure of the Anutan kindred 4. Symbolic structure of Anutan dwelling 5. Summary of Anutan binary organization 6. Map of Anuta Island, showing main terrestrial features, the fringing reef, and two visible offshore sea 7. Diagram of the twelve araavaka ‘canoe paths’ radiating outward from Anuta 8. Concentric rings around Anuta formed by reef and other bottom contours 9. Complete Anutan reef map as redrawn using computer graphics program 10. Cross-section of Anuta and its reefs 11. Pu Nukumanaia’s wind compass 12. Schematic representation of Anuta’s concentric model of physical and cosmological space

70

V.

1. A primitive sailing raft 2. The adze with a bend haft 3. The five part canoe 4. Attachment of the superstructure to the hull 5. A single outrigger travelling canoe of Satawal, in the Caroline Islands 6. A thamakau of Lakeba, eastern Fiji Islands 7. Sketch of a Madurese jukung 8. A Madurese jukung sailing downwind 9. Tacking a double outrigger 10. To tack a single outrigger 11. Map of the distribution of the mastless rig 12. The mastless rig today of a Madurese Janggolan

iv

71 72 73 73 75 76 77 78 79 81 81 86 87 88 89 90 91 92 92 92 93 93 95

13. Stern view of the Janggolan 14. Balinese double outrigger 15. The tripod mast and tilted rectangular rig 16. Rig of the Tuamotuan double canoe 17. Double canoe of Tahiti 18. Hawaiian rig on a small single outrigger 19. Early European rig of about 1500 AD 20. Two-boom triangular sail 21. Dutch spritsail 22. Gaff rigged cutter 23. Sail types

96 97 97 98 98 99 100 101 101 101 102

VI.

1. Map of French Polynesia and Raivavae 2. Comparison of dimension at maximum beam 3. Beached canoes 4. Moored canoes 5. Tensioning crossbeam lashing with a forked lever 6. Stages of assembly 7. Temporary stitching and wooden wedges 8. Vertical joint using synthetic cordage and epoxy glue 9. View of a canoe stern 10. Paddling a canoe, punting a canoe, canoe under power

107 108 111 111 113 115 117 117 118 119

VII.

1. The islands of Near Oceania and the Lapita expansion 2. Average summer and winter surface winds and currents 3. The velocity components of a drifting vessel 4. Shorelines of the islands of the Fiji group 5. The islands of the Lapita region in terms of simulation grid cells 6. Sample trajectories for vessels using the USCG and Levison parameters 7. Summary of simulation results 8. Time series of monthly crossing probabilities from the Solomonn-3 launch group to the Santa Cruz group 9. Time series of monthly crossing probabilities from Santa Cruz to the main Solomon target

122 123 125 127 128 130 131 137

1. The Polynesian triangle 2. The study areas of Levison et al. (1973) and Irwin (1992) and this simulation 3. Sailing parameter curve used in this simulation 4. The course of the Höküle‘a, from September 21 to October 9, 1999 5. A typical courseline generated using the stochastic wind model and sailing parameters from Levison et al. (1973) 6. A simulated voyage beginning on September 22, 1994 using the sequential weather model and the sailing parameters of Finney (1977) 7. Length of voyage from Pitcairn to Rapa Nui, using synoptic weather methodology, by month 8. Length of voyage from Pitcairn to Rapa Nui using synoptic weather methodology, by year

144 147

VIII.

v

138

149 150 151 152 152 153

LIST OF TABLES II.

1. Islands’ names

38

VI.

1. Dimensions of canoes in order of length 2. The different species used for canoe construction on Raivavae 3. Species of wood used for the construction of traditional canoes Today in Raivavae 4. Materials used in Raivavae canoes today

108 109 114

1. Overall crossing parameters for major island groups under the drifting and directly downwind sailing 2. Like table 1, but only for voyages that start during the Austral summer 3. Drift crossings into the Santa Cruz island group from sites in Near Oceania 4. Crossings into the Santa Cruz island group from sites in Near Oceania under the levison (sailing) parameters 5. Drift crossings into Fiji and Rotuma 6. Directly downwind sailing crossings into Fiji 7. Drifting crossings to Tonga and Samoa 8. Downwind sailing crossings to Rotuma, Tonga and Samoa

132

VII.

VIII.

1. A typical synthesized wind speed and direction matrix for August for a single grid cell 2. Distance travelled per day as a function of wind speed 3. T-tests of courseline attributes created using synoptic and synthesized weather methodologies 4. Coefficient of variation for courseline attributes created using synoptic and synthesized methodologies

vi

115

133 134 134 134 135 135 136 148 148 149 150

canoes had appeared in journals and accounts of voyages since the time of the earliest circumnavigations. But until Pâris, there had been no volume or descriptive annex to an account specifically dedicated to canoes.

Introduction Canoes of the ‘Grand Ocean’ by Anne Di Piazza*

In the 16th and 17th centuries, canoes were counted among the ‘curiosities’ of the islands but not until the 18th and 19th century did they become objects of classification and reflection. Captain Wallis was the first to describe Tahitian canoes. He distinguished three “different sorts”: outrigger dugouts “used chiefly for fishing”, larger plankbuilt canoes - either outrigger or double - for voyaging “far beyond the sight of land, probably to other islands…”, as well as very large paddling canoes “intended principally for pleasure and show” (Hawkesworth 1773, vol.1: 487). But the big question, nicely summarized by Captain Cook was “How shall we account for this nation spreading itself so far over this Vast ocean?” (Beaglehole 1955: 154). In fact, their observations and fascination with the canoes were, at least in part, directed at explaining the migrations of ‘this nation’.

In choosing a title, there is always the pleasure of the words, the fascination of a history, a geography; here for an Ocean; the ‘Grand Ocean’, a name coined by Lapérouse1, whose map “Carte du Grand Océan ou mer du Sud” 2 was published in 1797. Laperouse was never to know the popularity of his term, due to his tragic shipwreck on Vanikoro. In the 19th century, ‘Grand Ocean’ became the accepted term, at least in France, to designate the Oceanic space as opposed to its contents, the islands. In his introduction to the volume on the zoology of the Voyage of the Coquille under the command of Captain L. I. Duperrey, R. P. Lesson wrote: “Adopting the usage of many modern geographers, we will refer to the uncountable scattered islands within the Grand Ocean as Oceania3” (Duperrey 1826). Malte-Brun, in his Précis de la Géographie Universelle (1813: 228), popularised the terms ‘Grand Ocean’ and ‘Océanique’, the latter referring to “the 5th part of the world… entirely within the Grand Ocean, or the Ocean par excellence”. He went on to state that “this essential characteristic is not shared with any of the other parts of the globe; this character gives a particular physiognomy to its geography, its natural history and its civil history4”.

Session I Navigators as the principal actors in cross-cultural encounters The oldest sources for reconstructing the navigational knowledge of Oceanic peoples are the accounts left by European ‘discovers’ of the Pacific. Historians and anthropologists have been fascinated by the particular moment of these first encounters and first visions between peoples who until then had lived in complete ignorance of each other; a moment supposedly marking the beginning of a new era.

‘Canoes of the Grand Ocean’ is also a nod to Edmond Pâris, the “father of nautical ethnology” as developed by Reith in chapter III. Pâris was the first officer to systematically record canoes, carrying out the orders of his commander, Dumont D’Urville. Besides his plates for the official account of D’Urville’s voyages, his magnum opus consisted of two large volumes, one of text, the other an atlas of plates covering traditional watercraft of the world including the Pacific (Pâris 1843). Pâris was of course not the first to describe canoes of Oceania. Occasional commentaries, descriptions and drawings of

The actors on both sides of these first encounters shared a common trait: they were navigators, even if it was difficult for the Europeans to accord that status to the Other. In the Journals and accounts of voyages the island navigators remained ‘savages’, only incidentally mentioned, although their contributions were of the utmost importance, in particular for the discovery of islands. Today some of these native navigators are at the heart of historical and anthropological studies of first contact, to the point where one of them, Tupaia has become almost a legendary figure.

*

CNRS-CREDO, Marseille This is the earliest usage of the term known to the author. 2 The map is accessible at http://www.davidrumsey.com/ directory/ what/Atlas+Map/where/Pacific 3 “Adoptant la manière de voir de plusieurs géographes modernes, nous appelons Océanie les îles innombrables qui sont éparses dans le Grand Océan”. 4 “La cinquième partie de monde […] se trouve située toute entière dans le Grand Océan, dans l’Océan par excellence. Ce caractère essentiel ne lui est commun avec aucune autre division du globe; ce caractère donne une physionomie particulière à sa géographie, à son histoire naturelle, à son histoire civile”.

Sometimes anonymous, sometimes celebrated such as Tumai, Aotourou or Tupaia who respectively embarked with Quiros, Bougainville and Cook, these men went out to meet the caravels, frigates and sloops, helped them find safe anchorages, passes into the lagoon, and avoid reefs. They were of great help in these uncharted waters. Soon afterwards, these pilots, clearly expert in local seas were interrogated about their knowledge of neighbouring islands; knowledge keenly sought after by Spanish, French and English naval officers, hoping to discover and map the last great unknown part of the globe. Usually,

1

1

ANNE DI PIAZZA

beach. Holding a paper in his hand, with the compass before him, he began asking them many times respecting the lands to seaward; and all gave the same information. Other persons that same day put the same questions to other natives elsewhere in the island, always with the same result. So it appeared certain to us that the natives were truthful” (Cameron 1966: 167 cited by Maroto Camino 2005: 80-81).

their knowledge consisted of lists of island names whose directions were pointed to with their fingers, accompanied by parsimonious commentaries, indicating distance in sailing days, whether they were inhabited or not, and what they produced (foods, pearl shell, etc.). From the Europeans’ perspective, their first interest was to discover and acquire new lands, to enquire about their resources and richness for trade, to learn about the docility of the natives, as well as to determine whether these lands were islands or perhaps the long sought continent of Terra Australis. For example, when in 1567, the captive chief of St. Isabel in the Solomon Islands drew a map of his lands in the sand for Mendaňa and Quiros, they realised that St. Isabel was an island and not the Southern Continent (Beaglehole 1966:46, cited by Maroto Camino 2005: 80-81).

The objectives of the islanders are less easy to decipher. One may imagine political strategies aimed at creating alliances with these new strangers, to engage them in exchange networks or to get novel goods, sometimes doubtless to help the whites to go away. One striking thing about these cross cultural encounters is how often the islanders were willing to embark, ‘applying’ for temporary or long term jobs as pilots, sailors or geographers. Bougainville noted in his official journal: “An Indian who appears to be a chief has been aboard the Etoile since yesterday evening and does not seem to want to leave5” (Taillemite ed. 2006, vol.I: 315). In his published account of the voyage, we learn that this ‘Indian’ is none other than Aoturu, the first Tahitian who departed on a European ship, apparently of is own accord, and who almost completed a circumnavigation… before dying on the return voyage (Bougainville 1982: 245). Aotourou revealed to Bougainville the existence of numerous islands in the Tahitian archipelago, “some in confederation with Taiti, others always at war with her. The friendly islands are Aimeo [Moorea], Maoroua [Maupiti], Aca, Oumaîtia [Mehetia] and Tapoua-massou [Maiao]. The unfriendly ones are Papara [a district on Tahiti], Aiatea [Raiatea], Otaa [Tahaa], Toumaraa, Oopoa. These islands are as large as Taiti. The island of Pare, very rich in pearls, is sometimes an ally, sometimes an enemy. Enoua-motou and Toupai are two little uninhabited islands, covered with fruit trees, pigs, birds, and abounding in fish and turtles… these isles lie at different distances from Taiti. The most distant that Aotourou told me about is fifteen days sailing6” (Bougainville 1982: 267).

The official orders given to Quiros, chief pilot of the Alvaro de Mendana expedition in 1595 are revealing. “Learn from the natives whether there are other islands or extensive lands near, if they are inhabited, of what colour are the natives, whether they eat human flesh, if they are friendly or carry on war. Enquire whether they have gold in dust, or in small lumps, or in ornaments; silver worked or to be worked: metals, all kinds of pearls, spices and salt…” (Markham (ed.) 1904: 149). These instructions were followed to the letter during a later circumnavigation commanded by Quiros. Here is Ian Cameron’s report of the first encounter with the inhabitants of Taumaco in the Solomon Islands. “The Captain [Quiros] then asked Tumai whether he knew of other lands, far or near, inhabited or uninhabited. In reply Tumai pointed to his island, then to the sea, then to various points of the horizon; he then began counting on his fingers up to as many as sixty islands, ending with a very large land which he called “Manicolo” [Malakula, some 700 km to the south in Vanuatu]. The Captain wrote down the names, having the compass before him for noting the bearing of each island from Taumaco. To explain which were small islands, Tumai drew small circles, and for larger ones larger circles; while for the very large land to the south-west the opened both his hands and his arms as wide as he could. To explain which were the distant islands and which were nearer, he pointed to the sun, then rested his head on his hand, shut his eyes, and with his fingers counted the number of nights one had to sleep on the voyage. In a similar way he explained which people were white, black or mulattos; which friendly and which hostile. He gave us to understand, by biting his arm, that in one island they ate human flesh, and indicated he disapproved of such people. In this way what he said was understood. He repeated his information many times until he was tired, and showed a desire to return to his house; and then the Captain gave him many gifts for barter, and he departed after embraces and other tokens of love… Next day the Captain went to the village, and in order to corroborate what Tumai had said, he assembled the natives on the

These examples are far from isolated7. In reading the journals of the different voyages, it is clear that the ship captains almost systematically made use of the islanders’ ability to communicate their geographical knowledge, the 5

“Un indien qui paroit un chef est à bord de l’Etoile depuis hier au soir et n’en veut plus sortir”. 6 “les unes confédérés de Taiti, les autres toujours en guerre avec elle. Les îles amies sont Aimeo, Maoroua, Aca, Oumaîtia et Tapoua-massou. Les ennemies sont Papara, Aiatea, Otaa, Toumaraa, Oopoa. Ces îles sont aussi grandes que Taiti. L’île de Pare, fort aondante en perles, est tantôt son allié, tantôt son ennemie. Enoua-motou et Toupai sont deux petites îles inhabitées, couvertes de fruits, de cochons, de volailles, abondantes en poissons et en tortues… ces îles gissent à différentes distances de Taiti. Le plus grand éloignement dont Aotourou m’ait parlé est à quinze jours de marche”. 7 See Dening 1963 for more references.

2

CANOES OF THE GRAND OCEAN

names of the various islands, as well as their bearings and distance. This recourse to natives’ knolwedge became a proper strategy of navigation. When Boenechea, was dispatched to Tahiti in 1772, but lacking Wallis’s, Cook’s and Bougainville’s sailing directions to the island, “he picked up a man at Mehetia, who pointed to the west, and exclaimed, ‘Tahiti’” (Salmond, chapter II). Later, in 1774-1775, on another trip, when Boenechea decided to visit Ra’iatea, “the Aguila was guided by Puhoro, a navigator from Ma’atea, and Mau’a, the exiled high chief of Ra’iatea” (Salmond, chapter II citing Corney 1919).

Otto von Kotzebue “enthusiastically describes how at his request, the Marshallese readily converted their knowledge of the size, shape, and distribution of the islands in this group into ephemeral charts of some accuracy” (Finney, op. cit : 453). He got Lagediack, a navigator from Woje atoll, a chief on Maloelap, as well as the elderly Langemui to outline in the sand all the atolls of the Ratak and Ralik chains. Among other objectives, La Pérouse was explicitly ordered by Louis XVI, a passionate amateur geographer, to improve the maps of the Pacific. Disembarking on the island he named Ségalien, he wrote: “We have succeeded in making them understand that we want them to depict their country and that of the Mantchéoux. So, one of the elders rose up, and with the stem of his pipe, he traced the coast of Tartarie on the West, running nearly North and South. Facing it in East and lying in the same direction, he drew his island, and putting his hand on his chest, he made us understand that he had just drawn his own country; he had left a strait between Tartarie and his island, and turning to our vessels, which could be seem from the shore, he drew a line, indicating that one could go through… His shrewdness in guessing our questions was great, but still less than that of another islander, aged about thirty years who seeing that the pictures traced in the sand were disappearing, took one of my pencils and paper; he drew his island which he named Tchola and he indicated by a line the little river whose banks we were on…9” (La Pérouse [n.d.: 88] cited by Latour 1983: 227). The fact that both parties were able to understand each other through maps highlights their ability to truly communicate despite their different languages, perceptions, traditions and needs.

These men were not only pilots and geographers, and ironically, more erudite in these local seas than the European officers, they were also skilled cartographers, capable of drawing charts for their foreign colleagues. The most famous of these is the one known as “Tupaia’s Chart”, apparently Cook’s copy of Tupai’s original, depicting the location of 74 islands (Dening 1963, Lewthwaite 1970, Finney 1998, Di Piazza and Pearthree 2007). Other earlier native charts were recorded. In 1696 castaways blown off course while sailing from Lamotrek to Fais in the Carolines ended up on Samar, in the eastern Philippines, where there was a Jesuit mission. Recounting their story to the missionaries, “they also named the thirty-two islands that made up their ‘nation’ and later spread out pebbles on the beach to signify the locations of eighty-seven islands they claimed to have visited” (Hezel 1983 cited by Finney 1998: 453). Father Paul Klein had the chart redrawn on paper and archived. Another chart of the Carolines was drawn in 1721 by another Jesuit missionary, Father Cantova based on Guam, from directions given by Carolinian castaways (Finney op. cit.: 453-454). In his Journal, an officer on the Uranie, one of de Freycinet’s ships narrates an encounter with a navigator from the Caroline Islands, whose knowledge vividly impressed the crew. “His knowledge was so renowned among the sailors that M. de Freycinet wanted to meet him. He came and at the request of the commander drew a map of his archipelago. He located all the islands with an astonishing intelligence, listed all of them by their names and designated the particular kind of things each one produced. He showed a shell that he said served his companions as a compass, as well as the study of the stars, the direction of the currents… He depicted the different constellations with grains of maize8.” (Blais 2005: 132). Similarly, in his journal, the Russian explorer

Native navigational knowledge also appeared, although more tenuously, in somewhat comical situations. This happened to Aotourou who, while en route and wishing to visit an island a few days from Oumaitia, was indignant that Bougainville would not change his course, “ran to seize the ships wheel, of which he had already noted the use… to steer us toward the star that he had pointed

9

“Nous parvînmes à leur faire comprendre que nous désirions qu’ils figurassent leur pays et celui des Mantchéoux. Alors, un des vieillards se leva, et avec le bout de sa pipe, il traça la côte de Tartarie à L’Ouest, courant à peu près Nord et Sud. A l’Est vis-à-vis et dans la même direction, il figura son île; et en portant la main sur la poitrine, il nous fit entendre qu’il venait de tracer son propre pays; il avait laissé entre la Tartarie et son île un détroit, et se tournant vers nos vaisseaux, qu’on apercevait du rivage, il marqua par un trait qu’on pouvait y passer… Sa sagacité pour deviner nos questions était très grande, mais moindre encore que celle d’un autre insulaire, âgé d’à peu près trente ans qui, voyant que les figures tracées sur le sable s’effaçaient, prit un de mes crayons, avec du papier; il y traça son île, qu’il nomma Tchola, et il indiqua par un trait la petite rivière sur le bord de laquelle nous étions…”.

8

“La renommée de son expérience dans la marine fit que M. de Freycinet désira le voir. Il vint et traça à la demande du commandant la carte de son archipel. Il plaça toutes les îles avec une intelligence surprenante, les nomma toutes par leur nom et désigna l’espèce de production qui fournissait chacune d’elle en particulier. Il fit voir une coquille qu’il dit servir de boussole à ses compatriotes, ainsi donc à l’inspection des étoiles, à la direction des courants […]. Il traça avec des grains de maïs les différentes constellations”.

3

ANNE DI PIAZZA out10” (Bougainville 1982: 274). Polynesian seamanship was also noticeable upon the arrival of the Endeavour at Huahine, when Tupaia asked an islander to dive to make sure the depth was sufficient to allow a safe passage.

fishing grounds and other marine features. This is a nice example of the continuation of a ‘tradition’ born of the first contacts between outsiders and islanders who attempted to teach their geography to the foreigners by transforming their ephemeral ‘charts’ in the sand onto paper.

Part of this indigenous knowledge was lost on the officers, particularly when dealing with detailed navigational methods such as star sailing courses, star and wind compasses, etak systems, sea marks. Etc. None of the explorers went sailing with the islanders and only remarked generally on their understanding of winds, stars, clouds and weather. To apprehend the finer points of a knowledge system is not a simple matter of observation and experience… it may well require an anthropological or historical reflexive stance. Navigation is after all an art performed very differently by Islanders and Europeans. If the latter understood some of the scope and complexity of indigenous navigation, to go further, it is incumbent on us to put their knowledge into context; a task undertaken by the authors of the first four chapters.

Session II The many varieties of canoes Oceanic sailing vessels were called ‘canoes’ by the British and ‘piraguas’ or ‘pirogues’ by the Spanish and French. These terms were born of a context far from the Pacific: the isles of the Caribbean, where they designated dugouts without sails. To account the bewildering variety of Pacific canoes, they added qualifiers such as ‘outrigger’, ‘double canoe’, ‘double outrigger’. The only Austronesian word they borrowed is proa from the Malay prahu which originally meant simply ‘boat’, but which has come to mean single outrigger sailing craft in European languages.

Carlos Mondragón and Miguel Luque Talaván who, in addition to offering selections from Iberian journals and maritime diaries written between 1567 and 1606, give detailed accounts of first contacts in the Solomon Islands. They explore the intentions and purposes of the natives, who prior to this contact have already been engaged in exchange with ‘foreigners’. The authors’s translation allowed them to recognise that the large double canoe being described was probably a Tongan tongiaki, evidence of pre-European contact between Western Polynesia and Melanesia. Anne Salmond presents a rich synthesis of cross cultural contact on Tahiti, the preferred refreshment stop for English, French and Spanish ships during the 18th century, followed by missionaries in the 19th and whose inhabitants had to cope with foreigners such as Wallis, Bougainville, Cook and Boenechea. Investigating understandings, mis-understandings and exchanges of navigational systems between Islanders and Europeans, one enters into the complexities of two past knowledge systems. Eric Rieth takes us to the heart of the 19th century, and the objectives of Lieutenant (later Admiral) Pâris, who dedicated much of his career to recording extra-European vessels through his writings, paintings and plans as well as his policies as the conservator of the “Musée naval du Louvre”, also known as the “Musée de Marine”. One of his legacies is the series of scale models of Oceanic canoes still exhibited today. The chapter by Richard Feinberg examines the process of Anutan map making of the island’s reefs,

According to the missionary W. Ellis (1972: 293), the Tahitians described European boats as “islands inhabited by beings of a superior nature, upon whose orders lightning flashed and thunder rumbled”, a reference to muskets and cannons. The trader and consul J.A. Moerenhout believed, “O-Tahitians mistook it [Wallis’s ship] for a floating island seeing the masts as trees, the pumps as waterfalls…” (Moerenhout 1837: 389). In Teuira Henry (1962: 21), one can find that the Dolphin’s stern “was compared to a cliff”. These metaphors were not invented for European boats. They recall those used in Tahitian mythical chants where the island of Tahiti sailed to its location like a ship or pahi, that its eastern cliffs were compared to a carved canoe prow, and the island’s political structure to a canoe or va’a (Baré 2002: 116). In other myths, Tahiti is a fish, who was immobilized by a blow from a cultural hero, shaping its body into two peninsulas linked by a narrow plain, metaphorically the two hulls of a double canoe (Danielsson 1981: 51). These examples show that during the first contacts, the islanders integrated the European vessels into their own categories of thought and that metonymically, a vessel is an island and vice versa (Baré 2002: 113-118). At the same time, Europeans treated Oceanic vessels with no less astonishment and interest. The canoes were remarkable enough that two archipelagos were named after them. The Marianas were called the “Islas de las Velas from the great number of sails seen passing to and fro between the islands” (Haddon and Hornell 1975: 412). Samoa was named the ‘Iles des Navigateurs’ by Bougainville. He was doubtless influenced by seeing two sailing canoes 6 leagues (ca. 30 km) offshore and his reading of the Dutch navigators since he noted in his journal that: “Le Maire in more of less this same area encountered a double [canoe] out of sight of land with a

10

“courut saisir la roue du gouvernail, dont il avait déjà remarqué l’usage… pour nous faire gouverner sur l’étoile qu’il indiquait“…il est venu quantité de pirogue aussi à la voile allant d’une vitesse étonnante, manoeuvrant comme elles vouloient de façon qu’elle faisoient deux fois le sillage du navire fillant six nœuds: ils s’amusoit à passer sur l’avant à nous, à nous les faire croire vingt fois écrasés”. .

4

CANOES OF THE GRAND OCEAN shelter on the poop like those in Cythère [Tahiti]11” (Bougainville cited in Taillemite 2006, vol.I: 336).

devoid of scientific and literary skills, who visited Guam in 1710, carried a flying proa back to London, “thinking it might be worth fitting up as a curiosity on the canal in St. James Park, since we have none like it in this part of the world” (Haddon and Hornell 1975: 417).

Early circumnavigators highlighted the canoes’ size, speed, manoeuvrability, their ability to sail close to the wind, the facility with which they could be righted when capsized, their shallow draft and their light weight allowing easy landing. These remarks serve to point out advantages that the cumbersome western ships lacked. Reading the accounts, one can guess that the islanders were amused by these slow clumsy ships, and played with overwhelming skill, showing off the superiority of their own vessels. In Samoa, Vivez, the surgeon on the Etoile, described how “…there came a number of sailing canoes going at astonishing speed, manoeuvring as they wish in such a way that they make twice the wake [speed] of the ship which is doing 6 knots: they amused themselves passing in front of us, they made us believe they would be crushed at any moment12” (Taillemite 2006, vol.II: 251).

Both Haddon and Hornell, and Neyret made very full inventories of the great variety of Pacific canoes as reported by voyaging and missionaries accounts. The completeness and accuracy of these descriptions however depended on the curiosity and maritime acumen of the recorders, not all of whom were sailors or boat builders. But in their defence, one must recognise that the construction of these canoes is a complex art. Anson, who had plenty of time in 1742 to study a ‘flying proa’ owned by a Spanish officer in the Marianas took the canoe apart to better measure it and to make a detailed plan but was later unable to entirely put it back together again (Haddon and Hornell 1975: 413-415; Neyret 1974, vol.2: 159). Beyond the anecdote, is the question of the technological history of these canoes. How the interaction between bamboo rafts and dugouts came to produce the many types of Oceanic watercraft? If introduced technology such as the rudder, the fixed mast and the tilted rectangular cloth sail came to Oceania from the Indian Ocean, in chapter V Adrian Horridge develops, among other things the idea that “recently, European influences replaced or heavily overlaid the Austronesian designs all over the Pacific”. With the arrival of Spanish galleons in the Pacific following the Acapulco – Manila route, Oceanic canoes became the receptacle for technological transfer. After all as skilful carpenters, the islanders could be expected to learn aspects of European boat building techniques. And the great many varieties of canoes spread throughout the ‘Grand Ocean’ militate for a tradition that is not as conservative as is usually thought. It might be worth noting here that during his third voyage, while calling once again at Queen Charlotte Sound in New Zealand, Cook reported that to the Maoris, the Endeavour was Tupaia’s waka (Beaglehole 1967: 73).

Europeans did not always treat the canoes kindly. On some occasions, after firing upon their crews, canoes were brought alongside or taken aboard, with the idea of carrying them back to Europe. In 1767, Wallis took a small outrigger canoe from Nukutavake in the Tuamotus. Today conserved in the British Museum, it is the oldest Oceanic watercraft collected by Europeans (Haddon and Hornell 1975: 51). Bougainville did not take the trouble, and dropped the confiscated canoes back into the sea. Here is the incident which occurred off New Britain as reported by Louis-Antoine Starot de Saint-Germain, an official writer on board: “We stopped firing on them, and we thought that the number of dead amounted to 8 or 9 without counting the wounded. We captured 2 of their canoes which our ship’s yawls gloriously towed [back]. The prince [de Nassau-Siegen] wanted to go to shore to amuse himself with the one that our yawl had taken in tow, but it capsized with him and 4 men along the shore, from not knowing how to manage it and he almost drowned. The other canoe was abandoned at sea by the yawl from the Etoile13” (Taillemite 2006, vol.2 : 113). Other canoes ended up in European waters. Woodes Rogers, a pirate associated with William Dampier, not

Even today, the islanders continue to modify, adapt and incorporate new materials into their traditional canoes, as demonstrated in the chapter by Robert Veccella on the canoes of Raivavae. Such careful descriptions of the construction techniques of canoes are still too rare, just as they were in the past. And as Bougainville, Cook and others suspected, it is only by knowing more about them that we might eventually come to explain how virtually every island of the ‘Grand Ocean’ was occupied or at least visited by the Pacific islanders.

11

“Le Maire, à peu près dans ces mêmes parages, en rencontra une double et à chambre sur l’arrière, comme celles de Cythère, hors de vue de toute terre”. 12 “…il est venu quantité de pirogue aussi à la voile allant d’une vitesse étonnante, manoeuvrant comme elles vouloient de façon qu’elle faisoient deux fois le sillage du navire fillant six nœuds: ils s’amusoit à passer sur l’avant à nous, à nous les faire croire vingt fois écrasés”. 13 “On en cessa de faire feu sur eux, et on croit que le nombre des morts alloint de 8 à 9 sans compter les blessés. On enleva deux de leurs pirogues que nos canots remorquoint glorieusement. Mais le Prince voulant aller à terre pour s’amuser dans celle que notre canot menoit à la traine, il fut chaviré luy et 4 hommes sur la cotte, faute de savoir la conduire et falirent à se noyer. L’autre pirogue fut abandonnée en mer par le canot de l’Etoile…”.

Session III Paper canoes Since their earliest encounters, Europeans were quick to recognise the ingenuity of the many canoes of Oceania. This is one domain where indigenous people gained some measure of esteem. However, for Europeans, to accept

5

ANNE DI PIAZZA

that these embarkations could have sailed long distances to settle the islands was another matter altogether, either they considered these craft too frail, or that the navigational techniques lacked science and instrumentation. To account for settlement therefore they put forward far-fetched hypotheses on the geology of the ‘Grand Ocean’. According to Quiros, the people of the Marquesas Islands were “without skill or the possibility of sailing to distant parts”. He thus made the supposition that there existed a long chain of closely spaced stepping stone islands extending eastward from Asia (cited by Finney 1998: 445, note 9). Jacob Roggeveen, while sailing along the shore of Easter Island, “was at such a complete loss to explain how Stone Age people with only small, frail canoes at their disposal could be living on such an isolated island that he proposed they must have been separately created there by God” (Finney 1998: 445, note 11). A third, and probably the most common explanation until the late 18th century, was probably the existence of a sunken continent whose mountain tops formed the many isles of Oceania.

Oumaitia (Mehetia): “…we had an incontestable proof that the inhabitants of the isles of the Pacific Ocean communicate among themselves, even across considerable distances. The azur [clarity] of a cloudless sky lets the stars sparkle [through]; Aotourou after having attentively studied them pointed out to us the brilliant star in Orion’s shoulder, saying that we should steer our course by it, and in two days we would find a rich land that he knows of and where he has friends17” (Bougainville 1982: 274). More pragmatic, Fesche noted: “I am very persuaded that they sometimes make long voyages, otherwise what would be the use of their immense canoes 60 feet long, carefully preserved in expressly made hangars18” (Taillemite 2006, vol. II: 89). There grosso modo the matter rested with intermittent outbursts of doubt until the publication of Andrew Sharp’s book and the resulting symposium chaired by Golson in the 60’s-70’s. The debate concerned mostly whether the Pacific could have been settled by accidental or deliberate voyages, calling into question the accuracy of traditional navigation and the difficulty of holding a course over long distances. Sharp claimed that islands more than 300 miles couldn’t be settled, except by castaways onboard canoes blown off course, or refugees ready to leave their homes at any cost and trust their lives to the mercy of the winds and currents. This idea is not new. For Cook, influenced by the Forsters and later William Anderson, the ship’s surgeon “the story of Pacific migrations was to be told in terms of accidental voyages which peopled the scattered islands” (Dening 1963: 111).

Bougainville is generally accepted as the first European willing to credit Polynesian navigators the ability to perform deep sea voyaging (Dening 1963: 111). In his Journal (Taillemite 2006, vol. 1: 305), when approaching Akiaki in the Tuamotus, Bougainville was wondering about the settlement of the islands: “Who could tell me how they were carried here and what chain of communication links them to other peoples? The stones of Deucalion14 and Pirrha, were they cast as far as this mote of isolated earth?15”. In his published account, upon further consideration, he proposed an answer: “The knowledgeable people of this nation [Tahiti], without being astronomers, as claimed in our newspapers, have a nomenclature of the most remarkable constellations; they know their movements throughout the day, and they use them to direct their course in the open sea from one island to another. Of these voyages, sometimes more than 300 leagues [1600 Km], they lose sight of all land. Their compass is the course of the sun during the day and the position of the stars during the night16” (Bougainville 1982: 266). He further stated, two days after leaving

One may question, as did Turnbull, what Cook meant by “accidental”. He was the first navigator “to sail to a Pacific Island as an act of deliberate calculations”, provided with a Nautical Almanac, a precise sextant and Wallis’s latitude and longitude for Tahiti. “His predecessors had only found them accidentally or rediscovered them by sailing along the latitude” (Turnbull 2000: 60). Although, no one seems to refer to accidental voyaging when discussing Wallis’s or Bougainville’s ‘discoveries’ of Tahiti. One could even say that the only non-accidental discoveries made by any

14

Heroes in Greek mythology. Deucalion is the son of Prometheus and is married to Pyrrha. They were the sole survivors of a flood sent by Zeus and repopulated the earth by scattering stones that were transformed into men and women. 15 “Qui me dira comment ils ont été transportés jusqu’ici et quelle communication les lie à la chaîne des autres êtres? Les pierres de Deucalion et Pirrha ont-elles été lancées jusque sur cette motte de terre isolée?”. 16 “…Les gens [Tahitiens] instruits de cette nation, sans être astronomes, comme l’ont prétendu nos gazettes, ont une nomenclature des constellations les plus remarquables; ils en connaissent le mouvement diurne, et ils s’en servent pour diriger leur route en pleine mer d’une île à l’autre. Dans cette navigation, quelquefois de plus de 300 lieues, ils perdent toute vue de terre. Leur boussole est le cours du soleil pendant le jour, et la position des étoiles pendant les nuits”.

17

“…nous eûmes une preuve une preuve incontestable que les habitants des îles de l’océan Pacifique communiquent et entre eux, même à des distances considérables. L’azur d’un ciel sans nuages laissait étinceler les étoiles; Aotourou, après les avoir attentivement considérées, nous fit remarquer l’étoile brillante qui est dans l’épaule d’Orion, disant que c’était sur elle que nous devions diriger notre course, et que dans deux jours nous trouverions une terre abondante qu’il connaissait, et où il avait des amis”. 18 “Je suis très persuadé qu’ils font quelquefois des voyages de long cours, car à quoi serviroient les pirogues immenses de 60 pieds do long, soigneusement conservés sous des hangars fait exprès”.

6

CANOES OF THE GRAND OCEAN

of the European voyagers were when they had native pilots.

References

The debate following from the publication of Sharp’s thesis, served as the catalyst for studies of indigenous navigational systems wherein the calculative tradition is absent. The publication for example in 1972 of David Lewis’s book: “We, the Navigators”, now a classic, allowed a better understanding of orientation, dead reckoning, course keeping and techniques for making landfall. Other important works in the same vein are those of Gladwin (1970) and Thomas (1987). Furthermore, since 1976, Hokule’a, the replica of a Hawaiian double canoe, has sailed all across Polynesia without instruments or charts (Finney 1979). A parallel method of inquiry into the practice of traditional navigation is the use of computer simulations allowing virtual canoes to drift or sail under specific strategies (Levison et al. 1973, Irwin 1992, Di Piazza et al. 2007).

Anderson, A., Chappell, J., Gagan, M., Grove, R., 2006 “Prehistoric maritime migration in the Pacific islands: an hypothesis of ENSO forcing”. The Holocene 16 (1): 1-6. Baré, J-F. 2002 Le malentendu Pacifique. Des premières rencontres entre Polynésiens et Anglais et de ce qui s’ensuivit avec les Français jusqu’à nos jours. First ed. 1985. Paris: Editions des Archives contemporaines. Beaglehole, J.C. 1955 The Journals of Captain James Cook on his Voyages of Discovery, vol. I. Cambridge University Press, Cambridge. 1966 The Explorations of the Pacific. Black, London. 1967 The Voyage of the Resolution and Discovery 1776-1780. Hakluyt Society, Cambridge. Blais, H. 2005 Voyages au Grand Océan. Géographies du Pacifique et Colonisation 1815-1845. Paris: Ministère de l’Education nationale, de l’Enseignement supérieur et de la Recherche, Comité des travaux historiques et scientifiques. Bougainville, L-A. de 1982 Voyage autour du monde par la frégate du Roi La Boudeuse et la flûte l’Etoile. Annotated edition by J. Proust. Paris: Folio Gallimard. Cameron, I. 1966 Lodestone and Evening Star: The Epic Voyages of Discovery 1493 B.C. – 1896 A.D. New York: Dutton. Corney, B. G. 1919 The Quest and Occupation of Tahiti by Emissaries of Spain during the years 1772– 1776, vols 1-3. Hakluyt Society, London. Danielsson, B. 1981 Tahiti autrefois. Papeete: Hibiscus Editions. Dening, G. 1963 “The Geographical knowledge of the Polynesians and the Nature of Inter-Island Contact”. In J. Goslon (ed.), Polynesian Navigation. A Symposium on Andrew Sharp’s Theory of Accidental Voyages. Wellington, Sydney: Polynesian Society. Di Piazza, A. and Pearthree, E. 2007 A new reading of Tupaia’s Chart. The Journal of the Polynesian Society 116(3): 321-340. Di Piazza, A., Di Piazza, P. and Pearthree, E. 2007 Sailing virtual canoes across Oceania: revisiting island accessibility. Journal of Archaeological Science 34 (8): 1219-1225.

Today, if deliberate long-distance voyaging is no longer in question, doubts remain about the performance of early watercraft, such as their windward ability, and conjectural correlations between periods of relative climatic instability - more frequent ENSO episodes - and greater voyaging activity, whose westerlies favoured eastward crossings (Anderson et al. 2006). To better understand the strategies of Lapita voyaging, Chris Avis, Álvaro Montenegro and Andrew Weaver here propose a simulation where the sailing performances of the canoes are limited to drifting or sailing directly downwind. Their results suggest that the deliberate exploration and discovery of islands in the Lapita region did not require “advanced exploration schemes”, nor any windward capability when reconstructed climate variability is taken into account. Bradley Evans takes another tack. Using detailed synoptic wind data and sailing parameters derived from experimental voyages, he compares the course effectively sailed by Hokule’a with the one created by different methods of computation. The similarity of those courses, not only demonstrate the validity of the programming, but also answer common criticisms of computer simulation, their lack of utility and reality. If Oceanic canoes, by their very strangeness were surprising to the earliest European observers, it was not long before their descriptions shifted from being impressed, enthusiastic or fascinated and gave way to detailed observations, measurements, comparisons and representations. Canoes are also the means by which the islanders apprehend space. In this perspective, the study of canoes of the ‘Grand Ocean’ remains a vehicle for discovery.

7

ANNE DI PIAZZA

Levison, M., Ward, R.G. & Webb, J.W. 1973 The Settlement of Polynesia: A Computer Simulation. University of Minnesota Press: Minneapolis. Lewthwaite, G. 1970 “The Puzzle of Tupai’s Map”. New Zealand Geographer 26:1-19. Lewis, D. 1994 We, the Navigators: The Ancient Art of Landfinding in the Pacific. University of Hawai‘i Press. Honolulu Malte Brun, C. 1813 Précis de la Géographie Universelle, tome IV. Paris. Markham, Cl. (ed.) 1904 The Voyages of Pedro Fernandez de Quiros, 1596 to 1606, vol. I, series II). London: The Hakluyt Society. Maroto Camino, M. 2005 Producing the Pacific. Maps and Narratives of Spanish Moerenhout, J.A. 1837 Voyages aux îles du Grand Ocean… Paris: reproduction of Princeps edition, librairie d’Amérique et d’Orient, Adrien Maisonneuve. Neyret, J. 1974 Pirogues Océaniennes. Assoc. des Amis des Musées de la Marine. Paris, France. Pâris, F.-E. n.d. [1843] Essai sur la construction navale des peuples extra-européens ou Collection des navires et pirogues construites par les habitants de l’Asie, de la Malaisie, du Grand Océan et de l’Amérique dessinés et mesurés pendant les voyages autour du monde de l’Astrolabe, la Favorite, et l’Artémise. Editions Arthus Bertrand, Paris, 2 vol. Taillemite, E. 2006 Bougainville et ses compagnons autour du monde 1766-1769. Paris: Imprimerie Nationale Editions. Sharp, A. 1963 Ancient Voyagers in Polynesia. Angus and Robertson: Sydney. Thomas, S. 1987 The Last Navigator. Henry Holt and Company. New York. Turnbull, D. 2000 (En)-Countering knowledge traditions. The story of Cook and Tupaia. Humanities Research 1: 5576.

Duperrey L.I. 1826-1830 Voyage autour du monde execute par ordre du roi sur la corvette de Sa Majesté la Coquille, pendant les années 1822, 1823, 1824 et 1825. Zoologie par MM. Lesson et Garnot, 2 vols. In-4°. Paris: A. Bertrand Ellis, W. 1972 (1829-1831) Recherches Polynésiennes, 2 vols. Société des Océanistes 25. Paris: Société des Océanistes. Finney, B. 1979 Hokule‘a: The Way to Tahiti. Dodd, M., New York. 1998 “Nautical Cartography and Traditional Navigation in Oceania”. In D. Woodward and G. M. Lewis (eds.), The History of Cartography. Cartography in the Traditional African, American, Arctic, Australian, and Pacific Societies, vol. 2, Book 3, pp.443-494. Chicago, London: The University of Chicago Press. Gladwin, T. 1970 East is a Big Bird. Navigation and Logic on Puluwat Atoll. Harvard University Press, Cambridge. Haddon, A. C. & Hornell, J. 1936-38 Canoes of Oceania. Bishop Museum Special Publ. 27, 28, 29. Reprinted as one vol. 1975. Bishop Museum Press, Hololulu, Hawai'i. Hawkesworth, J. 1773 An account of the voyages ---Byron, Wallis, Carteret and Cook. 3 vols. London, Strahan & Cadell. Henry, T. 1962 Tahiti aux temps anciens. Tran. By B. Jaunez. Société des Océanistes 1. Paris: Société des Océanistes. Hezel, F. X. 1983 The First Taint of Civilization: A History of the Caroline and Marshall Islands in Precolonial Days, 1521-1885. Honolulu: University of Hawai’i Press. Irwin, G. 1992 The Prehistoric Exploration and Colonisation of the Pacific. Cambridge University Press: Cambridge. Kotzebue von, O 1821 A Voyage of Discovery into the South Sea and beering’s Straits… in the Years 18151818, 3 vols., trans. by H.E. Lloyd. London: Longman, Hurst, Rees, Orme, and Brown. Lapérouse, J-F. de n.d. Voyage autour du monde sur l’Astrolabe et la Boussole (1785-1788). Edition by M. de l’Ormerais. Paris. Latour, B. 1983 “Comment distribuer le grand partage”. Revue de Synthèse 3e série: CX: 202-236.

8

culture areas from which broader assumptions regarding voyaging practices are frequently inferred2. The particularities arising from authorship also extend to the presence of those Pacific islanders who became informants for the early observers of Oceanic technologies but whose apparent expertise or social standing may not always have been strictly representative of the core knowledge practices of their maritime region. More importantly, there has often been an underlying instrumentalist bias in the study of Oceanic watercraft that fails to recognise - and analytically disassociate itself from - the rationalist, post-Enlightenment prejudices of early Euro-American travellers whose interests leaned heavily towards form and function while marginalising the intricate value systems and local ontologies within and through which oceangoing vessels were constructed, used and conceptualised.

Chapter I Early European Descriptions of Oceanic Watercraft: Iberian Sources and Contexts By Carlos Mondragón* and Miguel Luque Talaván** Early Sources for the History of Oceanic Navigation When it first became a domain of specialist interest, research about Oceanic watercraft was directed to issues of geographical distribution and variety of form, i.e. styles and construction technologies (Haddon and Hornell 1975, Neyret 1974). Eventually, typologies of form which are directly relevant to the history of migration, movement and cross-cultural borrowings - gave rise to investigations that focused on function, with special reference to navigational techniques and maritime practices in specific localities (e.g. Gladwin 1970, Lewis 1994, Finney 1979, 1996). More recently, approaches to Pacific navigation have begun to incorporate temporal, social and ecological factors that are refining our understanding of long-term processes, transformations and the broader social significance of watercraft in island communities (e.g. Munn 1986, Irwin 1992, Kirch 2000, D’Arcy 2006). However, this accumulated body of research suffers from an important limitation that is directly relevant to the long-term historical reconstruction of Oceanic watercraft and maritime knowledge, namely a surprising dearth of historical descriptions regarding indigenous navigation systems and technologies.

One relevant effect of the problems mentioned above is that the study of ‘navigation’ in the Pacific has often been rendered as a seemingly universal - by which read culturally homogeneous - phenomenon of primarily technical nature that is characterised by ‘localised manifestations’ into which existing typologies and evolutionary progressions can be conveniently inserted. The overall result of these limitations has been the establishment of a circular logic in which the study of navigation becomes an exercise in addressing technical lacunae in the incomplete historical, archaeological and ethnographic records while effectively marginalising the multifarious value systems, ecological transformations and ontological nuances that are inherent to the existence of watercraft and traditional maritime knowledge (e.g. Feinberg 1988, Feinberg et al. 1995). The purpose of the present chapter is to present and discuss some of the earliest European descriptions of Oceanic maritime technology while exploring hitherto overlooked aspects of the context in which these sources were produced. The sources for these materials are of

The paucity of the historical record has had important consequences for the way in which traditional navigation is characterised and understood. An important example of this can be seen in the elaboration of long-term chronologies, in which the evolution of pre-modern voyaging is abstracted from a limited pool of data and often elides the particular circumstances within which specific descriptions arose (e.g. the style and biases of specific authors and contrasting value systems during cross-cultural encounters, all of which influenced the manner in which particular descriptions were constructed)1. For instance, it is well known, and the seminal overviews proffered by Lewis (1994) and Finney (1998) bear this out, that two specific regions, namely the Western Carolines and Tahiti, have become exemplary

2

In the case of the Carolines, a wealth of information was originally gathered in the early twentieth century by Wilhelm Müller in Yap and its immediate context (cf. Alkire 1984); this was later greatly supplemented by the in-depth studies of Gladwin (1970), Lewis (1994) and, more recently, D’Arcy (2006), whose cumulative insights allow a highly detailed reconstruction of navigational systems in Western Micronesia from the late eighteenth century to the present. Interestingly, Lewis has observed that, “The most detailed facts about a Pacific orientation system come from the Caroline Islands in Micronesia, where it can be studied in operation today. Unfortunately, no systems of comparable sophistication have survived, or have even been described, elsewhere in Oceania.” (1994: 167). By contrast the Tahitian data were first gathered in the eighteenth century as a result of Cook’s now famous interactions with Tupaia (cf. Salmond 2004; Di Piazza and Pearthree 2007), and while they represent one of the most prolific sources for the analysis and revival of navigation in Eastern Polynesia, they are nevertheless more fragmented than is the case with the Carolinian material.

*

El Colegio de México, [email protected] Universidad Complutense de Madrid 1 Inadequate attention to historical context is all the more evident in this case in light of the fact that it has long been addressed as a key subject of debate and theoretical elaboration by social and cultural historians working on a broad range of topics in Pacific historiography. **

9

CARLOS MONDRAGÓN AND MIGUEL LUQUE TALAVÁN

origins, which, as we will argue, may indicate the existence of ongoing contact between some of the Solomons outliers and far-flung maritime communities of Western Polynesian extraction during the early seventeenth century.

Early Modern Iberian authorship - i.e. written by late Renaissance Portuguese and Spanish observers - and appear in personal journals and maritime diaries that were composed between 1567 and 1606 during the course of three expeditions that sailed to the South Pacific from the Viceroyalty of Peru. The first two of these voyages were commanded by Álvaro de Mendaña y Neira, a Spanish notable who visited the Marquesas and Solomon Islands in 1568 and 1595, while the third voyage was commanded by Pedro Fernández de Quirós, a Portuguese navigator who first visited the South Seas as principal navigator for the second Mendaña expedition and subsequently returned to the Solomons, and ventured into north Vanuatu, while commanding imperial Spain’s last foray into the South West Pacific3.

Of Pig Tusks and Canoes In early 1568, the Spanish aristocrat Álvaro de Mendaña y Neira spent several months exploring a previously unknown archipelago in the South West Pacific that he named the Solomon Islands4. The impressions of the Spanish seamen about the people and landscapes of the region which they examined were registered in several accounts, the longest and most widely known of which was penned by Hernán Gallego, who was the pilot major (piloto mayor, chief navigator) of the expedition5.

However, the principal contribution of this chapter is not simply to offer a selection of scattered documents for the historical reconstruction of Oceanic navigation. Rather, it seeks to emphasise the ambiguous acts, contradictory motivations and frequently incommensurable value systems that informed the culture contacts that gave rise to the first European descriptions of Oceanic watercraft. In this respect, we hope to contribute to the ongoing revision and refinement of scholarly understandings and reconstructions of the Oceanic past, especially in relation to the period just prior to and following first contacts with European peoples (cf. Kirch and Rallu 2008).

In his chronicle Gallego describes various encounters with coastal communities from Santa Isabel and the surrounding areas. One of these contacts took place in May of 1568 on Furona, named San Jorge by the Spaniards, a large island located near the south west tip of Santa Isabel: […] at a distance of five leagues there was an island which the naturals [i.e. indigenes] call Varnesta, and its chief, Benebonafa. This island creates a channel with Santa Isabel […] running from Southeast to Northwest, in which there is a village with over three hundred dwellings. There were many pearls in this island, and the Indians do not give them much heed; they offered them all as ransom for a canoe [canoa in the original] which had been taken from them; they also brought some

In the sections that follow, readers will notice an obvious contrast between the content and length of the first and second excerpts presented below, and their accompanying analysis. The first of these texts consists of a succinct account about a brief crisis that was precipitated by Spanish seamen after their forceful seizure of an indigenous vessel on the island of Furona in the central Solomons (Figure 1). While this excerpt contains no details regarding the form and function of the vessel in question it offers a useful point of entry for an exploration of the possible social worlds, values and acts that surfaced in respect to Oceanic watercraft during the earliest encounters between Islanders and Europeans. By contrast, the second text offers a unique report, rich in technical detail, of a large voyaging canoe that was sighted in the South East Solomons by the Quirós expedition of 1606. In this case the meticulous attention given to the form and features of the indigenous vessel provides a singular opportunity to pinpoint its cultural

4

The name was in lyrical reference to the fabled land of Ophyr, from which King Solomon extracted his legendary trove of treasure - the equivalent of which Mendaña hoped to find and extract from the Austral regions of the Pacific. 5 The best known version of Gallego’s personal account - whose authorship is commonly attributed by mistake to Mendaña or Quirós - is a manuscript version that was transcribed in the eighteenth century from the original documents written by Gallego, which are now lost. This eighteenth century transcript is kept at the Real Biblioteca del Palacio Real in Madrid and provided the basis for subsequent lengthy and often incomplete versions of Gallego’s chronicle that were put together with the account of Quirós’s 1606 voyage. This compilation was eventually published as a single set of documents, see Justo Zaragoza (1876-1880), Historia del descubrimiento de las regiones austriales, hecho por el general Pedro Fernández de Quirós, 2 vols., Madrid: Imprenta de Manuel G. Hernández (a single volume facsimile of Zaragoza’s edition was recently published under the imprint of José Manuel Gómez-Tabanera, Madrid: Dove, 2000); in turn, the Zaragoza edition served as the basis for the English translation by Markham etc., but was revised, expanded and published as Pedro Fernández de Quirós (2000), Descubrimiento de las regiones austriales, with an introduction by Roberto Ferrando Pérez, Madrid: Dastin (Colección Crónicas de América 12). The materials from Gallego which we present in this section have been quoted and translated directly from the eighteenth century manuscript kept at the Palacio Real in Madrid.

3

These voyages are well known to Pacific historians and have been the subject of numerous studies (e.g. Jack-Hinton 1969, Hilder 1980, Spate 1979, Baert 1999, Luque and Mondragón 2005, Maroto Camino 2005, Angleviel et al. 2007); however, the set of excerpts offered below represent our particular selection, transcription and translation of various original manuscripts that are kept in Spanish and Portuguese archives and had been previously published in English but suffer from inadequate translation and have generally been inaccessible or marginalised in previous research about the early history of European culture contact in the Pacific.

10

EARLY EUROPEAN DESCRIPTIONS OF OCEANIC WATERCRAFT: IBERIAN SOURCES AND CONTEXTS

teeth, which seemed to be from a great animal; these they valued greatly, and said [to the Spanish] to take them and go back to their canoe [i.e. the patache, or launch, in which the Iberians had arrived] (Quirós 2000: 48-49).

absence of technical information regarding this ‘canoe’ it is worth pursuing the available historical, linguistic and ethnographic data about the region in which Furona is situated with a view to exploring some of the possible values and motivations that may have guided the actions of the ancestral inhabitants of this community during their encounter with the Iberian seamen.

Gallego provides no further details about the ‘canoe’, nor about the inhabitants of Furona and the circumstances under which their vessel was seized. However, in the

Figure 1: Island Melanesia, with the location of Taumako and the Central Solomons. contemporary social map of Santa Isabel and Furona - as reconstructed with particular reference to the anthropological writings of Geoffrey White - indicates that its pre-European coastal peoples inhabited a highly diverse sociolinguistic environment that was rich in interisland exchange, trading and raiding activities7.

In view of current demographic and settlement patterns, it is plausible to imagine that the Furonese may have belonged to an ancestral Bughotu language community with relations to south and central Santa Isabel6. The complicated history of internal migrations and forced displacement in this region makes it impossible to ascertain the precise origin of the coastal people who engaged with the Mendaña expedition on Furona. (According to at least one written account from the first period of sustained contact with Euro-American traders and missionaries, by the mid-nineteenth century “the infertile and deserted island of San Jorge” was no longer inhabited; cf. Jackson 1975: 69). However, the

7 A present-day snapshot of the linguistic differentiation that has characterised this region was produced by White during his first approach to Santa Isabel, in 1975, and is worth quoting here at length: “The sparse population [of Santa Isabel] belied the distance traveled in terms of linguistic diversity. Between our departure point in Bughotu [on the south east corner of Santa Isabel] and our subsequent arrival in Buala, thirty-five miles up the coast, we would traverse regions of three distinct language groups (of four spoken on the island); Bughotu spoken by almost one-fourth of the island’s population; Gao by a diminishing six percent; and Cheke Holo by over half of the population […]. Furthermore, we crossed a major linguistic divide, separating two language families located in the central and western parts of the Solomons archipelago. The Bughotu language is similar to languages spoken in Nggela, Guadalcanal and Malaita; whereas Cheke Holo, Gao and the fourth Isabel language, Zabana or Kia, are more closely associated with languages in the western Solomons” (White 1991: 23).

6 Geoffrey White believes that the very first islanders with whom the Spaniards encountered, on the east coast of Santa Isabel, belonged to the Cheke Holo family of language groups. At present, the Cheke Holo represent the majority language group in Furona itself, apparently as a result of recent migration, and are also present along much of central and north east Santa Isabel. Cheke Holo is subdivided into two groups, Maringe and Hograno, which are apparently the language of choice for trade in Santa Isabel and neighbouring islands; however, Cheke Holo differs markedly from Bughotu, and it is this division that currently marks the great divide between North West and SouthCentral Solomons languages, cf. White (1991), Gordon (2005).

11

CARLOS MONDRAGÓN AND MIGUEL LUQUE TALAVÁN

the other appealing to those powerful, incomprehensible and unpredictable foreigners to return to their boat and depart.

Importantly, these three forms of interpenetration would have relied on specific types of watercraft, which were likely to have been a special source of wealth and value. As White has observed, in reference to the ‘war canoes’ that were specific to headhunting expeditions: “In this complex of trading, raiding and ritualizing, the canoe was a highly strategic device that also acquired considerable symbolic significance. Canoes allowed the most intrepid raiders to travel across long distances, and so move anonymously to seek wealth outside the encumbrances of local networks of obligation or intrigue. The large war canoes were both a means to wealth and a symbol of power, particularly spiritual power that could be evoked and enhanced through ritual propitiation” (1991: 87).

As regards the ‘teeth’ on offer, they appear to have been circled pig tusks of the large spiral shape that represent an object of considerable value within the broader regimes of prestige and ritualised exchange that underpin the graded societies of the Solomons and Vanuatu9. It is well known that pig’s tusks function as markers of rank and power, which in the Solomons, and parts of North and Central Vanuatu is often related to the notion of generative potency, or mana10. But whatever the particular forms of social value that they carry, the point is that circled tusks are embodiments of social relations and power. Hence, whether the ‘teeth’ of ‘some great animal’ that were presented to the European seamen were pig’s tusks of the full-circle variety or belonged to another large animal such as a whale, they were clearly artefacts of considerable exchange value and their unilateral presentation constituted a significant ritual concession.

In view of the need to put values and processes in proper context, it is relevant to point out that the ‘considerable symbolic significance’ of ‘war canoes’, as described above by White, is a phenomenon that was exacerbated as a result of the increased cycles of ritualised violence that followed the first long-term Euro-American intrusions into the central Solomons during the nineteenth century (cf. Bennett 1987: 33-38). Nevertheless, the point here is to register the intense and ongoing ties of interdependency that have bound coastal societies across the Solomons since well before the sixteenth century, and that in the past these connections were sustained by a wide range of indigenous vessels. As is the case in other Pacific contexts, these vessels represented objects of great value, not only as a result of the significant effort and time that were invested in their construction, use and maintenance, but also in terms of the broader social, ecological and ritual associations that they embodied.

While the specific rationale that led to the offering of the ‘teeth’ must necessarily remain obscure, one further observation regarding the life-worlds of sixteenth-century Furona may help to shed light on the nature of indigenous maritime activity, and the form in which the islanders 9

The importance of pigs across Melanesia is well known and has been elaborated upon in numerous studies (e.g. Jolly 1984; Rodman 1996). Nevertheless it is worth noting that the region where the above contacts took place offers one of the highest levels of ritual elaboration in respect of pig exchange and the controlled flow and display of spiral tusks. What William Rodman has written about Vanuatu can equally be applied to the central Solomons: “[…] tusks are a product of culture rather than a given of nature. The magnificent tusked boars used in high-ranking and other ceremonies […] would not develop their tusks without human intervention in natural biological processes. […] tusks are a sign of the leader and a substance on which leadership is based. Europeans tend to think of the value of a pig as residing exclusively in its flesh; pigs are walking pork chops, protein on the hoof. Ni-Vanuatu see things differently. Most would tell you that, indeed, pigs are good to eat, but they also would say that the most valuable pigs are not the biggest, but those with the best-developed tusks. […] It takes seven or eight years for a boar to develop full-circle tusks. Relatively few pigs attain this ideal” (Rodman 1996: 160). 10 Mana is a pan-Oceanic term that is extant in the Bughotu language of central and south east Santa Isabel, although the neighbouring Cheke Holo languages of central Santa Isabel use a different term, nolaghi, to denote the same phenomenon (White 1991: 38). This form of productive capacity is primarily related to successful horticultural production and the nurture of immediate kin, but as White explains, “Beyond the narrow functionality of bountiful gardens and the like, maintaining a ritual dialogue with ancestral spirits was the primary means of acquiring spiritual power (nolaghi, mana) - the ultimate source of personal strength, well-being and success in wordly affairs”. (ibid.). For further discussions about mana see Keesing (1984), Shore (1989), Kirch and Green (2000).

That the ancestral inhabitants of Furona attributed significant value to the ‘canoe’ that Gallego mentions is not at issue. This is despite the fact that the precise nature of the vessel - was it a simplified outrigger, a dugout employed to ferry cargo, or a sophisticated ‘war’ canoe of the kind described by White? - will remain forever unknown. However, the ritualised exchange by which its ‘ransom’ was secured invites additional commentary. The key act in the exchange that took place on Furona was not the offering of pearls, whose value to the intruders may have been evident to the islanders from the outset, so much as the ‘teeth’ from ‘some large animal’ which were eventually brought out and offered to the Spaniards8. Whatever the particular circumstance that led to the offering, its main purpose appears to have been twofold, if one is to go by Gallego’s account: on one hand, it clearly aimed to propitiate the Spaniards into restoring the vessel that they had forcibly taken, while on 8

“It is likely that first encounters with European explorers produced anomie, in which conventions were quickly suspended or abandoned due to mutual incomprehension.” Meleisea and Schoeffel (1997:130).

12

EARLY EUROPEAN DESCRIPTIONS OF OCEANIC WATERCRAFT: IBERIAN SOURCES AND CONTEXTS

superstitious amazement of primitives. Although this presumption has gradually been discarded from the scholarly literature on ‘first contacts’, it continues to inform popular ideas about civilisation and backwardness, particularly in cultural regions that are characterised by small-scale agrarian societies, such as is the case for much of rural Melanesia12.

chose to react (i.e. ritualised exchange) when faced with unprecedented intruders. In respect to watercraft, it is possible to reflect on the fact that the Furonese may initially have associated the Spanish seamen with the Polynesian world, whose oceangoing vessels and regular approximation routes to Santa Isabel and Furona mirrored those of the Spaniards11. In effect, the size and form of the Spanish caravels may not have been completely alien to the Furonese, given the broad comparisons that could be drawn between a European caravel and a Polynesian waka - both being large oceangoing vessels equipped with sails and rigging. Despite the lack of hard evidence to back such a claim, this possibility is quite plausible in view of the links that bound Furona to the rest of the central Solomons and beyond. Four hundred years ago the peoples of the south and central Solomon archipelago were in regular contact with several Polynesian Outliers, and perhaps even some parts of Western Polynesia (Tonga, Samoa and Fiji, whose inhabitants had raided or colonized islands such as Anuta, Tikopia, Taumako, Bellona and Rennell). Hence, the Furonese may have had at least some previous experience of fair-skinned seafaring peoples and would therefore not have found the European intruders to be completely unfamiliar. Moreover, prior to their arrival in the Solomons, Mendaña’s ships had stumbled into Ontong Java, 250 kilometers to the north of Santa Isabel. According to Gallego’s account, the Polynesian inhabitants of Ontong pointed out the route that Mendaña should take in order to arrive at the Melanesian archipelago. Thus, in their approach to the east coast of Santa Isabel the Spanish ships first appeared to local inhabitants as large oceangoing vessels travelling precisely along that route.

At the Crossroads of the South West Pacific: Taumako and the Western Polynesian World One of the most peculiar canoe types found in the Pacific is known as ‘canoe of Lata’, or te puke in the language of the tiny Polynesian Outlier of Taumako, the largest of the Duff Islands in the South East Solomons. For a long time, perhaps centuries, the inhabitants of Taumako have been the exclusive guardians of the knowledge practices involved in constructing te puke. According to a recent study, “[m]ost canoes of this description have been constructed on Taumako on order from other communities in the Santa Cruz Islands and elsewhere in the Solomons” (Feinberg and George 2007: 3)13. The outstanding features of the te puke are its unique submerged hull and ‘slender foil’ or ‘delta wing’ sail,14 a recent description of which will serve as a contrast for the historical excerpt that follows: “Lata, a widely recognized Polynesian culture hero, is said to be the first person to build and sail a voyaging canoe of the type known as te puke. These are built with a roundish hull that is hollowed out from a single log. The top is boarded over and made watertight, except for the area enclosed by a riser box, which provides access for bailing and stowage of nonwater-sensitive cargo and ballast. When under sail the hulls are almost entirely submerged. The crew stays on top of one of two platforms that straddle the riser box and outrigger beams. Although there is one outrigger float (ama) per canoe, it may be fashioned from as many as four float timbers lashed together, depending on how

Although the implications of this scenario must necessarily remain speculative, they remind us that there are relevant alternatives to the tired but long-lived theories that hold that Pacific Islanders in first contact situations inevitably held European strangers to be supernatural and superhuman beings. Elsewhere I have argued that it is necessary to make a distinction between the idea that European intruders in early encounters were perceived as gods or ancestral beings, that is, spirits of a deific and superhuman nature (Mondragon 2006). This idea finds its origins in a centuries-old Euro-American presumption which holds that the sudden appearance of a ‘superior’ race amongst a technologically and politically ‘backward people inevitably provokes the naïve

12 The relevant point here is that even though these associations had some basis in certain encounters (e.g. some first encounters documented during the 1930s in the highlands of Papua-New Guinea (Connolly and Anderson 1987, Edward L. Schieffelin and Crittenden 1991), they did not necessarily erode local peoples’ ability to perceive themselves as capable agents throughout the course of any given encounter. 13 Exclusivity of construction and distribution/exchange rights for particular canoe types is not uncommon and occurs frequently in this region; e.g. Judith Bennett has observed that “[b]efore regular contact with Europeans, the inhabitants of Santa Ana, Santa Catalina, and the Surville Peninsula of San Cristobal [Makira] had obtained their large plank-and-rib canoes from other areas. Because the large canoes were expensive to purchase, smaller outriggers were made for everyday use” (Bennett 1987: 37). 14 Until recently this sail type was commonly known as a “crab claw” sail; however, the term “crab claw” denotes an asymmetrical shape, while recent research has determined that the delta wing model is in fact symmetrical (Mimi George, pers. com.).

11

In turn, it is likely that the people Furona had advanced notice, by way of those of their Cheke Holo and Bughotuspeaking neighbors who first encountered and had to deal with the Spaniards many weeks prior to their arrival on Furona, of the presence of aggressive and unpredictable pale seamen coming from Ontong Java. In response to the worrisome actions of the Spaniards along the east coast of Santa Isabel, where they had murdered several islanders and plundered some villages.

13

CARLOS MONDRAGÓN AND MIGUEL LUQUE TALAVÁN

much flotation is needed. The crew situates itself on whichever platform is most appropriate for sea conditions and the trim of the cargo. The predominantly submarine hull is subject to less surface tension and wave action than more conventional hulls, making them considerably faster. To the best of our knowledge, this hull design is unique to Taumako vessels. Like the vessels of the central outlier atolls, as well as some from the Polynesian heartland (e.g. the Tongan kalia), te puke are equipped with shelters on the outrigger platform. [Furthermore] Taumako sails are made from panels of plaited pandanus mats, sewn together vertically to form a graceful, slender foil, or delta wing shape. […] The design resembles ‘crab-claw’ sails of Tahiti and Hawai’i as depicted by artists on early European voyages of exploration, but its symmetrical shape, with long, slender, arcing wingtips, has few parallels in Oceania” (Feinberg and George 2007: 2-3)15.

construction and navigation in Taumako. The insights gleaned from these efforts appear to substantiate the possibility that te puke have been present in Taumako for the past two or three centuries17. However, the evidence collected thus far does not allow for a detailed reconstruction of the long-term evolution of canoe types in this locality. Moreover, as is the case with most other voyaging revival processes in the Pacific, contemporary claims regarding the exclusivity, uniqueness and antiquity of local knowledge must be approached with a critical view to actual historical evidence (cf. Finney 2003). In the case of the South East Solomons, these claims, and the scanty evidence available for the reconstruction of long-term trends in maritime knowledge practices underscores the fact that if ever there were a specific community for which it would be fortunate to possess accurate pre-modern data points in respect of canoe design, it would have to be Taumako.

Further information documented by George in collaboration with various local knowledge experts,16 suggests that the inhabitants of Taumako asserted a degree of control over the manufacture and dissemination of te puke-type watercraft across their local region in the recent past (Figure 2). Whether this has been a centuriesold practice is impossible to tell, but it is apparent that they currently perceive themselves to be the exclusive holders of the technical expertise and social values associated with the construction and use of this canoe type.

As it happens, the most significant description of an Oceanic vessel to arise from the early Iberian exploration of the Pacific was produced four centuries ago in this very island, but has not hitherto been the subject of specific analysis. It is to this unique historical description that we now turn.

The fact that the hull shape and sails of the te puke may be unique to Taumako has various implications for the history of maritime technology in the South East Solomons and beyond (Figure 3). On one hand, the obvious specificity of a given design invites questions regarding the length of time that it may take for particular elements to arise in relation to a canoe type, and subsequently change to the point where cumulative modifications give rise to a distinctively new kind of vessel. On the other hand, there is the matter of local knowledge, the evolution of inter-island networks and regional patterns of technological borrowing. How long have the submarine hull design and crab-claw sail of te puke been present in Taumako? More broadly, what can they tell us about the history and geography of cultural exchange in this region that sits uneasily at the frontier of several paradigmatic culture areas - namely Island Melanesia, Western Polynesia and Micronesia?

Figure 2: Two views of contemporary Taumako te puke. Both images by courtesy and © Marianne George and The Vaka Taumako Project.

In recent years there has been a small but valuable research initiative dedicated to understanding and supporting the building practices and cultural values associated with the indigenous revival of te puke

17 This research has been informed by the significant field experience in the South East Solomons of the anthropologist Richard Feinberg, and his recent fellow researcher Marianne George, who currently heads an ambitious recovery and revival effort known as the Vaka Taumako Project (cf. Feinberg 1995, Feinberg and George 2007, see also the VTP website, www.vaka.org, which constitutes the largest concentrated database on te puke design, building practices and cultural values).

15 I am especially grateful to Rick Feinberg and Marianne George for allowing me to quote from this advanced draft of their ongoing work on the te puke of Taumako. 16 Especially the te aliki Kaveia (Mimi George, pers. comm.).

14

EARLY EUROPEAN DESCRIPTIONS OF OCEANIC WATERCRAFT: IBERIAN SOURCES AND CONTEXTS

Figure 3: Idealised rendering of a Taumako te puke in motion. Note the submerged hull in contrast with the ama (outrigger float). Image courtesy and © Herb Kawainui Kane. In April of 1606, three ships commanded by Pedro Fernández de Quirós unexpectedly came upon the Duff Islands while vainly searching for the large island of Santa Cruz (Ndeni), which Quirós had previously visited when acting as chief pilot for Mendaña’s second Solomons expedition in 1595. Seeing that Taumako was the closest and largest of the visible islands in this group,

the Portuguese navigator ordered his men to drop anchor near to what he described as a “village on a small reef outcropping” (Quirós 2000: 235), which might perhaps correspond to the present-day small island of Tahua, which lies within the barrier reef off the South East coast of Taumako (Figure 4, 5).

Figure 4: “Sketch of the Duff’s Groupe”, published in A Missionary Voyage to the Southern Pacific Ocean, Performed in the Years 1796, 1797, 1798 in the Ship Duff commanded by Capt. James Wilson (London: T. Chapman, 1799).

15

CARLOS MONDRAGÓN AND MIGUEL LUQUE TALAVÁN

The ensuing cultural encounter with the local islanders lasted nine days. During this time the Iberians and the people of Taumako - represented by a chief identified as both Tumai and Falique, this last most likely being a corruption of the title ta ariki - played out a series of tense and asymmetric exchanges that were dominated by the Europeans’ coercive attempts at replenishing their onboard stocks of food and fresh water18. Eventually, to the islanders’ relief - as perceived quite explicitly by at least two of the Spanish seamen, namely Prado and Bermúdez - the Iberians departed after having secured their desired provisions, in addition to four Taumako boys who were kidnapped with the intention of eventual display at the imperial court in Madrid19.

Specifically, two important accounts of the encounter on Taumako have survived: they are Quirós’s official logbook, which was actually written and kept by his personal scribe, Luis de Belmonte Bermúdez, and the diary of Don Diego de Prado y Tovar, a Spanish aristocrat and military engineer whose artistic skill and trained eye gave rise to the following exceptional description, which we quote in full, with lengthy comments in respect of the measures that Prado employed and their contemporary metric conversions: “In this village we found some embarkations capable [of carrying up to] sixty persons, with which they navigate, very different they are from those of Europe and each one of them demands as much sail as one of our own vessels; they are [built] as follows: resting atop two thick logs [vigas, understood here as the two tree trunks that constitute the basic double hull catamaran structure of the vessel that he is describing], each log has a length of sixty feet [“pies”, approx. 16.68 mts] and a perimeter of ten and eight, by which I mean the circumference of each log [i.e. 18 feet, approx. 5 to 5.82 mts. in circumference]; above these [logs] they construct the body of their embarkation, which is made up of twelve beams [“latas” in the original Spanish, meaning slender beams],20 each of which lies [rests horizontally] from one log to the other, and each of which is thirty feet long and is bound to each of the two logs; each beam has a circumference of half a length [“media vara”, half a Castilian yard, approx. 417 mm]; above the beams they place an arched structure [made up of] thirty sturdy wooden poles, each of which is one palm’s length in thickness [approximately one hand’s length], such that they resemble a half-dome which is held together by eight poles that run horizontally from one side to the other [of the structure] and [each of] whose thickness and circumference measure one half of one foot [“medio pie” = 139 mm] and one foot and a half [417 mm] respectively [Prado refers to the circumference and diameter of the eight poles that act as roof beams holding the domed structure together]; these are tied with strong cords made from the fibres of the coconut shell; this arch is around ten and six [16] feet high; and above this domed structure they place a further quantity of slender sticks which are configured into a square shape that makes up the upper deck [“combés”] of the vessel on which those who must tend to the sail can travel. The floor [of this upper deck] is constituted by a strong layer of cane [Prado probably refers to bamboo] which is very well fastened and fashioned with a trapdoor by which people can enter and depart [from the inside of the arched structure below]; in the middle of the arched structure they place a further [smaller] table made of poles in the manner of planking, upon which they carry their food and drink, [this last of] which they decant inside hollow canes, each of which carries up to half a sack of liquid [“arroba”, the official

Setting aside the violence and repercussions of the unsettling presence of the Iberians on Taumako, the length of time during which they remained in close proximity to the local community - approximately twelve days - prompted some of the European seamen to produce extensive and detailed descriptions about the indigenous technology and material culture which they were able to observe.

Figure 5: Contemporary topographic chart of Taumako. 18 Quirós had been navigating blindly on dwindling rations through the Eastern and Central Pacific for several months, and by the time they disembarked in Taumako his crew were in no mood for exercising politesse with savages. See Luque and Mondragón (2005) for a detailed description of the criteria that led Quirós to trace the route which eventually took his ships to the Duff Islands. 19 Two of the prisoners quickly jumped ship within sight of Taumako, while a third escaped a few days later when the Spaniards sailed past Tikopia; a fourth Taumako youth remained aboard, was baptized as Pedro, and made it as far as Mexico City before succumbing to death.

20

At this time lata (also bao) was equivalent to an English “beam”, according to O’Scanlan de Lacy 2003: 333, and Appendix: 66) and the Vocabulario Maritimo of 1722.

16

EARLY EUROPEAN DESCRIPTIONS OF OCEANIC WATERCRAFT: IBERIAN SOURCES AND CONTEXTS

The exceptional detail that Prado provided in respect of the form, dimensions and individual components of the vessel that he observed invites detailed commentary, not least because it appears to point to a type of voyaging canoe that was rather different from the te puke, but no less surprising in terms of its possible origins and significance for ancestral connections and technological borrowings across the Solomons outliers during the early seventeenth century.

measure of the sacks in which cereals had to be transported to market in late sixteenth century Spain] and is very well fastened and placed on the beams which are in the water [Prado refers to the “latas”, or beams that lie horizontally across the two main trunks or hulls and make up the basic deck of the vessel], to which stones are also fastened in order to act as ballast; furthermore, from the logs in the water [the two outrigger trunks] long poles emerge which reach up the sides of the vessel, they being higher [than the floor of the upper deck] by one length [“vara”, or vara Castellana, approx. 835mm], and very well fastened, and upon these [the poles that stick up slightly from the upper deck] they secure their parapet from strong canes [Prado seems to refer to a reinforced handrail made of several long poles whose lower ends are fastened to the two hulls and whose upper ends are secured to the upper deck, and further reinforced with horizontal bamboo poles in order to form a guardrail which acts as protection for those standing on the uppermost part of the vessel]. The mast is made of three thick bambous, with a pulley and yard, and it is placed in the middle of the embarkation, which demands as much sail as our own vessel. On the extreme end of the [two hulls] trunks, the ends that lie on the poop of the vessel, they have high seats with two big oars placed above them, with which they govern the vessel. With the wind coming from the poop they must probably navigate well: if it is from the side they are worthless. They [the vessels] demand little water [i.e. not much depth in order to float] and have no need of a launch with which to disembark. Their bread is sahagu [sago], and further ahead [I] describe the manner of its manufacture. To the sides of the logs [hulls] they have some sticks and seats from which they row with canalets, which are a sort of small oar in the shape of paddles with which one can strike small balls; each is two lengths tall [“dos varas de largo”, approx. 1.67 mts] and is the same as those that are used all around these lands, from Borneo to the Malacas. (Prado, translated from old Castilian, as transcribed and presented in the appendix to Hilder 1990: 209)21.

First, however, it is necessary to clarify one of the greatest difficulties posed by Prado’s text, which arises from his peculiar style of technical description (for which see Mondragón 2007) and, most importantly, the criteria for the measurements that he employed. In effect, the size, shape and form (lengths and widths) that appear in the above excerpt were proper to a very specific moment in the historical record, namely the years 1590 to 1613, during which Imperial Spain first began to impose an official standard for measurements on all of the shipyards located within the Spanish peninsular territories. These measures applied equally to the written description of watercraft, and they appear to be the ones that Prado used, which is not surprising given the fact that he was trained as a military engineer, was clearly fastidious when it came to technical drawings and descriptions, and was very likely aware of the official standards for measurement required by the Spanish Crown. Prior to 1590 the two most common standards of measurement employed in the shipyards of Renaissance Spain hailed from the northern and southern coastal regions of the Iberian Peninsula. The northern shipyards, which lay along the coastline facing the Bay of Biscay, functioned according to the so-called Basque-Cantabrian atravesados en cuadrado que hacen el combez del navío donde va la gente que gobierne la vela. El suelo es una estera fuerte de cañas muy bien hecha con su postigo para entrar y salir la gente; en el medio de estos arcos tienen otros palos a modo de tablado donde llevan las comidas y bebida, la cual llevan dentro de cañutos que cada cañuto cabe media arroba de agua, muy bien atados y arrumados en las latas que están en el agua, atan algunas piedras para lastre; donde las latas de la agua salen unos palos hasta el combés, una vara mas altos, muy bien atados y en ellos arman su parapeto fuerte de cañas fuertes. El mástil es de tres cañas muy gruesas, con su polea y verga de caña, puesto en el medio de dicha embarcación que demanda tanta vela como nuestro navio. En los estremos de las vigas de la popa tienen asientos altos, con dos remos grandes, puestos en ellos, con que gobiernan el navio. Con el viento en popa caminarán bien; si es del lado a la bolina no valen nada. Demandan poco agua y así no han menester batel para desembarcarse. El pan de éstos es sahagu, como adelante se dirá de la manera que lo hacen. Por los lados de las vigas tienen unos filaretes y asientos de donde reman con sus canaletes, que son una suerte de remos hechos como palas de jugar a la pelota pequeña, de dos varas de largo cada una y de éstos usan por toda esta tierra hasta el Borneo e islas Malacas” (Prado, in Hilder 1990: 209).

21

In the interest of presenting the original Spanish version of Prado’s description to as wide a specialist audience as possible, we offer it herewith: “En este pueblo hallamos unas embarcaciones capaces de sesenta personas con que ellos navegan, muy diferentes de las de Europa y demandaban cada una de ellas tanta vela como nuestro navio; que son de la manera siguiente: sobre dos vigas gruesas, por dentro de cada una tiene de largo sesenta pies y de grueso diez y ocho pies, digo de cirumferencia cada una; sobre estas arman su embarcación, esto es doce latas atravesadas de una viga a la otra, que cada una es larga treinta pies y encajadas en las dichas vigas; y tiene de grueso cada una media vara; sobre estas vigas ponen unos arcos de palo grueso de un palmo cada uno y son treinta, a manera de una bóveda, con ocho palos atravesados por los lados, de medio pie de grueso y pie y medio de ancho; atados con fuertes cordeles de estopa de cáscaras de cocos; estos arcos son altos como diez y seis pies cada uno; sobre esta bóveda tienden otros palos más delgados

17

CARLOS MONDRAGÓN AND MIGUEL LUQUE TALAVÁN

system of measurement, which was the preferred system of measurement within the Spanish court prior to 1590. Those that were common to the south were associated

with the Seville-Guadalquivir (Atlantic) coasts and the broader maritime region of Andalucia.

Figure 6: Tongan kalia; note the double catamaran, the lower deck with shelter and the upper deck with reinforced railing, which is evocative of the vessel in Prado’s description. Image by courtesy and © Herb Kawainui Kane.

Figure 7: Earliest European rendering of a tongiaki, as depicted by Schouten and Le Maire in 1616. Note the long steering paddle being handled from the poop deck.

18

EARLY EUROPEAN DESCRIPTIONS OF OCEANIC WATERCRAFT: IBERIAN SOURCES AND CONTEXTS

described by Prado more closely resembles a Tongan kalia such as the one depicted in a recent idealised rendering by Herb Kane (Figure 6).

The growing confusion produced by these disparate systems in the context of imperial overseas expansion and increasing production of ships in late Renaissance Spain was first tackled in 1590, when King Phillip II attempted to unify Peninsular standards of measurement in his “Royal Decree [cédula real] of August 20, 1590”22.

The proximity of the double catamaran Tongan kalia pictured above with the one that was observed by Prado on Taumako is underscored, in particular, by the double catamaran structure sustaining an upper deck with reinforced railing. However, there is an additional feature in Prado’s description - the ‘high seats’ rising above the poop of the vessel, which he believed to function as a platform for the manipulation of two long steering paddles/rudders - that does not correspond to the overall design of a Tongan kalia. By contrast, these elevated platforms for steering paddles would seem to be analogous to an element that is present in another type of Tongan vessel known as a tongiaki (Figure 8, 9). What makes this canoe type particularly intriguing for the identification of the Taumako vessel is that there is documentary evidence for the active use of tongiaki in the Western Polynesian context only a few years after Quirós’s arrival in the Duff Islands. The source in question dates to 1616, when the Dutch voyagers Willem Cornelis Schouten and Jacob Le Maire encountered a canoe, identified as a tongiaki in the open sea to the north of Niatoputapu, between Tonga and Samoa. An engraving of this vessel, composed from spoken descriptions after the actual encounter, was published in 1619 in Schouten’s account of the voyage, and represents one of the earliest graphic renderings of an Oceanic voyaging canoe (Figure 7).

While it is beyond the scope of this contribution to provide detailed commentary regarding the contents of the Decree of 1590, several different historians have managed to accurately summarise the cédula’s complicated and multifarious conversions - which are essential to deciphering Prado’s description - as follows: • • • •

Pie de burgos [one foot]23 = 0.2786m / 278mm. Codo de ribera [one elbow’s length]24 = 0.5747m Vara castellana [one Castilian length] = 8359m / 835mm Toneladas [one tonne] = 1.518 m3.25

When these measures are applied to Prado’s 1606 text, it emerges that the dimensions of the vessel that he described are quite different from those of a Taumako te puke. Indeed, the most significant feature in the watercraft depicted by Prado, which is made clear from the outset of his text and determines the basic nature of the vessel in question, are the two ‘thick beams’ [vigas gruesas] that he identifies as constituting the main body and foundation for the rest of the features that they support - i.e. the lower deck, arched shelter, upper deck, mast. In effect, these ‘beams’ represent the structure of a double hulled vaka which, we would argue, appears to be most closely related to the vakalua (double hulled vaka) common to Tonga or the Solomons atoll of Sikaiana, whose presence in contemporary Taumako oral narratives points to the existence of past exchanges between Sikaiana and Taumako in direct reference to maritime knowedge and canoe building.26 Compelling as the Sikaiana data may be, however, on the basis of existing data we would posit that the double-hulled canoe

Conclusion Notwithstanding the strong resemblances in the features of both the Tongan kalia and the tongiaki in relation to the voyaging canoe described by Prado in 1606, it would be speculative, on the basis of existing evidence, to draw any direct identification between these three types of vessels. Nevertheless, two important conclusions can be drawn from this brief overview of Iberian historical sources.

22

This decree is known as the Cédula Real de 20 de agosto de 1590, and is discussed and reproduced in full in Casado Soto (1988: 84-94). 23 So called because it was a close adaptation of the pre-existing Cantabrian (hence Burgos) “foot”. 24 So called because it was also adapted from existing Cantabrian measures, hence the reference to ribera, or riverine deltas, which is a common term for bays and inlets along the North West Spanish coast (Cantabria, Asturias, Galicia). 25 See especially Casado Soto (1988: 58-72), and Rubio Serrano (1991: 93-96; 1998: 77), cf. also Veitia Linage (1672: 170, 182). 26 Specifically, Mimi George reports that “te aliki Kaveia and other elderly people from Taumako (mostly deceased in recent years) who have sailed extensively on te puke understood from their teachers that te puke design was an innovative improvement on vakalua […] that preceeded Lata’s invention of te puke.” (George, pers.com.).

First, that the canoe that the Spanish seamen encountered on Taumako was clearly not related to a te puke, insofar as Taumako te puke are known to us in recent times. Importantly, whatever the specific design and indigenous name given to the watercraft that Prado observed, it appears to have exhibited a number of basic structural features and elements (double hull, raised upper deck with railing, and steering paddles resting atop a poop deck) that strongly suggest a direct correlation with longdistance voyaging canoes of the vakalua type that were common in outliers such as Sikaiana and parts of Western Polynesia, particularly Tonga, during the early years of the seventeenth century. More importantly, the vessel described by the Spaniards may also have been of a kind in which elements of different canoes were present, but

19

CARLOS MONDRAGÓN AND MIGUEL LUQUE TALAVÁN

D’Arcy, P. 2006 The People of the Sea: Environment, Identity, and History in Oceania. University of Hawai’i Press, Honolulu. Damon, F. (in press)“On the Ideas of a Boat: From Forest Patches to Cybernetic Structures in the Outrigger Sailing Craft of the Eastern Kula Ring, Papua New Guinea”. In T. Kaartinen and C. Sather (eds.), Beyond the Horizon: Essays on myth, history, travel and society in honor of Jukka Siikala. Studia Anthropologica Fennica, vol. 2. Finnish Literature Society, Helsinki. Davenport, W. 1964 “Notes on Santa Cruz Voyaging”. Journal of the Polynesian Society 73: 134-142. 1968 “Social Organization notes on the Northern Santa Cruz Islands the Duff Islands (Taumako)”, Sonderdruck aus BaesslerArchiv, Beitraege zur Voelkerkunde 16: 137. Di Piazza, A. and E. Pearthree 2007 “A new reading of Tupaia’s chart”. The Journal of the Polynesian Society 116 (3): 321-40. Feinberg, R. 1988 Polynesian Seafaring and Navigation: Ocean Travel in Anutan Culture and Society. Kent State University Press, Kent OH. Feinberg, R. and M. George 2007 “Seafaring in the Polynesian Outliers”, in Helaine Selin (ed.), Encyclopedia of the History of Science, Technology, and Medicine in Non-Western Cultures, Second edition. Kluwer Academic Publishers, Amsterdam. Feinberg. R., ed., 1995 Seafaring in the Contemporary Pacific Islands: Studies in Continuity and Change. Northern Illinois University Press, DeKalb IL. Finney, B. 1998 “Traditional Navigation and Nautical Cartography in Oceania”, in G. Malcolm Lewis and David Woodward (eds.), The History of Cartography, Vol.3, part 2: Cartography in Traditional African, American, Arctic, Australian, and Pacific Societies. University of Chicago Press, Chicago. Finney, B. 2003 Sailing in the Wake of the Ancestors: Reviving Polynesian Voyaging. Bishop Museum Press, Honolulu. Gladwin, T. 1970 East is a Big Bird: Navigation and Logic on Puluwat Atoll. Harvard University Press, Cambridge Mass.

which is no longer extant, in which case it would constitute a fascinating exemplar of the ongoing transformations that have given rise to current canoe types across this region.27 In this respect, one of the most intriguing possibilities that arises from the above evidence is that at least one, and perhaps more, of the Polynesian outlier communities of the South East Solomons region were engaged in direct communication and technological borrowing with parts of the broader Western Polynesian world four hundred years ago. While they continue to be speculative, such a scenario may help to broaden the scope for discussion regarding the nature and historical development of interisland voyaging, borrowing and contacts in this fascinating maritime region of Oceania.

References Alkire, W.H. 1984 “Central Carolinian Oral Narratives: Indigenous Migration Theories and Principles of Order and Rank.” Pacific Studies 7: 2, 1-14. Angleviel, F., ed., 2007 Pedro Fernández de Quiros et le Vanuatu: Découverte mutuelle et historiographie d’un acte fondateur, 1606. GRHOC/Commission Européenne/République Française, Port Vila. Baert, A. 1999 Le Paradis Terrestre, un mythe espagnol en Océanie: Les voyages de Mendaña et de Quirós, 1567-1606. L’Harmattan, Paris. Bennett, S. 1987 Wealth of the Solomons: A History of a Pacific Archipelago, 1800-1978. Hawai’i University Press, Honolulu. Casado Soto 1988 Los barcos españoles del siglo XVI y la Gran Armada de 1588, Madrid: Editorial Naval. Colección de documentos y manuscritos compilados por Fernández de Navarrete (1971), Vol 23, Part 1. Kraus Thompson, Lichtenstein. Connolly, B. and R. Anderson 1987 First Contact: New Guinea’s Highlanders Encounter the Outside World. Viking, New York. 27

This intriguing possibility appears to be partially substantiated by a recent comment in which Mimi George reports that “Sikaiana people today told me that in their concept Sikaiana is the main hull and Taumako the ama. Sikaiana, they say, made and/or used vakalua, not te puke.” She further points out that this assertion is in contrast to a previous observation made by William Davenport in which “the close voyaging relationship between Taumako and Sikaiana is spoken of in terms of how Taumako was the main hull and Sikaiana was the ama.” (George, pers. com.; see also Davenport 1964).

20

EARLY EUROPEAN DESCRIPTIONS OF OCEANIC WATERCRAFT: IBERIAN SOURCES AND CONTEXTS

Figure 8: Two double-hulled tongiaki, moored and with their masts and sails lowered, at the bay in Tongatapu, Tonga, as pictured by Isaac Gilsemans, who travelled with Abel Tasman to the South Seas in 1643.

Figure 9: A contemporary rendering of a tongiaki from Tonga. Image courtesy and © Herb Kawainui Kane. Gordon, R.G. 2005 Ethnologue: Languages of the World, 15ª edición. Summer Institute of Linguistics International, Dallas. Haddon, A.C. and J. Hornell (1936-38 [1975]), Canoes of Oceania. Bishop Museum Press, Honolulu. Hilder, B. 1990 El viaje de Torres de Veracruz a Manila. Ministerio de Asuntos Exteriores, Madrid.

Irwin, G. 1992

The Prehistoric Exploration and Colonisation of the Pacific. Cambridge University Press, Cambridge. Jack-Hinton, C. 1969 The Search for the Islands of Solomon, 15671838. Clarendon Press, Oxford. Jackson, K.B. 1975 “Head-hunting and the Christianization of the Bugotu, 1861-1900”. Journal of Pacific History 10 (1): 65-78.

21

CARLOS MONDRAGÓN AND MIGUEL LUQUE TALAVÁN

Jolly, M. 1984

O’Scanlan de Lacy, T. 2003 Diccionario marítimo español. Facsimile edition. Originally published in 1831. Imprenta Real Museo Naval, Madrid. Purdue, B.K. 2002 “Tongan-European First Contact, 1616: Graphic Record to Scientific Image”. Bulletin of Tokyo Kasei Gakuin Tsukuba Women's University 6: 103-115. Rodman, W. 1996 “The boars of Bali Ha’i: Pigs in paradise”. In J. Bonnemaison et al. (eds.), Arts of Vanuatu. Crawford House Publishing, Bathurst, NSW. Rubio Serrano, J.L. 1991 Arquitectura de las Naos y Galeones de las Flotas de Indias (1492-1590), vol. 1. Ediciones Seyer, Málaga. 1998 “Las unidades de medida españolas en los siglos XVI y XVII”. Revista de Historia Naval 20. Schieffelin, E., R. Crittenden et al. 1991 Like People You See in a Dream: First Contact in Six Papuan Societies. Stanford University Press, Stanford. Serrano Mangas, F. 1992 Función y evolución del galeón en la Carrera de Indias. MAPFRE, Madrid. Shore, B. 1989 “Mana and tapu”. In A. Howard and R. Borofsky (eds.), Developments in Polynesian Ethnology. University of Hawai’i Press, Honolulu. Spate, O.H.K. 1979 The Pacific Since Magellan, Vol. I: The Spanish Lake. Australian National University, Canberra. Quirós, Pedro Fernandez de 2000 Historia del descubrimiento de las regiones austriales. Dastin, Colección Crónicas de América 12, Madrid. Veitia Linage, I. De 1981 Norte de la contratación de las Indias Occidentales. Facsimile edition. Originally published in 1672. Imprenta de Juan Francisco de Blas, Sevilla. Vocabulario Maritimo y explicacion de los vocablos, que usa la gente de Mar, en su exercicio del Arte de Marear…2000. Facsimile edition. Originally published in 1722. Imprenta Castellana y Latina de los Herederos de Thomas Lopez de Haro, Madrid. White, G. 1991 Identity Through History: Living Stories in a Solomon Islands Society. Cambridge University Press, Cambridge.

“The Anatomy of Pig Love: Substance, Spirit and Gender in South Pentecost, Vanuatu”. Canberra Anthropology (Special Volume: Pigs) 7 (1-2): 78-109.

Keesing, R. 1984 “Rethinking mana”. Journal of Anthropological Research 40 (1): 137-156. Kirch, P. 2000 On the Road of the Winds: An Archaeological History of the Pacific Islands before European Contact. University of California Press, Berkeley. Kirch, P. and R. Green 2001 Hawaiki, Ancestral Polynesia: An Essay in Historical Anthropology. Cambridge University Press, Cambridge. Kirch, P. and J-L. Rallu, eds., 2008 The Growth and Collapse of Pacific Island Societies: Archaeological and Demographic Perspectives. University of Hawai’i Press, Honolulu. Lewis, D. 1994 We, The Navigators: The Ancient Art of Landfinding in the Pacific, 3rd edition. University of Hawai’i Press, Honolulu. Luque, M. and C. Mondragón 2005 “Faith, Fidelity and Fantasy: Pedro Fernández de Quirós and the ‘foundation, government and maintenance’ of La Nueba Hierusalem in 1606”, Journal of Pacific History 40 (2): 133-148. Maroto Camino, M. 2005 Producing the Pacific: Maps and Narratives of Spanish Exploration (1567-1606). Rodopi, Amsterdam. Meleisea, M. and P. Schoeffel 1997 “Discovering Outsiders”, in D. Denoon (ed.), The Cambridge History of the Pacific Islanders. Cambridge University Press, Cambridge. Mondragón, C. 2007 “Ethnological Origins of the ni-Vanuatu “other”: Quirós and the Early Spanish Historiography of Asia and the Pacific”. In F. Angleviel (ed.), Pedro Fernández de Quiros et le Vanuatu: Découverte mutuelle et historiographie d’un acte fondateur, 1606. GRHOC, Commission Européenne, République Française, Port Vila. Munn, N. 1986 The Fame of Gawa: A Symbolic Study of Value Transformation in a Massim (Papua New Guinea) Society. Duke University Press, Durham NC. Neyret, J. 1974 Pirogues océaniennes, 2 vols. Association des Amis des Musées de la Marine, Paris.

22

Captain Cook; and three Spanish expeditions from Lima, sent by an anxious Viceroy of Peru who had been warned that the British intended to settle the island. As each of these expeditions landed, they raised their flags, ritually took possession of the island for their monarch and gave it a name - Wallis named it King George’s Island; Bougainville called it New Cythera, after the island of Aphrodite, the goddess of love; Cook called it Otaheite, its local name; and the Spaniards called it Amat’s Island, after the Viceroy - an illuminating choice in each instance.

Chapter II Voyaging Exchanges: Tahitian Pilots and European Navigators by Anne Salmond* In this paper, I will discuss exchanges of navigational knowledge between Tahitian navigators and European explorers in the mid-eighteenth century. Way-finding at sea is a complex art, and Polynesian and European voyagers accomplished this in very different ways, based upon divergent assumptions about the ocean, the cosmos, and persons. In Polynesia, various kinds of ancestors were called upon in navigational practice, whereas in European way-finding, instruments had a similar kind of agency. Yet during the early encounters between Tahitians and Europeans, island navigators were able to board European ships, and pilot them safely through the islands. In these collaborations, the survival of the ship was at stake; passages through coral reefs had to be negotiated, coral outcrops avoided, and “the stern test of landfall” had to be met (Lewis 1982) - either an island did or did not appear over the horizon when and where it was expected. On most occasions, this was successfully transacted. Here, I will investigate these collaborations between different knowledge systems, and how they were accomplished.

In this flurry of voyages, each expedition was the endpoint of a system of long-range control, described by John Law for an earlier period in Europe (Law 1986). Here, artifacts, people, texts and technologies were brought together in a single system, as European nations vied to protect or expand their imperium. The ships were robust and durable, capable of carrying a hundred or more men, and trade goods, armaments and supplies for a voyage of months, even years. These craft were owned and run by agencies of the Crown, which recruited, trained and paid the sailors; ran the dockyards in which ships were built and repaired; equipped them with gear, sails and supplies; and established a global network of bases. The backbone of this system was a chain of command which ran from the monarch; to the department of Government which drafted the instructions which guided the voyage, selected the ship and appointed its crews; to the captain, who had supreme command of the ship, within his orders and a strict set of naval conventions; to the officers and the petty officers; and down to the ordinary sailors. This chain of command in turn was upheld by a system of punishments and rewards in which court-martials, floggings and other penalties, or promotions were meted out; by the lay-out of the ship, which placed the ship’s commander in a Great Cabin, the officers in smaller cabins, and the men jammed together in their hammocks in a lower deck; and by constant drilling. The disciplined daily routine was ruled by the hourglass and the ship’s bell; the sequence of meals and watches (which determined who was on deck); the casting of the log and other instrumental observations which helped to determine the location and course of the vessel.

European Voyaging to Tahiti In the mid-eighteenth century, Tahiti and the surrounding islands were unknown to Europe. On contemporary charts of the world, the South Sea appeared as a vast expanse of ocean stretching between South-East Asia and the Americas, dotted with a few scattered islands, a partial coastline of Australia and a squiggle indicating the west coast of New Zealand. To the south of the ocean, it was thought that a large imaginary continent, Terra Australis Incognita, must exist in order to counterbalance the northern landmasses. From the 1760s, competing European monarchs sent a series of expeditions to the South Sea, hoping to set up strategic bases and claim Terra Australis for their own dominions.

The power of these ships came to a point in their weapons, cannons and muskets for fighting at a distance, pistols, pikes, swords and cutlasses for hand-to-hand combat. These weapons were used only against enemies, never in support of internal authority. If countries were at war, their ships were bound to attack and seek to destroy each other. If the islanders dared to attack a ship or its people, they became enemies by definition, and these weapons could be used against them. And as they sailed through the Society Islands, the sailors exercised another kind of power, charting the islands and surrounding

In 1767, when Captain Wallis on the Dolphin arrived at the island, Tahiti was ‘discovered’ and Wallis claimed it for the British monarch. Once news of the island was reported in Europe, it became an imperial crossroads. Over the next decade, the Dolphin was followed in quick succession by two French ships commanded by LouisAntoine de Bougainville; three British expeditions led by *

Distinguished Professor, University of Auckland.

23

ANNE SALMOND

of the horizon by the great arching bowl of the sky, and crossed by sea-paths between clusters of the known islands. It was also a marae, a sacred place where people went to cleanse themselves in times of spiritual trouble. The islands in this sea were fixed on a rock, Te Tumu or the ‘rock of foundation’, and below this rock and beyond the layered arches of the sky, there was Te Po, a cosmic darkness inhabited by the gods and ancestors.

ocean, transmuting them into space gridded by lines of latitude and longitude, stripped of substance and emptied of people. This alchemy was effective, because as successive European crews visited the archipelago, raising flags, conducting ceremonies of possession and giving new names to the islands and settlements, they often seemed unaware that they were in seas traversed for centuries by others, ‘discovering’ and claiming places that had long been inhabited.

An old Tahitian narrative explains that at the beginning of the world, a generative source produced space and then shooting stars, the moon, the sun and comets. As the stars emerged one by one, they sailed in canoes across the sky, and on their voyages, new stars were created. A star god eventually created ‘the kings of the chiefs of the earth… and the chiefs in the skies’, each with their own star, whose boundaries were marked by a marae, a great stone temple. The sky stood on star-pillars between the sky and the islands, and some of the carved boards on marae represented the stars which stood above them. When Tahitian navigators sailed off from a marae at the edge of the land, therefore, they were re-tracing the sky voyages of their star ancestors (Henry 1907).

Tahitian Voyaging For the Tahitians were also voyagers and explorers, whose ancestors had crossed the Pacific (Fig. 1). According to current scholarly accounts, they came in a great migration which began in Island South-East Asia perhaps five thousand years ago, rapidly moving eastward across the Pacific from island to island, and preceding the Viking oceanic explorations by about 2,000 years. This system of long-range exploration and settlement was made possible by fast, durable outrigger or double canoes, carrying perhaps fifty people or more and supplies for some weeks; a portable biota and a viral kinship system. Rather than a system of long-distance control, this was a system of exploration and settlement by kin-based replication.

It is thus not surprising that Polynesian voyaging was closely linked with the ancestor gods, and that priests and navigators were often the same people. According to Tahitian accounts, in the mid-eighteenth century, a marae at Opoa on the ancient island of Ra’iatea - Taputapuatea was dedicated to the war-god, ‘Oro. The island of Ra’iatea (also called Hawai’i) was described as the first land created by the ancestor Ta’aroa, and the homeland from which all other islands had been settled. The district around Taputapuatea was known as Te Po, the realm of darkness and ancestral power. When Oro’s image was consecrated at this great marae, it became the centre of a far-flung voyaging network (Oliver 1974). Feathered images of the ancestor together with stones from Taputapuatea, were carried across the archipelago, south to the Cook and Austral Islands, north to the Marquesas and east to the Tuamotu Islands, where new marae with the same name were established.

A Tahitian voyaging canoe might be 60 feet long or more, with a double hull, and a covered shelter on its platform. Such vessels were built and crewed by groups of kinsfolk, and carried supplies for a voyage of weeks, rather than months or years at sea, as was the case of European ships during this period. As they sailed, the navigator guided the canoe by means of chants to the ancestral gods, and navigational knowledge taught in the kin-based schools of learning. Voyages of exploration were often sailed upwind, allowing a safe and rapid downwind journey home. In unfamiliar waters a skilled navigator could identify and name new swells by studying the sea hour after hour, and the sequence of stars, the wind and current patterns and numerous other items of navigational information were memorised for the return voyage. During such expeditions the navigator slept as little as possible, ceaselessly scanning the sea and the night sky and kept watch for land clouds and homing birds. It was said that you could always recognize a star navigator by his blood-shot eyes1.

The cult of ‘Oro was led by the arioi, a society of orators, priests, navigators, travelling performers, warriors and famed lovers, distinguished by their tattoos and red barkcloth garments2. Large houses were built to accommodate them on their travels, and feasts staged for their entertainment. On their expeditions, a fleet of canoes assembled and travelled under ‘Oro’s protection. The missionary, Ellis, gave a vivid description of the arrival of an arioi flotilla: “[They] advanced towards the land, with their streamers floating in the wind, their drums and flutes sounding, and the Areois, attended by their chief, who acted as their prompter, appeared on a stage erected

Although Tahitian navigators crossed the ocean with confidence, the seas they traversed were quite different from those sailed by eighteenth century Europeans. According to early Tahitian accounts, their ancestors saw the Pacific Ocean as a flat plane, joined around the edges 1

For accounts of Pacific navigation, see Finney (1976, 1979, 2000), Gell (1985), Gladwin (1979), Howe (in preparation), Hutchins, (1995), Irwin (1992), Levison, Ward and Webb (1973), Lewis (1967, 1972, 1978), Taonui (1994.

2

An excellent account of the arioi in Tahitian and English is given by Rev. John Orsmond in his unpublished manuscript, ML Ms A2608, vol. 4, “The Arioi Wars in Tahiti”.

24

Figure 1: An engraving by Woollett of a fleet of Tahitian war canoes after William Hodges on Cook’s second voyage.

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

25

ANNE SALMOND

for the purpose, with their wild distortions of person, antic gestures, painted bodies, and vociferated songs, mingling with the sound of the drum and the flute, the dashing of the sea, and the rolling and breaking of the surf … the whole .. presented a ludicrous imposing spectacle” (Ellis 1859: vol. 1, 237).

Captain Wallis in Tahiti 1767 When the Dolphin arrived at Tahiti, the island was ‘discovered’ and the islanders entered European history. Equally, however, the Europeans entered Tahitian history, tangling these histories together. Wallis was searching for Terra Australis Incognita, hoping to inscribe its coastlines on the maps of the world, while the Tahitians thought that the Dolphin was a floating island, or perhaps a craft from Te Po, the realm of ancestors, the past and the future. At first, they were not certain whether the Europeans were people like themselves, ancestors or some new kind of being, although they had had prior warning of their arrival. When the Borabora warriors had first attacked Ra’iatea, they chopped down a great tree which sheltered the marae of Taputapuatea. Distraught at this desecration, a priest named Vaita entered a trance, and announced that a new kind of people were coming to the islands:

On this occasion, the fleet had arrived to entertain the locals. When arioi travelled en masse to the great ceremonies at Taputapuatea, however, it is said that their fleets of canoes carried images of the gods, and pairs of dead men and fish (including sharks and turtles) on their prows as offerings for ‘Oro. As they beached by the marae, drums and conch trumpets sounded, and some of these bodies were strung up in trees, while others were used as canoe rollers. According to early Tahitian accounts, in the mideighteenth century ‘Oro’s cult was in a state of rapid expansion. A ‘Friendly Alliance’ with Ra’iatea had been established with a far-flung network of islands, and their representatives met periodically at Opoa. At the heart of this expansionary history was Tupa’ia, an arioi priest who figured largely in the first European visits to the islands. Tupa’ia was a priest at Taputapuatea, and in about 1760 when warriors from Borabora killed the high chief of Ra’iatea in an attack on the island, the priests at Taputapuatea made a red feathered girdle and an image of ‘Oro and installed his son Mau’a as the successor. Soon afterwards Tupa’ia was despatched in the canoe of the god to carry these sacred treasures to Papara in Tahiti for safekeeping. Upon his return home the island was conquered, and in the final battle, Tupa’ia was wounded and his lands were taken. The young high chief Mau’a fled into exile in Papara to his mother’s people, and as soon as his wounds healed, Tupa’ia followed him. In Papara, Tupa’ia joined the local arioi, and soon became the lover of Purea, wife of the high chief of the district. Tupa’ia acted as high priest, counsellor and strategist for Purea and her family, and when they decided to install Purea’s son as the paramount chief of Tahiti, he initiated the construction of a great marae at Papara. In 1767, when the Dolphin brought the first Europeans to the island, they described Tupa’ia as Purea’s ‘right-hand man’; and Purea as the ‘Queen’ of Tahiti3.

The glorious offspring of Te Tumu will come and see this forest at Taputapuatea. Their body is different, our body is different We are one species only from Te Tumu. And this land will be taken by them The old rules will be destroyed And sacred birds of the land and the sea Will also arrive here, will come and lament Over that which this lopped tree has to teach They are coming up on a canoe without an outrigger4. According to this prophecy, while the bodies of ‘the glorious offspring of Te Tumu’ and the islanders would be different, they sprang from a single cosmic source. While they might differ in material appearance, the strangers were linked to the islanders by shared descent from Te Tumu. In Polynesia, the fundamental orders of the world were recounted in cosmological chants that derived all kinds of beings from a single generative origin. From this source, Te Tumu, different forms of life were created which came together to generate new forms, which were subsequently set apart in the world by acts of division. Within this shared network of kinship, all beings had form and spirit, but while their forms might be different, they shared the same spiritual substrate. This was quite unlike Western cosmological assumptions, which generally proposed a shared material substrate but distinct types of consciousness for different forms of life. One might expect such divergent ways of being to generate distinctive strategies in the first encounters between islanders and Europeans.

3

For an account of these migrations, see the unpublished manuscript history of Tahiti by Rev. R. Thomson, an English missionary in Tahiti, who died in 1857, (pp.13–17). Since he was able to interview old people who were alive at the time of the Dolphin’s arrival at Tahiti, his account of events from the 1740s onward is one of the most reliable available; although the even earlier account given by Morrison, one of the Bounty mutineers, is also very valuable. Thomson names Tupaia as the priest of “Oro who took the image of the god and the red feather girdle to Papara (p.16 of the manuscript); who later became Purea’s ‘paramour’ (p.36), describing him as ‘Tupaia, the priest of Oro who had accompanied the God from Raiatea, and who is reputed by the people themselves, as well as by Cook to have

And as it happened, when the Dolphin - a ‘canoe without an outrigger’, appeared off the coast of Tahiti, the been one of the cleverest men of the island”. (p.38, ATL Micro Ms Coll 2 Reel 169, London Missionary Society M660.). 4 Driessen (1982, vol. 17: 8-9). My thanks to Hank Driessen for access to his excellent unpublished thesis (1991).

26

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

islanders sought to engage with the strangers by means of ancestral power. Hundreds of canoes crowded around the ship, their crews ‘hallowing and hooting’ and holding up plantain branches (ritual representations of their own bodies). As they gazed at this strange vessel, Vaita’s prophecy was vindicated5. A tahua (priest) made a long speech, heralding the new arrivals, and threw his branch into the ocean. After this more canoes came out, and more speeches were made. The Dolphin’s crew, on the other hand, took it for granted that they could manage the meeting through material exchange, backed by the power of their weapons. They did not understand a word that was said, but stood with their guns at the ready and offered objects in return, holding up cloth, knives, beads and ribbons, and grunting like pigs and crowing like cocks to indicate what they wanted6.

scarlet jackets to ‘take possession’ of the island, a red pennant was hoisted on the beach. When they returned to the ship, islanders crowded around the pole, trembling when the pennant snapped in the breeze, and placing offerings beneath it. The next morning a huge crowd paraded with the red banner flying on a high pole, and surrounded the ship’s boats at the watering-place. Wallis, fearing that his men were about to be attacked, fired the ship’s guns again, killing many more people. George Robertson, the ship’s sailing master, wrote in his journal: “How terrible must they be shocked, to see their nearest and dearest of friends Dead, and torn to pieces in such a matter as I am certain they never beheld before. To Attempt to say what these poor Ignorant creatures thought of us, would be taking more upon me than I am able to perform” (Robertson 1955: 43).

The encounters which followed were tentative and violent. When islanders boarded the Dolphin and began to seize at the ironwork, Wallis took this as theft, and ordered a nine-pounder to be fired. The Tahitians leaped overboard and retreated in their canoes. This display of power was ominous, because according to Vaita’s prophecy, when the ‘canoe without an outrigger’ arrived, its crew would destroy the old rules and seize the islands. When a flotilla of canoes attacked the Dolphin in reprisal, cannons were fired into the canoes, causing havoc and killing many people.

From this time on, the power of Wallis and his men was no longer seriously challenged. The local people came out to make peace with the Europeans, and offered their women to the sailors. Arioi warriors were sexually voracious, and the sailors were probably being treated in this fashion. Purea, the ‘queen’ of Tahiti and a leading arioi, appeared with her ‘right-hand man’ Tupa’ia, intent on forging a close relationship with the Europeans. Purea spent a good deal of time with Wallis’s men, particularly the red-coated sergeant of marines, showering them with gifts and attention. She went on board the Dolphin on a number of occasions, and invited Captain Wallis to her home. In a ceremony at the great arioi house, she presented him with a plaited rope of her own hair, put a bunch of red feathers on his hat, and gave him a pregnant sow. The hair of a chief was sacred, a conduit to the ancestors, and the taumau or plaited hair bound Wallis into Purea’s lineage, while the other gifts evoked the actions of the god ‘Oro at his marriage, when he changed his two attendants into a sow in litter and a red feather bunch as gifts for his wife’s family, welcoming Wallis into the arioi society (Oliver 1974: 894-5, Handy 1971: 62). This ritual established a taio or ceremonial friendship between the English captain and Purea which in Tahiti amounted to a partial exchange of identities. When Wallis and his men left the island, the red pennant from the Dolphin was joined to the maro ‘ura or red feathered girdle brought by Tupa’ia from Ra’iatea. This girdle, worn only by high chiefs, was made of feathered squares each embodying the power of a great leader, stitched on a barkcloth sash. With this girdle, now incorporating the power of the British, Purea intended to install her son as paramount chief of the island.

It seems likely that the Tahitians associated the strangers with the war god ‘Oro. Red or ‘ura was a sign of his presence; thunder and lightening were signs of his power, and human sacrifices his tribute. When they attacked this great vessel, with its red-painted gunports and upper sides, there were claps of thunder and flashes of light, and many people were torn to pieces. Two days later, when Wallis sent an armed party of marines ashore in their 5

For illuminating reflections upon the first encounters between Melanesians and Europeans in Papua New Guinea, and the islanders’ lack of surprise, see Strathern (1990). 6 For accounts of the Dolphin’s voyage, see Robertson (1955), Rowe (1955) and Public Records Office London - Mss Adm 55/35 Captain Wallis’s Journal; Adm 51/4538/97 Lieutenant William Clarke; Adm 51/4541/95-6 Francis Wilkinson; Adm 51/4541/107-108 William Luke; Adm 51/4541/123-4 Anonymous; Adm 51/4541/125 Benjamin Butler; Adm 51/4542/109-10 George Pinnock; Adm 51/4542/111-2 Henry Ibbott; Adm 51/4542/113-4 Tobias Furneaux; Adm 4542/126-7 William Hambly; Adm 51/4539/102-6 George Robertson, Master; Adm 51/4543/115-6 Pender; Adm 51/4543/117-9 Samuel Horsnail; Adm 51/4543/128 Thomas Coles; Adm 51/4544/129 John Nichols; Adm 51/4544/131 Anonymous; Adm 51/4544/132 West; Adm 36/7580 Muster Roll of the Dolphin. Many of the men were ill during their stay in Tahiti (including Captain Wallis) and their journal entries for Tahiti were written after the events described, so that dates and details given for particular episodes are often in disagreement. On the whole, I have relied on the excellent account given by George Robertson, the ship’s master, for the basic chronology of events during the Dolphin’s stay on the island.

Bougainville and Ahutoru 1768 When Wallis sailed from Tahiti shortly after this ceremony, Purea was devastated. Nevertheless over the following months her people, led by Tupa’ia, continued to work on the great marae. In April 1768, their labours were interrupted by a brief flurry of excitement when a

27

ANNE SALMOND

visited Samoa, which at first Ahutoru mistook for France. Although Ahutoru warned him against the Samoans, Bougainville was impressed by the large outrigger canoes which zipped across the wake of the French ships at twice their speed, calling these the Navigator Islands. Ahutoru was the first Tahitian to voyage on a European vessel and survived the journey to France, where he became the focus of much popular curiosity and scientific interest.

French expedition commanded by Louis A. de Bougainville, a French officer, diplomat and mathematician, arrived at Hitia’a on the opposite side of the island. A young man named Ahutoru, a kinsman of the local chief, soon boarded Bougainville’s ship, and forged a taio relationship with the French commander, exchanging names with him. Bougainville stayed in Tahiti for only three weeks, and he and his men learned little about island politics.

Back in Tahiti, Purea and her people were completing the great marae at Papara. Emboldened by her alliance with Captain Wallis and his promises to return, Purea sent a messenger to invite the leaders of the island to witness her son’s installation as paramount chief. Enraged by her presumption, and lest their own lineages be subordinated, the chiefs from the south allied with the chiefs from the south. Before the investiture could take place, their warriors descended upon Papara, where they threw down the stones of the great marae and seized its sacred relics, including the red feather girdle into which the red pennant from the Dolphin had been woven. At the final battle at Papara in December 1768, so many people were killed that the beach was covered with their bones, and Purea, Amo and their high priest Tupa’ia were forced to flee to the mountains7.

Nevertheless, through the friendship with Ahutoru, Bougainville’s men became the first Europeans to realize that the islanders were capable of long sea voyages. As Fesche, one of the officers, remarked, “I am quite sure that they sometimes sail over long distances, for otherwise of what use would be the immense 60-foot canoes, carefully sheltered under especially made hangars?” (Dunmore 2003: series 3, vol. 9: 264). The French gave detailed descriptions of these canoes, which were made of two narrow hulls, joined by cross-pieces on which the mast was set, with a fixed sail. Each hull was crafted of two shaped logs lashed together, the first of which formed the prow and most of the hull, the second of which curved up to form the stern, five or six feet above the water. At the prow, a long flat plank twelve feet long was set, half of which covered the canoe from incoming seas, and half of which hung out over the water. At the stern, some of these canoes had portable shelters held up by carved columns. Sketches of these craft were published in Bougainville’s account of the voyage.

Captain Cook and Tupa’ia 1769 Four months later when another European vessel, the Endeavour, commanded by Captain James Cook8 and carrying a party of scientists and artists led by Joseph Banks, arrived at Matavai Bay, the local people were frightened. They had joined in the attacks against Purea’s people, and were acutely aware of her friendship with

When Bougainville’s expedition left the island, Ahutoru went with them, intent on visiting the King of France. During this part of the voyage, the first significant exchanges of navigational knowledge occurred between islanders and Europeans. Four days out from the island, Ahutoru gazed up at the stars and named them, pointing out the stars which indicated the bearing of Tahiti, and told his companions that by sailing NNW for two days, following “the bright star in Orion’s shoulder” they would reach an island where he had some family. When they ignored these directions, Ahutoru grabbed the wheel, and tried to steer the ship in that direction. According to the journals, as linguistic communications improved between him and the French, Ahutoru was able to name the stars in the night sky, describing the phases of the moon and predicting the weather with great accuracy. As Bougainville noted, “the better instructed people of this nation have a name for every remarkable constellation; they know their diurnal motion, and direct their course at sea by them, from isle to isle. In these navigations, which sometimes extend three hundred leagues, they lose all sight of land. Their compass is the sun’s course in daytime, and the position of the stars during the nights” (Dunmore 2003: series 3, vol. 9: 268).

7

Other, slightly different versions are given by Morrison, one of the Bounty mutineers in ed. Rutter (1937) The Log of the Bounty, Vols I-II, pp. 171-173 Golden Cockerel Press, London; see also Thomson n.d. pp. 26–27. The most authoritative general discussion is in Oliver (1974, vol.3: 1217–1225). 8 For accounts of James Cook’s life and the Endeavour voyage, see Beaglehole (1955, 1974), Hough (1995), Villiers 1967, Rae (1997). Beaglehole gives an excellent description of the crew, and a meticulous account of the preparations for the voyage in these two works. For primary sources see also Cook, James PRO Adm 55/40; BM Add Mss 27955, 27885; Ship’s Log BM Add Ms 8959; Hick, Zachary PRO Adm 51/4546/147-148; Log in Alexander Turnbull Library, Wellington; Monkhouse, W.B. BM Add Ms 27889; Molyneux, Robert PRO Adm 51/4546/152, Adm 55/39; Pickersgill, Richard Adm 51/4547/140-141; Wilkinson, Francis Adm 51/4547/149-150; Forwood, Stephen Adm 51/4545/133; Bootie, James Adm 51/4546/134-135; Monkhouse, Jonathan, Mitchell Library Log; Clerke, Charles Adm 51/4548/143-144; Anon Adm 51/4547/153; Adm 51/4548/154; Adm 51/4548/155; Banks, Joseph ML Journal; Auckland Public Library Grey Mss 47-75; Green, Charles PRO Adm 51/4545/151; Marra, J. (1967) A Journal of a Voyage Round the World in H.M.S. Endeavour, N. Israel, Amsterdam; Parkinson, S. (1773) A Journal of a Voyage to the South Seas in His Majesty’s Ship, The Endeavour. London, for Stanfield Parkinson.

The most distant island with which Ahutoru was familiar was fifteen nights sail from Tahiti. During their westward Pacific crossing, Bougainville’s expedition

28

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

In particular, Tupa’ia was enthralled with Solander and Parkinson’s sketches and drawings, and he soon learned to sketch, drawing many of the same subjects as these artists. Given the spiritual kinship forged between him and the Europeans, such exchanges were possible, and he quickly adopted a style which was indistinguishable from ‘naïve’ European sketching. A series of watercolours survive from the voyage which used to be attributed to Joseph Banks11, but recently a letter from Banks came to light which made it plain that one of these images (and by implication the whole series) had been drawn by Tupa’ia. It is suggestive that these sketches feature arioi themes, including marae, canoes, arioi musicians and dancers, and a chief mourner’s costume, and that they employ red, brown and black, the predominant colours of bark-cloth painting. Some male arioi (no doubt including Tupa’ia) were skilled in painting and dying barkcloth, while it is said that their tattoos featured naturalistic images of plants and people12. In return, Tupa’ia taught Parkinson about Tahitian dyes, and Parkinson and Banks both acquired arioi tattoos13.

Captain Wallis. If these new arrivals belonged to his group, they would surely try to avenge her. When Cook and his officers landed, therefore, hundreds of the local inhabitants prostrated themselves, holding up plantain branches in rituals of propitiation. Tutaha, the leader of the warriors who had defeated Purea’s people, hastened to establish his own relationship with the British. Some days later, however, a fleet of canoes arrived at Matavai Bay, bringing Purea, her husband Amo and their high priest Tupa’ia, seeking to re-establish their alliance with the British9. The ‘Queen’ went out to the Endeavour, where Cook invited her into the Great Cabin and presented her with a doll, whimsically saying that this was an image of his wife. Purea was delighted, and held it up as she returned ashore. Thinking this was a particular sign of favour, Tutaha was affronted until he was also presented with a doll, which no doubt seemed like an ancestral image10. He must have been terrified that as soon as the British met Purea and heard her story, they would attack his people. Cook tried to stay neutral in these struggles for power, but Joseph Banks was young and much less discreet; and over the following days, he began a passionate affair with one of Purea’s women. Relationships were increasingly intimate, but still volatile. When the astronomical quadrant, which was kept under close guard in the fort, was stolen, Tutaha, who had tried to flee the bay in his canoe, was captured and roughly handled. When Cook and Banks returned with the quadrant, Tutaha was released, seething with rage about the way he had been treated.

In addition to these artistic exchanges, Tupa’ia also began to teach Cook and Molyneux, the ship’s master, about Tahitian navigation. He began by dictating lists of islands in the seas around Tahiti, and during these sessions, Cook recorded a list of seventy-two14 and Molyneux a list of fifty-five islands15 with the approximate bearings of each island from Tahiti (see Appendix I). In his list, Molyneux recorded some details about the size of particular islands and their main resources, along with the sailing times from Tahiti for four of these places. Cook’s island list and Molyneux’s have only thirty-eight islands in common,

From this time on, Tupa’ia spent much of his time with the British, especially Banks and his party. He must have been gratified with the way that his enemy had been humiliated, and fascinated by Banks’s retinue, with their scientific and artistic equipment. Banks was wealthy and well born, with elegant clothes and an amorous disposition - just like an arioi; while Tupa’ia said to be one of the most intelligent and knowledgeable men in the archipelago, just like a natural historian. Nevertheless, without the ceremonial friendship forged between Purea and Wallis, it is unlikely that Tupa’ia would have shared his knowledge with Banks’s party. As their mutual command of each others’ languages improved, he tried to teach Banks and his companions about Tahitian beliefs and customs, including sacred knowledge, taught in the schools of learning. In their increasing intimacy, a new depth of communicative exchange became possible, and not only through the medium of language.

11

For instance, in Joppien and Smith (1985). See Henry (1928: 236). It could be worth checking whether Tahitian dyes, rather than English water-colours, were used for any of Tupaia’s paintings. 13 See Parkinson’s account of dyeing cloth (Parkinson 1773: 3738, 40-41, 44, 46). The Quaker artist also had himself tattooed (p. 25), so his interest in Tahitian art was genuine. It is also interesting to note William Wales’s comment after his first visit to Tahiti during the second voyage: “Since Europeans have come amongst them they sometimes print [bark-cloth] in diverse figures by diping the End of a Bambo, cut properly, into the juice, in imitation of our Handkerchiefs; but they seldom ever wear it thus printed (in checkers?) themselves, at least I never saw them do it”. (Wales in Beaglehole ed. (1969, vol. II: 799). 14 Although Cook’s island list is inserted in his journal later in the voyage, it seems likely that Tupa’ia passed on much of this information before he left the island. After sailing from Tahiti, Cook wrote that “These people have an extensive knowledge of the islands situated in these seas, Tupia as well as several others hath given us an account of upwards of seventy, but as the account they have given of their situation is so vague and uncertain I shall refar [refrain from] giving a list of them until I have learnt from Tupia the situation of each island with a little more certainty”. (Cook in Beaglehole, 1955: 138). 15 Molyneux, Robert, Ship’s Log, Adm 55/39, 62. Molyneux’s island list is inserted after the entry for 13 July 1769, the day that the Endeavour sailed from Tahiti. 12

9

Banks in Beaglehole ed. (1962: 266). See Molyneux’s account of the meeting with Purea and Tupaia, in Beaglehole ed. (1955: 554). 10 See Henry (1928: 90-1), when she discusses two dolls given by a Russian navigator to a high chieftainess, who treasured them as the representations of two deceased women from her own family.

29

ANNE SALMOND

In the event, despite their rudimentary grasp of each other’s languages, Tupa’ia and Cook managed to communicate sufficiently well to avoid these dangers.

indicating that during his conversations with these two men, the high priest-navigator shared different fragments of his navigational knowledge. More than half of the island names on these lists can be linked with islands on contemporary charts (55 from a combined list of 89 island names), although the lists also include islands that have no cartographic equivalent - for instance Tumu-papa, a name which refers to the union of Tumu (Ta’aroa’s phallus) with Papa, the Earth Mother; and those names beginning with Hiti-, evidently drawn from the story of the ancestral voyager Rata (Hiti-teare, Hiti-tautaureva, Hiti-tautaumai, Hiti-poto, Hiti-te-tamaruire, etc.)16, for instance. The identifiable islands are spread across a vast oceanic expanse around the Society Islands, from the Marquesas and Tuamotu Islands in the east; the Austral and Cook Islands in the south; and the islands of Samoa, Tonga, Tokelau and Fiji in the west - an east to west sweep of more than 5,000 kilometres; although Tupa’ia made it clear that he had visited just twelve of these places himself - eight in the Society Islands, two in the Australs and two in the Tongan Archipelago.

This accomplishment was far from trivial, given the differences between Tahitian and European ways of voyaging. During his training, a navigator like Tupa’ia mastered chants to call upon ancestral power, including the sea gods who controlled the ocean; star gods, who guided the canoe in the darkness; and the ancestors who controlled the winds, some benevolent, others malevolent in their intentions. He learned lists of known islands, with the bearings from one particular island to another and the ‘starpath’ between them, and some details about each of these islands. During a voyage, he oriented himself during the daytime by reference to a ‘wind compass’ with the cardinal points fixed by the rising and setting of the sun and the shadow of the canoe’s mast at noon; and the other points named after the winds, all controlled by ancestors. As the sea bounced off the hull, particular swells, and their deflection patterns off islands were identified. At night, familiar seas were marked out by the ‘star-paths’ between known islands, defined by a succession of named stars which rose or set in succession at a point on the bearing of the destination island. As the navigator followed the starpath, he estimated his stage in the journey by dead-reckoning, counting elapsed time in nights or ‘po.’ As a destination island approached, he noted birds flying out at dawn or returning to land at dusk; swells bouncing off the land; and clouds piling up over the land-mass.

Cook found the locational information given in these lists too “vague and uncertain” for practical use, however, and before they sailed from Tahiti, he and Banks also began to work with Tupa’ia on a chart of the islands towards which they were heading, trying to translate his knowledge into cartographic space. Quickly grasping the idea of mapping, Tupa’ia produced a remarkable chart of Ra’iatea (his home island), Tahaa, Borabora and Maurua showing their coastlines, passages and reefs, which includes images of mountains drawn without perspective (like his sketches of marae); with many placenames of islets, passages and settlements written around the coasts in Banks’s handwriting. In part, this chart seems to have been based on the sketches of coastlines that island navigators drew in the sand or with charcoal; but it also looks like a practice exercise in Western cartography, since the outline of Ra’iatea is cross-hatched or shaded in various styles.

A Tahitian navigator and his canoe thus formed a single navigational device. In discussing South American shamanism, Vivieres de Castro has suggested that when a shaman dons an animal mask and clothing, he activates the power of a different body17. In a similar fashion, in Polynesia, the navigator brought ancestral knowledge taught in the navigational schools together with the prow, masts and the rigging of the canoe to follow the stars and find north and south, and the hull as a swell-gauging instrument. Over long years of voyaging which began with an apprenticeship to an older kinsman, the navigator learned to read the sea, stars and winds, until this knowledge became reflexive and embodied. His canoe was also often named after an ancestor, and the power of various ancestors was built into it, enhancing the navigator’s capabilities. When Tupa’ia chanted to the ancestors, he became a channel for their power, guiding the voyage to a successful conclusion.

From this chart and Tupa’ia’s descriptions, Cook had gained some information about the Leeward Society Islands to the north and west of Tahiti, many of which had not been previously visited by Europeans (at least as far as Cook knew); and he was keen to explore the archipelago and add more islands to his charts of the Pacific, while Tupa’ia was eager to visit his homeland. When the Endeavour sailed from Matavai Bay, after three months in Tahiti, Tupa’ia sailed with them, acting as the ship’s pilot through the Society Islands. It is interesting that Cook was prepared to hand over the navigation of his ship to Tupa’ia, but these were uncharted waters, with unpredictable coral outcrops and atolls, and he had become convinced of the priest’s navigational knowledge.

17

Viveiros de Castro (1998, vol. 4, (3): 482); see also Pedersen, Morten (2006), “Talismans of Thought: Shamanist Ontology and extended cognition in Northern Mongolia”, in ed. A. Henare, M. Holbraad, and S. Wastell, “Thinking through Things: Theorising Artefacts in Ethnographic Perspective”, UCL Press, London.

16

Henry (1928: 477) gives the names of Hiti-au-revareva, Hititautau-atu, Hiti-poto, Hiti-uta, Hiti-ta’i and Hiti-marama as clans or islands encountered by Rata.

30

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

became available, estimates of longitude were still quite approximate18.

On board the Endeavour, on the other hand, the crew were not kinsfolk but wage labourers, paid according to seniority and status, and largely trained at sea. The sailors and marines were organized into ranks and messes, and their ‘wooden world’ was ruled by naval custom and routine, and informed by its own store of seafaring memories and knowledge. Although experienced sailors (and especially the ship’s master) also acquired an embodied knowledge of the sea, stars and winds, the officers largely guided the ship by the instrumental observation embedded in its daily routines, and by reference to a language of mapping, measurement and calculation.

On the Endeavour’s voyage of discovery, new coastlines were charted using surveying techniques and their locations fixed by astronomical observation, while the depth of the coastal seabed was measured with the lead, especially near the approaches to harbours or lagoons. Coastal profiles were also sketched, showing views of the island from the deck of the ship on particular bearings. Accuracy was paramount, so that subsequent expeditions could use these graphic records as way-finding devices. The charts were at once computational devices in which the knowledge of generations of voyagers was embedded; and objects of power, kept secret from imperial competitors or published to enunciate territorial claims19. The instruments, the ship and the crew, and the charts and profiles thus came together in a single navigational system, extending the navigators’ senses, capabilities, memories and networks of communication.

As Charles Frake and Edwin Hutchins have argued (Frake 1985, Hutchins 1996: 112-114), navigational instruments were computational devices in which generations of voyaging experience were embedded, and which required delicate calibration to local conditions. It took long practice and precise co-ordination to use them effectively. Ship’s time was measured by an hour glass, and corrected at noon when the sun was at its zenith. Ship’s time thus ran from noon to noon, and the hours were recorded in the first column in the ship’s log, the official record of the voyage. The speed of the ship was measured by the chip log, a device with a triangular piece of wood (the ‘chip’) on the end of a rope knotted at regular intervals, which was thrown overboard until it floated beyond the wake, when a half-minute glass was upturned and the rope was let to run freely from the reel, so that the number of tagged knots unreeled when the glass ran out could be counted. These measurements were taken every hour, and the ‘knots’ and fathoms per hour at which the ship was travelling were recorded in the next two columns in the log. Course or direction were determined by reference to the compass, which was oriented to magnetic north (although magnetic variation caused major problems), and recorded in a fourth column, while the direction of the wind was recorded in a fifth column.

Given these very different voyaging systems, it is not obvious how Cook and Tupa’ia, and other European captains and island navigators were able to communicate sufficiently to safely guide the ships. Perhaps the only way to investigate this issue is to examine surviving information from the voyages. In the case of the Endeavour, when the ship sailed from Tahiti, according to Pickersgill, “by Tobias Directions we stood to the Wt in Quest of some Islds which he said lay that way not far distant”. Unfortunately he did not describe how these directions were transmitted - perhaps by pointing, or instructions in a mixture of Tahitian and English. The next morning, when the wind died, he chanted to Tane, asking him for a good breeze. On this occasion Joseph Banks, the young naturalist from the Royal Society, mocked Tupa’ia’s efforts, saying that “I plainly saw he never began till he saw a breeze so near the ship that it reachd her before his prayer was finishd”. Nevertheless, a pleasant wind sprang up that evening which carried them directly to Huahine.

Each noon, the captain and officers estimated the ship’s latitude by measuring the altitude of the sun above the horizon with a sextant. During the Endeavour voyage, the officers estimated the ship’s longitude by measuring the angle between the Moon and some other heavenly body at night, at an exact time recorded by the ship’s watch. The rate of the going of the watch was then estimated, and the time corrected. An officer then looked up the time at Greenwich when these bodies were the same distance apart in the Nautical Almanac, just published before the voyage, with its tables of the lunar distances from certain bright stars, calculated for every three hours during the year, and these two times were correlated. These calculations enabled the track of the ship to be plotted on a chart, gridded for longitude and latitude and depicted from a vantage-point high above the vessel, although during this voyage, before the Harrison chronometer

Upon their arrival at Huahine, Tupa’ia sent an islander to dive down to check the depth of the ship’s rudder, so he could guide the Endeavour safely through a passage in the reef. When they anchored in the lagoon, Tupa’ia led Cook and his party ashore. On the beach, he chanted for some time and then presented the local high chief with gifts including a black silk handkerchief and two bunches of red feathers. After receiving ritual gifts in return, Tupa’ia went immediately to the local marae to thank Tane for their safe passage.

18 For accounts of British navigation in the eighteenth century, see Taylor (1968), and May Commander (1973). 19 For an intriguing discussion of the work that charts and maps do, see Turnbull (1993).

31

ANNE SALMOND

a European ship had been wrecked and all the crew killed by the Boraborans. By now, communications were evidently quite sophisticated, because one of the sailors was able to record Tupa’ia’s account of the Boraboran conquest of Ra’iatea, a narrative that is almost indistinguishable from oral histories recorded by the first missionaries to the island. After several days of cruising, Tupa’ia guided the Endeavour to Hamanino, his home harbour on the west side of Ra’iatea, now controlled by a chief from Borabora. There they were entertained by the arioi, with skits, dancing and music. When they set sail again several days later, Banks commented sardonically, “We again Launched out into the Ocean in search of what chance & Tupia might direct us to” (Beaglehole 1955: 151n).

After several days at Huahine, Tupa’ia guided the ship to Ra’iatea. When they arrived, he led his companions through the reef to Taputapuatea, the headquarters of the arioi society. On the beach, he conducted the rituals of arrival, and then Cook ordered the Union Jack to be hoisted and took possession of the island. After this, Tupa’ia took his companions to the great marae. As they approached one of the fare atua or god-houses, Joseph Banks put his hand inside and tried to prise apart the bindings on one of the ancestral figures. According to Parkinson, “[the local priests] behaved so coolly that the captain did not know what to make of them. Toobaiah, who was with him, seemed to be quite displeased. We did not know the occasion of their reservedness” (Parkinson 1773: 69). This was obtuse, because the god-houses were intensely tapu. The to’o or images were the ancestors themselves, bound with sinnet cord to intensify their sacred power, and to try to tear apart the binding was a terrible desecration.

In fact, however, Tupa’ia and Cook and Banks were on diverging trajectories. During this journey, Tupa’ia must have spoken with ‘Oro in his dreams, and watched the night voyages of his star ancestors. Perhaps he hoped to take the cult of ‘Oro to new islands. As he piloted the Endeavour through the Society archipelago, he was taking the Endeavour on an arioi voyage. At each island they visited, Tupa’ia conducted the proper rituals upon landing, and visited the local marae. When he took his European companions to Taputapuatea, the marae where he himself had been trained, this was an act of power. At this same marae, the most powerful in the islands, Vaita had prophesied that ‘a canoe without an outrigger’ was coming, bringing a new kind of people. In the company of these powerful strangers, and in defiance of the Borabora conquerors, Tupa’ia had accomplished this prediction.

Wandering around the coast the next morning, Banks saw many large arched canoe-houses. Ra’iatea was famous for its canoes, and the men in these houses were making or repairing pahii, the double sailing canoes on which they voyaged to other islands, working with a dexterity that amazed him as they dubbed out the hulls with their stone axes. These pahii were large, some as much as seventy feet long, with planks sewn onto a solid keel, and Spöring and Parkinson both completed detailed drawings of these canoes with their high curved prows and crab-claw sails; while Banks described how they were built, adding that “we saw several of them at Otahite which had come from Ulhietea (Ra’iatea) and Tupia has told us they go voyages of twenty days” (Banks in Beaglehole 1962: 320). During his discussions with the navigators at Ra’iatea, Banks was told that they preferred to use the middle-sized pahii for their longer voyages, sometimes staying at sea for several months (although they had to land on an island every fortnight or twenty days to get fresh food and water, which they carried on board in large bamboos). During this visit Cook arrived at a view of Polynesian origins which accords closely with contemporary scholarship: “In these Pahee’s [pahi], these people sail in those seas from Island to Island for several hundred Leagues, the Sun serving them for a compass by day and the Moon and stars by night. When this comes to be prov’d we Shall be no longer at a loss to know how the Islands lying in those Seas came to be people’d, for if the inhabitants of Uleitea have been at Islands laying 2 or 300 Leagues to the westward of them it cannot be doubted but that the inhabitants of those western Islands may have been at others as far to westward of them and so we may trace them from Island to Island quite to the East Indies” (Beaglehole 1955: 154).

At the same time, it suited Cook’s purposes for Tupa’ia to pilot the ship through this dangerous passage. As they sailed round the archipelago, Cook and his officers were able to draft a detailed chart of the islands, Banks collected new plants and artifacts, and the sailors recovered from the venereal infections they had caught in Tahiti. Although Cook and his men recognized and respected Tupa’ia’s technical ability as a navigator, however, it seems that they never accepted his role as a high priest of ‘Oro. His navigational advice came with the guidance of the gods, and the authority of their power. It is not surprising that Cook found this difficult to grasp, and eventually impossible to follow. As soon as the Endeavour left the Society Islands, Tupa’ia lost control of the voyage. Although he urged Cook to sail to the west where he said there were plenty of islands, some of which he had previously visited in a voyage that “took 10 to 12 days going thither, and 30 or more coming back”’ (Cook in Beaglehole 1955: 157) (which Cook identified from his description as ‘Uiha and Vava’u in the Tongan archipelago), Cook refused to sail west and turned south instead, in keeping with his orders to search for Terra Australis. He was curious about Tupa’ia’s story, however, asking how he had managed to

Upon sailing from Taputapuatea, Tupa’ia guided the ship through a maze of reefs; and as the Endeavour approached Borabora, the home of his enemies, he urged Cook not to land there, because in his grandfather’s time,

32

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

There has been much debate about this chart, with some scholars arguing that Cook misunderstood the Tahitian words for north and south (which identify south as the direction towards which the north wind blows and vice versa), thus reversing the positions of a number of islands; while others have regarded the chart largely as a mythological projection20. Both kinds of explanation try to grasp Tupa’ia’s depiction in purely European terms, however, the first trying to translate it into cartographic space while the second characterises his chart as fantastic, dismissing the kind of world in which Tupa’ia navigated. A fascinating new study that comes closer to this cosmos demonstrates that Tupa’ia’s map has little to do with cartographic space, but rather draws upon the ways in which navigators in the Society Islands conceived of the oceanic environment. Their knowledge of the surrounding seas was organised into lists of islands with sets of bearings (elaborated as ‘star paths’) radiating out from each island, linking it with other islands. The study argues that Tupa’ia’s chart is a combination of such ‘plotting diagrams,’ each with a different orientation, only one of which corresponds to the cardinal points added by Cook; which allow at least 40 islands on Tupa’ia’s chart to be reliably identified (Di Piazza and Pearthree 2007).

sail home again from the western islands against the prevailing winds. In response, the high priest told him about westerly wind shifts in November to January which his people used to sail to these islands, returning home when the winds changed back to an easterly direction. When Cook asked Tupa’ia whether there were any islands to the south, and whether he had heard of any large land masses on that bearing, Tupa’ia responded that although the most southerly island that he had visited was Moutou (the old name for Tubuai in the Austral Islands), south-east of Tahiti, his father had told him about islands further south, but none of these were very large as far as he knew. If Cook continued to sail south, however, he would find an island called Manu’a a little to the east of their current course (according to the island lists collected a little later by the Spanish, Manu’a was a large, uninhabited island with a lagoon). The winds were unfavourable, however, and they passed this island without sighting it. Tupa’ia said that Manu’a was now ‘e topa,’ a term also used for the setting of the sun. On their current bearing, he said, they would soon arrive at an island called ‘Hiti-roa’ that he had visited twenty-three years earlier, although these people had had no further contact with the northern islands. At noon the next day when ‘Hiti-roa’ (Rimatara, one of the Austral Islands) loomed over the horizon, Tupa’ia gave Cook the names of a number of other islands to the south and the south-west, most of which also began with Hiti- (the islands to the south he named were Tubuai and Raivavae in the Austral Islands, and those to the south-west were associated with the story of the ancestor Rata’s great voyage).

When they reached New Zealand, Tupa’ia’s presence on the Endeavour made a great impression on local Maori. He was a high priest from Ra’iatea, a fabled homeland of their ancestors, who could understand their language. During their six-month’s circumnavigation, he negotiated the initial encounters in each place they visited, and often spoke with the local priests. These must have been marvelous conversations. When Cook returned to New Zealand on the second voyage, and the canoes came out to the ship, their crews cried out for Tupa’ia, and wept bitterly when they heard he had died in Batavia. It seems that Maori thought the Endeavour was Tupa’ia’s ship, and they mourned him and remembered him for generations.

After their brief visit to Rurutu, everybody on board was deceived by a cloud on the horizon which they mistook for an island or the Unknown Southern Continent. Tupa’ia, who had offered a name for this ‘island,’ was deeply chagrined when he discovered that it was a mirage. Now that he had reached the limits of his voyaging experience, the high priest began to pass on more of his geographical knowledge to Cook and Banks, describing a total of about 130 islands (including some further south) and sketching a chart of the islands around Tahiti which no longer survives. Cook drafted a final version of this chart, however, which is closely linked to the lists of islands that Tupa’ia had earlier dictated (Fig. 2). Although many of the islands on the chart and the list can be identified with islands on contemporary charts, outside of the Society Islands the bearings are mostly inaccurate. Cook also translated the distances traveled between islands, described in ‘nights’ by Tupa’ia, into cartographic distance, resulting in further inaccuracies. In addition, there are islands on the chart which do not exist in cartographic space, which is not surprising since island navigators dwelt in a world in which ancestor gods were real and islands could swim like fish, behaving in ways that were fundamentally different from the world described by Western science.

When the ship arrived on the east coast of Australia, however, and Tupa’ia could no longer communicate with the local people, his value to the voyage was over. Although he impressed the sailors by always being able to point accurately towards Tahiti, this sort of information was no longer useful (Beaglehole 1955: 147n). When he died in Batavia, Cook remarked that “He was a Shrewd Sensible, Ingenious Man, but proud and obstinate which often made his situation on board both ÿisagreeable to himself and those about him” (Beaglehole 1955: 442). During the voyage, Cook and his men had often chafed at their dependency upon the high priest; and in successive histories of the voyage, too, Tupa’ia and his contributions to the voyage are rarely mentioned in any detail.

20 See for example Thomas (1997: 4, 2001: 127-128), Finney (1998), Turnbull (2000).

33

ANNE SALMOND

Figure 2: Cook’s chart after Tupa’ia ‘s lists of islands preserved in the British Library. Copyright British Library Board, all rights reserved (Add Ms 21593.C). return to Tahiti, Puhoro agreed to sail on the frigate as a pilot for the return voyage to Lima, and Mau’a decided to go with him.

The Spanish Explorers and Puhoro 1774-5 Although Cook often visited Tahiti during his second and third voyages, he never worked again with an island navigator. Once accurate charts of the surrounding islands had been drafted, he no longer needed local assistance. By contrast, the Spanish expeditions dispatched from Lima after news of Wallis’s and Cook’s voyages reached the Spanish court had no reliable charts to guide them through the islands. Cook and Bougainville had kept their charts and sailing directions a secret, but nevertheless the Spaniards heard about Tupa’ia’s navigational feats, and were eager to seek similar assistance.

One day during this passage, while Hervé, the Aguila’s pilot, was working on a chart of the islands that they had sighted during their voyages, Puhoro asked him what he was doing and examined the chart intently. When he realized that it represented the islands east of Tahiti, he told Hervé that he had visited all of these places, identifying each island by the passages through its reef. He reported that there were many more islands in that direction, although he had visited only 18 of them. When Hervé tried to explain about the four cardinal points, the navigator told him which of these islands had pearls in their lagoons, remarking that Me’etia, one of the islands they discussed, had a large supply of pearls. This island was two days’ sail south-east of Tahiti, and periodically its inhabitants sailed to Tahiti to exchange their pearls for bark-cloth garments. According to Puhoro, Tapuhoe also had pearls in its large lagoon; and all of these islands were inhabited except for Tekokoto.

In this way, the Spanish commanders gained extensive experience with island navigators. When Boenechea was first dispatched to Tahiti in 1772, he picked up a man at Mehetia, who pointed to the west, and exclaimed, ‘Tahiti.’ They followed his directions, and when the island came into sight, the man pointed at each of the bays and headlands in turn and named them. During Boenechea’s return visit to Vaitepiha Bay on the south coast in 1774-5, he decided to visit Ra’iatea, and the Aguila was guided by Puhoro, a navigator from Ma’atea, and Mau’a, the exiled high chief of Ra’iatea21. Upon their

During these and other conversations, Puhoro dictated a list of 15 islands to the east of Tahiti, including most of the NW Tuamotus; and 27 islands to the west, including many of the Society Islands, and Atiu and Rarotonga in the Cook Islands (see Appendix: I). As he listed each

21 For accounts of the Spanish voyages to Tahiti, see Corney (1919).

34

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

island, the navigator described its topography and reefs, its main produce, whether or not the island was inhabited, the ferocity or otherwise of its inhabitants, sometimes the name of its ari’i, and how many days it took to sail there from other named islands. Significantly, this island list, as compared with the lists dictated by Tupa’ia to Cook and his companions, included several new islands ‘Pounamu’ which was said to have very high mountains and a barren landscape, with plenty of fish, inhabited by cannibals (which must be Te Wai Pounamu, the South Island of New Zealand); ‘Fenua Teatea,’ inhabited by light-skinned people who could speak Tahitian, which was a larger and more fertile island (no doubt Aotearoa, the North Island of New Zealand); and ‘Vaitahu,’ a high, very populous and productive island (in the Marquesas) all of which had been visited by Hitihiti, a young man from Borabora who sailed with Cook to these places during his second Pacific expedition. Evidently, the island lists taught in the schools of navigation were constantly updated with information from new voyages.

Conclusion

During these conversations, the Spaniards also learned that in Tahiti, navigators were known as fa’atere. These men made long journeys, using a wind compass that divided the horizon into sixteen parts, with the cardinal points fixed by the rising and setting of the sun; and y Varela recorded the name of each direction. When setting out from port the navigator would test the direction and strength of the wind, and assess the direction of the swells. Using these signs and a feather and palmetto bark pennant on the mast to show any wind shifts, he would set his course, using the stars at night and the sun during the daytime to fix his bearings. A fa’atere could distinguish the planets from the fixed stars, and they named the stars used to voyage from one island to another after those islands. According to y Varela, one of the officers, they were brilliant navigators: “Not only do they note by [the stars] the bearings on which the several islands with which they are in touch lie, but also the harbours in them, so that they make straight for the entrance by following the rhumb of the particular star that rises or sets over it; and they hit it off with as much precision as the most expert navigator of civilised nations could achieve” (Corney 1919, vol. II: 284-287).

As they began to learn each other’s languages, Tahitian navigators were able to recite lists of known islands to the explorers, along with brief descriptions of their topography, inhabitants and produce. At sea, ostension or pointing at landmarks and destinations, and naming these places provided a first step in sharing navigational information. As the navigators became familiar with each other’s vessels and their capacities, this experience provided bridgeheads for practical collaboration. Eventually, Tupa’ia drafted a chart of the islands around Tahiti, perhaps drawing upon the island practice of drawing the relative positions of islands in the sand; and by indicating the bearing and elapsed nights of voyages between named islands, he was also able to help Cook to draft a chart, although this led to some confusion.

From the evidence of these voyages, it seems that the exchanges between Tahitian and European navigators were characterized by a kind of rough intelligibility. In the beginning, each drew upon their own familiar practices to make sense of the other. For the Europeans, the maritime practice of using local pilots to guide ships into unknown harbours gave a precedent for drawing upon the expertise of island navigators. For the Tahitians, an ancestral relationship had to be established before the navigators could share their knowledge with the strangers. This was achieved by forging a taio, or ceremonial friendship, in which part of the being of each was bound into the other. By the ritual presentation of names and gifts, the taio shared part of each other’s identities, including kinship networks and alliances. Through this exchange of personhood, and perhaps only in this way, the gift of navigational information to the Europeans could be accomplished.

At the same time, much was lost in these partial and approximate exchanges. There is no evidence, for instance, that Tupa’ia or Puhoro ever mastered European navigational instruments, or grasped the mathematical calculations by which the Europeans charted the islands. They understood enough about the power of the strangers and their systems of long-distance control, however, to decide to travel to the European homelands to meet their viceroys or monarchs. No doubt they hoped to recruit these leaders as taio, binding them and their power into their own kinship networks. At the same time, the Europeans were not able to record much Tahitian navigational knowledge (for instance the starpaths between particular islands), probably because the translational task was too difficult. It is also clear that for them, the ancestral strategies and purposes of the Tahitians remained largely unintelligible. The ways in which island navigators called upon ancestors to guide their craft fell into the realm of superstition, and this attitude is replicated in most historical accounts of the voyages. Interestingly enough, the use of ancestral power is also largely absent from the literature on Polynesian

y Varela also remarked on the uncanny accuracy with which Puhoro predicted the next day’s weather each evening, “a foreknowledge worthy to be envied, for, in spite of all that our navigators and cosmographers have observed and written about the subject, they have not mastered this accomplishment” (Corney 1919, vol. II: 287). When they reached Lima, Puhoro and Mau’a lived there for some months, and when the Aguila sailed again, Puhoro went with them but Mau’a, the exiled high chief of Ra’iatea, adamantly refused to return to the islands.

35

ANNE SALMOND voyaging22 (although not, by all accounts, from voyaging practice). In discussing anthropological fieldwork and writing, Marilyn Strathern has reflected upon the nature of complex engagements between different fields of action. Her reflections could equally apply to the early encounters between Tahitian and European navigators, and the distinct fields of action and assumption in which they operated; “Each is an order of engagement which partly inhabits or touches upon, but does not encompass the other. Indeed, each may seem to spin off on its own trajectory. Each point of engagement is thus a replacement or reordering of elements located in a separate field of engagement altogether. And the sense of loss or incompleteness which accompanies this, is common anthropological experience” (Strathern 1999: 2).

The old rules will be destroyed And sacred birds of the land and the sea Will also arrive here, will come and lament Over that which this lopped tree has to teach They are coming up on a canoe without an outrigger”. References Beaglehole, J.C. 1955 The Journals of Captain James Cook on His Voyages of Discovery: Vol. 1. The Voyage of the Endeavour 1768–1771. Cambridge University Press for The Hakluyt Society, Cambridge 1974 The Life of Captain James Cook. The Hakluyt Society, London. Corney, B. G. 1919 The Quest and Occupation of Tahiti by Emissaries of Spain during the years 1772– 1776, vols 1-3. The Hakluyt Society, London. Di Piazza, A. and E. Pearthree 2007 “A new reading of Tupaia’s chart”. The Journal of the Polynesian Society 116(3): 321-40. Driessen, H. 1982 “Outriggerless Canoes and Glorious Beings”. The Journal of Pacific History 17. 1991 From Ta‘aroa to ‘Oro. M.A., thesis. Australian National University, Canberra. Dunmore, J. 2003 trans. and ed. The Pacific Journal of LouisAntoine de Bougainville, 1767-1768, series 3, vol. 9. The Hakluyt Society, London. Ellis, W. 1859 Polynesian Researches during a Residence of Nearly Eight Years in the Society and Sandwich Islands, vol. I. Henry G. Bohn, London. Finney, B. R. (ed.) 1976 Pacific Navigation and Voyaging. Polynesian Society, Wellington. 1979 Hokule‘a: The Way to Tahiti. Dodd, M., New York. 1998 “Nautical Cartography and Traditional Navigation in Oceania”. In D. Woodward and G.M. Lewis (eds.), Cartography in the Traditional African, American, Arctic, Australian and Pacific Societies, vol. 3, Part 2. The University of Chicago Press, Chicago. 1999 “The Sin at Awarua”. The Contemporary Pacific 11: 1-33. Frake, C. 1985 “Cognitive Maps of Time and Tide among Mediaeval Seafarers”. Man 20: 254-270. Gell, A. 1985 “How to Read a Map: Remarks on the Practical

I have this sense of loss about these early encounters. It is not possible to talk with those who participated in them, who are long since dead; only the oral histories, chants, logs, journals, sketches and charts survive to speak of their experiences. One can try to enrich the account by delving into background information from a variety of sources, but this cannot be elicited at will. Too often, no-one at the time recorded crucial details of what happened, and all of the oral histories and some European sailors’ accounts are retrospective, shaped by hindsight; while Tahitian and European accounts are shaped by very different interests and questions. In trying to understand Cook and his men, the Spaniards and their crews, and the various Tahitian navigators they encountered, and their mutual impacts on each other, there are doubled and redoubled hazards of distance in time, and between the different forms of life they inhabited. These were very great voyages, however, and Cook’s voyages in particular now have almost mythical status, both in Europe and in Polynesia. At the same time, European myths, social hierarchies and technologies of power are still visible in many contemporary histories of these expeditions. Often, the islanders are rendered passive, mastered by their own myths or acting as bitplayers in European dramas; while their own accounts are trivialized or ignored. In these histories, one can still see European systems of long-range control at work, obscuring Polynesian intentions and purposes. For these expeditions were collaborative accomplishments, to which men like Tupa’ia and Puhoro and their knowledge systems made significant contributions. It seems better history, and better anthropology to try to do justice to the complex, many-sided dynamics of these engagements, but this is not simply an academic matter. In the postcolonial Pacific, Vaita’s prophecy still echoes: “Their body is different, our body is different We are one species only from Te Tumu. And this land will be taken by them 22

Logic of Navigation”. Man 2 (2): 271-286.

For a striking exception see Finney (1999).

36

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

Pedersen, M. (in prep.) “Talismans of Thought: Shamanist Ontology and extended cognition in Northern Mongolia”. In A. Henare, M. Holbraad, and S. Wastell (eds.) (2005) Thinking through Things: Theorising Artefacts in Ethnographic Perspective. UCL Press, London. Rae, J. 1997 Captain James Cook Endeavours. Stepney Historical Trust, London. Robertson, G. in ed. O. Warner 1955 An Account of the Discovery of Tahiti, From the Journal of George Robertson, Master of H.M.S. Dolphin. Folio Society, London. Rowe, N. 1955 Voyage to the Amorous Islands: The Discovery of Tahiti. Andre Deutsch, London Strathern, M. 1990 “Artefacts of History: Events and the Interpretation of Images”. In J. Siikala (ed.), Culture and History in the Pacific. The Finnish Anthropological Society, Transactions 27: 25-43. 1999 Property, Substance and Effect: Anthropological Essays on Persons and Things. The Athlone Press, London Taonui, R. 1994 Haerenga Waka: Polynesian origins, migrations and navigation. M.A. thesis, University of Auckland. Taylor, E.G.R. 1968 “Navigation in the Days of Captain Cook”. The Journal of the Institute of Navigation 21: 256-276. Thomas, N. 1997 In Oceania: Visions, Artifacts, Histories. Duke University Press, Durham. 2001 “Cook and Tupaia, a Tale of Cartographic Meconnaissance?” In M. Lincoln (ed.), Science and Exploration in the Pacific: European voyages to the Southern Oceans in the Eighteenth Century. National Maritime Museum, Woodbridge, Suffolk. Turnbull, D. 1993 Maps are Territories: Science is an Atlas. University of Chicago Press, Chicago. 2000 “(En-)countering Knowledge Traditions: The Story of Cook and Tupaia”. Humanities Research 1. Australian National University, Canberra. Villiers, A. 1967 Captain Cook, the Seamen’s Seaman: A Study of the Great Discoverer. Hodder and Stoughton, London. Viveiros de Castro, E. 1998 “Cosmological Deixis and Amerindian Perspectivism”. The Journal of the Royal Anthropological Institute 4 (3): 469-4.

Gladwin, T. 1979 East is a Big Bird: Navigation and Logic on Puluwat Atoll. Harvard University Press, Cambridge Mass. Henry, T. 1907 “Tahitian Astronomy. (Recited in 1818 at Porapora, by Rua-nui (Great-pit), a clever old woman). Birth of the Heavenly Bodies”. Journal of the Polynesian Society 16: 101-4. 1928 Ancient Tahiti. Bishop Museum Bulletin, Honolulu. Hough, R. 1995 Captain James Cook: a Biography. Hodder and Stoughton, London. Howe, K. (ed.) (in prep.) Waka Moana. David Bateman Ltd. Hutchins, E. 1995 Cognition in the Wild. M.I.T. Press, Cambridge Mass. Irwin, G. 1992 Prehistoric Exploration and Colonisation of the Pacific. Cambridge University Press, Cambridge. Joppien, R. and Smith, B. 1985 The Art of Captain Cook’s Voyages: Volume I. The Voyage of the Endeavour 1768-1771. Oxford University Press, Melbourne. Law, J. 1986 “In Power, Action and Belief: A New Sociology of Knowledge?” Sociological Review Monograph 32: 234-263. Routledge and Kegan Paul, London. Levison, M., R.G. Ward, and J.W. Webb 1973 The Settlement of Polynesia: A Computer Simulation. Australian National University Press, Canberra. Lewis, D. 1967 Daughters of the Wind. Gollancz, London. 1972 We, the Navigators. Australian National University Press, Canberra. 1978 The Voyaging Stars: Secrets of the Pacific Island Navigators. Collins, London. Marra, J. 1967 A Journal of a Voyage Round the World in H.M.S. Endeavour. N. Israel, Amsterdam. May, Commander W.E. 1973 A History of Marine Navigation. G.T. Foulis and Co. Ltd., Henley-on-Thames, Oxfordshire. Oliver, D. 1974 Ancient Tahitian Society, vols. 1-3. Australian National University Press, Canberra. Parkinson, S. 1773 A Journal of a Voyage to the South Seas in His Majesty’s Ship, The Endeavour. For Stanfield Parkinson, London.

37

ANNE SALMOND

Appendix I: Island lists COOK’S LIST 1769

MOLYNEU X’S LIST 1769

TUPA’IA’S CHART 1769

GAYANGOS’S LIST 1774

ANONYMOUS SPANISH LIST 1774

The names in italics are unknown, but given in modern transcription

NORTH-EAST QUARTER Between N and NE 1. Oopate

2. Ooura

CONTEMPORARY NAMES

Aowra [NE & E of Tahiti]

3. Teoheow

Oopati

Opatay [small, low, reef. Inhabited, some pearls; I day from Jooau [E of Tahiti]

Oura

Auroa [Large, not as big as Tahiti; low, plenty of coconuts, yams, dogs, pearls. Many inhabitants; 1 day from Tayaruro. E of Tahiti] Otiehao [small, flat, reefs. Coconuts, yams, much fish. Good people, make excellent mats; within sight of Orairoa, E. of Tahiti]

Teoheow

Opatay (small, low, lagoon and reefs; coconuts, yams, some pearls, inhabited. 1day from Toaau] Aurua [Large, smaller than Tahiti; coconuts, yams, dogs, pearls. Low, many reefs. I day from Tayaruru]

Apataki, W.Tuamotus [Corney 188]

Toaau [small, low, lagoon and reefs, coconuts and yams, 9 days from Tahiti]

Tikehau / Ti’ehau, W. Tuamotus [Corney 189]

Orayroa [Large, low with reefs. 3 inlets for small boats, but not deep. Only a few coconuts, yams and dogs; many pearls. Arii called Oheiruria]]

Rangiroa / Rairoa, W.Tuamotus [Beaglehole 293]

‘Au’ura / Kaukura, W. Tuamotus [Corney 188; Finney 449]]

4. Oryroa

Oraieeroa [NE and E]

Oryeroa

5. Ohevapato

Heewapotta [SW & NW] / Oheewapoto [Inhabited; NE and E] Ota [NE & E of Tahiti]

Ohevapoto

Takapoto, King Georges Islands, N. of Tuamotus [Di Piazza & Pearthree]

Otaah

Oheevaroa [large, inhabited; NE and E]

Ohevaroa

Tahanea, W. Tuamotus [Di Piazza & Pearthree; Finney 449] Hivaroa / Hivaoa, Marquesas [Motteler; Finney 449]

6. Otaah 7. Ohevaroa

8. Temanno

Temanno

9. Oota NE 10. Whareva

Ootto

11. Whaterreero

Owharawa [NE & E]

Whareva

Amanu, Tuamotus? [Salmond] Uta / Uto Guaraba [Same size as Tapuhoe, identical in all respects. 2 days from Tapuhoe; E of Tahiti].

Whatterreero

Huaraba. The size of Taphuoe, but the canoes take another two days]

Fakarava,W. Tuamotus [Corney 189; Finney 449] Ha-te-riro? Hereheretue, S. Central Tuamotus [Di Piazza & Pearthree]

38

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS 12. Tetioo

Oteeohiaow [NE & E]

Oateeu

13. Tetineohva

Tetineohvah

14. Terouwhah

Terdouuha

NNE 15. Whaoa 16. Whaterretaah

Owhao [NE & E] Eohatetirreeto oa [large, inhabited; NE and E]

17. Whaneanea EBN 18. Ohevatoutuai

SOUTH-EAST QUARTER South to SE 19. Moutou

20. Toometoaroaro 21. Tennowhammeata ne 22. Ohetetamaruire

23. Ouropoe

ESE and East 24. Mytea or Oznaburg Is [visited by Tupa’ia]

26. Ohirotah

Whaou

Hao, S. Tuamotus [Salmond; Finney 449] Ha-tiri-tua?

Whaterretuah

Whaneanea

Tahanea, Tuamotus [ Finney 449]

Ohevatoutoua i [maa te ta ta pahu rahie ote pahei no Brittane]

Hiva-tautau-ai

Omaowtaow [ large, inhabited; NE and E] Toomootoaria ro [inhabited; SW & NW of Tahiti] Ineehaewham eatane [large, inhabited; NE and E] Heeteetoomar oeiree [inhabited; SW & NW of Tahiti] Aowroopou Oaiarota [large, inhabited; NE and E]

Moutou

Maiatea [large, inhabited; NE and E]

Mytea

25. Ohevanue

Atiu, Cook Islands [Salmond; Finney 449] Te Tai o Nu’uhiva or Nukuhiva, Marquesas [Di Piazza & Pearthree; Finney 449] Te Rua, Ua Pou, Marquesas [Di Piazza & Pearthree]

Tupuai [Know only that it is inhabited. W of Tahiti]

Tubuai, Austral Is [Di Piazza & Pearthree]

Tomotoaroaro

Tumu-to-aroaro?

Tenewhamme atane

Te Nifa-mea-tane?

Ohietemaruir u

Hiti-te-tamaruire?

Ourupoe

Uru-pou

Matea [The one we call San Diego, mountainous, reefs, coconuts, bananas, fish and some pearls.

Ohevanue Oaiarota [ large, inhabited; NE and E]

Tupuay [Inhabited, abundant pigs, coconuts and bananas]

Mehetia, E. Society Is [Salmond]

Ahu-nui? S. Tuamotus [Finney 449] Rio-tahi? Hiro-tahi?

Oirotahi

39

ANNE SALMOND SOUTH-WEST QUARTER WBS and WSW 27. Imao or York Island [visited by Tupa’ia]

28. Tapooamanue or Saunders I [visited by Tupa’ia]

Between S and SW 29. Manua [visited by Tupa’ia]

30. Honue 31. Oheteroa [visited by Tupa’ia]

Imao

Tabeoamanno o [small high island 1 day’s sail from Otahiti]

Tapooamannu [Meddua no te taboona no Tupia pahei toa].

Tupaemanu [small, low, reefs, coconuts, plantains, breadfruit, dogs, pigs, fowls, good water. People very tractable; arii named Oahau. [W]].

Mannoa [ large, inhabited; NE and E]

Mannua

Manua [larger than Morea. Coconuts, plantains, plentiful uru. Uninhabited, whirlwinds. W of Tahiti]. It was sighted by us.

Honue

Onowhea

33. Otootooera

Otootooera

34. Opooroo 35. Ooouow

Owporrow [inhabited; SW & NW]. Oawaow [inhabited; SW & NW]

Eimeo/Mo’orea, Society Islands

Manua [The Isle of Birds. Quite high, plentiful coconuts, bananas, breadfruit and dogs. Reefs, a deep inlet called Atanua. Uninhabited, because blasts of furious winds sink the canoes].

Manu’a, Samoa [Di Piazza & Pearthree; Corney 192]

Tapuaemanu/Maiao, Society Is [Corney 191; Finney 449].

Honue Hitiroa / Rurutu, Austral Is [Motteler; Finney 449]

Oheteroa

32. Onawhaa

Moorea [The one we call Santo Domingo. Mountainous, fertile, very well populated. At war with the Taiarapu people]. Tupuemanu [The Shorn Island; small, low, reefs, abounds in coconuts, bananas, breafruit, hens, pigs and dogs. Good water, tractable people. Arii is called Oahari].

Nofea, ‘Uiha, Tonga [Di Piazza & Pearthree; Finney 449]. Tutuila, American Samoa [Finney 449] Upolu, Western Samoa [Finney 449].

Opooroo Oouow

Auau, Vava’u, Tonga [Di Piazza & Pearthree; Finney 449]

36. Terororomatiwhat ea

Teorooromati watea

Te Orooro-ma-tiwhatea?

37. Tetawhite

Teatowhete

Te Tafiti, Swains Island, American Samoa [Di Piazza & Pearthree]

Oheavie

Savai’i, Western Samoa [Di Piazza & Pearthree; Finney 449] Purea-matehea

Between the SW and WSW 38. Oheavie 39. Pooreomathehea 40. Tiamoorohete 41. Ohetetaeteare

Ohiawai [inhabited; SW & NW]

Heeteetaitirre [inhabited; SW & NW]

Teerrepooop mathehea Teamoorohete Oheteiteare

Te Muru-hiti Hiti-te-are – Atafu, Tokelau [Di Piazza & Pearthree].

40

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS 42. Ohetetareva

Heeteetaitirri va [inhabited; SW & NW]

43. Ohetetoutoumi West 44. Mooenatayo

Ohetetoutoure va

Hiti-tautau-reva / Pitcairn [Beaglehole 294] Hiti-tautau-mai / Mururoa [Beaglehole 294]

Ohetetoutoum i Tea moanitte [inhabited; SW & NW]

Moenatayo

Fakaofo, Tokelau [Di Piazza & Pearthree]. Te Tu-patu.. ?

46. Oheteteutenatu

Tetupatupaea how Ohetetoutouatu

47. Ohetepoto

Ohetepoto

45. Tetupatunaeohew

NORTHWEST QUARTER NBW 48. Tethuroa [visited by Tupa’ia]

49. Oonnah

50. Oboha 51. Maataah Between N and W 52. Huiheine [visited by Tupa’ia]

Tetiroah [uninhabited], matoo/low island

Oanna [NE & E]/ Owanna

Oanna [Tupia tata re pahie matte]

Hiti-tautau-atu / Crescent Islands [Beaglehole 294] Hiti-poto Te Poto / Mangareva [Beaglehole 294]

Teturoa [small, low, reefs, coconuts, pearls plentiful but not good, fish. Belongs to the arii Tu; W of Tahiti]

Teituroa [is the one we called Three Brothers. Small, low, reefs, abounds in coconuts, pigs, dogs, fish, some dull pearls].

Tetiaroa, N.Society Is [Corney 190]

Oana [small, low, reefs; coconuts, yams, dogs, a few pearls. 2 days from Maropua. E of Tahiti].

Oana [Like Morapua, but it has good people]

Ana’a, W. Tuamotus [ Finney 449; Corney 188] Puha Mataa, Mataiva, W. Tuamotus [Finney 449]

Maataah Huaheine

Oahine [Plenty of uru, plantains, coconuts, pearls, pigs and fowls; dense population, 2 bays good for big ships. The arii is Tahuoha, tributary to Opuni. W of Tahiti].

53. Ulietea [visited by Tupa’ia]

Ulietea [Tuboona no Tupia pahie tayo]

54. Otaha [visited by Tupa’ia]

Otahah

Oraiatea and Taha [2 islands united by a reef; a good bay on the west side. Both productive of uru, coconuts, plantains, pigs, fowls and good water. Inhabitants are well disposed, keep up a close communication with Tahiti. The arii is named Opuni. W of Tahiti]. “

41

Oahine [The beautiful island. Very fertile, breadfruit, bananas, coconuts, hens pigs and dogs. Well populated, docile inhabitants. The arii is Tapuaoha, and is tributary to the arii of Ra’iatea] Isles of Princess and Taha. Two islands united a reef; quite high, well populated, abundant resources. Arii is Opuni.

“

Huahine, Society Islands

Ra’iatea, Society Islands

Taha’a, Society Islands

ANNE SALMOND 55. Bolabola [visited by Tupa’ia]

56. Tubai [visited by Tupa’ia] 57. Maurua [visited by Tupa’ia]

Bolabola

Toopwai

Maurua

Opoopooa

59. Opopatea Between NBW and West 60. Whennuaouda

Opopotea

62. Oourio 63. Orurutu 64. Oatuu 65. Oahooahoo

66. Oweha 67. Orotuma 68. Tenuna 69. Orevavie / Oryvavai

Porapora {St. Pedro; small and high, surrounded by reefs. Well populated, abundant resources, some pearls. Under the domination of the arii Opuni]

Tebooi

58. Opoopooa

61. Motehea

Porapora [Small and elevated, surrounded by reefs. Passage on S. side through which a ship can enter. Uru, coconuts, plantains, dogs, pigs in plenty; a few pearls. Well populated, belongs to Opuni. W of Tahiti].

Borabora, Society Islands

Tupai, Society Islands Maurua [Low, three high hills, reef, produces coconuts, uru, plantains, dogs, pigs and a few pearls. Has fresh water, inhabited. Belongs to Opuni. W of Tahiti].

Maurua [San Antonio. Small, with three high mountains, surrounded by reefs, well populated]

Maurua / Maupiti, Society Islands

Pukapuka, Cook Is [ Finney 449] Popatea

Whannoeah Aowra [low island, inhabited, 10 days sail from T.] Omatea [ large, inhabited; NE and E]

Whennuaouda

Genuaora [middle size, flat, reef. Coconuts, birds. Uninhabited. W of Tahiti]

Motuhea

Matea [high, reef. Breadfruit, coconuts, plantains, fish, pearls. 2 days to reach Maitu. E of Tahiti]

Orooheoo [inhabited; SW & NW] Orarrtoo [inhabited; SW & NW]

Oureu

Uriu

Orurutu

Rurutu, Australs [Salmond]

Ohaowahaow [inhabited; SW & NW] Oweeha [inhabited; SW & NW]. Owratoomoo [inhabited; SW & NW] Tainoonna [inhabited; SW & NW] Whoraivewai [inhabited; SW & NW].

Ooto Oahoo-ahoo

Oahuahu [inhabited, abounds in coconuts, plantains and pigs; W of Tahiti].

Genua-ura [medium size, low, with a reef. Uninhabited]

Fenua Ura / Manuae, Scilly Is., W. Society Is [Corney, 193]

Ma’atea, Makatea, W. Tuamotus [Corney 189]

Bahuahu [Inhabited, abundant in pigs, coconuts and bananas]

Utu Ahuahu [Oahu? Corney 193], Mangaia, Cook Is [Di Piazza & Pearthree].

Oweha

‘Uiha [Tonga Islands?]

Orotuma

Rotuma, Fiji [Salmond; Finney 449].

Tinuna

Tai-nuna

Orivavie [too miti to te rara te vietea]/ Oryvavai [2X].

Oroibaay [Santa Rosa; divided in two, arii named Serabaroy].

42

Ra’ivavae, Austral is [Motteler: Finney 449]

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

70. Toutepa 71. Orararhoa 72. Oahourou

Taowtipa [inhabited; SW & NW].

Toutepa

Tau-tipa

Woahaowroo [ large, inhabited; NE and E]

Oahourou

Rara-roa Oahu-roa, Hawa’i? [Salmond]

NOT LISTED BY COOK 73. Waowrea [inhabited; SW & NW]. 74. Mobehaa [low, large island 10 days sail from Tahiti]

Olematerra [On Cook’s map only]

Oo-Ahe [On Cook’s map only]

Tebooi [On Cook’s map only]

75. Oreematarra [Inhabited; SW & NW]. 76. Otaanooe [inhabited; SW & NW]. 77. Oneewarroa [inhabited; SW & NW]. 78. Teeteepitoop. ohaow [inhabited; SW & NW]. 79. Oraieeroa [NE and E]

80. Aowahei [NE and E]

Aurea Mopiha [middle size, low, reef, coconuts, pearls. Uninhabited, many large, flightless birds. 2 days from Maurua. W of Tahiti]. Orimatara [Inhabited, abounds in coconuts, plantains and pigs. W of Tahiti].

Mapisa [Medium size, low with a reef, depopulated. Only has coconuts, some pearls and many birds of extraordinary size. 2 days from Maurua].

Maupiha’a, W. Society Is.

Rimatara, Australs [Corney 193]. Ta-nui Niwa-roa

Te Tipitupuhau?Nukunonu, Tokelau [Di Piazza & Pearthree] Oryroa

Orairoa [large, low, reefs. 3 bays; coconuts, yams, dogs, fish & pearls. People very tractable. E of Tahiti].

Ooahe

Ra’iroa, W. Tuamotus [Corney 189].

Ahe, King George Is, N. Tuamotus [Motteler; Finney 449]. Po-atea-ta’ora?

81. Pooatea ta’owra [Inhabited; NE and E] 82. Otoomoobapa [ large, inhabited; NE and E]. 83. Tippoowai [ large, inhabited; NE and E] 84. Heeteehaniani a [ large, inhabited; NE and E]

Tumu-papa [the sexual union of Tumu and Papa]. Tipu-wai, Tubuai, Australs?

Hiti-ha-niania

43

ANNE SALMOND 85. Teeteeneehee vo [ large, inhabited; NE and E]

Tetineoheva

Te Tini-o-heva?

86. Tomanowhola [ large, inhabited; NE and E] 87. Maowroeati [small high island 1 day’s sail from Tahiti] 88. Opoitai [inhabited; SW & NW] 89. Tirreetaowtap atanee [inhabited; SW & NW]

To-manawa-ora

Maurua [St Antonio]

Maurua-iti

Opoitai Tiritautapatane

90. Oremaroa Oremaroa [On Cook’s map only]

Rima-roa 91. Otemateroa 92. Maatah

93. Orarothoa

Mataiba [small, low, reefs. Fowls, dogs, yams, much fish, some pearls. Good people. In sight of Otiehao. E of Tahiti]. Rarotoa – know nothing more, but it is inhabited; W of Tahiti] 94. Tabau [small, flat, reef, inhabited. Barren, but plenty of fish and pearls. 1 day from Joaau; E of Tahiti]. 95. Tayaruro [small, flat, reefs, 1 day from Jooau, E of Tahiti] 96. Oarutua [very small, flat with reefs, plenty of fish and pearls, few inhabitants, near Auroa; E of Tahiti] 97. Tapuhoe [plenty of coconut and yams, some pearls. Bad lot of people; E of Tahiti] 98. Maemo [small, low, coconuts, yams, dogs, fish, plenty of pearls. Reef. Inhabitants mild. 3 days to Guaraba. E of Tahiti]

44

Matayba [In every way like Otiesero, has some hens]

Te Materoa Mataiva, W. Tuamotus [Corney 189; Finney 449].

Rarotonga [Corney 193; Finney 449] Tautau [small, low, uninhabited; rich in fish and pearls. The canoes from Toaau go there to fish, I day]. Tayaruru [Small, low, uninhabited, rich in fish and pearls. 1 day from Tautau] Oarutua [Very small and flat with reefs, abounds in fish and pearls. Few inhabitants, very close to Aurua] Tapuhoe – the one that we call Todos Santos, whose inhabitants are regarded as bad in Tahiti. Maemo [small, low, abundant coconuts, yams, dogs, fish and pearls. Surrounded by reefs, docile inhabitant. 3 days from Huarapa]

Niau or Kauehi? W Tuamotus [Corney 188].

Taiaro, W. Tuamotus [Corney 188] Arutua, W. Tuamotus [Corney 188]

Tapuhoe= ancient name of Anaa, W. Tuamotus [Corney 188]

Makemo, central Tuamotus [Corney 189]

VOYAGING EXCHANGES: TAHITIAN PILOTS AND EUROPEAN NAVIGATORS

99. Maropua [small, low, reefs. Coconuts, yams, dogs and fish plentiful. Bad people. 2 days to Maemo. E of Tahiti. 100. Joaau [small and low, lagoon & reef, [plentiful coconuts & yams, pearls. 9 days E. of Tahiti] 101. Oaiyu [like Orimatara; W of Tahiti] 102. Puatireaura [know only it is inhabited; W of Tahiti] 103. Te Miromiro [No people, plenty of birds; W of Tahiti]

Maropua [Like the previous island, but the people are bad. 2 days by canoe from Maemo]

Marutea? Central Tuamotus [Corney 189]

Toaau [small, low, lagoon and reefs, coconuts and yams, 9 days from Tahiti]

Toau, W. Tuamotus [Corney 187]

Oayyu [Inhabited, abundant pigs, coconuts and bananas]

Atiu, Cook Islands [Corney 193] [? Corney 193]

Tomiromiro [Is only inhabited by an infinity of birds]

Bellingshausen’s Is , just W of Tahiti? [Corney 193]

104. Hoaituputupu: ditto 105. Marere [has people, W of Tahiti] 106. Ponamu [Is peopled and has excessively high mountain peaks; is very barren of fruits, but has fish in the greatest plenty, and this is the only support of its inhabitants, who are very rude and live in caves. The people of other islands near it are afraid to land there, because, they say, there have been several instances of strangers being killed and eaten by the natives. W of Tahiti]. 107. Genua teatea [New Zealand; is inhabited by a white people who speak the same language as those of Amat’s, and have the same appearance, but it abounds in produce, and is larger]. 108. Teone tapu [Inhabited; W of Tahiti]

Toaytuputupu [Ditto]

Waitaputapu - Waitapu [Corney 193] Marere

109. Uritete [large, productive, people very well disposed. W of Tahiti].

Uritete [Large, fertile, and its inhabitants are good].

45

Marere [Is inhabited] Ponamu [Is inhabited, has very high mountains, very barren in fruits but abundant in fish, which is the only food of its inhabitants, who live in caves and are very fierce, because some canoes having arrived on one occasion the crew were eaten]

Pounamu, Te Wai Pounamu, South Island, NZ [Salmond]

Genuateatea [large and fertile, inhabited by white people who speak and dress like those of Otaheti].

Whenua Teatea, Aotearoa NZ [Salmond]

Teonetapu [Populated]

Te One tapu Tongatapu? [Corney 194] Uritete, Manuai? Corney 194]

ANNE SALMOND

110. Oaitaho [largest island they know of; high, populous, productive; people speak same language as Tahitians, have same appearance. W of Tahiti]. 111. Oaurio [only its existence is known. W of Tahiti]. 112. Oaupo [only its existence is known. W of Tahiti]. 113. Genua baro [only its existence is known. W of Tahiti]. 114. Teputuroa [only its existence is known. W of Tahiti].

Oytaho [The largest of Vaitahu, Marquesas which we have [Corney 194] information, high and fertile and very populated, and its inhabitants dress and speak like those of Otaheti. Oauriu [only know Auriu that this island exists] Oaupo [only know Upo that this exists] Genua-varo [ditto]

Fenua raro

Teputuroa [ditto]

Teputuroa

115. Huaraba [The size of Tapuhoe, and identical to it. The canoes take another 2 days]. 116. Otiesero [Small, low, with reefs. It has coconuts, yams, dogs, fish, some pearls. Docile inhabitants, who make excellent mats. It is in the mouth of Orayroa]. 117. Vuimatara [Inhabited, abundant in pigs, coconuts and bananas]. 118. Puariro-aura [Inhabited, abundant in pigs, coconuts and bananas]

Huarapa

Taihero [?]

Huimatara?

Puariro-a-ura

Di Piazza, Anne and Pearthree, Erik, give some identifications of island names in “Cook and Tupaia, the encounter of two cartographers”. Finney has a map which identifies some of the island names on Tupa’ia’s chart in Finney (1998). Beaglehole gives identifications for some of the island names in Beaglehole, J.C. 1955, The Voyage of the Endeavour 1768-1771. Cambridge University Press, published for the Hakluyt Society, pp. 291-294, Cambridge. Other names are identified in Motteler, Lee S. (2006), Pacific Island Names: A Map and Name Guide to the New Pacific, Bishop Museum Press, Honolulu.

SOURCES: Cook’s island list is in Cook in Beaglehole (1955: 291293). Molyneux’s island list is in National Archives, UK, Molyneux, Robert, Ship’s Log, Adm 55/39, 62. Cook’s final version of Tupa’ia’s chart is in British Library, Add Ms 21593 C. Gayangos’s island list is in Corney (1919: vol. II: 187194). The Anonymous Spanish list of island names is in ‘Descripción de las isles del Occeano Pacifico, Por Dn. Domino de Boenechea, en los anos de 1772 y 1774,’ Museo Naval Madrid Ms 476, ff. 84-101, trans. from original by Gwyn Fox.

46

maritime vocabulary, marine paintings and, of course, nautical ethnography through his studies of traditional European and extra-European watercraft. In the latter domain, the “Essai sur la construction navale des peuples extra-européens”2 played an absolutely fundamental role. It was published3 as two large volumes, one of text the other as an atlas of plates, with a section dedicated to canoes of Oceania.

Chapter III Traditional Oceanic Canoes as seen by Admiral Pâris by Eric Rieth ∗ F

To account for Pâris’ scientific contributions to knowledge of canoes of Oceania, we will first examine the context and the conditions in which he assembled his documentation, in particular for Oceania. Then, in a second part, we will consider, using Oceanic canoes, how Pâris, in terms of historical problematics, thought of extra-European boats as objects of History and Memory.

François-Edmond Pâris was born in Paris on March 2, 1806. In 1820, young Pâris passed the entry exam of the Royal Maritime College (College Royal de la Marine) located in the town of Angoulême (Charente). After two years of theoretical training in the peaceful charentaise countryside, and four years of practical training at sea on various sailing ships, he received his first post. In 1826, he embarked on the Astrolabe, a corvette commanded by Dumont d’Urville for a three year circumnavigation that would be decisive for his future. During the long months onboard the Astrolabe, he joined his naval duties with those of a researcher in nautical ethnography. He accumulated copious data, such as notes and architectural drawings of traditional craft that he encountered at sea and during the Astrolabe’s calls in foreign ports.

The documentation of Oceanic canoes There were two important epochs in Pâris’ professional life that correspond to two periods of data collection, analysis and publication of his ethnographic documentation. For the most part, his materials are archived in the research library or collections of the “Musée national de la Marine du Palais de Chaillot”, in Paris.

Numerous other postings followed both on land and at sea, eventually leading to his promotion to Rear Admiral (on August 13, 1864). When he reached retirement age on March 6, 1871, Pâris left behind a long carrier as a sailor. His forty five years of navy service also included research in different domains, including nautical ethnography. This scientific activity, exemplified by his extensive publications, was crowned by election to the French National “Académie des Sciences” in 1863, to the seat once occupied by Auguste Bravais, a renowned physicist and mineralogist.

The first period was relatively short. It lasted for only ten years during which Pâris made his three circumnavigations. His first voyage was aboard the Astrolabe, under the command of Dumont d’Urville, from April 22, 1826 to April 2, 1829. Pâris’ principal duties were to participate in hydrographic surveys with his shipmates: Faraguet, Guilbert, Jacquinot and Lottin. Pâris’ name appears at the foot of numerous nautical charts, including those of Tabou (Tonga), Havre Carteret (New Zealand) as well as islands and straits of northern

On March 12, 1871, less than a week after leaving active duty, Rear Admiral Pâris succeeded Morel Fatio, the maritime painter, as the conservator of the “Musée naval du Louvre”. Also known as the “Musée de marine”, it had been founded on December 27, 1827. Until his death, on April 8, 1893, Pâris devoted his energies to making the “Musée naval” a museum encompassing watercraft and maritime traditions of the past and present, from France and the rest of Europe as well as from extra-European countries1.

2

The complete title is: “An essay on the naval architecture of the extra-European peoples or an Inventory of ships and canoes made by the inhabitants of Asia, Malaysia, the Great Ocean and America, measured and drawn during the circumnavigations of the Astrolabe, the Favorite and the Artémise”. “Essai sur la construction navale des peuples extra-européens ou Collection des navires et pirogues construites par les habitants de l’Asie, de la Malaisie, du Grand Océan et de l’Amérique dessinés et mesurés pendant les voyages autour du monde de l’Astrolabe, la Favorite, et l’Artémise”. For a critical edition of all the Plates (plans and lithographs) of the Atlas, see (Rieth 1993). 3 The Essai does not bear a copyright date. However numerous documents in Pâris’ personal dossier, archived at the Service Historique de la Défense, département Marine, at Vincennes (cote CC7/1908), suggests that the edition likely dates to the beginning of 1843. In a letter dated December 7, 1842, Captain Pâris (capitaine de vaisseau) clarifies that “…The publication of the Essai on the naval architecture of the people foreign to Europe… is already well underway and will soon be completed. I hope that you will allow me to express my wish to return to sea” (Rieth 1993: 18-19, 30). Note that in this letter, Pâris evokes the title of the Essai by referring to “people foreign to Europe” and not yet to “extra-European people”.

Pâris, sailor, researcher and museum conservator, left an imposing body of scientific work touching on steam propulsion, propellers, armoured ships, marine life saving, handling of both sail and steamships, as well as ∗

CNRS, Laboratoire de Médiévistique Occidentale de Paris, Musée national de la Marine. 1 Shortly after his appointment to the direction of the Musée naval, Pâris wrote an article defining his scientific position, one that would still be current today (Pâris 1872).

47

ERIC RIETH

Sulawesi. But every time the opportunity presented itself, Pâris would study the local boats and make, whenever possible, architectural drawings, as he had been ordered by Dumont d’Urville to “draw all the different kinds of canoes that we will find in the savage countries”. Therefore, during the Astrolabe’s layover in Tonga Tabou

(Tongatapu) from late March to early May 1827, Pâris measured and drew plans of numerous canoes, of which four manuscript plans are conserved in the “Musée national de la Marine” (Figs. 1-4). In the legends of the figures, Pâris’ original titles are in quotes and preceded by their catalogue number.

Figure 1. MnM Inv. B9n-8154, “Tonga-Tabou Pirogue de mer nommée Vaca” (Tonga-Tabou, Sea canoe known as a Vaca).

48

Figure 2. MnM Inv. J9n-8195, “Pirogue ou Tafahanga de Tonga-Tabou” (Canoe or Tafahanga from Tonga-Tabou).

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

49

Figure 3. MnM Inv. B9d-8005, “Tonga-Tabou Pirogue double nommée Calié” (Tonga-Tabou, Double canoe known as a Calié).

ERIC RIETH

50

Figure 4. MnM Inv. J9n-8226, “Calié de Tonga-tabou” (Calié of Tonga-tabou).

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

51

ERIC RIETH

During the voyage of the Astrolabe, Pâris’ studies of native watercraft were sometimes conducted under circumstances quite different from the apparently paradisiacal conditions of Tonga Tabou. The layover in Vanikoro is a good example. In mid-February 1828, the corvette came in sight of Vanikoro (Solomon Islands) looking for traces of the Lapérouse expedition. The high temperature and humidity of the climate was particularly disagreeable and the inhabitants were less than

welcoming. With the help of an islander the site of a shipwreck, presumed to be that of Lapérouse’s Astrolabe, the first ship to bear that name, was located. The typical remains of a modern ship, including anchors, cannons, iron cannonballs, ceramics… were scattered about under several meters of water. A few of these remains were collected and taken aboard the second Astrolabe. During this mission that ended the mystery of the loss of the Lapérouse expedition in 1788, Pâris made

Figure 5. MnM Inv. J9n-8219, “Radeau à balancier de Tevai, Vanikoro” (Outrigger raft from Tevai, Vanikoro). drawings of numerous vessels4. Only one manuscript plan, that of a raft with an outrigger, is today preserved in the “Musée national de la Marine” (Fig. 5).

hydrographic skill he had demonstrated on the Astrolabe; although the principal objective of la Favorite was diplomatic. Her mission was to show the French flag along the coasts of east Africa, India, Malaysia, Indonesia, the Philippines, China, Indochina, Australia, New Zealand and South America. Besides these political and commercial ends intended to counter increasing British colonial power, Laplace was also to continue the hydrographic mapping started by his predecessors. Pâris’ role was decisive in this domain and earned him the rank of “chevalier de la légion d’honneur” on April 30, 1833.

Pâris’ second circumnavigation, on board the corvette la Favorite commanded by Captain (capitaine de frégate) Laplace, took place between December 30, 1829 and April 20, 1832. Pâris was ordered on this voyage for the 4 Plate 114 of the Essai shows the plan of a “grande pirogue du village de Manévé dans l’île Vanikoro” (figs. 1-8) (large canoe from the village of Manévé on the island of Vanikoro), that of a “pirogue de Tevai” (canoe from Tevai) (figs. 4-6) and the plan of a “radeau à balancier de Tevai” (outrigger raft from Tevai) (figs. 7-8). Lithographs of two (pl. 161-162) of Pâris’ drawing of canoes from Vanikoro were published in d’Urville (1833).

In a note dated April 10, 1833 signed by Baron Tupinier, the director of Ports and Arsenals, it is emphasized that “… Mr. Pâris had already made great services to hydrography during his voyage on the Astrolabe, when he

52

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

Figure 6. MnM Inv. B 178, “Pirogue du Grand Ocean vue à Valparaiso” (A canoe of the ‘Grand Ocean’ seen at Valparaiso).

53

ERIC RIETH

On April 22, 1839, barely ten months after the accident in Porto Novo, the Artémise struck and grounded on an uncharted coral reef in the entrance to Papeete. The damages were important: the after part of the keel, the garboard and the second strake were torn off. Major flooding ensued, necessitating repairs to the hull. The frigate was towed into Papeete bay by the ship’s boats, then careened ashore to fix the damage. This delicate and impressive operation was carried out by Pâris with the assistance of an Irish sea captain resident in Tahiti, several captains of whaling ships and many Tahitians. More than six hundred men were necessary to careen the frigate6.

embarked on the Favorite”. Tupinier concluded: “I cannot praise this officer’s zeal and talents too highly” (Rieth 1992: 42). From the Favorite’s voyage around the world, Pâris brought home, among other documents, an impressive collection of one hundred and eleven water coloured drawings and three pen and ink drawings bound into an album, today conserved in the “Musée national de la Marine”5. Contrary to the route followed by the Astrolabe, the Favorite did not sail through Oceania. And yet, plate 113 of the album drawn during a port stop in Valparaiso (Chile) between November 14, and December 10, 1831, shows, according to Pâris a “Pirogue du Grand Océan vue à Valparaiso”, (“a canoe of the Great Ocean seen at Valparaiso”) (Fig. 6). It illustrates a single canoe with an outrigger to port. The dugout hull is raised by sewn planks. The prow is elongated by a narrow horizontal platform. How did this manifestly Polynesian outrigger canoe come to be found in Valparaiso? Was it brought back by a merchant ship or a whaler who had stopped on a Polynesian island? Was it sailed to Chile by Polynesians? Pâris gives no explanation on the origin of this typically Oceanic craft.

During the nearly two months of the Artémise enforced layover in Papeete, Pâris was able to collect rich documentation on canoes 7 . In total six canoe plans were drawn in Tahiti among which three manuscript plans are conserved in the “Musée national de la Marine”8 (Figs. 78). F

F

Upon returning to France, Pâris was ordered to serve, first in Brest, later in Paris at the “Dépôt des cartes et plans de la marine” (The Naval depository of charts and plans) as assistant to commandant Laplace to edit the official account of the Artémise’s voyage. While in Paris, he also worked on the edition of his “Essai sur la construction navale des peuples extra-européens” which featured plans of canoes from New Zealand, Melanesia, Micronesia and Polynesia on eighteen plates (Pâris 1843: pl. 105-106, 109-114, 116, 118-119, 121-127) and five lithograph plates (Pâris 1843: pl. 107-108, 115, 117, 120) illustrating canoes in their nautical environment.

Pâris’ third and last circumnavigation was made under Captain (capitaine de vaisseau) Laplace on board the Artémise between January 21, 1837 and May 11, 1840. The primary politico-commercial objective was analogous to that of the Favorite. However the Artémise followed a different route that carried her through the waters of New Guinea, Polynesia, and Hawaii, allowing Pâris to complete his documentation of canoes of Oceania. But before discussing this subject, it is important to briefly recall a sombre episode in the life of lieutenant (lieutenant de vaisseau) F.-E. Pâris.

After publication of the Essai, it was a long time before Pâris was again able to devote time to the study of extraEuropean watercraft. It was not until 1871 and his nomination to Conservator of the “Musée naval du Louvre” that he again began research on traditional watercraft of Europe and beyond. During this last period of professional activity, Pâris dedicated himself to completing documentation of extra-European canoes from a museographic perspective. From a strictly scientific point of view, his work on the naval architecture of extra-European boats had been completed with the publication of the Essai in 1843 and the construction of a series of canoe models based on his plans, for the Musée naval du Louvre, then under the

During the Artémise’s port stop in Pondichery (India), in June 1838, Pâris was sent ashore on a fact finding mission to Porto-Novo. He was to investigate the machinery in a factory as part of the politico-commercial objective of the expedition. During his visit, which seems to have involved some industrial espionage, Pâris was severely injured, either through imprudence or bad luck. Whatever the case, his injury required the hasty amputation of his left forearm. This grave injury in no way interfered with his participation in the mission of the Artémise, nor with his career as a naval officer as indicated by his subsequent commands at sea. During the Artémise’s long layover in Tahiti, Pâris exhibited willpower, courage and a remarkable capacity for recuperation.

6

For a technical description of this operation, see Jean Boudriot, “La frégate l’Artémise et son abattage à Tahiti en 1839”. Neptunia, 1976 (121): 37-44. 7 See E. Rieth “L’amiral Pâris (1806-1893): un regard ethnographique sur les embarcations de la Polynésie française”, to appear in 2007 in a collective volume coordinated by Tara Hiquily, Musée de Tahiti et des Îles. 8 Ref. Musée national de la Marine, Paris (service recherche): B9n-8033 (see fig. 7), B9n-8036 (see fig. 8) and B9n-8141, “Tahiti. Grande pirogue de pêche” (Tahiti. Large fishing canoe).

5 Ref. Musée national de la Marine, Paris (service recherche): B 178. For a critical edition of this album see: Le Voyage de la Favorite. Amiral Pâris, text and legends by Eric Rieth, Editions Anthèse, Paris, 1992.

54

Figure 7. MnM Inv. B9n-8033, “Ancienne pirogue de Tahiti l’Ivahae du temps de Cook” (Ancient canoe from Tahiti, an Ivaha from Cook’s time).

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

55

ERIC RIETH

Figure 8. MnM Inv. B9n-8036, “Archipel Po-motou. Pirogue double mesurée à Tahiti en 1838” (Po-motou archipelago, double canoe measured on Tahiti in 1838). workshop in 18469 (Fig. 9). In his 1872 paper, where Pâris discussed his museographic program, he noted that:

direction of Naval Engineer Lebas. Pâris served as scientific consultant for these models beginning in the years 1840-1845. One of them is the double canoe or “calié” [Kalia] from Tonga built in the museum

9

The model was built after the plan manuscript B9d-8005 (see Fig. 3); ref. Essai, pl. 118 and 121 (figs. 1-3). The model, at a

56

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

documentation resulting from Pâris’ impressive work, which he continued right to the end of his life, is extremely rich.

“…the ingenious constructions of the savage peoples are barely represented by a few models built during Mr. Lebas’ time and in the meantime they have disappeared from most of the islands where thirty years ago one could still admire the native’s skill at overcoming their total lack of metal such as we need to fashion and connect all the parts of our boats”. In reference to Polynesia, he specified “Who would now suppose that in Tahiti, king Otou once had a fleet including a double canoe 110 feet long and that in Tongatabou, in 1826, there were canoes 80 feet long, partly sewn together that were stored in slipways covered with elegant ogives” (Pâris 1872: 980). Faced with the lack of models of such canoes in the museum collections as well as the dearth of models of steam and iron ships, Pâris spelled out his museographic choices: “It is overall toward these two kinds of vessels [extra-European canoes and iron and steam ships] that the feeble resources of the Musée de marine, reduced to three workers, and most of the conservator’s salary, are being used”.

What are the main characteristics of this body of work? First off, this was original data, acquired during field studies carried out either by Pâris or by his correspondents, whom are known by name for the most part. This basic source material consists of a series of manuscript plans, drawn (and sometimes coloured), dated, located and archived in the Musée national de la Marine13. Pâris as a good hydrographer was very particular about the accuracy of these plans. For the canoes of the Caroline Islands he insisted for example on “…the preference that we have always given for precise plans. At Umata, on the island of Guam, we have measured numerous Pros that came from the island of Satahoual [Satawal] among which we have chosen that of pl. 106 to show all the details of this curious canoe” (Pâris 1843: 97). With his concern for scientific rigour and honesty, Pâris often explicitly noted lack of precision in measurements or observations that occurred during his field studies. When the Astrolabe was coasting along the north shore of New Guinea, a conflict ensued when canoes loaded with armed men came alongside. Pâris took this opportunity to follow d’Urville’s instructions to “…draw all the different kinds of canoes”. These instructions were carried out in rather unusual circumstances, which may explain a certain imprecision in the record. Pâris explains: “During the time they [the canoe men] spoke among themselves, we climbed up to the mast top, from where it was possible to draw a plan of their canoes; into which dimensional errors have perforce slipped” (Pâris 1843: 93).

Shortly after defining his museographic program, which today one would call a scientific and cultural project, Pâris began construction of a series of models, calling on personnel of the museum workshops as well as on independent model makers. It is within this context that the Tuamotu double canoe model was built, after the plan drawn onboard the Artémise in 1839. Baude, an independent modelist made it for 320 francs, a sum that Pâris paid out of his own salary10 (Fig. 10). Pâris donated the completed model to the “Musée naval”. The model of the large Tahitian fishing canoe has a similar story. It too was drawn in 1839 during the enforced layover of the Artémise in Papeete, and built by Baude “a skilled shipwright from Cherbourg”, in 1875 for the sum of 65 francs11.

To these original data, should be added the published work, the “Essai sur la construction navale des peuples extra-européens”, wherein the great majority of the plates relating to canoes of Oceania were taken from his manuscript plans. Where this was not the case, Pâris took care to cite the original source. An example is plate 111 illustrating the canoes of the Caroline Islands, after the plan manuscript B9n-8134 (Fig. 15). Pâris noted: “This plate is copied from the voyage of the Coquille, commanded by Mr. Duperrey”, referring to the circumnavigation by lieutenant (lieutenant de vaisseau) Duperrey on board the Coquille from August 11, 1822 to March 24, 1825. Similarly, Pâris notes that on plate 125, the plans of a “Grand Pahie de guerre de Taïti” (Grand war Pahi of Taïti) and those of a “… pirogue du port de la Résolution dans les îles de Nouka-Hiva) were copied after Captain Cook’s second voyage.

Pâris based the construction of the models on his documentation collected during the years 1829-1840. He filled gaps in his data by calling on a vast network of correspondents and informants, both civilian and military, that he had established during the years of his missions and postings. One of them, particularly valuable for nautical questions, was a Mr. Amourous (or Amouroux), a surveyor who had made scaled drawings of two New Caledonian canoes which became the basis of two of the models12 (Figs. 11-14). As one can see, the scale of 1/16.6, was made in 1846 in the naval museum workshop. 10 This model was built after the plan manuscript B9n-8036 (see Fig. 8); réf. Essai, pl. 126. 11 Réf. Musée national de la Marine, Paris (service collections): inv. 33 EX 9; ref. Plan manuscript (service recherche): B9n8141; ref. Essai, pl. 123, figs. 4-6. 12 There are two plans for each canoe model. The canoe model 1 (inv. 35 EX 1) was built after the plan manuscript J9n-8205 (see Fig. 13) and J9n-8207, “Nlle Calédonie. Pirogue des naturels d’après M. Amouroux” (New Caledonia. Canoe of the natives by M. Amourous). The canoe model 2 (inv. 35 EX 2) (see Fig.

14) was built after the plan manuscript B9n-8114 (see Fig. 11) and J9n-8196 (see Fig. 12). 13 This documentation was digitized in 2006 to facilitate consultation and preserve the originals from too frequent handling.

57

ERIC RIETH

Figure 9. MnM Inv. 31 EX 3, “Calié, pirogue double de Tonga-Tabou” (Calié, double canoe from Tonga-Tabou).

Figure 10. MnM Inv. 33 EX 15, “Pirogue double des îles Tuamotou” (Double canoe from the Tuamotu Islands). About the Polynesian canoes of Tonga, Pâris explains with admirable clarity, how “The Caliés [Kalia], as do all the outrigger canoes, tack like the Carolinian Pros… When one changes tack, one carries it [the steering oar] to the end that will become the stern. We have attempted to show on plate 120 all the operations of this swiftly carried out manoeuvre; one sees the men pulling on the opposite stay and others busy shifting the steering oar; we have only represented the small Vaca on which we tacked, since too many crewmen would lead to confusion

Besides the plans, the Essai also contains lithographs that can be read on two levels. The first level, most direct and subjective, is that of aesthetics. Each lithograph portrays one or more canoes in their nautical context. The sea and the men give a sense of scale, less abstract than strict geometric lines to the eye of a reader interested in voyages and exotic horizons, but not especially knowledgeable about extra-European marine architecture. The second level is strictly documentary, illustrating and complementing the plans, showing the particularities of certain manoeuvres.

58

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

Figure 11. MnM Inv. B9n-8114, “Nouvelle Calédonie. Pirogue double d’après les dessins de Mr Amourous” (New Caledonia. Double canoe after drawings by M. Amourous). about the role that each man had to play…” (Pâris 1843: 119).

sails of the Tongan archipelago, Pâris points out that these “…sails, as simple as they are ingenious, set very well, handle the wind better than those of our boats, this we verified during a passage of numerous miles on a Vaca rigged with such a sail” (Pâris 1843: 119).

The text volume of the Essai describes, analyses, compares and interprets its associated Atlas of plates, comprising plans and lithographs. It is unknown whether the manuscript and the preparatory notes for the Essai have been conserved. We are left with the published text, still a fundamental reference more than 150 years after it was written. This work has been declined in five verbs; observe, draw, describe, analyze and compare (Rieth 2001). A sixth verb should be added: experience since Pâris, as both sailor and researcher, sailed on some of the boats detailed in the Essai so as to better understand how their rigs functioned. Apropos the triangular woven mat

Finally, the documentation also includes the collection of models built after Pâris’ plans. The “curriculum vitae” of these models is for the most part known. That is the sources on which they were based, their dates of manufacture, their scales, the name and status of the model maker (museum staff or not), their cost, and how they were acquired by the museum (gift, purchase)...

59

ERIC RIETH

Figure 12. MnM Inv. J9n-8196, “Nouvelle Calédonie. Pirogue d’après Mr Amourous” (New Caledonia. Canoe by M. Amourous). ashore were limited. They varied from a few days up to two months at the longest. In any case port calls were unusual. During these stopovers, he also had to accomplish his primary duties of hydrography intelligence gathering, or careening the Artémise during the 1839 layover in Tahiti. In other words, his time ashore was far from exclusively devoted to a systematic, methodical research program. In all likelihood, these conditions limited the scope of his scientific research to an architectural and descriptive approach leaving little opportunity for more profound observations. A second problem relates to the drawings and the plans that constitute the base of the documentation. Without lingering over the question of possible errors in the plans, such as the almost unavoidable inaccuracies linked to redrawings and despite Pâris’ preference for “…exact plans”, one may question the precise nature of the manuscript plans - whether originals or copies - in the Musée national de la Marine. In this regard, Pâris gives a few clues.

One of the exceptional characteristics of Pâris’ collections and one of its major interests for researchers studying extra-European watercraft and those of Oceania in particular, is the opportunity to use its three interconnected levels of documentation together - the manuscript plans, the plates and text of the Essai, and the scale models. All three are accessible at the “Musée national de la Marine” in Paris. However, reading this remarkable scientific material requires a critical eye. One of the problem is that his 1829 to 1840 field research was supported and encouraged but also subordinated to other priorities by the military hierarchy. It appears that Pâris had a good knowledge of general history and had as well acquired solid grounding in the history of extra-European culture. The latter was central to his work at the time. Still, the fact that his research spanned virtually the entire globe led him inevitably to remain something of a “generalist”14. Moreover, his stops 14 Thus explaining certain approximations and superficial approaches.

60

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

Figure 13. MnM Inv. J9n-8205, “Pirogue à balancier d’après les dessins de Mr Amourous” (Outrigger canoe after M. Amourous’ drawings).

61

ERIC RIETH

Figure 14. MnM Inv. 35 EX 2, “Pirogue double de Nouvelle Calédonie” (Double canoe from New Caledonia). In the field, each vessel first had to be carefully measured. Given Pâris’ preference for scientific rigour, one expects great precision in his plan. Yet, his comments on the Tuamotu canoe plan raise certain questions. He notes: “These canoes are far from symmetrical, as indicated on the plan and sectional views, whose lines were traced with care to the nearest decimetre by abscises and ordinates; the canoes themselves prove how useless it is to carry precision this far, it serves merely to highlight irregularities invisible to the eye and only noticeable with a ruler and compass in the hand” (Pâris 1843: 135).

Once the measurements were taken, Pâris made his finished drawings on board ship, going from his tables of coordinates to line drawings of plan, horizontal and transverse views. These two stages of the work are clearly mentioned in the introduction to the Essai: “…all measurements were first taken with care, then transferred onto plans drawn on board the Astrolabe, the Favorite and the Artémise” (Pâris 1843: 4). Despite careful analysis of each manuscript plan drawn and signed by Pâris, it is difficult to be sure whether they are the originals drawn on board ship or whether they are copies made at a later date for the plates to be published in the Essai or for use by the model makers. Many of them seem to have been redrawn for the latter. Certainly the risk of errors, while not an immutable law, augments with the lapse of time from taking the measurements to the original drawings to making the final copy(s).

Did he always keep to this same rigor in his measured drawings by abscises and ordinates; a rigor which he seems to have considered of secondary importance for understanding naval architecture? Indeed, given his comments on the irregularities of the Tuamotu canoe hulls, in particular their asymmetry, was he not sometimes - justifiably in his eyes - led to take less care in his measurements? How can one therefore evaluate ‘precision’ from ‘imprecision’, ‘exact’ from ‘inexact’?

Let us make a last comment about Pâris’ plans. Beyond their scientific value, they also possess a memorial dimension as emphasized, probably not by accident, in the last four lines of the Essai (1843: 156): “It is perhaps no less interesting to examine these ingenious efforts, and

62

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

These vessels, whose history has been too long ignored were introduced into the field of research by Pâris. He recognized that, although these boats were primarily used for fishing and local transportation, they represent privileged subjects of ethnographic study due to the structural and morphological diversity of their hulls and of their modes of propulsion and steering. These boats, in symbioses with their nautical environment and their local or regional socio-economic milieu are, according to Pâris “…in these places and [as] handled by the inhabitants…one could never find better” (Pâris 1843: 2).

moreover to preserve them for the future, since in a time not far from now the only traces of most of these canoes [will be] these plans”. Such are the principal characteristics of this unique documentation on extraEuropean naval architecture. Pâris was not simply content to record data and one of the originalities of his work, which was a profoundly novel approach at the time, was his use of historical problematics in analysing and interpreting the data. Historical problematics for the study of traditional watercraft

Before returning to his ethnographic approach, it may be useful to recall here, because the fact seems little known, the opposition between the relative uniformity of vessels conceived for the open sea and the great diversity of those built for coastal navigation. This was expressed by the celebrated Swedish engineer-shipbuilder F. H. af Chapman in his 1775 treatise on Naval Architecture that was published in French in 1779. Pâris was probably inspired by Chapman, whose work he was certainly familiar with. Concerning high seas navigation, Pâris noted: this “navigation in general, that must take account of any eventuality… led little by little to uniformity”. In any case, Pâris amplified Chapman’s idea, put it into an historical perspective, and extended it to extra-European cultures. Pâris went well beyond the purely technical and architectural aspects, integrating an anthropological dimension: “In countries that lack our technical means, one sees clear-cut variation; everyone had first off to adapt their inventions to their resources, and they succeeded. Their boats are all of a well marked type, they are simple and perfectly adapted to their locales and needs” (Pâris 1843: 3).

Pâris was not only a privileged observer on his three circumnavigations, but also a scientist who opened a new field of study, that of nautical ethnography. If it is traditional to qualify the French historian Augustin Jal (1795-1875) as the “father of maritime archaeology” due to his outstanding publications, such as the two volumes of his Archéologie navale published in 1840 and his very large Glossaire nautique published in 1848, it is as justifiable to qualify Pâris as the “father of nautical ethnography” even if this expression is contemporary and is not explicitly mentioned in Pâris’ work. The main themes of this new field of research were formulated with great pertinence in the introduction to the Essai. Later, notably after his nomination to the head of the Musée naval, Pâris improved, completed and nuanced his ethnographic ideas, while remaining ever faithful to the scientific choices made during his youth, from 1830 to 1840. The introduction to the first volume of his “Souvenirs de marine conserves” published in 1882 could in a certain manner be substituted without notable modifications for the introduction to the Essai15. Even so, this profound stability of thought should not be confused with a fossilization of ideas.

Pâris gives numerous examples of an anthropological approach to the naval architecture of canoes of Oceania. One example concerns the large double hulled war canoes called “caliés” [kalia] of Tonga (Pâris 1843: 117-119, pl. 121). About their socio-cultural dimension, Pâris notes that each high chief owned several “caliés”, each built by carpenters whose “… hereditary profession confers a kind of noble title”. The canoe, as a representation of power, not just that of its owner, but that of its builder as well, is put forward. Still, it is of course on the canoe itself, its material, its structure, its form, its function… that Pâris focuses his entire attention.

His thinking was that extra-European watercraft, those of ‘savage’ or ‘native’ peoples, to use 19th century terminology, might be objects of history like other material expressions such as religious structures which by their prestige had long been recognized as such. In the introduction to the Essai, which serves as a kind of scientific manifesto, Pâris not without a certain lyricism, notes that a canoe represents “…the most beautiful masterpiece of the human spirit, no monument, no invention equals its marvellous totality, and although having become commonplace, as anything one sees every day, they deserve no less admiration than that heaped so prodigiously upon other objects”. He adds: “We are in the most complete ignorance of the maritime world of [these] people about whom we know the details of their clothing, weapons and the most common utensils they used in the past” (Pâris 1843: 2).

The first point is that of their means of production. In this regard, one of the biggest problems encountered by the people from Oceania seems to be the absence of iron minerals and metallurgy disallowing the use of iron tools, nails or metal pins to join the wooden parts of their boats. Yet the “caliés” of Tonga studied by Pâris were more than 10 meters long, constructed of elements held together and assembled by lashings of coconut fibres covered with a plant resin similar to pitch. The lacing holes were made with tools of mammal or fish bone or of fire hardened wood. About these lashings, Pâris notes

15

At least in regards to basic principles since in “Souvenirs de marine conservés”, Pâris concentrated his efforts on traditional boats of Europe.

63

Figure.15. MnM Inv. B9n-8134, “Archipel des îles Carolines. Pirogue de l’île Duperey dessinée à bord de la Coquille. Archipel des Carolines. Pirogue de Rotouma” (Caroline archipelago. Canoe from Duperey Island drawn onboard the Coquille. Caroline archipelago. Canoe from Rotouma).

ERIC RIETH

64

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

that: “…the sewing [is] the most ingenious that we have seen. We could not understand how it was done other than by the signs that the natives were glad to show us”. This remark reveals how, whenever it was possible, Pâris would record the various steps of manufacture, or chaîne opératoire, leaving out no detail.

paddling; the Malays have [even] used them for their ocean going ships fitted out with two or three ranks of oars” (Pâris 1843: 4). Again one finds the importance of the nautical environment perfectly expressed as is the adaptation between the milieu and the technical choices of the extra-European boat builders.

The second point is that of the management of the boats. This is another originality of Pâris’ work. He was not content simply to carry out a ‘static’ architectural study, but he enlarged his field of analysis to the boat’s behaviour at sea and the way it was handled. In regard to the sailing characteristics of the Tongan “caliés”, he notes that the movements of the two unequal length hulls linked by a platform are not synchronous and that this provokes “…considerable torsions to the platform; [the canoes] crack in a disquieting manner whenever the swells get rough and from time to time their sewn joints open, letting the water in”. The problem raised here by Pâris, that of the forces working on the linkages between the hulls, is still a difficulty faced by modern racing multihulls. Two choices are available to today’s builders: that of a light and flexible coupling as used by canoes in Oceania and that of a heavy rigid coupling. The first option is usually seen on modern trimarans while the second appears to be favoured on catamarans.

Although Pâris notes the importance of sail for single outrigger Oceanic canoes, he does not neglect to point out certain defects. Thus he emphasizes the tendency of the Tongan “vacas” to luff when the wind decreases (Pâris 1843: 121). He goes on to point out that this ‘defect’ doubtless explains the position of the mast to windward and the steering oar to leeward which decreases the tendency to round up (Pâris 1843: 122). The question of canoes stabilized by either one or two outriggers, formulated in terms of a techno-cultural marker, thus not limited to the scale of a community or an island, but enlarged to the Malay and Oceanic cultures as a whole, touches on peoples and racism. There is indeed a racial aspect to Pâris’ thinking, although one should look at it in light of the pre-colonial ideological context of the era, as well as the importance of the race concept in 19th century anthropological thought. In this regard certain citations extracted from the Essai are very revealing.

The Essai includes a series of monographic studies, as exemplified by the chapter on Tongan canoes, each more or less follows a similar outline based on Pâris’ characteristic anthropological approach to watercraft. According to the area under consideration, certain aspects are more developed than others. This is the case, for example, of a fundamental topic for Oceanic naval architecture, the outrigger “…an ingenious artefact by which an external weight maintained by levers at a distance resists the force on the sail allowing a very large increase in the sail area while the canoe alone could not even float [upright] and would capsize without the aid of its outrigger” (Pâris 1843: 4).

In his introduction to the canoes of the ‘Grand Ocean’, Pâris first off recalls that sailors and researchers “…have observed general analogies between peoples grouped into distinct races” (Pâris 1843: 94). For example, he notes that the Oceanic people of the Caroline Islands “…have handsome features; straight hair, well proportioned limbs” and “… although they are natural warriors, they are less vicious than the negros” (Pâris 1843: 95). Making a comparison based on physical anthropological criteria of his time, Pâris remarks that the people of the Santa Cruz Islands “…the black race of the Pacific Ocean [have a] close resemblance with African negros”. But then adds that these men of the Pacific isles “…are less lazy; seem to love navigation and build canoes that lack nothing of the grandeur of those of other archipelagos” (Pâris 1843: 110). Apropos the people of the Tongan Islands, Pâris felt that these “…islanders have merely to occupy themselves with their pleasures, and since nature has instilled them with an active character, they have not used these gifts as have the negros, who consider laziness as the greatest of benefits” (Pâris 1843: 115).

The various systems of single outriggers are described in the Essai as a function of the length, form and number of cross beams on one hand, and the length and form of the float on the other, and finally by the presence or absence of a balance platform. Within the diversity of types of single outriggers from various archipelagos, there are underlying similarities that led Pâris to distinguish two techno-cultural worlds, that of the people of Oceania and that of the people he called the Malays. The latter were associated with the double outrigger, like modern trimarans, and the former with the single outrigger, or proas: “Each has its advantages: the first serves in calm uniform seas, among the countless islands of the great Asian archipelago; the second supports more sail area, follows the movements of a rough sea better, causes less wear and allows crossing the fairly long distances that separate islands in the south seas. It is well adapted to sailing whereas the other is better for rowing or

Without focussing any longer on the racial, or even racist, content of such a position, that make one shudder today, three particular aspects deserve comment. Firstly, Pâris is quasi explicit in establishing a relationship of influence, even determinism, between the landscape, climate and resources of Africa and the African people. This link, negative here, between nature and culture, is emphasized throughout the Essai and concerns, primarily the canoes. Certain statements are

65

ERIC RIETH

without finding a single [canoe] house; most of the villages possessed only one or two canoes” (Pâris 1843: 133). In short, this assessment is dramatic. In Tahiti, Pâris measured and drew a total of five fishing or transport canoes of different types. Besides the large Tuamotuan canoe with two dissymmetric hulls nearly 13 meters long (Pâris 1843: 135-137, pl. 126), the truly Tahitian vessels consisted of: a double fishing canoe with two unequal length hulls, a bit more than 6 meters long with a fitting for torch fishing on its bow (Pâris 1843: 134, pl. 122, figs. 15-16), a small single outrigger fishing canoe just over 5 meters long (Pâris 1843: 135, pl. 123, figs. 5-8), a single outrigger fishing canoe slightly more than 8 meters long (Pâris 1843: 135, pl. 124, figs. 1-3), another single fishing canoe similar in length to the preceding one (Pâris 1843: 134, pl. 124, figs. 4-6) and finally a large outrigger canoe about 11 meters long (Pâris 1843: 133-134, pl. 123, figs. 1-4). According to Pâris, the latter, in poor condition, had doubtless been abandoned for many years. It was the largest they observed on Tahiti and may well have been a voyaging canoe. Pâris’ drawing shows a hull built of planks laced together without caulking. The poop is raised into a curve while the prow is low and horizontal. The hull is fitted with a single outrigger whose forward crossbeam extends on the side opposite the float, forming a balance platform. This canoe seems to have been the sole naval architectural testimony of the great inter-island voyages once made from Tahiti still extant during Pâris’ time.

eloquent, nearly caricatures since their logic seems so simplified, even simplistic: “…they [the Africans] feed themselves with what nature provides without working… It is therefore not surprising that there, navigation has remained completely hopeless: their canoes are nothing but hollowed tree trunks; they lack all the ingenious additions used everywhere else and are only suitable for rivers and coastal bays” (Pâris 1843: 6). Secondly this profoundly negative attitude toward Africa is expressed in a diametrically opposite manner for Oceania, its peoples and its canoes. Pâris, as did others before and after, yielded to a fascination for the Pacific islands. We will come back to this. Thirdly, the relation between nature and culture, sometimes positive, other times negative, could have perhaps been formulated by Pâris into ‘people of the forest’ and ‘bad natives’ on one hand and ‘people of the sea’ and ‘good natives’ on the other. Pâris puts the Oceanic peoples, whose boats and nautical knowledge represented the highest signs of their culture, into the second category. Under Pâris’ quill, Polynesia seems to have had a particular attraction among the various cultural components of Oceania. Certainly his stay in Tahiti, following the grounding of the Artémise in 1839, was one of the longest stops during his three voyages around the world. It doubtless enabled him to increase his knowledge of the island and its inhabitants. Moreover Pâris never forgot the wonderment that Cook and Bougainville had felt discovering the countryside and people of Tahiti when he evokes the “…temperate climate favouring the most prodigious and rich gifts of nature; [the people] worry about nothing but their dances and pleasures”. Proof that it was a kind of Polynesian Eden, an Eden in part reconstructed by the explorers, where the only difference separating the social classes, the rich and the poor may have been, according to Pâris, their decorations of flowers and feathers, “…the only sign that distinguished the upper classes since nature was the same for all the people; in a country where the land is so rich, poverty could not exist” (Pâris 1843: 123).

Pâris who developed a new anthropological approach, that of nautical ethnography, often went beyond strictly objective scientific observations. From the very beginning of the Essai, Pâris did not hesitate to take positions and transformed the introduction into a kind of manifest for the study of extra-European watercraft not only as objects of History but also of Memory. And it is this memorial dimension, beyond that of the science of history, that led Pâris not only to observe, describe, compare, analyze and interpret, but also to judge, sometimes harshly. About the people of the Caroline Islands, he wrote “…[their] weapons and canoes are what they have most perfected”, he clarified that the “Europeans have unfortunately come and disrupted everything in the name of civilisation”, causing all the traces of material culture to disappear, “…these objects of which there remain only a few scattered drawings, were burned by fanatics or traded to be sold as curios: this fact follows naturally from the Europeans’ behaviour, who upon becoming the masters exploit everything for their own profit” (Pâris 1843: 95). As a consequence of their devastating attitude, Pâris observed that no more canoes were being built in the Caroline Islands. From that time onward the people seemed to prefer to acquire old boats from visiting whalers.

But this society perceived as quasi-egalitarian - we are perhaps not far from primitive communism - is not in reality that which Pâris would in his turn discover nearly seventy years after Bougainville. “…It is no more the Taïti so gay, so brilliant of Bougainville and Cook; the numerous population has already disappeared, today the land is partly deserted” (Pâris 1843: 124). Travelling across the island, Pâris observed villages practically abandoned, their populations becoming acculturated, living in misery. As for the voyaging, fishing and war canoes, Pâris noted that their number had been considerably reduced. He deplored: “…the long stay that we made among them [the Tahitians] was used for [studying] the canoes; [but] all the large ones have disappeared” (Pâris 1843: 125). And he added: “…We have sailed around the island, navigating nearly always inside the reefs, usually within a few meters of the shore

If relations with European and American sailors, notably whalers, common in the Pacific at that time, were in part responsible for the acculturation of the Oceanic peoples,

66

TRADITIONAL OCEANIC CANOES AS SEEN BY ADMIRAL PÂRIS

D’Urville, M.J. Dumont 1833 Voyage de la corvette l’Astrolabe exécuté pendant les années 1826-1829 sous le commandement de M. Jules Dumont D’Urville, capitaine de vaisseau. Atlas Historique. J. Tartu, Paris. Rieth, E. 1993 Voiliers et pirogues du monde au début du XIXe siècle. Essai sur la construction navale des peuples extra-européens de l’amiral Pâris (1843). Editions Du May, Paris. (in press) “L’amiral Pâris (1806-1893): un regard ethnographique sur les embarcations de la Polynésie française”. In Tara Hiquily (ed.), Musée de Tahiti et des Îles, Papeete.

Pâris felt that the heaviest responsibility lay with the Anglican missionaries. Regarding the degradation of the Tahitians’ conditions of life, he forcefully emphasised that “…the propagators of the Anglican faith in the Grand Ocean unite fanaticism with the most active mercantile spirit…They reign as masters over the incredulous natives” (Pâris 1843: 125). Indeed, while condemning the practices of the missionaries, Pâris exhibits an evident paternalism vis a vis the ‘natives’, considered ‘big innocent children’ in the classic colonialist scheme of thought. Through his criticism of the impact of European culture on that of the extraEuropeans, Pâris seems to have been unafraid to appear as what today is called an ‘engaged researcher’. Conclusion

Note Pâris’ intellectual engagement for the defence of extraEuropean societies subscribes to a certain logic. Indeed, Pâris the ‘father of nautical ethnology’, devoted an important part of his activities as a sailor, scientist and conservator to the study of extra-European boats and the preservation of their memory. As noted above, his attitude as a researcher went beyond a supposed scientific ‘objectivity’, his subjectivity often becoming almost ‘militant’, an ardent defender of vernacular naval architecture.

All figures are from photographs by the Musée National de la Marine, Paris.

For Pâris, boats among which the canoes of Oceania take pride of place, are significant material expressions of a society and its particular culture. Following from this anthropological perspective, the knowledge and defence of extra-European boats cannot be separated from knowledge and defence of their builders and users. This is one indicator, among others of the modernity of thought and deed of François-Edmond Pâris, sailor, researcher and maritime conservator. References Boudriot, J. 1976 “La frégate l’Artémise et son abattage à Tahiti en 1839”. Neptunia 121: 37-44. Pâris, F.-E. 1872 “Le musée de Marine”. Revue Maritime et Coloniale 3: 974-983. n.d. [1843] Essai sur la construction navale des peuples extra-européens ou Collection des navires et pirogues construites par les habitants de l’Asie, de la Malaisie, du Grand Océan et de l’Amérique dessinés et mesurés pendant les voyages autour du monde de l’Astrolabe, la Favorite, et l’Artémise, 2 vols. Editions Arthus Bertrand, Paris. 1992 Le Voyage de la Favorite. Collection de bateaux dessinés d’après nature, 1830-1832. Text and legends by E. Rieth. Editions Anthèse, Paris.

67

My own thinking about Polynesian spatial orientation has followed a similar trajectory. Since the early 1970s, I have engaged in fieldwork on Anuta, a Polynesian community in the eastern Solomon Islands (Fig. 1). Based on that experience, I have suggested that Polynesian social relations are often represented spatially as an elaborately involuted system of hierarchically ordered binary oppositions, organized in a predominantly linear pattern. For Anutans, east is superior to west; right is superior to left; front is superior to back; high superior to low; up to down; and seaward to landward. These oppositions are organized in such a way as to create an integrated system that articulates with social rank - the relationship between senior and junior; chiefs and commoners; clergy and congregation; men and women; and adults and children. I presented the Anutans’ commitment to a system of dual symbolic organization as pervasive, arguing that it has shaped their perceptions of and interactions with outsiders and enabled them to make sense of their changing world (Feinberg 1980, 1982a). I contended that each time the Anutans’ binary structures were contravened as a result of historical accident, the community reorganized its social and physical arrangements so that the binary system was quickly reimposed (Feinberg 1980). And I have suggested that the well known tendency for Polynesian atolls to be more egalitarian than volcanic islands might be understood in part by the difficulty of mapping a linear model of social space onto the circular geography of an atoll environment (Feinberg 1988a).

Chapter IV Polynesian Representations of Geographical and Cosmological Space: Anuta, Solomon Islands* by Richard Feinberg** For over a century, and in works encompassing most of the globe, anthropologists have examined symbolic representations of social, physical, and cosmological space. Much of this writing has been concerned with binary structures, as may be seen in theoretical publications dating at least to Durkheim and Mauss (1963), Hertz (1973), and Drukheim (1914), as well as important ethnographic reports by Codrington (1891) on Melanesia, La Flesche (1916) on North America, Hocart (1929) on Fiji, Granet (1933) on China, and many others. In 1949, Lévi-Strauss (1969) explored the implications of moiety systems for marital exchange and alliance. He later argued (Lévi-Strauss 1967) that dual organization is more complex than had been generally recognized but retained the conviction that binary structures are fundamental to human thought - a view reflected as well in works ranging from Ortiz’s (1969) analysis of Tewa Pueblo cosmology to Sahlins’s re-analysis of Fijian house and village organization. And, applying these insights to archaeology, Kirch (1996) has argued for the continuity of largely binary semiotic structures on Tikopia from prehistoric times onward1.

During my most recent visit to Anuta, in 2000, I pursued a project that involved mapping the island’s reefs, fishing grounds, and other marine features. I was mightily impressed at the islanders’ detailed knowledge of their marine environment, and the experience gave me a new perspective on Anutan understandings of geographical and social space. That project forced me to modify my view of Anutan cosmology. As part of the study, I compiled a map of approximately 300 marine features within a two to three mile radius of the island. In the process, I learned that the Anutans’ binary order is crosscut by an organization of the world as a series of concentric rings which are, themselves, divided into quadrants. Within those rings and quadrants are the hundreds of coral heads and other fishing areas upon which Anutans depend for their livelihood.

In recent decades, interest in dual organization has been largely superseded by concern with cognitive mapping and related systems of spatial orientation2. This work tends to be grounded in a recognition that cognition is complex, incorporating not only multiple structures but also models based on senses other than sight (Gell 1985; Widlock 1997)3. *

This chapter is a revised version of a paper submitted to American Anthropologist. ** Kent State University 1 Among the many other studies of symbolic dualism in Polynesia, see Eyde (1969) and Hooper (1981) on Tikopia; Shore (1976, 1982) on Samoa; Biersack (1982) on Tonga; Borofsky (1987) on Pukapuka; and Feinberg (1980, 1982a, 1988a) on Anuta and Nukumanu. 2 “Cognitive maps,” an idea first introduced by Tolman (1948), are “internal representations or stored memories of experienced environments” (Golledge 1999:xi; Kitchin and Freundschuh 2000). Like drawn or printed maps, they depict the physical space that one must negotiate, whether in the course of daily life or under extraordinary circumstances. 3 The “map” metaphor has been applied or assumed by students of Pacific navigation at least since Gladwin’s (1970) pathbreaking work on Polowat Atoll. Recently, however, the notion of cognitive mapping with its implication of static visual

representations has been called into question (e.g., Golledge and Stimson 1997, Kitchin and Blades 2002, Levinson 1996). Even in the 1970s, Lewis (1972, 1976) suggested that mental maps are dynamic and are continually updated as one’s location and direction change. Others (Cornell and Heth 1999, Heft 1996, Ingold 2000) have argued that spatial orientation is created through an interactive process of engagement with the environment. The relevance of multisensory perception is well illustrated by the way in which Pacific Island navigators orient themselves at sea by feeling wave patterns; indeed, Gladwin (1970) goes so far as to offer an image of the Pacific navigator “feeling” his way along a chosen course.

69

RICHARD FEINBERG

Figure 1: Map of central Pacific Ocean showing relationship of Anuta to other islands and island groups (adapted from Yen and Gordon 1973: 3). While this article echoes themes that can be found in other recent anthropological treatments of physical social space, several features also set it apart. Unlike most contributors to Bennardo (2002), as well as my own earlier treatments, this inquiry does not rely primarily on linguistic analysis. Rather, it emphasizes the Anutans’ detailed empirical knowledge of specific locations in their immediate - and sometimes not-so-immediate surroundings. Nonetheless, it also deals with the abstract conceptual model in terms of which such spots are located. In some respects, it resembles work of Basso (1996), Marshall (2004), and Stewart and Strathern (2004), as well as many contributions to Stewart and Strathern (2003), that treat geographical features as a source of cultural signification. However, I address marine features, most of which are normally hidden from direct human perception. Like work on Pacific wayfinding (e.g., Gladwin 1970, Lewis 1972, Finney 1977, 1994, 2003, Thomas 1987, Feinberg 1988b), it addresses people’s ability to locate places that are inaccessible to sensory observation until one reaches

This is not to say that my earlier description of the Anutan cosmos as associated with dual symbolic organization is in error, but that other structures intersect the binary model. In a sense, this is reminiscent of LéviStrauss’s depiction of Bororo socio-spatial symbolism as containing elements of both diametric and concentric dualism (Lévi-Strauss 1967) and Sahlins’s description of Fijian symbolic structures as containing triadic and quadratic as well as binary elements (Sahlins 1976). The Anutan system does not precisely replicate those of the Bororo or the Fijians. Rather, the point is that multiple models of social and cosmological space may exist simultaneously, and we risk distorting our ethnographic accounts if we assume an exclusive commitment to a single system. This article will outline the process by which a group of islanders and I created the map of Anuta’s marine geography, and it will use that experience as a way to explore the articulation of the binary/linear, concentric, and quadratic models which structure the Anutans’ environment, enabling them to find their way within it and providing them with a sense of their place in the cosmos.

70

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS

one’s destination, and it places the home island in relation pinpoint local fishing grounds and de-emphasizes wayfinding to distant locales.

to a larger universe. But it examines the ability to are immediately subdivided into two types: chiefly (nga maru) and commoner (pakaaropa) (Fig. 2). The kindred, for ritual purposes, is divided into four groups, but these are constituted on the basis of two intersecting binary divisions: mother’s vs. father’s side and central vs. peripheral (Fig. 3) (Feinberg 2004: 129-132).

Anuta’s dual symbolic organization Dualism, in many ways, pervades Anutan cultural and social life. The island has four kainanga ‘clans’, which

Figure 2: Founders of Anuta’s Commoner and Chiefly ‘Clans’. Leaders of the non-chiefly kainanga are referred to as the pokouru ‘heads’ of their respective clans; the leader of a domestic unit is often called the pokouru of his unit; and the senior male and female of each generation in a given domestic unit are termed the urumatua, or sometimes simply te uru, the latter being an abbreviated form of pokouru. The junior member of the most junior line is the lowest ranking person, and termed te murimuri pakaoti ‘the last [one] in back’.

Perhaps more notably, social relationships - especially those associated with rank - are often represented in spatial terms. Mua ‘before’, ‘in front’ is superior to muri ‘behind’, ‘in back’; runga ‘up’, ‘on top’ is superior to raro ‘down’, ‘below’, ‘on bottom’; and maurunga ‘high’ is superior to mauraro ‘low’. For this reason, the senior chief is known as Te Ariki i Mua ‘The Chief in Front’, and his junior colleague is Te Ariki i Muri ‘The Chief in Back’. The leading clan is Te Kainanga i Mua ‘The Clan in Front’, while the Kainanga i Pangatau, the third ranking clan, is sometimes called Te Kainanga i Muri ‘The Clan in Back’ because it is said, unlike the other clans, to be descended not from a man but from a pair of women. The Bishop of Melanesia is known as Te Ariki Maurunga o te Rotu ‘the high chief of the church’, and the High Commissioner for the Western Pacific (the chief executive officer of the British Solomon Islands Protectorate during the pre-independence period) was also Te Ariki Maurunga.

Much interaction among persons occupying divergent positions in the social hierarchy is informed by the front/back and high/low dichotomies. A chief’s head should be higher than those of his subjects, and no one may stand in a chief’s presence when the chief is seated. One should not turn one’s back on a chief or social superior; therefore, in walking on a narrow path, the individual of highest honor walks in front, while those of lower rank follow behind. If one is seated with a chief inside a dwelling house and should decide to leave, one must crawl backward, continuing to face the chief until outside. When a chief is seated outdoors with a group of other people, he is elevated physically and ritually by sitting on a mat, stone, coconut, or other object, while the rest of the assembly sits directly in the sand; and when a

Corresponding with the metaphorical use of height to convey ritual priority, the head is the most tapu ‘sacred’ part of the body, so that one’s head may not be touched by an inferior. The feet are given the lowest ritual esteem, demeaning anyone or anything toward whom they point.

71

RICHARD FEINBERG

the right hand is the one skilled enough to manipulate a fish hook. In accordance with this view, Anutans build canoes with the outrigger on the port side, so that fishing is generally done over the starboard gunwale. The starboard side is called te paai katea or paai pai ika, the latter meaning ‘fishing side’; port is te paai ama ‘outrigger side’. The left hand is known as te nima tema; tema, as far as I am aware, has no referent other than left. The left side of an object other than a canoe is te paai tema; the right side is te paai matau.

chief leaves the island or returns from overseas, he is carried to (or from) his vessel. He also may be carried as a token of special honor in various ceremonial situations.

Central Peripheral

Father’s Side

Mother’s Side

pai maatuaa pai makitanga*

pai tuatina pai tupuna

Figure 3: Diagram illustrating the binary ritual structure of the Anutan kindred. *Te pai makitanga is peripheral in the sense that she is not normally of the pai maatuaa’s unit. Te pai tupuna in the sense that the prestation is optional, but when it is included it always is presented to the same unit as the pai tuatina.

Male is superior to female in both the political and ritual domains. Formal decision-making power is in the hands of men. Men of the two chiefly clans are known as maru ‘protector’, ‘person of rank’. Women are not maru, regardless of their clan affiliation. Nor do women have a right to speak at the pono, the general assembly of the community. Other things being equal, women are expected to defer to men and obey their instructions. A woman may not touch a man’s head without his permission, sit on his sleeping mat, use his headrest, or sit with her legs pointing toward a man. In church, the congregation is segregated according to sex, with men sitting on the right side of the aisle and women sitting on the left.

When persons of approximately equal rank meet after a long absence, they typically press noses as a form of greeting. If the relationship is one of slight inequality, the inferior may press his nose to the superior’s neck; if the discrepancy is great - as in a commoner greeting a chief the former presses his nose to the latter’s foot or knee. Similarly, a common gesture of contrition involves crawling on hands and knees to the offended party and pressing one’s nose to the other’s knee as a token of submission. Should a social superior particularly wish to honor an inferior, on the other hand, he may express his wish by neutralizing or reversing many of the normal gestures of respect.

Ngaatai ‘seaward’ is superior to ngaauta ‘landward, inland’. Only men fish in the ocean beyond the fringing reef, while most garden work is done by women (Yen 1973). Both men and women fish on the reef flat. On Anuta, as in Fiji (Hocart 1929:126, Sahlins 1976), all of these relationships may be observed in the symbolic structure of the dwelling house (Fig. 4). The island is more or less circular, but houses are limited to an erratic line running from east to west along the southern shoreline. The buildings are rectangular, with doors at either end and on the inland side. The side facing the sea has no entranceway. It is called mataapare, the ‘face (i.e., front) of the house’ - the sacred side and place of honor. Only adult males sit there.

Anutans often picture the world as a huge flat plane, sloping gradually downward from east to west. The east side of the island, the first to greet the sun each morning, is called matapenua ‘the face (i.e., front) of the island’. The west side is muri penua, ‘the rear portion of the island’. East is higher than (i.e., superior to) west, and all lands to Anuta’s east are referred to as i runga ‘above’. Those to the west are said to be ‘below’ (i raro). To the east are the Polynesian islands, which constitute Anuta’s ancestral homelands, and Europeans, whom Anutans admire because of their technology, light complexions, and, to a degree, because they originally ‘descended’ (ne ipo) to Anuta from the east (‘above’). To the west are the Melanesian islands, whose people Anutans hold in less esteem. On Anuta itself, the easternmost noporanga ‘village’ is known as Mua ‘Front’; to the west is Muri ‘Rear’.

The inland side of the house is called tuaumu (tua or back; umu or oven). Most houses have a fireplace for small-scale cooking; and this is located at tuaumu, the profane side of the house. When women and children sit inside, they position themselves i tuaumu, while men occupy mataapare. A chief sits i mataapare toward the structure’s eastern (“upward”) end. When lying down inside a house, a man must point his head toward mataapare and his feet toward tuaumu. A woman either lies on the left, facing in the same direction, with her head a little further from mataapare than her husband’s, or she lies at right angles to him with her head pointing toward his left leg.

As is true in many cultures (e.g., see Needham 1973), right is deemed superior to left, so that in formal situations when the two chiefs sit together, the ‘Chief in Front’ is seated on the right while the ‘Chief in Back’ sits to his senior partner’s left. The right hand is known as te nima atamai or nima matau. Atamai means ‘mind’ or ‘meaning’; thus, the designation indicates that the right hand is the one, in a sense, that knows what it is doing. Matau means hook; therefore, nima matau suggests that

The overall integration of Anuta’s symbolic dualism comes into focus if we visualize the island’s two churches which, like the dwelling houses, are rectangular and

72

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS

SUPERIOR high front east male seaward right sacred

situated in an east-west direction. The catechist, the ‘head’ (pokouru) of the Anutan church, speaks from the pulpit in the ‘front’ (i mua, the eastern end) of the church, while the congregation sits facing him. The men sit on the right side of a central aisle; the women on the left. Sitting in the congregation, one faces not only the catechist, but matapenua, the ‘front of the island’. To the right is mataapare and the ocean, the domain of male activity; to the left is tuaumu, inland, the site of agriculture, and all that is symbolically female. Figure 5 provides a summary of these pervasive oppositions.

INFERIOR low back west female inland left profane

Figure 5: Summary of Anutan Binary organization.

Figure 4: Symbolic structure of Anutan Dwelling community, and over 200 miles from the provincial capital - has the resources to support a commercial fishing business without depleting the resources on which it depends for subsistence. When I posed the question to Anutans, I was immediately assured that their island has an extensive network of reefs. In fact, my friends insisted, their reef is far larger than those of much bigger islands such as Tikopia, Anuta’s nearest populated neighbour with a land area covering about 1.5 by 3 miles, and Patutaka, an uninhabited island 30 miles away which may

Quadratic and concentric structures Although I still understand Anutan spatial symbolism to have a strong binary component, I became aware of competing models during my most recent visit to the island in 2000. I had gone intending to explore Anutan development strategies, particularly the plans put forth by several leading islanders to create a commercial seafood industry. My initial question was whether the tiny island a half mile in diameter, 70 miles from the neighbouring

73

RICHARD FEINBERG

occupied with the construction of a new canoe, and he referred me to Pu Tongotere, a competent reef finder.

approach a mile in length. Indeed, Anutans presented their discussions in terms that made the supply of reefs and reef fish appear virtually inexhaustible.

In that first session, I began by drawing a map of the island (Fig. 6) in the center of a sheet of notebook paper. Pu Tongotere listed for me the names and locations of about three dozen distinct coral heads within about twomiles of the island, plus two more distant reefs - Te Aonga and Te Akau Motu. That was without citing anything on the extensive northern shelf, with which he reported himself to be less familiar. Both he and, later, the senior chief referred me to Pu Paiaki, whom they indicated is Anuta’s premier repository of information about the reefs and their indicators (pakamaironga).

I remained sceptical of those claims in large part because of the island’s character as a single volcanic outcropping with no barrier reef, lagoon, or outlying islets. If one stands on shore and looks to sea, the surf breaks on a fringing reef that extends no more than a few hundred yards at its widest point. Beyond the surf line is the deep blue sea which, from the perspective of a Western observer, appears to be an undifferentiated and virtually bottomless expanse of brine. From my experience fishing, swimming, diving, paddling, and sailing with Anutans, I was aware that the sea bottom extended several miles before falling off to depths of many hundreds - or even thousands - of feet, and that the intermediate area was broken up by numerous coral heads. However, I did not appreciate the number of marine features, the systematic way in which they were patterned, or the extraordinarily detailed knowledge of the ocean floor that some Anutan fishermen possessed. As it turned out, I devoted many hours during my 2000 visit to creating a map of Anuta’s reef system, based on the memories of the community’s most renowned sea experts. In collaboration with my friends Pu Paiaki, Pu Rangituteki, and Pu Nukuriaki, I mapped the size, shape, location, and relative depth of approximately 300 marine features within a radius of two to three miles of the island. I was told that this was just a fraction of the features that Anutans recognize and utilize, but that we had no paper large enough to accommodate them all and still be legible.

Later that day, my old friend Pu Nukuriaki went over the map that I had compiled with Pu Tongotere. He made a few corrections and added a number of reefs and place names, bringing us up to about 60 features. At that point, he insisted that we stop so that I could make a larger map, providing more room to draw in all the reefs that were still missing. I spent the rest of that afternoon and the next morning redrawing Pu Tongotere’s map on a larger sheet of paper and writing up additional notes. The following day I took the new, enlarged map to Pu Nukuriaki, who was waiting at his house with Pu Rangituteki and Pu Aatapu. They all seemed interested in the project and added a number of reefs and sand banks, bringing the total to around seventy. But they decided that too many features on my original chart were out of alignment. The inshore reefs as I had drawn them, they said, were too small in relation to the size of the island; and there were a great many reefs that had not yet been recorded at all. One reason for the misalignment, they suggested, was that the map (even in the larger second and third iterations) had become too crowded to put all the reefs in their proper places. They asked me to double the size of the paper (to four feet wide by six feet long!!), and only draw in the island, very small, in the center of the paper. They promised to get Pu Paiaki and suggested that on the following Sunday, when the church prohibits economically productive labor, we redraw the entire map once more - correctly this time. I spent the rest of the morning and early afternoon recording notes and preparing yet one more edition of the map for our forthcoming Sunday session.

This experience raises the question of how Anutans are able to keep track of such extensive, detailed knowledge. Part of the answer is simple and obvious: some Anutans have very good memories which they work hard to develop and maintain, and they do so because the payoff is critical to their comfort and survival. Their memories, however, are aided by the systematic way in which they compartmentalize their surroundings, superimposing an abstract mental schema on the physical environment. As we produced the map, step by step, over the course of several days, I became increasingly aware of their abstract models of the world and its spatial divisions, and how quadratic and concentric models of geographical space articulate with the binary and linear world-view that had been impressed upon me during my earlier visits.

On Sunday, Pu Paiaki, Pu Rangituteki, Pu Nukuriaki, and I spent most of the day mapping the reefs. Pu Rangituteki began by dividing the ocean into quadrants, then trisecting each of them. The result was to divide the ocean into twelve great regions, known as araavaka ‘canoe paths’. They are also termed nga paai moana ‘the sides (or sections) of the ocean’. Working counterclockwise and beginning in the southwest, they are:

The mapping process I initiated the project by mentioning my interest in mapping Anuta’s reefs to my friend, Pu Penuaika, an important leader and capable fisherman. He referred me to his father, Pu Nukumanaia, who is regarded as the island’s best navigator and overall seaman, and with whom I had worked closely in my earlier study of Anutan seafaring (Feinberg 1988b). Pu Nuku, however, was

Te Rongona South-southwest Mua Poranga (‘Front of Poranga’) South

74

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS

Tua Nga Rau Akau (‘Back of the Coral Heads’) South-southeast Te Araavaka (‘The Canoe Path’) East-southeast Te Uru Puko (‘The Top of the Puko Tree’) East Te Pura o Te Ana (‘The Pura of the Cave’) East- northeast Mua Uru Paa (‘Before the Top of the Pandanus Palm’) North-northeast

Te Akau Roa (‘The Long Reef’) North Te Uru Kauariki (‘The Top of the Kauariki Tree’) North-northwest Mua Vai (‘Front of Vai’) West-northwest Mua Pao (‘Front of Pao’) West Te Uru Kauariki (‘The Top of the Kauariki Tree’) West-southwest

Figure 6: Map of Anuta Island, showing main terrestrial features, the fringing reef, and two visible offshore sea marks.

75

RICHARD FEINBERG

Most of these names refer to the locations on the island that first come into view when one approaches Anuta from the specified region of the sea, or the last to disappear below the horizon when heading away from the island along the designated path. Uru Paa is a section of Anuta that is named for te paa, a type of pandanus palm which bears an edible fruit. Uru Kauariki is a section of the island named for the kauariki, a tree that bears a small, edible nut. Uru Puko is named for the puko, which includes two varieties of large tree with light wood, Hernandia peltata (te puko pakatangia) and Cordia subcordata (te puko vai). Vai ‘Water’ is the name of the spring, which flows from the hillside in the northwestern section of the island. Te Araavaka ‘The Canoe Path’ is named in recognition that this is the route to Patutaka, an uninhabited island thirty miles to the southeast, which Anutans visit at least once a year to hunt birds. Te Ana is a cave at the base of a monolith on the southwestern beach, directly across the island from Te Pura o Te Ana4. A pura is the point from which some landmark first rises into view. Te Rongona, Nga Rau Akau, and Te Akau Roa are the names of marine features. There is some redundancy of naming, so that araavaka is both a general term for the great regions of the sea and the specific name for one of those regions. And two separate araavaka are called Uru Kauariki. Such redundancy is found in many areas of Anutan language and culture, including the names for a number of marine features. Speakers infer the meaning from context and insist that the potential ambiguity is not a source of confusion. Figure 7: Diagram of the twelve araavaka ‘canoe paths’ radiating outward fromAnuta.

Once the twelve araavaka were labeled and identified, Pu Rangituteki drew concentric circles around the island to orient himself (Fig. 7). Inside the first circle were nga toka, and inside the second were nga rau akau. Surrounding nga rau akau is te punga o te akau ‘the top of the reef’ (sometimes shortened simply to te akau ‘the reef’), an essentially uninterrupted circle of coral and rock, approximately five miles in diameter. Beyond that lay te muri akau ‘the back of the reef’ and te roto ‘the interior’, a term referring to the open sea. Then, starting in the southwest and working counterclockwise around the island, Pu Paiaki and Pu Rangituteki filled in well over 100 toka, rau akau, ainaina, apenga, motu, pupua, patu, papa, and para (see below for a description of each of these types of marine feature). We worked for four hours and filled in the sea up to the second concentric circle (Fig. 8). The following day, we agreed, we would continue our project and map nga akau o te moana ‘the reefs of the ocean proper’.

We got about half way through nga rau akau o te punga o te akau ‘the coral heads at the top of the reef’ and decided to continue the following morning. At this point we had mapped about 250 offshore geographical features, all but two of which are invisible from shore. On Tuesday we finished the chart, adding approximately fifty more features (Fig. 9). Pu Rangituteki informed me that all of the rau akau on the map - 300 or so - are rau akau vanevane, named for a type of small fish (perhaps a variety of grunt or sea perch) that is especially abundant in those locations. There are many other reefs, he told me, where Anutans go for other kinds of fish; there simply was not room or time to map all of the rau akau! Anutan Terms for Marine Features In the process of mapping, Pu Paiaki and Pu Rangituteki distinguished approximately a dozen separate types of marine feature. These include the following:

The next day the trio sent for me at 8:30 AM, and we spent an hour and a half working on the map.

The akau penua ‘land reef’ is the first concentric circle beyond the surf line (te pati o ngaru). This falls to a depth of about thirty feet before dropping off to te para. The para penua is the second concentric circle beyond the surf line. It is a large expanse of sandy bottom which

4 There are actually two ana ‘caves’, Te Ana Tii ‘The Small Cave’ and Te Ana Rai ‘The Large Cave’, located within about two hundred yards of one another on the southwestern shore. The cave to which this name applies is probably Te Ana Rai.

76

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS

A tapatapa is the end of a reef shelf. It may form a border between two marine features or two distinct sections of ocean floor. On the map, the position of te tapatapa o te muriakau (the last concentric line in figure 9) is only approximate. Pu Paiaki and Pu Rangituteki indicated that they are confident of the precise locations of all the reefs and other features drawn within the tapatapa boundaries. They stated that the map compresses the size of the reef because even the largest available paper was too small to incorporate everything and accurately represent all distances. Te tapatapa o te muriakau, they said, should have been drawn about three inches farther out than it was on our four by six foot sheet of paper. The Akau Motu, if drawn, would have been about four inches off our map - again because of the size of the paper.

continues outward from the island until it reaches another ring of reef. The akau ‘reef’ or te punga o te akau ‘the top of the reef’ is the third concentric circle beyond te pati o ngaru (the penultimate concentric line in figure 9). The para penua drops off gradually until it reaches te akau, at which point the bottom rises much closer to the surface, forming a ring of relatively unbroken reef. Beyond the punga o te akau is the deep sea, te roto. This constitutes the fourth concentric circle beyond te pati o ngaru. The overall appearance of Anuta’s reef, as shown in figure 10, is almost that of a complex atoll, perhaps resembling Chuuk in Micronesia, with a ring of coral and a volcanic island in the center. The difference is that Anuta’s outer reef never comes closer to the surface than forty or fifty feet and has little discernable effect on wave patterns.

An apenga is a spot on the ocean that successors to an original explorer go to with great frequency. Taunga is a synonym for apenga.

A motu is a very small coral head. This is also the term for a small island or a ‘satellite’. An islet close to a larger island is that island’s motu, and a small coral head near a larger marine feature is its motu as well. A rau akau is a more expansive coral head than a motu. It is fairly flat and, in contrast with a papa (see below), a rau akau stands directly on a sandy bottom. Such a sandy stretch of submarine terrain is known as te kau aapua. A papa is a coral patch that is larger than a rau akau and, unlike a rau akau, it does not stand on the bottom. It is either perched on te punga o te akau or it stands on the top of a rau akau. A toka is an expansive patch of coral that extends close to the surface. The top of such a feature is close enough to the surface that waves break over it during storm conditions. There are several types of sandy bottom. A pupua is a small patch of sand surrounded by a circle of coral; an ainaina is a somewhat larger patch of sandy bottom surrounded by reef. Also, an ainaina tends to be oval in shape, in contrast with a pupua, which is more round5. A kau aapua is a patch of sand surrounded by reef, similar to an ainaina. A para, by contrast, is a very large expanse of sandy bottom, out beyond the akau penua. It also refers to a large region of ocean in which such features are found. The named para are deeper than the surrounding area. Red broken lines on the original map indicated either coral borders or sudden changes in depth.

Figure 8: Concentric rings around Anuta formed by reef and other bottom contours.

A pura is the point from which some landmark first rises into view. 5

I should note, however, that the one pupua I inspected visually while diving in the ocean looked like a not-very-big rock with, as far as I could tell, no sandy spot in the middle.

77

RICHARD FEINBERG

Figure 9: Complete Anutan reef map as redrawn using computer graphics program. Compiled by Pu Paiaki, Pu Rangituteki, Pu Nukuriaki, Richard Fienberg (pu Tokerau), Ute Dymon, Mattyhew Rollins and Brian George.

78

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS

Figure 10: Cross-section of Anuta and its reefs. Often more than one reef or marine feature has the same name; if called upon to specify which Puinga or which Patiaitu is in question, the speaker may indicate the general location or araavaka in which the feature is found.

Naming reefs and marine features Most marine features are named either for the first fisherman to locate and fish at the designated spot or for the fisherman’s canoe. Kapivaka, Kavakimoana, Puinga, Poraukiapi, Kaponotangata, Potuao, Vatauia, and Patiaitu, for example, are all named after canoes, while Pu Maatopa, Pu Maevatau, Pu Raropuko, Pu Pareaatai, and Pu Pita are named after fishermen. In the case of Patiaitu, both the reefs and the fisherman’s canoe were ultimately named for a powerful spirit of the tupua penua ‘land spirit’ variety (Feinberg 1995, 1996) who was believed to help cure sick people during the pre-Christian period.

Some features are named for their appearance, as is Te Papa Tea ‘The White Papa’. One reef is called Tangiao ‘Man in the Moon’, which Anutans think that it resembles in appearance. Many reef names incorporate and indicate some relation to another marine feature, as in the just-mentioned Motu o Pu Pareaatai. Maungarua ‘Two Hills’ is a reef found in a spot from which one may sight the two monoliths (nga pungaana) on the beach in the southwestern section of the island. Literally, maunga means hill or mountain, and it is the proper name of Anuta’s 212-foot hill. The two monoliths, when viewed from the ocean above this reef look to Anutans like a pair of hills. And Te Turanga ‘The Standing Place’ is a reef whose name indicates the relationship between it and Tupenua, the cliff face on Anuta’s northern coast.

On occasion, reference to the explorer may be oblique. For example, Nga Moumou refers to a group of rau akau discovered by Pu Raropuko Moumouipenua. Pu Raropuko was the fisherman’s ‘marital name’ (ingoa pakamaatua), while Moumouipenua was his ‘personal name’ (ingoa tangata or ingoa pouri) - a type of name that every Anutan has and which is bestowed at or shortly after birth6. Or a reef may be named for some distinguishing characteristic of the first fisherman. A couple of reefs, for example, are named Kaipaariki ‘Ringworm’ in recognition of a fisherman who suffered from an extensive ringworm infection.

Visual inspection On Wednesday morning, the day after my associates and I completed our map, Pu Paiaki and Pu Rangituteki took out a canoe to give me a visual introduction to some of the reefs we had charted. We left the island at 9:30 and first visited Te Toka Rai ‘The Large Toka’, a submerged boulder about a half mile out to sea from Te Ava Tii, the main point of egress across the fringing reef. My companions located the coral heads by sighting on distinctive landmarks known as kaavenga ‘carriers’ or pakamaironga ‘indicators’. The primary kaavenga or pakamaironga of the Toka Rai is Te Pungaana Rai, a 75-

Sometimes, designations for the type of marine feature are combined with names of people or canoes. Thus one relatively large coral head is called Pareaatai after the fisherman, Pu Pareaatai. Nearby is a much smaller coral head, known as Te Motu o Pu Pareaatai ‘Pu Pareaatai’s Satellite’.

6 See Feinberg 1982b, 1983 for a discussion of Anutan personal names.

79

RICHARD FEINBERG

depth here, but Pu Rangituteki estimated that it was comparable to Pupua.

foot-tall monolith on the southwestern beach, which stands directly inland from the toka. I made two dives down to the reef and measured the depth at 25 to 27 feet. The tide at the time was almost at its height; were it lower, the distance from the top of the boulder to the ocean’s surface would have been less. That made the toka deep enough that it poses no danger to canoes or ships, but shallow enough to cause waves to break in storm conditions - the primary criterion for being identified as a toka. The ainaina or sandy bottom from which the toka rises looked like it went down about another ten feet. I estimated the rock, which was more or less circular, to be approximately 50 feet in diameter; but I had no way to measure its precise horizontal dimensions.

The next stop was Te Muri Patu Tuei ‘The Old [Reef] Behind the Stone’, the stone in question being the Patu o Mangoo ‘Shark Rock’, a boulder that projects ten feet or more above the ocean surface. The Muri Patu Tuei is slightly to the west of the Patu o Mangoo and is, itself, quite large - perhaps 100 or more feet across. We measured the depth of the para around it at 90 feet. Pu Rangituteki advised against lowering the depth gauge to the rock itself for fear that it would get caught and be lost7. It was still raining, so again the main kaavenga were the Patu o Mangoo and the contours on the bottom.

From there we headed southeast another half mile or so to a true rau akau, Pupua. The ‘landmarks’ (nga kaavenga i ngaauta) for Pupua are the Pungaana Tii and a coconut palm on top of the Pungaana Rai. This reef is also about the same distance from shore as the Patu o Mangoo, which is used as another point of reference.

This was the last reef that we explored from the water. But we went from there to the Muri Patu Pou ‘The New [Reef] Behind the Stone’, to do some serious fishing8. This coral head was a little closer to the Patu o Mangoo. It was more recently discovered than the Muri Patu Tuei, and it is just inland from both the Muri Patu Tuei and the Patu o Mangoo. Fishing proved to be slow, so we soon abandoned that spot and began moving toward the island, hopping from one rau akau or motu to the next along the way. Before long we were back to Te Akau Penua, so the sandy patches were now ainaina rather than para. Around noon, we pulled up the our lines, and Pu Rangituteki switched bait; he replaced the thick pieces of octopus tentacle that we had been using with a long narrow strip from the tentacle’s end, removed the sinker, and we trolled from there to Te Pati o Ngaru.

At some point along the way, Pu Paiaki declared that we were now paddling over Te Para. I looked down with my face mask and could not tell the difference between the para and nga ainaina; so I asked if we had passed some tapatapa ‘boundary’. Pu Rangituteki replied that the main difference between the Para and an ainaina is that when one reaches Te Para the sea gets deeper; more than anything, it is that major drop-off which marks the change. The depth looked about the same to me, but he seemed confident that there was a substantial difference. Since it is hard to judge distances to the bottom when looking down from the surface, and Pu Rangituteki had a great deal of experience fishing in that area, his judgment seemed plausible.

Conclusions Anutans have several distinct systems for categorizing physical and social space. Those systems intersect and form a complex network of cosmological structures; the one or ones upon which an Anutan draws at any given time depend upon the immediate purpose.

I expected Pupua to be smaller than Te Toka Rai. Still, I was surprised at just how small it was; no more than a rock sticking up from the bottom and only, it seemed, a few yards across. I dove 33 feet, and the top of the rau akau appeared to be another ten to fifteen feet below me. Rather than attempt a second dive, I had Pu Rangituteki attach my depth gauge to his fishing line and lower it. The para on which Pupua stands was 80 feet below the surface. Obviously, either the rock rises much farther from the bottom than appears to be the case when looking at it from above, or I was a good deal farther from the top of the rau akau than I had thought. At this point, we were something like a mile from shore.

One system not discussed in this paper is the wayfinder’s wind compass, examined in my book on seafaring techniques (Fig. 11). A second is the linear/binary system in which east / west, up / down, right / left, chiefly / commoner, male / female, sea / land. This system is particularly important as it relates to social rank, ritual honor, and political authority. A third system involves the canoe paths (araavaka) that radiate out from the island and divide the world into twelve great regions, each one defined in relation to the specific points on the island and their alignment with particular fishing banks or other islands to which Anutans might sail. Such radial conceptualizations of space have been reported from

From there we continued eastward to Pakautu. Along the way, it began to rain very hard, so we could barely see the island and could make out no landmarks on shore. The major kaavenga, then, was the Patu o Mangoo and the contours of the ocean bottom. In fact, most of the process of locating reefs seems to be done by looking at the bottom after triangulation gets the fishermen to the appropriate vicinity. We made no attempt to measure the

7

Ninety feet is well beyond the free-diving range of anyone on the island. 8 This effort was more or less obligatory since Anutans rarely go to sea without attempting to catch fish.

80

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS

Figure 12 Schematic representation of Anuta’s concentric model of physical and cosmological space. Figure 11: Pu Nukumanaia’s wind compass (from Feinberg 1988b: 99).

overseas, Anutans say that he is following his vaevaerangi. Lastly, I would note that if one turns this concentric model of horizontal space on its side, it corresponds precisely with the ten levels of heaven (Rangi Tua Tai through Rangi Tuangapuru) cited in Anutan tangikakai mythological tales of gods and culture heroes from the pre-Christian period - giving Anutan cosmology a threedimensional symmetry. The roughly similar threedimensional complexity of cosmological space is well described by Keller and Kuautonga (n.d.: 269) in their analysis of oral traditions from West Futuna, a Polynesian outlier in Vanuatu. They say with respect to spatial movement in Futunese stories known as hkai: “Movements were valued according to their inward/outward trajectory along the radii of the cosmological sphere or the circular spaces encompassed by it rather than along a two dimensional vertical axis. Connecting inner and outer in the narrative discourses are transoceanic routes, trajectories to and from the sky, and island pathways including garden trails, and tracks to and from the reef. Following a centrifugal course toward the Underworld or the cloudy upper home of the spirits is fraught with danger while returning to the land is the West Futuna version of a happy ending. Movement toward the reef by contrast with centripetal movement conveys similar values. Descent is associated with uncertainty and trial while ascending from reef to land offers promise and stability. Movement to and from the distant horizon periphery carries more conflicted associations….”.

many other parts of Oceania (e.g., Handy and Pukui 1972, Bennardo 2002)9. A fourth system, which intersects that formed by the canoe paths, is the concentric one in which Anuta occupies the center. If one begins with the island itself and looks outward, the concentric model of horizontal space provides exactly ten distinct rings: (1) te penua ‘the island’; (2) te vaevaitai ‘the seashore’; (3) te aropitai ‘the reef flat’; (4) te pati o ngaru ‘the surf line’; (5) te akau penua ‘the land reef’, which is a ring of inshore reefs; (6) te para penua, a ring of deep, relatively undifferentiated sandy bottom; (7) te punga o te akau ‘the top of the reef’, a ring or coral and stone that bulges up beyond the para toward the surface; (8) te muri akau ‘the back of the reef’, that is, the drop-off into the deep sea; (9) te roto ‘the open sea’; and (10) te vaevaerangi ‘the horizon’ (Fig. 12). The vaevaerangi in one sense, of course, does not constitute a discrete ring around a particular spot since it is relative to the viewer’s position. From the perspective of the island, however, it includes not just the point where the sky appears to meet the surface of the ocean, but also everything beyond. When one leaves to seek his fortune 9

Cf. Handy and Pukui’s description of traditional Hawaiian land-holding patterns: “The complete ‘ili was a narrow strip of ahupua‘a, continuous from shore to mountain top” (1972[1958]:4).

81

RICHARD FEINBERG

Anutan representations of space and their cosmological significance differ in specifics from those of West Futuna. Yet on Anuta no less than Futuna, complex spatial understandings articulate with values, symbols, meanings, and empirical perceptions as expressed in myth and oral history; in ritual positioning; in social and political relations; in wayfinding techniques; and in the mental maps on which islanders depend in order to procure a living from the land and sea.

1982b

References

1983

Basso, K. 1996

1988a

1982a

Wisdom Sits in Places: Landscape and Language Among the Western Apache. NM: University of New Mexico Press, Albuquerque. Bennardo, G., (ed.) 2002 Representing Space in Oceania: Culture in Language and Mind. Pacific Linguistics 523. Research School of Pacific and Asian Studies, Australian National University, Canberra. Biersack, A. 1982 “Tongan Exchange Structures: Beyond Descent and Alliance”. Journal of the Polynesian Society 91: 181-212. Borofsky, R. 1987 Making History: Pukupukan and Anthropological Constructions of Knowledge. Cambridge University Press, Cambridge. Codrington, R. H. 1891 The Melanesians: Studies in their Anthropology and Folklore. Dover, New York. Cornell, E. H. and C. Donald Heth 1999 “Route Learning and Way Finding. In R. Kitchin and S. Freundschuh (eds.), Cognitive Maps: Past, Present, and Future”, pp. 66-83. Routledge, London and New York. Durkheim, E. 1966 The Elementary Forms of the Religious Life. New York: Free Press. (First published in 1915). Durkheim, E. and M. Mauss 1963 Primitive Classification. Translated and Introduction by R. Needham. University of Chicago Press, Chicago. (First published in 1903). Eyde, D. 1969 “On Tikopia Social Space”. Bijdragen Tot de Taal-, Land-, en Volkenkunde 125: 40-63. Feinberg, R. 1980 “History and Structure: A Case of Polynesian Dualism”. Journal of

1988b

1995

1996

2004 Finney, B. 1977 1994 2003

Anthropological Research 36 (3): 331378. “Structural Dimensions of Sociopolitical Change on Anuta”. Proceedings of Conference on Evolving Political Cultures in the Pacific Islands, pp. 124142. Gloria Cronin (ed.). Institute for Polynesian Studies, Lā‘ie. “Some Observations on a Polynesian Naming System: Personal Names and Naming on Anuta”. Journal of the Polynesian Society 91 (4): 581-588. “What’s in a Name? Personal Identity and Naming on Anuta”. Central Issues in Anthropology 5 (1): 27-42. “Socio-spatial Symbolism and the Logic of Rank on two Polynesian outliers”. Ethnology 27 (3): 291-310. Polynesian Seafaring and Navigation: Ocean Travel in Anutan Culture and Society. Kent State University Press, Kent, OH. “Christian Polynesians and Pagan Spirits: Anuta, Solomon Islands”. Journal of the Polynesian Society 104 (3): 267-301. “Spirit Encounters on Anuta, Solomon Islands. In J. M. Mageo and A. Howard (eds.), Spirits in Culture, History, and Mind”, pp. 99-120. Routledge, New York and London. Anuta: Polynesian Lifeways for the Twenty-First Century. Waveland Press Long Grove, IL. “Voyaging Canoes and the Settlement of Polynesia”. Science 196 (4296): 127785. Voyage of Rediscovery: A Cultural Odyssey through Polynesia. University of California Press, Berkeley. Sailing in the Wake of the Ancestors: Reviving Polynesian Voyaging. Bishop Museum Press, Honolulu.

Firth, R. 1969

“Tikopian Social Space: A Commentary”. Bijdragen Tot de Taal-, Land-, en Volkenkunde 125: 64-70.

1985

“How to Read a Map: Remarks on the Practical Logic of Navigation”. Man 20: 271-286.

Gell, A.

Gladwin, T. 1970

East Is a Big Bird: Navigation and Logic on Puluwat Atoll. Harvard University Press, Cambridge, MA. Golledge, R. G., (ed.) 1999 Wayfinding Behavior: Cognitive Mapping and other Spatial Processes.

82

POLYNESIAN REPRESENTATIONS OF GEOGRAPHICAL AND COSMOLOGICAL SPACE: ANUTA, SOLOMON ISLANDS

Kitchin, R. and S. Freundschuh 2000 Cognitive Mapping: Past, Present and Future. Routledge, London and New York. La Flesche, F. 1973 Right and Left in Osage Ceremonies. In R. Needham (ed.), Right and Left: Essays on Dual Symbolic Classification, pp. 32-42. University of Chicago Press, Chicago. (First published in 1916) Levinson, S. C. 1996 “Relativity in Spatial Conception and Description”. In S. C. Levinson (ed.), Rethinking Linguistic Relativity. Cambridge University Press, Cambridge. Lévi-Strauss, Cl. 1967 “Do Dual Organizations Exist”. In Cl. Lévi-Strauss (ed.), Structural Anthropology. Basic Books, New York. 1969 The Elementary Structures of Kinship. Revised edition. Translated from the French by J. H. Bell, J. R. von Sturmer, and R. Needham. Beacon Press, Boston. (Original: 1949). Lewis, D. 1972 We, the Navigators. University of Hawai‘i Press, Honolulu. 1976 “Observations on Route Finding and Spatial Orientation Among the Aboriginal Peoples of the Western Desert Region of Central Australia”. Oceania 46: 249-282. Marshall, M. 2004 Namoluk Beyond the Reef: The Transformation of a Micronesian Community. Westview Press, Boulder, CO. Needham, R., (ed). 1973 Right and Left: Essays on Dual Symbolic Classification. University of Chicago Press, Chicago. Ortiz, A. 1969 The Tewa World: Space, Time, Being and Becoming in a Pueblo Society. University of Chicago Press, Chicago. Sahlins, M. D. 1976 Culture and Practical Reason. University of Chicago Press, Chicago. Shore, B. 1976 “Incest Prohibitions and the Logic of Power in Samoa”. Incest Prohibitions in Micronesia and Polynesia. Journal of the Polynesian Society (Special Issue) 85: 275-296. 1982 Sala‘ilua: A Samoan Mystery. Columbia University Press, New York. Stewart, J. P. and A. Strathern

Johns Hopkins University Press, Baltimore and London. Golledge, R. G. and R. J. Stimson 1997 Spatial Behavior: A Geographic Perspective. Guilford Press, London. Granet, M. 1973 “Right and Left in China”. In R. Needham (ed.), Right and Left: Essays on Dual Symbolic Classification, pp. 4358. University of Chicago Press, Chicago. (First published in 1933) Handy, E. S. Craighill and Mary Kawena Pukui 1972 The Polynesian Family system of Ka-‘u, Hawai‘i. Charles E. Tuttle Company, Rutland, Vermont and Tokyo. (First published in 1958). Heft, H. 1996 “The Ecological Approach to Wayfinding: A Gibsonian Perspective”. In J. Portugali (ed.), The Construction of Cognitive Maps, pp. 105-132. Kluwer Academic, Dordrecht. Hertz, R. 1973 “The Pre-eminence of the Right Hand: A Study in Religious Polarity”. In R. Needham (ed.), Right and Left: Essays on Dual Symbolic Classification, pp. 331. University of Chicago Press, Chicago. (First published in 1909). Hocart, A. M. 1929 Lau Islands, Fiji. Bernice P. Bishop Museum Bulletin Number 62. Bishop Museum Press, Honolulu Hooper, A. 1981 Why Tikopia has Four Clans (with a comment by Raymond Firth). Royal Anthropological Institute of Great Britain and Ireland, Occasional Paper 38. Royal Anthropological Institute, London. Ingold, T. 2000 The Perception of the Environment: Essays in Livelihood, Dwelling and Skill. Routledge, London and New York. Keller, J. D. and T. Kuautonga n.d. Nokonofo Kitea: We Keep Living This Way. Unpublished manuscript. Kirch, Patrick V. 1996 “Tikopia social space revisited”. In J.M. Davidson, G. Irwin, B. F. Leach, A. Pawley, and D. Brown (eds.), Oceanic Culture History: Essays in Honour of Roger Green, pp. 257-274. New Zealand Journal of Archaeology Special Publication. Kitchin, R. and M. Blades 2002 The Cognition of Geographic Space. I. B. Taurus, London.

83

RICHARD FEINBERG

2003 2004

Thomas, S. 1987 Tolman, E. C. 1948 Widlock, T. 1997

Yen, D. E. 1973

(eds.) “Landscape, Memory, and History: Anthropological Perspectives”. Pluto Press. Empowering the Past, Confronting the Future: The Duna People of Papua New Guinea. Palgrave Macmillan. The Last Navigator. Henry Holt and Company, New York. “Cognitive Maps in Rats and Men”. Psychological Review 55: 189-208. “Orientation in the Wild: The Shared Cognition of the Hai‘om Bushpeople”. Journal of the Royal Anthropological Institute 3: 317-322. “Agriculture in Anutan Subsistence”. In D. E. Yen and J. Gordon (eds.), Anuta: A Polynesian Outlier in the Solomon Islands, pp. 113-149. Pacific Anthropological Records, Number 21. Bishop Museum Press, Honolulu.

84

The constraints

Chapter V

The Austronesians left no written history of their origins and so our conclusions must be drawn from a wide variety of other evidence. Much of the argument depends on their achievement in transporting people and agriculture over great distances and the natural constraints for a hull or rig to survive at sea. The essential considerations that appear repeatedly throughout this account are the climate and prevailing winds; water temperature; surf, coral reefs or mud; the availability of food and fresh water on coral islands and at sea; availability of timber, fibre and material for tools; design in relation to stresses; strength and fatigue fracture of materials; the compromises relating to the scale effect of the size of the canoes; decay and replacement of parts; and the different requirements for fishing, fighting, exploration, or colonization. These constraints effectively determined the designs of rigs and hulls, the directions and distances sailed, and which designs of hulls, rigs and sails would survive. Each deserves a more comprehensive treatment than is possible here.

Origins and Relationships of Pacific Canoes and Rigs by Adrian Horridge∗ Although the origins of the basic structures and rigs are lost in the prehistoric past, a survey of a wide variety of examples and their known history shows that Pacific outrigger canoes were originally as homogeneous as the Austronesian people, but later they were influenced by readily recognizable introduced technology. The earliest transport was probably a raft of large bamboo stems, with a rig consisting of a two-boom triangular plaited mat sail supported on a loose prop, as survived into modern times in several places. The canoe hull evolved from a dug-out tree trunk, to which side planks and stem and stern pieces were sewn. The interaction between the raft and the dugout produced the outrigger canoes and the double canoes that made possible the Austronesian conquest of the Pacific. Outrigger booms placed across the hull of the enlarged canoe were lashed down to pierced lugs that were left projecting in the bottom of the hull. When large trees were not available, this technique could bind together a plank boat made of drift wood. Limited by beliefs in magic and ancestors, besides engineering constraints, designs were extremely conservative, but the spread of cultivated seeds and roots was rapid and widespread and the social organizations became diversified.

A warning There are several accounts that describe the details, distribution and uses of the outrigger canoes and other indigenous boats of the Pacific region. Dodd (1972) gives a very sympathetic account of Polynesian canoes with excerpts from the original accounts of individual explorers and missionaries, an extensive bibliography, and numerous illustrations. My diagrams illustrate only the principles that I wish to explain. The most comprehensive reference books are by Haddon and Hornell (1975) and Neyret (1974). There you will find most of the factual data about the boats that was collected, without having to search the numerous primary sources. They were compilers, however, and their interpretations of origins and relationships are now highly suspect because they made conclusions from the data that was available at the time, mainly from the boats themselves. Others, notably Needham, Bowen, Doran, also have faulty interpretations of the evolution of the structures and rigs. Still others, notably Best, Buck, Holmes, Horridge, Malinowski, refer to local regions. There are other essential references to special periods or topics, notably Bellwood on the history, Lewis on navigation, Buck and Emory on cultures, Dodd (based on Cook’s artists) and Pâris for old illustrations. There are other large and scattered topics where more research is needed, mainly on experimental sailing, ancient weather, local vocabularies for boat parts, and the genetics of domesticated plants, animals and men1.

In the west Pacific, the influence of the Chinese was negligible, but from the Indian Ocean the stern rudder, the mast fixed in a thwart and the tilted rectangular cloth sail spread eastwards before 800 AD. From about 200 BC, the Austronesian triangular sail spread westwards across the Indian Ocean and became the lateen, which continued to the Mediterranean and eventually to Portugal by the 14th century. Spanish sailors shipwrecked in the Tuamotus in 1526 AD account for planked hulls with solid stiff ribs, a fixed mast with stays, a fore-and-aft tacking rig, a rudder mounted on a pin, and other European details in eastern Polynesia. Recently, European influences replaced or heavily overlaid the Austronesian designs all over the Pacific.

∗

Australian National University, Box 475 P.O., Canberra, Australia 2601. Email: [email protected]

1

In this brief and crowded account, I can do little more than draw attention to the main trends.

85

ADRIAN HORRIDGE

Austronesians were sufficiently skilled to make pottery, to weave, to transport useful plants, had sufficient astronomy to navigate, and probably were themselves immune to diseases that were lethal to the indigenous islanders. Once under way, they discovered the bonus of discovering pristine islands with sea bird colonies, turtles and dugong that provided easy new food supplies, most of which were soon eaten so they had to move on.

The beginning Java was inhabited by some of the earliest hominids, who were probably not sailors; they came from Africa but we know nothing about their transport. One puzzle is that for 500,000 years early man occupied Java, and presumably much of coastal south east Asia, now flooded, but apparently traveled no further for a long time. The first boats were probably of reed bundles, as used for millennia on the Indian Ocean and surviving into the 19th century in Tasmania, New Zealand and the Chatham Is. There were easy crossings all the way to Australia, especially via New Guinea at times of minimum sea level. The first colonists reached Australia more than 50,000 years ago, and later made bark canoes.

How long rafts had been in use in the region of the Sunda Shelf no-one can now say, but the short and recent time scale of the rapid Austronesian expansion was insufficient for the development of a complete new tropical agriculture and marine technology. Almost certainly, during this period the triangular sail was a mat that was heavy when wet, so it was supported by 2booms, pushed up with a loose prop because they had no pulley, and the prop was held up with a stay on the windward side. Probably this was not an Austronesian invention, but it was an early solution to the problem of raising a heavy, wet and therefore weak mat sail on a raft. On Mangareva in E. Polynesia, Beechey found a raft with this rig carrying large numbers of people in 1826 (Haddon and Hornell, 1936 vol. 1: 93, fig. 64). The balsa wood rafts of the Peruvian coast were similar and probably of Asiatic or Austronesian origin.

Stone tools suitable to make dug-out canoes have not been found older than about 20,000 years. Therefore the earlier occupation of Australia and Island South East Asia as far as the Solomon Islands was almost certainly made with rafts. Bamboo was abundant in S.E. Asia, can be cut without an adze and is readily bound together with rotan to make large rafts that are seaworthy for months but not years. At least two attempts to sail rafts with the northern current from Asia to America have been almost successful in recent times2. Living along the tropical coasts is a healthy life style free from malaria, with abundant food, and anyone can see today that the coastal villages are not short of children. A long period with raft transport allows enough millennia for the first inhabitants to develop tropical gardens, the uses of local plants, pig husbandry, fishing techniques, and other crafts in use today in New Guinea, and to grow in numbers. Then, the great shrinking of the land area of the Sunda Shelf caused progressive flooding of coastal lands3 that must have forced the pre-Austronesian people to move repeatedly. Until they developed tools to build dug-outs, they must have traveled by raft. We might ask what made the Austronesians migrate through Island South East Asia, when we know that most of today’s Austronesian cultures there are sedentary. First, they were driven from behind by population growth on the mainland of Asia that accompanied the earliest Chinese (Han) agricultural revolution about 8,000 years ago. Secondly, like the Celts who moved into Europe, the Romans into Gaul, or the Vikings to Scotland, they had a technology superior to that of the indigenous inhabitants. They brought agriculture based on millet, rice, taro and other robust foods that could be dried and stored for long periods. As inferred from root words held in common in many branches of their languages, the earliest

Figure 1. A primitive sailing raft before the stone adze was available to make a dugout canoe. The rig is the twoboom triangular sail supported on a loose prop and held up by a rope on the windward side. The hull could have been of logs or bamboo lashed together. The leaboards adjust the course, which is steady when the centre of effort of the sail is ahead of, and in line with, the centre of drag. Plenty of such rafts have sailed in historic times.

2

Bisschop (1959) and Tim Severin (1994), the latter with a Vietnamese ghe bè; Piétri (1949:89). 3 Bellwood (1978:422). Three million square kilometres of land was lost between 14,000 and 7,000 years ago.

A survey of the sailing rafts in use in historic times reveals a standard construction but a variety of rigs. The

86

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

hull was slightly curved up at the ends; the poles were lashed with rotan to cross pieces; they were put on a course by vertical dagger boards that were moved to bring the centre of drag behind and in line with the centre of pull of the sail (Fig. 1). The tak pai of Taiwan4 had a square sail (Fig. 23E); the balsa rafts of the Peruvian coasts (Johnstone,1980:224-228) used a two-boom triangular sail: the bamboo rafts (ghe bè) of Haiphong Bay, Vietnam (Piétri 1949:89), used a cotton lug sail (Fig. 23X) or the low rounded junk sail of the southern Chinese. Rafts in Fiji and Santa Cruz (Neyret 1974:48,81; Haddon and Hornell 1936: vol.1:30) and Mangareva (Gambier Is.) (Haddon and Hornell 1936: vol.1:9) had the local mastless two-boom triangular mat sail (Fig. 1).

pottery, in the Santa Cruz Is. and Fiji, is dated 1300-1500 BC. These pots were decorated in a simple geometric style, and have been found in a long arc eastwards from the Solomon Is. Examples are still made in the Kai Is (Lat. 6º S; Long. 133º E), but the Polynesians lost the art of making them.

The Austronesians Non-Austronesian languages from earlier periods persist in Timor, Papua-New Guinea, parts of Halmahera, and of course Australia5. The Austronesians are defined as the speakers of a particular family of languages, with their own set of crafts, myths, taro, beaten tapa bark for cloth, breadfruit, sugar cane, banana, and other agricultural plants6, house designs, weaving, pottery7, an artistic style of carving and of tattoo8, who spread into the region of Taiwan (Bellwood et al. 1995: chapt. 5) about 4,000 years ago, then southwards to the Philippines and Indonesia. The most primitive of the surviving Austronesian languages are thought to be those of central Taiwan. They replaced older Australoid peoples related to Papuans and Australian aborigines. Their present distribution is almost identical with that of the outrigger sailing canoe, but not all Austronesian canoes have outriggers.

Figure 2. The adze with a bent haft, a suitable tool for making a dug-out canoe and carving flat planks. The binding tightens as it is used. The Austronesians diversified into the individually isolated and distinct highland cultures of central Taiwan, Luzon, Borneo, Java, Sumatra, Nias, Sulawesi, Flores, and Sumba. The Chams of the Vietnamese coast were Austronesians. Wood from Sumatra drifts naturally to the coast of Madagascar, which was first reached by Indonesians about 200 BC, probably associated with the direct transport of cinnamon to the coast of Africa (Miller 1969: chapt. 8). Later groups followed over a period of 1,000 years. The Austronesians who moved eastward had a stone tool industry, their adze had a characteristic bent haft (Fig. 2), a style of building houses and rat-proof granaries that originated in Asia, together with weaving, food grinding, unglazed pottery not made on a wheel, agriculture with millet and later rice, bark cloth of tapa, large stone megaliths, genealogy and ancestor worship, customs concerning wife giving, pigs, chickens, dogs and taro. Later, the inland Indonesian peoples were not much interested in boats or the sea, although they retained myths about their own origins from over the sea, and frequently they echoed boat structures in their coffins or house styles. Examples are the boat coffins in Sulawesi and Sumba, the council platform like a stone boat at Sangli Dol on the east coast of Tanimbar (Lat. 7º 51' S, Long. 131º 21' E), and the boat terminology for the parts of the house in Sumba. Maritime villages dependent on fishing or trade developed less on coasts that were liable to be inundated by tsunamis, a precaution that has been ignored in modern times.

They carried with them a common theme in Austronesian art that can be traced from the earliest bronze castings (Heine Geldern 1932) of the Zhang dynasty of China, 3,200 years ago, which themselves were based on earlier wood carvings. The double spiral and other motifs can still be seen in the decorations on houses and canoes, on body tattoo and wood carving of the Maoris, and as far east as the Marquesa Is9. Their pottery was not made on a wheel but by beating the clay while holding a smooth round stone within the pot. When sand is mixed with the clay, such pots can be used for cooking. The earliest 4

Nishimura (1925). There is a copy in the library, Nat, Marit. Mus. Greenwich, London. 5 For numerous cultures of S.E Asia and the Archipelago before the Austronesians, see Bellwood (1978: chapt. 3). 6 See Barrau (1963); Horridge (1985: 2nd Ed.:100, Appendix 1). 7 Spriggs, M. in Bellwood et al., (1995). For dates of Lapita pottery, see Green in Jennings (1979:33). 8 See http://www.hawaii.gov/hidocs/tattooing.htmlef for references to tattoo. 9 For Marquesan tattoo patterns, see http://wiki.bmezine.com/index.php/Marquesas Islands.

One or more groups of these Austronesians went islandhopping eastwards against the prevailing winds and currents. Avoiding malaria and lands already occupied, they passed to the north of New Guinea, and colonized

87

ADRIAN HORRIDGE

island after island, working eastwards, making many island bird populations extinct as they went. Eventually they became the Polynesians and also contributed to the stock of Melanesia and Micronesia. Most languages of the Solomon Is. are Austronesian, but dark colour and fuzzy hair are not Polynesian traits. With an interesting exception close to Taiwan and others in the Solomon Is., the Austronesian boats were dug-out canoes with one outrigger on the windward side, or double canoes. The details of the designs were adapted to sailing among shallow reefs in warm seas without harbours. For their over-water migrations they relied on larger outrigger canoes that carried families, fire and agricultural supplies. All branches of these migrations used boats based on the five part hull, the two-boom triangular sail with the mastless tilting rig with no pulley and no fixed rudder (see below), or later modifications of these. Eventually they reached the islands of the Eastern Pacific, and coastal South America10. They took with them the Asiatic chicken, and there is considerable discussion as to whether they took a few useful plants to America and brought others back into Polynesia11.

Haddon and Hornell 1975:191). They made rope and string by plaiting sennet. They carried with them the chicken, dog, pig and a long list of plants from Asia that reached Tahiti before 1770, including taro, yam varieties, pandanus varieties, hibiscus, sugar cane varieties, breadfruit, banana varieties, gourds, the Abrus bead, fish poisons Barringtonia and Wikstroemia, the tapa tree (Broussonetia), ratans, coconut, a Piper species for making kava, the Tahitian chestnut (Inocarpus), medicinal plants such as Urenia and the dapdap tree (Erythrina variegata), the oil nut (Hernandia) and the putty nut tree (Parinarium). Bamboo roots, sago roots, betel nut with associated Piper betle, and rami fibre (Boehmeria) were transported from Asia as far as Melanesia. All these must be kept out of salt water. Following each step in the eastward colonizations, there must have been many return trips to fetch these plants, and the enormous feat of transportation reveals the quality of the boats.

At first, travelling eastwards through the region of the annually reversing monsoon winds near Asia was easy. The first major hurdle was the stretch of 900 km (550 miles) of open water between Vanuatu (New Hebrides) and Fiji against the prevailing wind. This is exactly the right situation to encourage the bold to experiment with new types of rig and boat, because explorers could see flotsam coming from land upwind and could expect to return home again downwind. Accounts of raft journeys on the open ocean show that they could no longer use rafts. Most of the cultivated plants had to be carried as shoots or roots because they were propagated vegetatively, but they can be damaged by sea water. To carry families and plants, they probably had to wait for the development of the large double canoe, 30 m long, which could carry 100 people or several tons of cargo12. Weaving and pottery were eventually lost because the raw materials were lacking. Plaited pandanus leaves, split rushes or palm leaves, coconut fibre, and beaten tapa bark, were substitutes for woven fabric. They also developed techniques for making cooked, fermented and dried breadfruit, a sugary nutritious food that keeps for long periods at sea, ways of carrying fire, seeds and living shoots out of contact with sea-water, and numerous techniques for fishing and sewing (Best 1925: 122 and onwards; Buck 1957;

Figure 3. The five part canoe. (a) Double outrigger type. (b) The stem and stern pieces and the two planks. (c) The hull, showing spreaders fitted under lugs. The 5-part canoe The designs of the canoe hulls were fundamentally conservative although very diverse in detail. The Austronesian cultures revered the ancestors, or the founders of the village, and they believed people returned in a cycle after death in different flesh and bones. In the same way, the canoes never perished; they were the same canoes as the ancestral ones but made from different trees. There were other factors at work in the preservation

10

See note 17, and earlier work by Doran and by Jett, S.C. in Riley et al. (1971). 11 Barrau (1963); Bellwood (1978); Langdon (1989). The problems are the sweet potato, maize, and cotton going westwards; the banana, coconut, pottery and adze styles going eastwards, and bottle gourds from Africa. See note 17. 12 Fijian ndrua; See Haddon and Hornell 1975:319.

88

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

of designs, the fear of making a fatal change, the fear of being accused of witchcraft to take advantage of neighbours, the secrecy of the boatbuilding guild, and their long detailed training in which every detail of the technology was memorized.

6 equal lengths of x centimetres, which is usually 100 cm in South Bali. The front surface of the mast is placed x cm back from the inside edge of the bows. The length of the mast is 2x+10 cm and the stern outrigger boom is x/2 from the inside end of the stern. This proportional system enables the boatbuilder to build to any desired size, depending on the basic unit. As well as the builder’s own body measurements, an appropriate bamboo stick with equally spaced nodes acted as a measuring stick. In Bali, following the small asymmetry in the human body, the forward outrigger boom is always 4 or 5 cm longer on the left side. The total length of the front outrigger boom is equal to the distance between the two booms, and so on for all measurements. There was no concept of hydrodynamic efficiency, but each village has its own style, with slight differences, so there were improvements by trial and error.

Design by proportions The base of the dug-out canoe hull was hollowed from a single log, that was built up with planks in various designs in different parts of the Pacific. Where large trees were not available, irregular shaped planks were carefully carved from driftwood, fitted and sewn together. The main hull had the sides raised by an additional plank on each side, and forked stem and stern pieces were added at the ends. This basic five-part canoe (Fig. 3) was almost identical in design from Madagascar, through Indonesia and the Pacific to Tahiti and Hawai'i. The preferred timber was Calophyllum inophyllum13, a widespread tree that occurs near the shore wherever Austronesian people colonized. Tourists sit in its shade along the beaches of Bali and Fiji. Other timbers were magnolia, jackfruit, breadfruit or several hardwood trees that were abundant in high islands. Outrigger floats were made from the light wood of a common garden tree, Erythrina, of which a very widespread species was found in Tahiti by Cook’s botanists. Every step in the construction was bound by traditions that were learned by heart and supported by myths. The dug-out hull or the keel is the female part of the canoe; the forked stem piece is the male part. The two were fitted together in a marriage ceremony that persists today in some Indonesian cultures, especially Balinese (Horridge, 1979), Buginese (Horridge, 1987) and Makassarese, called the kawinan ceremony. A tiny piece of gold, or a paper bearing the name of the canoe or ship, is placed in the mortice in the female part before the tenon on the stem part is inserted. Unfortunately, similar ceremonies that must have been customary in Polynesia have been lost or not yet revealed.

Figure 4. Attachment of the superstructure to the hull. (a) in a 5-part canoe, (b) in a sewn planked canoe, (c) ditto, with dowels between the planks. Dowels were not used until a suitable iron drill became available.

In Austronesian cultures the traditional units of measurement were taken from the dimensions of the human body, and related ideas of perfect ratios originated in ancient India. The traditional boats, like the houses, are still built following a system of numbers that give the proportions of every component (Horridge 1987: fig. 8b). The trolling juking of the Straits of Lombok measure 3 fathoms (depa) and three hands between the inside of the ends. The depa is the span of the arms of the canoe builder, officially 1.7 metres. The thickness of the hull sides and bottom are carefully measured by using a stick as a depth gauge. The internal hull length is divided into

The traditional method of holding the boat together was to put the hull into compression like an inverted arch by tightening the lashings between a thwart and the flexible ribs lashed to projecting lugs (Horridge, 1982) on the inside of the hull. Small cross bars (called spreaders) were fitted inside the hull in many cultures (Fig. 4). The outrigger booms must be lashed down very firmly. The lashings were of non-extensible rotan or other vines that tightened up when wet. We can now see that the hull of the 5-part canoe lends itself naturally to the design of the outrigger attachments, because the cross piece that clamps down the side planks was extended as the

13 Calophyllum: a genus of tropical trees useful for timber, oil and medicine. See Burkill (1966).

89

ADRIAN HORRIDGE

outrigger boom (Figs. 3, 4). The whole structure was readily dismantled for repairs.

second smaller hull was given the same name, kata or katir, as the outrigger float. The double outrigger canoe (Figs. 7, 8), with both floats in the water at all times, is specialized for inshore fishing, limited in size and not suitable for travelling in the open ocean. Even with superior materials, a modern family-size trimaran is a dangerous boat in a storm. Sewn boats appear to have been the rule until iron tools made it possible to insert strong dowels between the planks, and even then sewing with coir persists in the Indian Ocean. All components were lashed or stitched together, and sewn joints between planks were sealed with resin or the pounded nuts of the putty-nut tree (Parinarium glaberrimum). This forms a strong and rigid joint, so the planks do not move relative to each other when it is set. If the planks were fixed to rigid ribs, the seams would open up when the boat was taken out of the water and dried, but flexible lashings allowed for this shrinking.

Inshore fishing canoes had the outrigger booms lashed over an open hull. Some offshore travelling canoes had the hull sealed with dammar-type resin and with pounded putty nut, which was extremely resistant (Pâris 1843: plates 114, 118). Their superstructures were raised above the waves on a waterproof rectangular box with a hatch so that goods could be kept dry inside. A platform between the outrigger boom and the hull was extended on the leeward side to accommodate passengers, often in a small shelter (Figs 5, 6). Big double canoes also had a central raised platform. Exact designs varied (Haddon and Hornell 1936-38; Neyret 1974) but the same principles were observed over huge areas and time. In the Pacific generally, all sea-going canoes had a single outrigger on the windward side or a second hull. The

Figure. 5. A single-outrigger travelling canoe of Satawal, in the Caroline Is., Micronesia. This boat was encountered by the French corvette, the Astrolabe, in 1826 and the print was published by Pâris in 1843 (plate 107). The outrigger is on the far side, beneath a little hut with a square door. The sail is held up by a stay to the wide beam on the lea side, which also supports a plaited shelter.

90

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

Figure 6. A thamakau of Lakeba, eastern Fiji Is. In 1826 (Lat. 18ºS; Long. 179ºW). The people were typical Melanesians. The hull is about 1 m deep, and can be kept dry. The man on the left holds a long thin steering paddle. The method of supporting the sail, and the lacing, can be clearly seen. To go about, the tack of the sail is carried to the other end of the boat. Similar craft are still in use (Pâris 1843: plate 117). Ropes and lines were plaited sennit, of fibre from palm trees, the baru tree (Hibiscus), or beach morning glory (Ipomoea pes-caprae). A seaside shrub (Wikstroemia foetida) was used for making fishing nets and fish poison. Sails were of small squares of plaited rushes or split palm leaves, sewn together and edged with a sennet boltrope.

They were laced to flexible booms along two edges. Clearly, the sail was a weak component that determined much of the design and dimensions of the whole boat. When not in use, the sail with its booms was taken off and stored in a dry boathouse. There the whole canoe or lashed-lug boat could be disassembled for the

91

ADRIAN HORRIDGE

replacement of the lashings, often accompanied by special ceremonies.

The design of the rig was highly constrained by the strength and elasticity of materials, which were particularly sensitive to wear and fatigue fracture. The design probably originated in Island South-East Asia before the Austronesians arrived, as indicated by its absence from the Asian mainland. The sail was pushed up with a pole, which pivoted freely at its base, removing all bending stresses. It was supported by a rope to the windward outrigger (Figures 7, 8). It was not pulled up on a halyard, which requires a much stronger rope and a pulley, and doubles the load on the mast. The pulley was unknown in Polynesia until the arrival of the European explorers. The doubtful example in Anson (1748) on an illustration of a flying proa that was reproduced by Haddon and Hornell (196: vol. 1:414) was an error14.

Figure 7. Sketch of a Madurese jukung with the classical mastless rig. Note the position of the two-boom triangular sail when heading into the wind, and the forward extent of the outrigger floats to prevent nose-diving. The sail was also measured by secret ratios. In Bali, the leech (the free edge) is two thirds of the length along the upper boom and is one hand span more than the length along the lower boom. The exact curves along the edges, to give the sail a shallow belly, are decided by the owner after much discussion. There are many arguments about the correct details and every canoe is identical in each fishing village, but from time to time adjustments are made, and they may adopt a new design from outside, although claiming that it is exactly like that of their ancestors. The belief that the existing canoes were the same canoes as those of the ancestors, but made out of different trees, meant that even if the designs had changed over time, it would not be admitted.

Figure 9. Tacking a double outrigger going upwind. The sail lies in front of the mast and therefore there can be no forestay. To go about, the boat is turned downwind and the sail, together with the sheet, allowed to come right around to the other side over the bows. The sail was weak and heavy when wet. Therefore loads along the edges of the sail were distributed by lacing or sewing to wooden or bamboo booms that prevented destructive flapping. The stay holding up the sail ran to the outer end of the outrigger boom, so that loading on this stay was minimal. On a single-outrigger canoe a strong gust of wind on the sail easily lifted the outrigger, and the flexible ends of the sail booms assisted in spilling the wind. The sail pivoted on its tack (the lower forward corner) and was pulled back and down to go into the wind, like a modern windsurfer (Figs. 1, 7, and on the left in Fig. 9), but was tilted forwards and across the boat to go downwind (Fig. 8, and on the right in Fig. 9). The sail

Figure 8. A Madurese jukung sailing downwind. Note the position of the two-boom triangular sail. Design of the mastless tilting rig

14 Absence of a pulley is negative evidence; no pulley has been found in valid artifacts collected by Cook.

92

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

cannot swing over the hull to the other tack because the prop is in the way. To change on to the other tack, the bows were pointed downwind and the sail was allowed to swing right round the front to the other side (Fig. 9), accompanied by the sheet (called “wearing ship”).

The canoes were the most important possessions of the Austronesian maritime cultures, requiring great expense in time to make and maintain, and demanding remarkable expertise to sail and navigate safely. It was important that the numerous spirits of the timber trees, the fish and the sea were placated by ceremonies and offerings so that the canoes and their crews would return safely. Therefore, as for objects of significance in all cultures, activity around them was accompanied by ceremonies that were religious in nature and magical in practice. There is a variety of evidence that everywhere the details of construction were embodied in myths and ceremonies15. Malinowski (1922: chapt. 5) describes the ceremonies at each stage during the building of a canoe. The lashings were renewed with great ceremony before every voyage. Horridge (1979) describes the surviving ceremonies in the building of an Indonesian prahu and those for Balinese canoes (Horridge 1987:66-72); Ellis (1831) those of Tahiti, and others are reported in Haddon and Hornell (1975). Sadly, most of this pre-contact material culture of the Pacific Islands has been lost.

Figure 10. To tack a single outrigger canoe with a twoboom triangular sail, first the sail was closed upwards, then the tack of the sail was carried to the other end of the boat and the sail was opened again. This manoeuvre takes the boat along the dotted track.

Outrigger canoes in historical times The double outrigger canoe of Madagascar, Malaysia, Philippines and Indonesia (Figs. 7, 8, 9) is a specialized fishing canoe for relatively calm seas. The design limits the size because the leeward outrigger digs into the water, but large versions are used for carrying cows (Horridge 1987:21) or schoolchildren (Horridge 1985: plate 7) in Indonesia. A common form of fishing is to leave a float anchored on a long length of rotan into which palm leaves have been fixed, and then each day put a net around the fish sheltering among the fronds. The outrigger booms carry the large coils of plaited rotan on the canoe. The double outrigger provides stability when the owner pulls fish into the boat single-handed, or for trolling for tuna with a line on each side. Examples are the colourful 1 or 2 man fishing jukung of Bali and Lombock (Horridge 1987), the fast pangkur of the Mandar people of West Sulawesi (Horridge 1985: plate 9), the line fishing canoes of Java (Horridge 1987:119 and onwards) and the fast fish transports of the Madurese (Horridge 1987:81-115.). In contrast, in these areas the single outrigger canoe is very small and specialized for inshore fishing by one man with the throw-net, and recovering the loaded net over the uncluttered side.

Figure 11. Map of the distributions of the mastless rig in historical times, within the dotted lines. In many places the tack of the sail was lifted bodily and carried to the other end of the hull (Fig. 10), as was done on the single outriggers of the Marianas, the Carolines, Fiji, Tonga, and the double canoes of Fiji and Tonga, all of which were reversible end for end. To close the sail, the lower boom was raised by a long halyard, called tutup in Indonesia, that passed over the upper boom and back to the lower boom (Figs. 5, 14). Sails were closed upwards and lifted off the boat complete with their two booms. These are general features common to most of the area of the mastless rig (Fig. 11). Although large steering paddles were used, they were not mounted on pivots, as in Arab, Chinese, Indian or European cultures. Ceremonies

Performance of the oceanic single outriggers depended on the exact rig, and was superb with the wind on the quarter. Many eye-witness accounts describe how they could run rings around a European ship under full sail (1843), just as a modern wind-surfer can frequently outpace a racing yacht. When heading straight down wind, the two-boom triangular sail is unstable, as it then 15 For example, Best (1925), Buck (1957), Malinowski (1922), Ellis (1834).

93

ADRIAN HORRIDGE

lacks lateral stability and is sensitive to squalls. When sailing into the wind with the sail pulled well back, however, the performance was hindered by the lack of a keel. The flying proa (Anson 1748; Haddon and Hornell 1936: vol. 1:412-421) of the Marianas did better with a flat lea side on the hull, but none of the outrigger canoes with one sail and no keel could sail closer than about 70º to the wind, because of drift. For most purposes this was good enough.

The speed of the change shows that plaited mat sails were a pain to make and use, although they had been a principal influence on the design of the rigs. Rotan lashings were replaced by galvanized fencing wire. Ropes were originally all plaited sennets, so they also were weak and not easily hauled through a deadeye. They were rapidly replaced first by hemp and sisal, then by nylon and later by cheaper polymers. A mast fixed in a thwart with wire stays and a masthead pulley and halyard was introduced, and the new materials allowed the sails to be larger. This, in turn, promoted the transition to other more efficient and more convenient rigs. Even so, the twoboom triangular sail persisted in many island groups.

Long distances were traveled. Until the early nineteenth century regular trips were made from Yap and Palau in the western Carolines northwards to the Marianas Is. and eastwards past Truk to the eastern Carolines, and then separately from there on to the Marshall Is. (Haddon and Hornell 1936: vol. 1:439). From there the route (usually of raiding parties, and always broken into sections) was southwards to the Gilbert and Ellice Is. (Haddon and Hornell 1936: vol.1:440; vol. 2:43) then to Wallis Is. and on to Samoa. Along a line further south, there was a route from the New Hebrides to Fiji and on to Tonga. From Tonga there was a route to the Cook Is. and on to Tahiti (Haddon and Hornell, 1936: vol. 1). Before 1800 there was regular trade from the Tuamotus in far Eastern Polynesia to Tahiti16, and north to the Marquesas. Much of this involved sailing eastwards against the prevailing winds and currents. In fact, the signs of land to windward in the flotsam and migrating birds, the necessity of getting upwind and the high probability of getting back home if in difficulties, have contributed to the development of boats and rigs on coasts with onshore prevailing winds, notably the Chams, Vikings, Portuguese and the English.

The mastless rig with a tilting sail was lost or strongly modified in Indonesia and the Philippines under the influence of Buginese traders with the tilted rectangular sail and the early colonists with jibs, staysails and yards. Curiously, the old design survived in Madura relatively unchanged, where it demonstrates how effective these rigs were (Horridge 1987:116-119). Versions of a single outrigger, the Fijian thamakau, are still used in the outer islands of Fiji. The Madurese fast canoes for transporting fresh fish (Horridge 1985: plates 8, 9) have the tilting rig with double outriggers and a strong rudder support. The Madurese jangollan, a heavy cargo boat up to 100 tons, is an modern example with the rig modified by addition of a short stubby mast (Figs. 12 and 13). These are the largest vessels now sailing that are available to give us some impression of how the rig of the large Fijian journeying canoes was handled at sea. They are very steady when heavily loaded. There appear to be no experimental measurements of their sailing performance, and it may soon be too late.

Sailing north or south is relatively straight-forward in the Pacific. Every few years the Polynesians had the opportunity of sailing long distances eastwards with favourable winds17. In Island South East Asia, the winds are dominated by the monsoons, and reverse twice a year. Sailors there can make regular trips and expect to sail home in comfort downwind when the monsoon turns.

Around Bali, the rig has been modified by supporting the upper boom on a short stubby fixed mast which is set one sixth of the way back from the bows, close to the forward outrigger boom, and firmly wedged in a strong thwart. The upper boom of the sail is slung from the mast at a big mechanical disadvantage (Fig. 14), and the tack of the sail is held rigid in the centre of the bows, so that the upper boom is held in the same plane as the midline of the hull. This arrangement presumably gives a better performance when sailing upwind, but the chief advantage is that the rig looks after itself when the single man crew is otherwise engaged with the fishing. The slope of the upper boom, and therefore the centre of action of the sail, is adjusted to give more weather helm by moving the tack forwards. In some villages there is a special smooth channel in which the tack slides. In Madura and Java, the mast is longer (Horridge 1987: figs. 65-68.). The Madurese leti-leti rig, with a large single sail (Horridge 1985: fig. 12; plates H, J, K.), is similar but has the tack on a fixed pivot in the bows, and the upper boom supported on a very stiff mast. To tack with these rigs, the boat is turned downwind to wear ship and bring the sail right around over the bows (Fig. 9). On large boats, there

The influence of other cultures on Pacific canoes The original mat sail was heavy, weak and soon rotted, especially when wet. Sails of bamboo slats were limited to the area of Chinese influence. European sails were introduced everywhere when woven cloth became obtainable, often sewn together from long strips of the strong cotton fabric used for mattress covers. Modern canvas followed.

16 Pahi described by Neyret (1974); Haddon and Hornell 196: vol.1: fig. 89; Langdon 1988:180). 17 Under the influence of changes in the El Niño Current which crosses the Pacific. See Irwin (1992) and the Polynesian Voyaging Society at http://pvs.kcc.hawaii. edu/evolution.html.).

94

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

is often a trimming sail (Fig. 23P) in the bows supported on a short mast (Horridge 1986a: fig.15). In several parts

of Indonesia, the influence of the Dutch is apparent in the spritsail rig on small boats (Horridge 1985: plate 24).

Figure 12. The mastless rig today. This is a Madurese Janggolan sailing downwind heavily loaded with salt. The main is a two-boom triangular sail supported by a pole to the stern and stays to the beams on each side. The fore sail is similar with a sliding ring on the bowsprit. The vertical pole in the stern is a support for the lowered rig. Note the two huge rudders that act as a keel.

95

ADRIAN HORRIDGE

Figure 13. Stern view of the janggolan in figure 12. Note the four stays to the beam on each side, and the way that the shape of the belly of the sail is controlled by the lacing. square or rectangular sails (Fig. 23E, R). The opening of the cinnamon and pepper trade through the Malacca Straits about 200 BC coincided roughly with the introduction of the triangular sail into the Indian Ocean, and the square or rectangular sail (Fig. 15) or the lugsail rig (Fig. 23V, W) from the Indian Ocean into Island South-East Asia. Fixed masts and square sails (Fig. 23E) are depicted on carved murals of the 8-9th centuries on the temples of Borobodur and Ankor Wat. The large ships of the spice trade, up to 300 tons, that the Portuguese found at Malacca in the 16th century had several masts with huge sails of this type (Manguin 1980; Horridge 1985:2).

Early modernization Alexander the Great had a fleet on the Indian Ocean, but he must have been using existing boats of Indian or Persian construction, probably derived from the ancient trade between the Persian Gulf, the Indus valley civilization, and Egypt18. At that time all Indian Ocean ships were of sewn planks and so far as is known had 18

Ancient Egypt used the rectangular sail hanging on a single yard on a fixed mast, with many controlling sheets. ref. Casson, Chap.2.

96

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

the eighth century AD, the Austronesian maritime empire of the Chams was based on the port of Hoy An, near Da Nang, and their sewn boats and triangular sails persisted until recently along this coast. From there they traded as far as Japan in the north and Sumatra in the south. Trade in metal spread eastwards from the Indian Ocean and from mainland Asia. Bronze drums Bellwood (1978:180-191,222) were carried from the region of Vietnam to the Eastern limits of Indonesia 2,000 years ago, soon to be followed by iron smiths and other traders. Iron tools, and their names, together with some Sanskrit words, spread with traders from mainland Asia by at least the 5th century AD. Iron was smelted in large quantities in Borneo and Sulawesi in the 10-12th centuries. Iron metal for foundry work by local smiths was traded along the coasts of New Guinea and to the eastern edge of the Philippines. The iron tools eventually influenced the boat building techniques and larger boats were made by adding planks in a set pattern that was constant for centuries, with a name for every plank. The curved chisel20 made round holes accurate enough for hardwood dowels to be inserted between the abutting edges of the planks instead of, or as well as, stitches and lashings (figure 4c). The dowels resist sheer stresses between the adjacent planks and overcome the main problem in building larger boats with many rows of planks. As before, the hull was held together by flexible ribs lashed down to projecting lugs on the planks, in a way that could be easily dismantled. This was the lashed-lug boat (Horridge 1982), which is known to be the standard planked boat of the East Asian Archipelago and the basis of power for the numerous sultanates of the Archipelago for about 1,000 years up to about AD 1800. Usually with outriggers, and with numerous men rowing or paddling, the kora-kora was a ship of war that could go directly into the wind across shallow reefs, and could be picked up and carried inland for safety. Many isolated areas, such as the Kai, Aru, Tanimbar, Solor, Sangir islands, kept this design into colonial times. Pirates with these vessels were finally defeated only when the steam gunboats arrived.

Figure 14. Balinese double outrigger trolling jukung, with a short fixed mast and a two-boom triangular sail. For tacking with this rig, see figure 9.

Figure 15. The tripod mast and tilted rectangular rig that was characteristic of the Indonesian trading boats from the 8th to the 19th centuries, and still persists today where there was Makassarese influence.

From about 1500 AD onwards, Buginese traders from South Sulawesi reached along the coast of New Guinea as far east as Biak in the north and the Aru Is. in the south, and later they fished for trepang off the north coast of Australia. They hung their traditional quadrilateral tilted rectangular sail on a fixed tripod mast (Fig. 23Q), which was frequently copied on the boats of cultures they encountered throughout central Indonesia, but their rig spread only as far as the trade goods from Asia, that is to

The Chinese rigs and boat building designs had negligible influence on the Pacific, or indeed anywhere outside China, but the Austronesians, with their language, customs and maritime tradition colonized the coasts of Vietnam and even the Mergui Archipelago on the west coast of Burma. Chinese rigs meet Austronesian rigs on the island of Hainan and the border with Vietnam19. In

20

The curved chisel, sharpened on the inner surface to act as a drill, was mentioned by Fr. Alcina in 1668 as a lokob (Horridge, 1982). It was the prized tool of the prahu builders of S. Sulawesi (Horridge, 1979).

19 Vietnam rigs in Piétri (1949). Ref to southern limit of Chinese rigs in the end-map in Paris P. (1955).

97

ADRIAN HORRIDGE

the Bismarck Archipelago and the Torres Straits. The Buginese are still moving by sea and establishing new colonies in Eastern Indonesia, with a mixed life-style as traders and exploiters. Their introduced diseases wipe out the local people on isolated islands. They build houses on the beach and later take over the land, as any tourist can see on the north coast of Bali.

Other findings, too numerous to be all coincidences, indicate that the crew of the San Lesmes had a lasting influence on the local population. The fishing net knots and the creation myths were European. Many of the earliest explorers noted the dress fashions and European appearance of the chiefs. The shipwrecked Spanish dogs of the Tuamotus had longer hair in more colours than the Polynesian dogs. When Capt. James Cook arrived in Tahiti in 1769, he found that prized dog hair was traded from the Tuamotus to Tahiti for the decoration of capes and robes. In 1767, Capt Wallis brought back from Nukutavake in the Tuamotu Archipelago the hull of a small canoe 3.8 m long (Haddon and Hornell 1936: fig.45) with no sign of the lashed lug or 5-part canoe designs. Later explorers described large double canoes with two masts called pahi, up to 30 m long, made of sewn planks fixed to shaped ribs with wooden pegs, with a keel and surmounted by a gunwale. Several of these double canoes, with distinctly European details in the construction, were sketched in the 18th and early 19th centuries. They were symmetrical and could sail either way by turning round the sails. The steering oar pivoted on a wooden pin. In Cook’s time they ran a regular service between Tahiti and the Tuamotus via Ana’a and Mehetia, transporting labourers and goods of all kinds.

Figure 16. Rig of the Tuamotuan double canoe, from a drawing by Capt. Wallis in 1767. Similar rigs were introduced by the Portuguese to many parts of their trading empire, notably the coasts of Brazil (Pâris 1843: plate 131; Bellec 1993). European introductions into E Polynesia A fascinating series of events appears to have led to the early introduction of a European boatbuilding technique and tacking rig in far eastern Polynesia, spreading from the Tuamotus to Tahiti. As recounted in books by Bob Langdon21, a Spanish caravelle of the Loaisa expedition, the San Lesmes, ventured far out into the Pacific in 1526, and eventually was wrecked on the island of Amanu, in the Tuamotus. One of its iron cannon was recovered from the reef there in 1929. Some of the Basque crew survived and had children, as shown by recently surveyed genes and numerous clues discovered by later explorers, from de Quiros in 1606 onwards. Captain Wallis of the Dolphin22 who was the next European to visit Amanu in 1767, made a drawing of a boat with two masts and two triangular sails (Fig. 16) and described a large planked boat being constructed frames first on a beach. Many details, notably ribs, stays, straight keel and plank pattern, were European in style.

Figure 17. Double canoes of Tahiti with a tacking rig. These copies of paintings by Lt George Tobin (with Bligh on the Providence) and others by Cook's artists, suggest that the masts could be tipped forwards or back to trim the direction of sailing, but the sail pressed against the stays when sailing downwind. They look inadequate for long sea voyages. The rigs of the Tahitian single outrigger and double canoes were well illustrated by the artists Parkinson, Henry Roberts and Webber, who traveled with Cook, and by Tobin with Bligh23. They had one or two curious

21

Langdon (1975, 1988). A splendid ripple on the world of scholarship; very educational. 22 See Langdon (1975), plate 20, for a copy of Wallis’sketch (from the National Library of Australia). Abstracts of Wallis’ journal (unpublished copy in Public Record Office, London), were published in Hawkesworth (1773).

23 Pahi double canoe in Dodd (1972:94). All Cooks artists and Lt George Tobin (with Bligh on the Providence) illustrated the

98

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

vertical tacking sails, often called the ‘Oceanic Spritsail’. The masts were of bamboo, supported by four stays of twisted hibiscus bark on each side, with fore and aft stays in some examples. The tall sails were each supported by a long curving boom along the trailing edge, and were slung behind the mast so that they tacked in the European way. To my eye, they look unseaworthy. This rig differed in principle from those outside Eastern Polynesia because the sail was behind the mast. There was no obvious way to control weather helm, and the relatively short stiff mast required about 8 stays that must have greatly increased the loading on it. The sail extended far above the top of the mast, with a sharp leading edge for greater efficiency (Fig. 17). When sailing down wind, the sail pressed against the stays, hardly a recipe for safety in the open sea. All of Cook’s artists illustrated the sails hoisted on both masts at the same time on the double canoes. None of these details were to be expected in eastern Polynesia.

communication with Tahiti, also had a tacking rig (Fig. 18) with a single outrigger25. Japanese style pottery found on the coast of Peru has been dated at 4,000 to 4,700 years old (Riley et al., 1971). Recently, a catamaran from Japan largely drifted 15,000 km to San Francisco in 51 days, continuing 6,000 km to Ecuador in 27 days. The discovery of DNA tested bones of the Asiatic chicken on the coast of Chili, carbon dated to 1200 AD (Storey et al. 2007), together with probable transport of banana and coconut, justified the work of Langdon (1989) and also suggests that someone really reached S. America from somewhere. This calls for a reconsideration of the boats that could span 5,000 km, and new re-enactments with truly antique craft. Clearly, the Tahitian and Hawaiian canoes of 1770 are unlikely candidates, but around 1200 AD they must have had better craft to reach Tahiti and Hawai’i. On the other hand, large Chinese junks were lost in the same system of currents about the same time. Interesting problem.

Further consideration of plaited mat sails suggests that they would disintegrate if allowed to flutter or flap for long. They were laced to thin flexible booms and cut away at the leach. This would account for the curious thin boom along the trailing edge of the Tahitian rig, as a solution to the problem of making a tacking rig with a plaited mat sail (Fig. 17). We know that long distance journeying stopped. My own view is that their mat sails and sennet ropes were too weak to support a tacking rig for distant travel, so they had cut off their options. Discussions in my own previous works omitted these European influences. Based on the varied and detailed revelations, one must conclude that the triangular tacking sails of the far eastern Pacific, and many other curious details, are strong evidence of Spanish intrusions 250 years before other explorers appeared. The broader significance of these findings is that we cannot infer that the rigs and all the other peculiarities of the tacking canoes were the most ancient by the “age distribution method” i.e., the idea that the oldest went the furthest. The San Lesmes was not the only ship to disappear in Polynesia before 1770. William Ellis (1834, chapter 16) recorded three accounts of foreigners landing at Kealakekua on Hawai’i before Cook. New evidence appeared in the late 1950’s of Spanish castaways from a galleon lost in the 1570’s, that would explain the pale complexions, helmets, short mantles and iron daggers noted by Cook. The burial casket of a deified chief deposited in a cave about A D 1600 contained a 2 m piece of European sail cloth and a iron tool like a chisel24. Ancient laid ropes that could be Spanish flotsam have also been found. The Hawaiians, who had some early

Figure 18. Hawaiian rig on a small single outrigger fishing canoe.

The triangular sail in the Indian Ocean

rigs of Tahitian canoes. For Tobin’s journal, see Schreiber (2007). 24 Reports of the casket are filed in the Bishop Museum Library under the heading ‘Spanish problem’. See also Stokes (1930).

25

Hawaiian tacking rig in Holmes (1981) p.65, and Buck (1957). See also (with reservations) Polynesian Voyaging Society at http://pvs.kcc.hawaii.edu/evolution.html.

99

ADRIAN HORRIDGE

About 200 BC trade between Arabia, India, Java and China developed. Cinnamon, cloves, nutmeg and pepper from Indonesia, and sandalwood and whale products from Timor, were shipped to Crivijaya and Malacca by Indonesians, then northwards to China, or westward to the African coast, as indicated by their eventual arrival in small quantities in the Roman Empire from that direction (Miller 1969: chapt. 8). There was an opportunity for the triangular sail to spread into the Indian Ocean and interact with rigs with fixed masts and square sails (Fig. 23E) or lug sails (Fig. 23X). Out of this union was born the Arabic Lateen26 which is an approximation to a triangular sail by sailors who already had a mast fixed in a thwart, the pulley, rope and rigging for a square sail (Fig. 23E). Actually, the origin of the lateen rig in the Indian Ocean is obscure, as would be expected if it was derived from early Austronesian traders.

mizzen mast (Fig. 19) and later became a gaff sail (Fig. 22). This was the origin of the fore and aft rig in Europe28. The two small caravelles which accompanied Columbus in the Santa Maria set out for America with two masts, each with a large triangular lateen, which was the rig that was later carried by the San Lesmes into the Eastern Pacific. So the triangular sail itself completed the circumnavigation of the world. Colonial influences Hull construction With the arrival of the European colonists and traders in the Pacific, the old rigs were quickly exchanged for new ones, sometimes in a very short time, as rigs are easily seen and copied. In Hawai’i, for example, square sails (Fig. 23E) and topsails were adopted almost at once. In Indonesia and elsewhere, boats with the mainsail in front of the mast and no forestay were fitted with a jib-sail, which had to be taken down every time that the boat went about (Wallace 1869: chapt. 28). The ancient technology with a tilting prop put only the weight of the sail on the hull, but now in a strong wind the stays and large sail generated unbearable stresses, 10 times the weight of the sail, that could pull out stays and push the mast through the bottom of a traditional hull. Therefore the whole technology had to change together; a long straight keel that was angled on to the stem and stern posts, instead of a gradual curve, the use of thick ribs to which the planks were solidly and permanently attached with tree nails, in contrast to the lashed ribs that compressed the hull, the use of thwarts, the transom and stern rudder, and the stepping of the mast on the keel.

Figure 19. Early European rig of about 1500 AD, with the triangular lateen sail on the mizzen mast. Austronesian influence is apparent today in many places around the Indian Ocean, with outrigger canoes in Sri Lanka, the west coast of India, and the Comores Is. (East Africa). Madagascar was first occupied by the Austronesian traders that arrived early in the first millennium AD, and their agriculture, house styles and language are still predominant there. The tilted triangular sail, now on a fixed mast, spread rapidly in the area dominated by the monsoons, reaching the Mediterranean in late Roman times (Casson 1971: 268-269)27. The triangular sail of the Nile gaiassa is of unknown age but not from ancient times BC. The lateen sail was rare in Roman records but common in the eastern Mediterranean. from the 9th century. The Crusader ships in the 13th century had two lateen sails as did numerous merchantmen of Pisa, Genoa and Venice. The ottoman lateen sail was adopted eventually by the Portuguese of the 11-12th centuries, first as the steering sail on the

Most of the Polynesian boat building and sailing traditions slowly disappeared in the 19th century, but some have been recently revived in modern guise for heroic re-enactments29. You should look carefully, however, for modern detail in modern "ancient" boats. Although the copies in nylon and fibreglass may look similar to the traditional ones, they usually have technically significant innovations, and are less likely to be worked to bits by a rolling sea. An interesting compromise between old and new was the hoisting of a rectangular lug sail (Fig. 23X) or a twoboom triangular sail upon a mast fixed in a thwart, so that one of the booms lies parallel to the mast, but projecting above it, while the other acts as a boom (Fig. 20), as on 28 ‘The Fore-and-Aft Rig in America’ by E. P. Morris (1927) is a collection of rare information. 29 e.g., Finney, B (1979); Jennings (1979) Chapt. 14. See also Polynesian Voyaging Society at http://pvs.kcc.hawaii.edu/evolution.html.

26

See Fig. 23 I. The sail was loose-footed when cloth was used. For contemporary illustrations, see Bellec 1993. Columbus' caravel rig, ref Landström 1961: 106. 27

100

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

the English Mirror dinghy. We call this a gunter lug sail (Fig. 23O). Judging from widespread illustrations from Thailand to Tahiti, there were many rigs of this type in the 19th century. At Singapore30, throughout Indonesia31, on the central Vietnamese coast32 and in Micronesia33, this system produced a fast convenient tacking sail, but it required a strong fixed rudder and a jib for trimming the rudder, because the sail will not tip forwards or back.

Dutch spritsail rig, or a relation (Fig. 21), as found at Ende in south Flores, Indonesia34 in Jakarta Bay35, and once common in the eastern Pacific36. In Island south-east Asia there are still plenty of small boats with a trapezoid or a European rectangular lug sail (Fig. 23X) suspended on a fixed mast by two halyards ending at different distances from the centre of the upper boom37 and a lateral rudder that rises conveniently over coral reefs.

Figure 22. Gaff rigged cutter, still in use in parts of Indonesia.

Figure 20. Two-boom triangular sail attached to a fixed mast, as on the 19th century Singapore fast boat. Now called the Gunter-lug sail in England and layar nadé in Indonesia (Fig. 23O).

Even the adoption of the European fore-and-aft rig took place in parts, and at first imperfectly, starting at major ports and places where European ships met local traders. The Singapore fast boat (Fig. 19) was developed at the end of the 19th century to bring passengers from ships and to deliver fresh fruit and vegetables to that port. About 1900, the European yacht rigs were introduced for the same purpose, speed. Even fast racing yachts purchased by local wealthy colonists had an immediate effect upon local rigs that slowly spread away from the main ports. The gaff rig38 was everywhere copied from European schooners (Fig. 22). A careful search for the forebears of the Bugis pinisi of 1960-1980 with two curtain gaff sails and two jibs, suggests that the rig was probably copied from American schooners, which had given up the square topsails in the early 19th century, and instead used gaff topsails. For example the L.A. Dunton, built in 1921, and now preserved at Mystic Seaport, Connecticut, was one of the last of a long line of Cape Cod fishing schooners,

Figure 21. Dutch spritsail. In corners of the former colonial empires one still finds fishing canoes or even large sailing transports with a

34

Endé rig in Burningham (1990: fig. 10). Jakarta Bay rig. (Horridge 1985: plate 24). 36 Spritsail in the Leeward Is. (Haddon and Hornell 1936:vol.1: fig. 84). 37 Some illustrations in Warington-Smyth, Chapt. 11. 38 Gaff rig lambo See Horridge The Prahu, figure 32. The lambo gaff rig was that of the English pearling boats of early 20th century. 35

30

For a scale drawing of a Singapore fast boat see Warington Smyth (1929:408). 31 History of the lambo and nadé rig (Horridge1985: chapt.18). 32 The Ghe Câu of Phan Rang (Piétri 1949:plate 43). 33 Micronesian modern rig (Knox-Mawer and Carmichael 1968:40d).

101

ADRIAN HORRIDGE

traders and whalers, with almost exactly the rig adopted in South Sulawesi by about 1930.

lug in Europe, also spread from Hong Kong, Singapore and Batavia. In some places the tilted rectangular sail was abandoned in favour of a triangular tacking sail, as in the southern Philippines, on the west coast of Sulawesi, and at Pasir Putih in Eastern Java. At the end of the 19th century, there was a hotch-potch of colonial influences superimposed on the declining indigenous traditions, quite difficult to unravel.

Sometimes progress stopped at the use of a rope and pulley to raise the triangular sail, as in Madura, sometimes progress stuck at the use of the Dutch spritsail, as in the Aru Is. or at the square sail (Fig. 23E), as in the Bismarck Archipelago and Torres Straits. Later the nadé rig (Horridge 1985: fig. 33) (Fig. 23O) called the gunter

Figure 23. Sail types, with arrows suggesting relationships. In each case, the first name given here is preferred because it refers to the shape or place and carries no overtones of previous theories. In the box on the left are the tilting mastless rigs; in the box on the right are the European rigs.

102

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS

Conclusion

A, Madurese jukung rig; Primitive Oceanic Lateen of Haddon and Hornell (1936); Crane Sprit of Doran (1981). Suitable for rafts or outrigger canoes without pulley. Ancient Peruvian Balsa rafts. B, Polynesian Oceanic Rig; Simple Oceanic Spritsail of Haddon and Hornell (1936); Oceanic Spritsail of Bowen (1953) and Doran (1981). The main boom stands on a notch, supported by stays. C, Two-boomed triangular sail on a fixed mast. Mechanically similar to K. D, Indian Ocean Loose-footed Lugsail; Trapezoid Lateen. E, Square sail; Ancient Egypt, Mediterranean. F, Old Micronesian Rig; True Oceanic Lateen (Haddon and Hornell 1936; Crane Sprit of Doran (1981). G, As in B but going downwind. Marquesan Claw Sail (Lewis 1972). H, Tilted Rectangular Sail probably from the Indian Ocean to Indonesia, or Trapezoid Sail (Coramandel Coast). Common S.E. Asia, also near New Guinea; (Haddon and Hornell 1936: vol. 2:162, 174, 246). I, Arab or Mediterranean Triangular Lateen. J, Gaff Rig, 18th century English and American; 19th century in Indonesia. K, Balinese Jukung Rig; Madurese Leti Rig. The upper boom is fixed at the head of a short mast and the tack is fixed in the bows. L, Hawai’I Rig; Boom Spritsail (Haddon and Hornell 1936); Oceanic Spritsail (Doran 1981; Holmes 1981); Society Island Spritsail (Bowen 1953). M, New Guinea Crab Claw; Crab Claw Spritsail (Bowen 1953). N, Tilted Elliptical Sail on fixed mast. New Guinea. O, Gunter Lug Sail; entirely modern, English origin in the West Pacific. P, Madurese Foresail in two positions. The upper boom is on a short mast. A ring on the tack slides along the bowsprit. Copied from the Dutch. Q, Tilted Rectangular Sail on a tripod mast. Introduced into Java before the 8th century. Still used by Buginese, Makassarese and others, called Layar Tanja. R, Square Sail on Tripod Mast. A suitable rig for reed boats that were once widespread. Ancient Egypt, Sumeria, Indus Valley cultures. S, Leg of Mutton Sail or Bermuda Rig. T, Indian Ocean Double Spritsail; Proto-Oceanic Spritsail (Bowen 1953). Sri Lanka, India, Madagascar. Possibly a recent adaptation for hauling a trawl downwind. U, Melanesian Spritsail. Relationships unknown. V, Boomed Spritsail. Dutch. W, Loose-footed Spritsail. In Java called layar suduk (dagger rig). Common in Indonesia from the Dutch influence. X, Lug Sail. Related to E. Ancient Middle East, Indian Ocean, Scandinavia. Now world-wide.

Making sense of the Pacific rigs is based first upon the absence of the pulley, the mast fixed in a thwart, stays, the fixed rudder and the square sail (Fig. 23E), none of which spread primitively into the Pacific. The Polynesians had a more appropriate rig for use with mechanically weak materials. They used sennet and pushed the sail up on a loose prop, which in turn supported the mechanically weak mat sail that was laced to booms on two sides. They tilted or rotated the sail on its tack to steer like a windsurfer, so they never needed to invent the fixed rudder or the jib sail. The Austronesian rig later contributed to the development of sailing upwind worldwide, and was itself modified and then replaced (Fig. 23) when tougher and stronger materials made possible larger and faster designs. The great variety of Pacific canoes was rather exhaustively described by Haddon and Hornell (1936) and by Neyret (1974). However, the evolution of Pacific hulls and the relationships between the rigs suggested by these authors, or Bowen (1953) or Needham (1971), or others that copied from them, does not correspond at all with that given here. They did little first hand research among sailors, and were unaware of the recent evidence. More importantly, they were not engineers, sailors or boat builders and failed to elucidate the over-riding constraints that governed the limited range of designs and size of boats in the Pacific before the European explorers arrived. The mechanically weak mat sails limited the rig and the size. The over-riding consideration in the hull design was the need to avoid fatigue fractures and relieve the sheer stresses caused by flexing and twisting in the waves. Canoe superstructures were built with lashings so that they could accommodate stress concentrations. The planks were sewn edge to edge and the flexible ribs lashed to lugs kept the planks pressed together like an inverted arch in compression. You have to get away from the European ideas that the square sail (Fig. 23E), fixed mast and rudder are primitive and that sails were made of sail cloth. The best introduction to an understanding of Pacific rigs is to learn the art of windsurfing with a tilting mast and study the performance of the rigs currently in use by the Madurese (Figs. 7, 12, 13), then consider the engineering uses of the relatively weak natural materials that were available, and the limited options that allowed a safe design with the best compromise of speed and permissible size. References Anson, Baron George 1748 A voyage round the world (1740-44), compiled by Richard Walter. Reprinted (G. Williams, Ed.) 1974 University Press, Oxford.

103

ADRIAN HORRIDGE

Bellec, F. 1993

Finney, B. R. 1979 Hokule'a: The way to Tahiti. Dodd, Mead & Co, New York. Haddon, A. C. & Hornell, J. 1936-38 Canoes of Oceania. Bishop Museum Special Publ. 27, 28, 29. Reprinted as one vol. 1975. Bishop Museum Press, Hololulu, Hawai'i. Hawkesworth, J. 1773 An account of the voyages ---Byron, Wallis, Carteret and Cook. 3 vols. London, Strahan & Cadell. Heine-Geldern, R. 1932 Urheimat und früheste Wanderungen der Austronesier. Anthropos 27: 543-619. Holmes, T. 1981 The Hawai'ian canoe. Editions Limited, Hanalei, Hawai'i. Horridge, G. A. 1979 The Konjo boatbuilders and the Bugis prahus of South Sulawesi. Monogr. 40. Maritime Mus. Greeenwich, London. 1985 The prahu. 2nd edn. Oxford University Press, Kuala Lumpur. Page numbers in the notes refer to this (preferred) edition. 1982 The lashed-lug boat of the Eastern Archipelagoes. Monogr. 54. Maritime Mus. Greenwich, London. 1986a Sailing Craft of Indonesia. Singapore Oxford in Asia. 1986b “The evolution of Pacific canoe rigs”. Journal of Pacific History 21: 83-99. 1987 Outrigger canoes of Bali and Madura, Indonesia. Bishop Museum Press, Honolulu, Hawai'i. Irwin, G. 1992 The prehistoric exploration and colonisation of the Pacific. University Press, Cambridge. Jennings, J. D. (ed.) 1979 The prehistory of Polynesia. ANU Press, Canberra. Jett, S. 1971 “Diffusion versus independent development”. In C.L. Riley, J.C. Kelley, C.W. Pennington, R.L. Rands (eds.), Man across the sea. Problems of Pre-Columbian Contacts. Austin. Univ of Texas Press. Johnstone, P. 1980 The sea-craft of prehistory. Routledge & Kegan Paul, London. Knox-Mawer, J. and Carmichael, P. 1968 A world of islands. Collins, London. Landström, B. 1961 (English Ed.) The ship. Allen & Unwin, London. Langdon, R. 1975 The lost caravel. Pacific Publications. Sydney.

Nefs, galions et caraques. Editions Chandeigne, Paris. Barrau, J. (ed.) 1963 Plants and the migrations of Pacific Peoples; a symposium. Honolulu, Bishop Museum Press. Bellwood, P. 1978 Man's conquest of the Pacific. Collins, Auckland N.Z. Bellwood, P. 1995 “Austronesians Prehistory in Southeast Asia: Homeland, expansion and transformation”. In P. Bellwood, J.J. Fox, and D. Tryon, D. (eds.), The Austronesians: Historical and Comparative Perspectives. Dept. of Anthropology, Australian National University. Canberra. Bellwood, P., Fox, J. J. and Tryon, D. (eds.) 1995 The Austronesians: Historical and Comparative Perspectives. Dept. of Anthropology, Australian National University. Canberra. Best, E. 1925 The Maori Canoe. Govt. Printer. Wellington, New Zealand. Bisschop, E. de 1959 Tahiti Nui. New York. Bowen, R. Le Baron 1953 Eastern sail affinities. American Neptune 13: 81-113 and 185-214. Buck, P. E. 1957 Arts and crafts of Hawaii. Also in Bishop Mus. Special Publ. 45 Section 6 (1964). Burkill, I. H. 1966 A Dictionary of the Economic products of the Malay Peninsular. Publ. Govt of Malaysia. Kuala Lumpur. Burningham, N. 1990 Stemless boats of Ende Bay. The Beagle. Records of the Northern Terr. Mus., Darwin 7: 105-119. Casson, L. 1971 Ships and seamanship in the ancient world. University Press, Princeton, N. J. Dodd, E. 1972 Polynesian seafaring. Dodd, Mead & Co, New York. Doran, E. 1981 Wangka: Austronesian canoe origins. Texas A & M University Press, College Station. Ellis, W. 1834 Polynesian researches. 2nd Edn. London, Fisher & Jackson. Green, R. 1979 “Lapita”. In J.D. Jennings (ed.), The prehistory of Polynesia. ANU Press, Canberra.

104

ORIGINS AND RELATIONSHIPS OF PACIFIC CANOES AND RIGS 1988 1989

The lost caravel re-explored. Brolga Press, Canberra. “When the blue-egg chickens come home to roost”. Journal of Pacific History 24: 164192.

Historical and Comparative Perspectives. Australian National University, Canberra. Stokes J.F.G. 1930 Burial of King Keawe. Papers of the Hawaiian Historical Society, Honolulu, 17: 63-73. Storey, A. A, J.M. Ramirez, D. Quiroz, D.V. Burley, D.J. Addison, R. Walter; A.J. Anderson, T. Hunt, J.S. Athens, L. Huynem, E.A. Matisoo-Smith 2007 “Radiocarbon and DNA evidence for a PreColumbian introduction of Polynesian chickens to Chile”. Proc. Nat. Acad. Sci. USA 104(25):10335-10339. Wallace, A. R. 1869 The Malay Archipelago. Macmillan, London. Warington Smyth, H. 1929 Mast and sail in Europe and Asia. Blackwood, London.

Lewis. D. 1972

We, the navigators. Canberra. Australian National Univ. Press. Malinowski, B. 1922 Argonauts of the western Pacific. Routledge & Kegan Paul, London. Manguin, P.-Y. 1980 “The Southeast Asian ship: An historical approach”. Journal of Southeast Asian Studies 11: 266-276. Miller, J.I. 1969 The Spice Trade of the Roman Empire. Clarendon Press, Oxford. Nishimura, S. 1925 Ancient rafts of Japan. Publ. Soc. Naval Architects, Tokyo. Morris, E. P. 1927 The Fore-and-Aft Rig in America. Newhaven; Yale Univ. Press. Needham, J. (ed.) 1971 “Civil engineering and Nautics”. In Science and Civilization in China, vol.4(3). Cambridge University Press. Neyret, J. 1974 Pirogues Océaniennes. Assoc. des Amis des Musées de la Marine. Paris, France. Pâris, F. E. 1843 Essai sur la construction navale des peuples extra-Européens. Arthus Bertrand, Paris, France. Paris, P. 1955 Esquisse d’une ethnographie navale des peuples annamites. Publ. Mus. Land- en Volkenkunde Prins Hendrik, Rotterdam. N°3(2):1-86. Piétri, J. B. 1949 Voiliers d'Indochine. S.I.L.I., Saigon. Riley C.L, J.C. Kelley, C.W. Pennington, R.L. Rands 1971 Man across the Sea. Problems of PreColumbian Contacts. Austin. Univ of Texas Press. Schreiber, R. (ed.) 2007 Captain Bligh’s second chance. The journal of Lt. G. Tobin. University of New South Wales Press. Severin, T. 1994 The China Voyage. London. Little, Brown and Co. Spriggs, M. 1995 “The Lapita Culture and Austronesian prehistory in Oceania”. In P. Bellwood, J.J. Fox and D. Tryon (eds.), The Austronesians:

105

geographic situation one part of the island is covered with luxuriant forest. The other Austral Islands have certain of these characteristics but not all at the same time. Modernisation has always been resisted and the airport was only built after years of negotiations and the election of younger representatives. Electrification has been a sensitive subject where sentiments about the observation of the Sabbath weigh heavily. The piped water supply is not potable and projects to tap the aquifer have failed. It should be noted that certain areas of the island are totally avoided by the inhabitants. A mixture of religious belief and the past confinement of lepers have made areas prohibited to the point where no inhabitant may enter therein (Edwards 1993: 41). The Raivavae language is characterised by differences in usage as well as a singsong pronunciation. Raivavaean has resisted the domination of Tahitian and is used by all the natives of the island. In these conditions, their canoes continue to satisfy the needs of the fishermen and they are built by hand as they were in the past.

Chapter VI Dugout and sewn plank canoe construction on Raivavae, Austral Islands by Robert Veccella* The island of Raivavae covers 16 km² and has a thousand inhabitants. It is 730 km South-southeast of Tahiti (Fig. 1). An airstrip opened in 2002 ended two centuries of ‘isolation’ and will certainly lead to the disappearance of what remains of the ancestral traditions. The inhabitants are quasi self sufficient but economic difficulties have been increasing for years. Raivavae is unusual in that it has a very large lagoon, accessible to ships, surrounded by numerous motu or sandy islets. Because of its

Figure 1: Map of French Polynesia and Raivavae ____________________________ *

Archaeologist in charge of the “Underwater Archaeological Research Group” in Tahiti.

107

ROBERT VECCELLA

During my first research trip in 1998, I was surprised by the characteristic canoes of this island which seem to have come from another era. These canoes have received little attention in the literature. This paper reports on the current state of traditional canoes in Raivavae as well as a review of unpublished ethnographic descriptions from the 1930’s.

Canoe Length number 24 6,35 32 6,80 30 6,90 5 7,05 46 7,25 31 7,90 34 7,90 45 8,00 40 8,15 23 8,20 37 8,40 29 8,50 2 9,10 39 9,20 6 9,30 15 9,30 27 9,30 50 9,35 25 9,40 41 9,40 49 9,40 43 9,55 4 9,60 47 9,65 21 9,80 51 9,82 38 10,20 42 10,25 22 10,30 36 10,40 20 10,50 13 11,05

Canoes on Raivavae In 1999, I recorded fifty two canoes along the entire north coast of the island. Canoes of the south coast were not studied. Of the 32 canoes that were measured, 50% were between 9 and 10 metres long; 90% were between 8 and 11 metres (Table 1, Fig. 2). Nearly half of them had the same height and width at their point of maximum beam, that is they were ‘square’ in cross-section. The others varied only an inch or two, exceptionally 3 (2.5-7.5 cm). In 60% of the canoes, the lower portion of the hull was in two pieces. The remainder were in one piece. No three piece hulls were observed. Mounting the outrigger on the starboard side of the hull which has been considered a hallmark of Raivavae canoes was in fact only observed in 60% of the cases. When outboard motor brackets were installed 80% of the time they were on the port side, 55% of the time on the starboard side (the overlap is because in some cases the bracket allows the motor to be on either side). One atypical canoe had a transom stern just aft of the rear crossbeam and this innovation may become more common in the future. Motorization is not at present the norm and although 65% of canoes could carry motors since they have motor brackets, the rest are propelled uniquely by hand. All finished canoes are painted. The dominant colours are white and a range of blue, light blue and blue-green.

Internal width 0,45 0,40 0,30 0,42 0,40 0,40 0,35 0,53 0,45 0,40 0,45 0,40 0,45 0,38 0,52 0,50 0,60 0,45 0,45 0,43 0,45 0,53 0,52 0,48 0,54 0,50 0,45 0,40 0,45 0,48 0,50

Depth 0,40 0,38 0,30 0,40 0,45 0,40 0,35 0,48 0,45 0,48 0,45 0,40 0,42 0,45 0,50 0,55 0,58 0,45 0,45 0,45 0,45 0,50 0,52 0,40 0,46 0,45 0,40 0,45 0,45 0,48 0,45

Table 1: Dimensions of canoes in order of length.

The earliest descriptions There is little in the literature about the canoes of Raivavae. Haddon and Hornell (1936:153) note, citing Corney (1913: vol.2: 313), that according to Andia y Varela, who discovered the island with Gayangos in 1775, Raivavae had the same type of canoes as Rurutu. Samuel Stutchbury gave a short description of the canoes in 1826, making reference to those of Rapa but his sketch seems to have been strongly influenced by canoes from New Zealand (Branagan 2001: 72-73). One must wait until the 1920’s to find the first detailed descriptions of Raivavae canoes. Stokes (1929-1930) an ethnologist at the Bishop Museum, spent six months on the island in 1922 and carried out the first archaeological excavations on Raivavae but they were never published. His text is simply descriptive, the section on canoe manufacture was not completed and only a few species of trees are cited. The most complete information is gathered by Stimson (1928-1931) but again his preliminary report remains in the form of a transcription of notes.

Figure 2: Comparison of dimension at maximum beam.

108

CONSTRUCTION OF DUGOUT AND SEWN PLANK CANOES ON RAIVAVAE, AUSTRAL ISLANDS

Outside of the polemic surrounding certain of his texts, Stimson was a careful observer who meticulously recorded an enormous amount of data. For example he described no less than 71 different species of trees on Raivavae, giving their vernacular and scientific names and their primary and secondary uses. His data was collected from multiple informants. In general Stimson’s work has been considered problematic, notably because of certain unfounded interpretations. This has been complicated by the liberal use that Marshall made of Stimson’s note.

Botanical name

Concerning Raivavae canoes, Stimson divided his work into five parts; one man dugout canoes, sewn canoes, sailing canoes, rafts and racing model canoes. This paper will be limited to sewn canoes. Outrigger fishing canoes are from 20 to 30 feet (6-9 m) long. They are narrow and have a flat bottom. They are made of hollowed logs and planks, sewn together with plaited coconut (Cocos nucifera) husk fibre (sennit). Stimson felt that only in the Austral Islands, had the old way of making and assembling canoes out of numerous pieces, as well as the plaiting of Pandanus (Pandanus tectorius) been maintained without change.

Common name Beach hibiscus

Vernacular name Pūrau

Cocos nucifera Linné

Coconut

Tumu ha’ari

Calophyllum inophyllum Linné Barringtonia asiatica (Linné) Kurz Mangifera indica Linné Artocarpus altilis (Parkinson) Fosberg Casuarina eguisetifolia Linné Thespesia populnea (Linné) Solander exCorrea Dodonaea viscosaJacquin Pandanus odoratissimus Linné

Alexandrian laurel Fish-poison tree Mango Breadfruit

Tāmanu or ‘ati Hotu ou hutu Tumu Vī ‘Uru

Wood used for bow and stern covers, fruit fibres for caulking, hulls. wood used for bow and stern covers Sap used for caulking and glue.

Ironwood

‘Aito

Rear crossbeam.

Pacific rosewood

Miro ou āmae

Dowels, bailer.

Apiri Fara

Rear crossbeam. Plaited mat sails.

‘Ō’iri* (?)

Hibiscus tiliaceus Linné

Pandanus

Uses and notes Wood used for all parts of the canoe, especially the nose (bow cover), outrigger float, forward crossbeam, paddle, bailer and mast. Bark for cordage. Husk fibre for lashings. Oil for caulking. Nut shell to block lashing holes. Stem and leaves’s fibres for assembling sails. Bailers, hulls.

Ficus prolixa Forster

Banyan tree

‘Aoa ou ‘ōrā

According to Stimson, it seems that the small round red seeds with black ‘eyes’ of these shrubs were sued to decorate the lashings by mixing their juice with breadfruit sap. Cordage.

?

?

Hoto (?)

Paddles.

Table 2: The different species used for canoe construction in Raivavae in 1930 according to Stimson. *This word has different meaning such as stone adze blade lashed with coconut sennit, the action of lashing an adze head to the haft. A synonym that designates adzes used for hollowing canoes is Orna. It appears that Stimson may have confused this term. The timber for canoes was always chosen from the centre of a healthy tree. The species varied as a function of the piece being made (Table 2). Beach hibiscus (Hibiscus tiliaceus) due to its strength and light weight could be used for any part of the canoe but particular attention was given to the quality of the wood used for the bow section.

Tamanu (Calophyllum inophyllum) and Pacific rosewood (Thespesia populnea) were second choices due to their weight and possible defects. They were nevertheless often used for the bottoms due to their resistance to abrasion when canoes were launched or pulled out of the water. The fish-poison tree (Barringtonia asiatica) and the

109

ROBERT VECCELLA

Except for the number of people, certain materials and compressing the canoe during sewin, Stimson’s description has not changed very much to today.

mango (Mangifera indica) were sometimes utilised for the upper parts although not the bottoms even though these woods are sensitive to insect attack. The lightness of breadfruit wood (Artocarpus altilis) was useful for the upper strakes or gunwales which are not exposed to the same stresses as the other parts and which allowed lightening the canoe. The hull bottoms were carved in a U shape whose bottom was 2 inches (5 cm) and whose upper edges were 1 inch (2.5 cm) thick, surmounted by sewn on gunwale strakes. The different parts, seven according to Stimson, were in the past carved with stone adzes while the wood was still green, since the arrival of Europeans they are carved with steel tools. Smoothing and polishing the skin was carried out with coral abraders and shell scrapers. Stimson noted that the pieces were assembled with wooden pins, caulked and sewn together. In 1930 the joints between the upper and lower parts were reinforced by Thespesia dowels inserted every arm length. These dowels, two inches long and a half inch in diameter (5 cm by 12 mm) were driven half their length into the upper edges of the hull bottom once the holes in the upper parts were ready to receive them. The fore piece was reinforced by a thick wooden post or gudgeon to resist the weight of harpooner. It’s curved, flaring shape also contributed to the solidity of the assembly. Stimson noted that this may be an innovation borrowed from recent Tahitian or European methods of construction. Also according to Stimson the two ornamental ears at the front contributed to the solidity of the assembly1. The rigidity of the sides was ensured by the profile or lip carved into the upper edge or the gunwale, by a thwart and by the crossbeams tied between the two sides.

Henri Jacquier (1955: 494-496) published an article entitled “Note sur les pirogues actuelles de l’île de Raivavae (Notes on the current canoes of Raivavae). Clairvoyantly he ended his article by soliciting the competent authorities of the time to pay closer attention to these canoes. Although the article contained numerous errors, it was nevertheless a cry of alarm for protection of Raivavae’s maritime patrimony. Following his stay in 1957, Marshall (1961) wrote “Ra'ivavae, an expedition to the most fascinating and mysterious island in Polynesia”. It does not include a chapter on canoes, but it does discuss the community work groups involved in their construction. It gives little information about techniques, materials, dimensions or form. Still the accompanying sketch is extraordinarily current (Marshall 1961: 156). As a student of Stimson, Marshall gives great importance to the work of his master2. For him, canoe makers were the sole link with past traditions. In each village there were one or two masters who were willing to build canoes for money. These men did not work full time; their price seems to have been too high. These experts still preserved the knowledge of carving the planks, and the techniques of assembling the seven pieces to produce a graceful sewn canoe with its extended flat nose or bow piece. But Marshall already noted in the beginning of the 1960’s that the inhabitants of Raivavae were beginning to use metal bars and imported glue when they could. Marshall claimed that these sewn canoes had a solididy and load capacity greater than those from other islands made from douglas fir (Pseudotsuga menziesii) planks or of simple dugouts. Marshall claimed to have been stunned when he saw these sewn canoes, the largest in the islands, for the first time. With their extended prows, they seemed to have come from out of the distant past. Like many he had only known them from the drawings of the early explorers. He noted that on Raivavae the outriggers could be set up on either side contrary to the practice on other islands where they were only found on the left side.

The caulking used in ancient times was a heated mixture of breadfruit sap, wood ashes and coconut oil. This glue was spread along the parts to be joined and covered by a layer of bark cloth (tapa). The entire canoe was compressed while the glue dried, Stimson noting that it was encircled by bands of firmly tensioned hibiscus bark for one night. The sewing was carried out with a forked stick, the cord being tensioned on one side while the stitch was beaten on the other to eliminate the slack. Starting on the interior the cords were passed three times through each hole. The remainder of the hole was stuffed with coconut husk. The outrigger float was always made from a straight trunk of light beach hibiscus wood about 6 inches (15 cm) in diameter, supported by two crossbeams. The rigid forward crossbeam is carved from the resistant heartwood of beach hibiscus. It continues outboard of the float and curves upward to avoid shocks from the waves. The rear crossbeam is a flexible bow of ironwood (Casuarina equisetifolia) that allows the canoe to navigate in rough seas, giving it more stability.

During his stay, Marshall did not observe the construction of a canoe but he did record one being refitted. The work took place behind the owner’s house. The canoe was 5 years old and the lashings had not been retightened since its construction. The plank joints were no longer tight and it had begun to leak making it unsuited to long trips on the lagoon. The owner’s two daughters had married sons of the only canoe builder in the village. Together the two families shared the work of rebuilding the canoe. The women made cordage by stripping the bark of young

1

These ornamental ears were only seen once by Stimson and Stokes. They are present on the Raivavae canoe that Jacquier had had made in 1961, now in the “Musée de Tahiti et des îles”.

2

It seems that Marshall’s chapter on canoes was based on Stimson’s unpublished field notes from 1929

110

CONSTRUCTION OF DUGOUT AND SEWN PLANK CANOES ON RAIVAVAE, AUSTRAL ISLANDS

Figure 3: Beached canoes. Note the characteristic overhanging bow piece (nose).

Figure 4: Moored canoe.

111

ROBERT VECCELLA

exhibition catalogue was written. According to him Raivavae remained the only island where canoes were made all alike of seven pieces joined by eight sewn seams. For Brunor, canoes made in this manner, but larger, were the kind used by the ancient Polynesians to conquer the entire Pacific and to reach the most distant islands.

Beach hibiscus trees for the temporary lashings. Meanwhile several men spun coconut fibre string by rolling it on their thighs to plait into the sennit for the lashing. Others carved ironwood dowels. The most skilled men re-cut the plank joints to their new fit. Under the eye of the old builder, whose intervention was limited to an occasional grumble or wink, the entire team worked like a well oiled machine. When all the preliminary work was finished, the more precision task of assembly could begin. The women and the young children retired, leaving the men to work, but the older boys were left to observe, to learn and help out with small tasks. Each piece was trimmed and the new dowels were installed. The men cut strips out of old jute sacks the width of the joints. The upper surface of the hull bottom was spread with a malleable black substance, a strip of sacking was laid on and covered with the same product. Then the upper plank was carefully installed.

One must wait until Stimson, whose studies were quasi contemporaneous with those of Stokes, to find the first drawings and good descriptions of the canoes. We don’t have any reliable descriptions prior to 1930. Comparing their descriptions to my own observations, it appears that neither the form nor the construction have changed much in the past 70 years. Certain traits of these canoes, notably the bow platform, are not unique to Raivavae but are shared with the Leeward Society Islands. The carpenter

The different pieces were temporarily bound together with strips of beach hibiscus bark passed through the lacing holes. These lacings were tensioned with wedges of soft wood. During the trial assembly of the different pieces another team was occupied making the braided coconut fibre sennit. The next task was to lift the partially assembled canoe and to pose it on sections of coconut trunk half buried in the earth. A large curved branch of beach hibiscus was placed crossways over the canoe and another underneath. They were then joined by numerous turns of beach hibiscus bark that were then tightened by a stick like a big tourniquet3. Once the glue had set, the temporary lacings were carefully removed and replaced by the definitive braided coconut lashings. This difficult step required a series of meticulous operations carried out by a team of specialised workmen. The cord was passed though the holes using a needle made of coconut leaf midrib. The cord was tensioned by one of the men applying all his weight on a forked stick while a second man hammered on the cord on the other side of the hull to aid in tensioning it. A carved wooden pin was then hammered into the hole to keep the cord from slipping.

Yves Lemaitre (1973: 116) defines tāmuta as: carpenter, mason, also one who carries out corresponding work. The tāmuta rā’au is the carpenter, the cabinet make. According to the dictionary of the Académie tahitienne (1999:449), it is a noun borrowed from the English ‘carpenter’. There is also the term tahu’a va’a referring to canoe builders (op. cit.: 1999:439). Joseph Tchong, an archaeologist and linguist by training, believes that tahu’a and tāmuta, despite the English derivation of the latter have the same meaning. That is, an expert, one who knows and owns certain knowledge, one who is the authority in the domain. In his notes from Raivavae, Stimson (1928-1931) never speaks of tāmuta, nor does he use this Tahitian word, only canoe-maker or shipwright. He does not appear to emphasise the symbolic dimension of the role. However in texts written in Raivavaean, his informant does use tāmuta. On Wallis (Uvea) a traditional carpenter is called tufuga. It is both a title and a familial statute transmitted from father to son through the generations (Simutoga 1992: 166). It was the same in Tahiti in ancient times (Henry 2000: 162). On Raivavae the father, the canoe builder, passed his knowledge and his tools to his sons or sons-in-law who worked with him. Canoe building is not a full time occupation, the carpenter does other things as do all the islanders, he goes fishing, he tends his plantations of taro (Colocasia esculenta), bananas (Musa spp.), etc. These activities fill in time when they are not working on canoes and are part of the daily rhythm. Certain carpenters make model canoes for sale but these are not simply models for tourists. They probably issue from the recorded ancient practice of racing model canoes (Henry 2000: 287). They may also serve as a means of transmission of knowledge.

Here again Marshall’s observations are entirely up to date. Martin A. Brunor (1969) made numerous voyages through the Australs. Starting in 1920 he collected hundreds of archaeological and ethnographic objects that have enriched the collections of the Peabody Museum in Salem, Massachusetts. Notably he brought back a sewn canoe from his travels that was shown along with numerous other objects in an exhibit in the late 1960’s. During his first voyage he had noted that one could still find sewn canoes on five islands in East Polynesia, unfortunately he didn’t note which ones. At the time the

The tools The tools used by the ancient canoe builders have left few traces. Stimson listed their wood working tools as four

3

This technique is not used anymore except when making a beverage from ti (Cordyline terminalis (L.) Kunth) roots.

112

CONSTRUCTION OF DUGOUT AND SEWN PLANK CANOES ON RAIVAVAE, AUSTRAL ISLANDS

types of adzes, an axe, a gimlet and a chisel. There are two types of axes on Raivavae, those used for felling trees and those used to carve canoes. Their cutting edge, parallel to the helve allows working in relatively narrow spaces, as on the flanks of the canoe, where the movement of the arms follows the longitudinal axis of the hull (Stimson 1928-1931: 723). Stimson indicates that piercing tools were once made from stone or a very hard variety of coral (op. cit.: 720). The change in tools is the most remarkable indicator of the evolution of canoe construction. As we will see, the gestures and operations have remained little changed from the past while the tools have changed radically. All the ancient tools have been replaced by manufactured ones. On Wallis, carpenters have retained their ancient style of adze hafts upon which they lash a steel blade (Simutoga 1992). In general, steel adzes have replaced the traditional stone bladed adze, yet they still remain the symbol of the canoe builder as well as a vestige of the ancient stone bladed adzes. Their cutting edges are perpendicular to the haft, which distinguishes them from axes, whose edges lie in the same plane as the haft. These tools are relics of a past where woodworking was a part of daily life. Nowadays carpenters have problems replacing their steel bladed adzes because they can no longer be found in hardware shops in Tahiti. The carpenter’s square serves to trace right angles, but more usually as a rule during canoe construction. It is in steel with two legs, unequal in both length and thickness. The long end measures two feet (60 cm). For someone not looking for it the tensioning lever passes unnoticed. It is a simple forked stick, 60 to 130 cm long depending on its use, cut in the vicinity of the work site and stripped of bark with a machete. Short ones are used on the inside of the hull and long ones to tighten the lashings on the exterior. Finally, a special mention should be made for the strips of beach hibiscus bark that serve equally as measuring cords, hauling ropes and as temporary lacings for mounting and adjusting the hull pieces.

estimated durability of 15 to 17 years and finally albizia (Albizia falcataria) which does not exceed 8 years. The red-bead tree is a rare species and therefore seldom used. Breadfruit and mango although of lower quality are more abundant and are the most commonly used. They are used for all parts of the hull except for the gunwale which is usually beach hibiscus. The float is always hibiscus and also usually the forward crossbeam. The rear crossbeam is a branch or young trunk of ironwood. Albizia which has only recently been used in canoes is placed last due to its weakness and vulnerability to insects (Tables 3, 4).

Figure 5 Tensioning crossbeam lashing with a forked lever.

Measurements The carpenter’s main measuring tool is very simple, it is his own body. The thickness of the bottom is two inches, two digits of the carpenters thumb. The outrigger float is one fathom, an arm span, away from the hull. The carpenter lies on the crossbeam to get an accurate measure. The only instrument is the carpenter’s square used for laying out multiples of two feet (ca. 60 cm). Foreign systems of weight and length have been adopted since 1848, with the arrival of the first Europeans (Henry 2001: 333).

Building a canoe today Community work groups are the rule on Raivavae; what Marshall called the “work gang” (Marshall 1961: 146). In his work, he used the Tahitian word pupu, team, group of people (Lemaitre 1973: 103). It was reported several times to me that these work groups are a means of transmission of knowledge from the elders to the youth. Mango trees may be felled in any season but the other species are cut in the austral autumn. Midyear and year end are not however very favourable for tree felling because of the holidays at those times. The choice of the tree is made by the owner depending either on the prospective use of the canoe or to benefit from a tree cut for some other reason. The upper planks and the

The woods used today Tenoo Flores, a Raivavae canoe builder ranks the species in the following order: the red-bead tree (Adenanthera pavonina) which has an estimated durability of 20 to 30 years, breadfruit, beach hibiscus, and mango with an

113

ROBERT VECCELLA

this operation is carried out with a chainsaw. Contrary to the usage elsewhere, on Raivavae fire is not used to aid hollowing canoe hulls. The spot where the tree is to fall is chosen to aid its removal or the work to be carried out at the site. Once on the ground, the branches are removed. The log is crosscut into a billet the length of the bottom or of one of the pieces of the bottom. The remainder of the trunk and the branches that are large enough are used for the other pieces. The billet is chosen from the best part of the log, from the stump to the first branches. It should be healthy wood without flaws. A specific curvature is not sought to facilitate forming the sheer. If the site is not adapted to the work of squaring the log, the rough billets are transported to a favourable location.

crossbeams are usually carved out of species other than mango. The width of the canoe is limited by the diameter of the trunk and visa-versa. As an example, an 80 to 90 cm diameter tree gives a canoe 60 cm wide inside. One 60-70 cm in diameter yields a canoe only 40 cm wide. In the former case the canoe would be in the over 9 meter long category. In the latter it will be from 6-9 m long. All the work, indeed all daily activities are punctuated by blessings. Felling the tree does not escape this immutable rule. The tree is given over to the lord’s grace prior to felling as well as at the beginning of each day’s labour. Before the arrival of the Europeans, felling was carried out with a stone bladed adze. Since the 18th century, the steel axe has rapidly replaced the traditional tool. Today

Botanical name Hibiscus tiliaceus Linné Coffea arabica Linné Cocos nucifera Linné Calophyllum inophyllum Linné Barringtonia asiatica (Linné) Kurz Inocarpus fagifer (Parkinson) Fosberg Mangifera indica Linné Artocarpus altilis (Parkinson) Fosberg Casuarina equisetifolia Linné Thespesia populnea (Linné) Solander ex Correa Paraserianthes falcataria (Linné) I. Nielsen Synonymous Albizia falcataria (Linné) Fosberg Adenanthera pavonina Linné Melia azedarach Linné Pseudotsuga menziesii (Lindley) Carrière

Common name Hibiscus Coffee Coconut Alexandrian laurel Fish-poison tree Tahitian chesnut Mango Breadfruit Ironwood Pacific rosewood Albizia

Vernacular name Pūrau Taofe Tumu Ha'ari Tāmanu Hotu Māpē Tumu Vī 'Uru 'Aito Miro (Falcata)

Red-bead tree

Paina Tīra

Douglas fir*

Table 3: Species of wood used for the construction of traditional canoes today in Raivavae. *This species does not grow on Raivavae, it is imported from North America. The billet is turned over and over so that the carpenter can choose the best profile. Then it is wedged in place and the four sides are squared with a chainsaw. Once the sides of the billet have been dressed, the carpenter lays out the centre line. The exterior contour is traced by eye in a succession of segments to obtain the desired profile. One curve is drawn ad hoc on one side of the axis and another one, symmetrical to it, is laid out by means of perpendiculars measured4 from the axis. As a rule, lengths are not measured but are reproduced by means of a strip of beach hibiscus bark. The pieces are measured in arm spans (fathoms). These are subdivided, for example for the emplacement of thwarts and (later) the stitches, with the carpenter’s square. Once the contour has been traced, the carving is begun to obtain the desired form. This work is principally carried out with a large axe. The carving of the exterior is done rapidly with much precision. The carpenter uses an adze to dub out the flat bottom, the sides and the turn of the bilge. The dexterity with which he uses this tool requires little sanding to 4

finish the surfaces. Except for the bottom pieces, the others pose no problem in manipulation and they are frequently turned over while being worked. At this point the exterior of the bottom pieces need to be sufficiently finished so as to require no further attention. This is even more important when the bottom is in one piece. Once these pieces are carved they are turned over once and for all. The carpenter lays out the interior contours on the upper surface of the baulk with a marking gauge leaving a margin of the desired thickness. The bottom is usually left two inches (5 cm) thick and the sides one inch (2.5 cm). The hollowing of the hull is begun by making transverse and longitudinal cuts with a chainsaw. Once crosscut, the waste is removed by knocking out the large chunks. This is repeated level by level until the desired depth is reached. The end of this phase marks an important step in the construction of the canoe because the weight has been reduced enough to allow it to be more easily transported5. 5

Approximately 85% of the timber in the billet is removed during carving of the hull. The percentage of waste would be even higher if account was taken of the entire tree.

Usually by eye, although a carpenter’s square may be used.

114

CONSTRUCTION OF DUGOUT AND SEWN PLANK CANOES ON RAIVAVAE, AUSTRAL ISLANDS

Breadfruit Ironwood Alexandrian laurel Fish-poison tree Tahitian chesnut* Mango Coffee Coconut Beach hibiscus Pacific rosewood Albizia Red-bead tree Tīra Douglas fur Steel rebar Galvanized pipe Nylon Corrugated iron

(**)

Legend: Alternative (*) not used for hulls Substitute Usual

(**) preferred timber for gunwales. In certain cases, the entire canoe may be made from one species: red-bead tree, mango, red-bead tree, albizia, Pacific rosewood, etc.

Preferred Table 4: Materials used in Raivavae canoes today.

115

Lashing lever

Aft stanchion Forward stanchion Rear crossbeam Forward crossbeam Motor bracket Sun protection Lashings

Dowel

Thwart

Bow

Material

Stern

Use

the hull with an overhand knot and with a clove hitch to a crosspiece wedged between the sides. In certain cases the rough hull is transported by water, towed by another canoe.

Gunwale stakes Outrigger float Bottom

It may now be taken to the site where it will be finished and assembled. This does not always happen immediately. The roughed out bottom piece is hauled by means of a rope of beach hibiscus bark tied to the point of

ROBERT VECCELLA

d) Fitting and temporary lashing of the gunwale timbers. e) Fitting and temporary lashing of the stern piece. f) The canoe is temporarily assembled. g) Permanent lashing of the hull pieces. h) Permanent lashing of the bow piece. i) Permanent lashing of the gunwales. j) Permanent lashing of the stern piece.

a)

b)

After the joints between the pieces are rectified, assembly begins with the hull bottom. The site had been first cleared and timbers placed to obtain a flat work surface. In the case where the hull bottom is in two pieces, the two finished parts are placed in alignment. Wooden pins are driven in place on the edge of one piece and corresponding holes are made in the edge of the other. The luting is installed to assure water-tightness. The pieces had already been pierced to receive the stitches. The same holes are used for both the temporary and final sewing. Once the pieces are in place the temporary lacing is carried out with strips of bark in the same way that one laces ones shoes but from both sides. These lacings are tightened firmly by hand before being tied off. Finally once the joints are well pulled to, wedges are driven firmly in between the lacings and the hull. The stitches are left proud on the surface, they are never let into the wood. In the past, the construction of a canoe took twenty people for the operation of clamping the pieces together. Today only six men are needed for the heavy tasks on a new canoe and just two men for regular maintenance. Ten strong fellows will be required for the launching. Today tensioning the temporary lacings with wooden wedges has simplified the operation and reduced the number of participants. Water-tightness is assured by the use of a combination of two materials: asphalt and strips of woven sackcloth have replaced breadfruit sap and the spathes surrounding coconut inflorescences used in the past. These materials are not caulking, driven into the seam after assemblage, but a system of luting put in place at each step of the assembly. The luting is made of black asphalt or coal tar generally used to waterproof joints in tin roofing. A block of tar is heated in a saucepan to make it malleable. As needed, the softened material is rolled into strips or cords. After cleaning the surface, a cord of coal tar is applied to the joint. It is covered with a band of sacking the width of the joint. Another cord of tar is placed on top. The second section of the hull is immediately put in place compressing the material and squeezing it over the surfaces of the joint. These materials are already becoming out of date and are being replaced by silicone caulk and epoxy glues.

c)

d)

e)

f)

g)

h)

i) j)

The cordage used in sewing the canoes is made from coconut husks. Mature nuts are used because they give fibres that are both flexible and strong. The nuts should as well be long and narrow to give the best fibres. The husks are removed and soaked in sea water for several weeks. The next step is to isolate the fibres by beating out the corky interstitial material. The bundles of fibres are then

Figure 6: Stages of assembly a) Principal of assembly. b) Fitting and temporary lashing of the dugout hull pieces. c) Fitting and temporary lashing of the bow piece.

116

CONSTRUCTION OF DUGOUT AND SEWN PLANK CANOES ON RAIVAVAE, AUSTRAL ISLANDS

rinsed and dried. This preparation renders the fibres nearly rot proof. The length of the bundles depends on the length of the nuts and may reach 30 centimetres. The cordage is made in two steps. First the fibres are rolled on the thigh to make single ply strands of some 20 fibres. Next three of these are plaited into a three-part braid or sennit. The person braiding the sennit keeps it under tension by pulling it with his toes during the plaiting. When one strand runs out another is twisted onto it to continue the braid. There are only a few aged people who are still able to make this coconut fibre sennit. The required quantity of sennit for each lashing is estimated in fathoms and then cut off. The quantity of sennit required varies depending on the thickness of the braid. It takes about 100 meters of sennit to build a 28 foot (9 m) canoe. Each vertical double stitch takes about 3 feet (1 m) of sennit. A vertical joint takes not less than 8 feet (2.5 m) and lashing the forward crosspiece to the hull and the outrigger float requires 33 feet (11 m). Synthetic cordage is coming more and more to replace the traditional fibres. The preparation and braiding of the sennit is long and fastidious and results in cordage less strong than synthetic fibres.

between the legs of a forked stick and wrapped several times around the handle being held in the carpenter’s hands. The sennit is hammered upon while being tensioned to insure its tightness. The lever is pulled against the hull to tension the cord during the hammering. Once the desired tension has been reached, a wooden pin is driven into the hole to keep the sennit from slipping. The lever is freed and the tail of the cord is passed through to the other side. This new half turn is tensioned and hammered as was the precedent. Another wooden pin is placed, allowing the lever to be freed. The first pin is removed while the second rests in place permitting the passage of another half turn and completing the second full turn. The cord is tensioned and hammered, once again blocked by a pin while the other one is removed. The sennit is again passed to the exterior and so on. Once three full turns have been made, the cord is passed diagonally from bottom to top to the next set of holes and the same operations are repeated for three turns. At the end of this operation, the cord is simply passed under itself, pulled tight and cut off. The same procedure is carried out for each (double) stitch.

Figure 8: Vertical joint using synthetic cordage and epoxy glue. After the different pieces of the canoe have been individually finished, they are carefully fitted together, one against the other. The joining surfaces are trimmed to give perfectly smooth joints. The dowels are fitted and the stitching holes are made. Assembly begins with the two pieces of the hull bottom. This is the most difficult joint due to their weight. The pieces are doweled, the luting is put in place and the two pieces are driven together. Blocking under the hull is of primary importance to assure the stability of both pieces during assembly. Temporary stitching of the two vertical sides is then carried out. Now the upper portions of the hull can be put in place and temporarily sewn to the bottom. There does not seem to be a strict order as to whether the bow or stern piece are installed first, but once it is fitted the

Figure 7: Temporary stitching and wooden wedges Sewing The sewing is carried out by at least two people. The carpenter supplies the tension and the apprentice threads, hammers and binds the sennit following the directions of the master. The sennit is passed through the lacing holes from the interior toward the exterior. A turn or loop is made between two adjacent holes on either side of a joint. The first turn overlaps the end of the sennit to bind it on the exterior. On the interior side the sennit is passed

117

ROBERT VECCELLA

gunwale strakes are then added and temporarily sewn to it and to the bottom pieces. Vertical dowels are fitted between the upper and lower portions. No horizontal dowels are used except in the joint between the two bottom pieces. The last of the stitches in the vertical joint between the two bottom pieces is made into the gunwale strake. The last of the upper parts is installed, doweled vertically and temporarily sewn to the rest. The canoe has now been completely assembled and the joints are waterproof. The only thing that remains is to replace the temporary bark stitching with the permanent sennit.

be a straight stick with a rectangular cross section. It is lashed to the hull with a diagonal or square lashing. The rear crosspiece is cut from a branch or the trunk of a young ironwood tree. It is stripped of bark and the secondary branches are removed but otherwise it is left rough. It is usually cut with a machete and the tool marks remain visible. It’s upper curve is arched like a bow with a short counter curve where it is attached to the hull. Like the forward one, it is lashed to the hull with a seizing. Pieces of rubber inner tube are placed between the crosspieces and the gunwale. The forward crosspiece is attached to the float by two stanchions or 8 mm diameter steel reinforcing rods or the steel stirrups used on electric poles. Whatever system is used, the bars are driven into the float and the crosspiece is tied on above. This lashing is made of strips of rubber inner tubes. Some of these timbers can last for years. Old rear crosspieces may be reused during reparation or rebuilding but the forward crosspiece is replaced. The float is made of the providential beach hibiscus. The diameter and length vary depending on the size of the canoe. They are from 12 to 20 cm. in diameter. Their length is the distance from the prow to the rear crosspiece. The float is generally a fathom (1.8 m) from the hull. It is positioned empirically, and the side it is placed on is a function of the natural lean of the hull although the left side is traditionally preferred. The first sea trials of the canoe allow it to be balanced by varying the distance between the hull and the float. With time the shape and characteristics of the float may change forcing the owner to make new adjustments. The forward crosspiece and the float are linked by a U shaped metal support and two rope stays. The U is of galvanised steel pipe, 20 mm in diameter, bent twice at 90° and driven into the float.

The horizontal and vertical joints are simple butts or carvel joints, to use the marine terminology. Other kinds of joints such as half laps or scarfs may be used. Mechanically there seems to be no obligation to put stitches on the flat bottom. The risk that these two pieces will open up is minimised by the vertical stitching between them and the horizontal stitching to the upper strakes. The two parts of the bottom do however have a tendency to open up. Shear stresses between them are taken up by the horizontal dowels while the vertical sewing on each side resists rocking forces. This arrangement is completed by the upper works which tie the whole assembly together. The gunwale is the piece under the greatest stress, working in both tension and compression. The lashings between the different pieces are flexible rather than rigid allowing the canoes to flex depending on their load and external forces. The tar luting is a flexible glue that responds to movement. In contrast, the newer materials used to glue the joints and coat the planks are rigid and their use risks causing new kinds of problems which should begin to appear in a few years.

The motor support is a simple crossbar made of a piece of timber mounted on edge extending outboard on one side. It is braced by two other thick pieces of wood to resist the heavy asymmetrical forces Storage and maintenance The launching of a new canoe is no longer an occasion for a feast but there is a Christian benediction. Numerous people assist in the launching. A 7 to 8 metre long canoe requires about 10 people to carry it from its construction site to the closest beach, usually only dozen of meters away. When not in use, the canoe is taken out of the water. Always at night, if the landing site allows it, almost without exception on Sundays and certainly if it will not be used for longer periods. If it has a motor it will be removed. The storage site is preferably shaded. If not then coconut leaves are laid over the canoe to protect it from the sun. Except for the case of very small canoes or those without outriggers, the hull is never turned over or tipped on its side. The canoe is brought ashore by pushing it up the beach, logs keeping it above the ground. The mooring line tied to the ‘nose’ (ihu) is passed around the overhanging branches and tied to the rear crossbeam.

Figure 9: View of a canoe stern: note the rear crossbeam and bracket allowing motor to be mounted on either side. Once the different parts of the hull are assembled, the next step is mounting the outrigger. The carpenter cuts the forward crosspiece out of a milled plank or hews a square timber, generally beach hibiscus. The crosspiece is usually carved with several curves and counter curves, which add to its rigidity. On small canoes it might simply

118

CONSTRUCTION OF DUGOUT AND SEWN PLANK CANOES ON RAIVAVAE, AUSTRAL ISLANDS

Where the beach is reduced to a narrow strip of sand or absent due to the configuration of the terrain, an ingenious system of supports and slides is used. The owners regularly maintain their boats. Traces of repairs such as pieces of planks sewn on are visible on some of them especially on the gunwales which are made of light wood that must be regularly replaced. Nowadays epoxy glue is usually used for repairs. During major refits, after years of service and normal maintenance, these modern multi-component glues are used to replace the coconut fibre lashing and the tar between the various parts of the canoe. Part of the island benefits from a concrete coastal road and the rest by a crushed coral road so that trips between villages are now made by land. Before these improvements the canoe was the principal means of transport. Today canoes are giving way to automobiles. Only fishing on the lagoon, travel to the islets and the transport of materials is still done by canoe

limited to inventory surveys and the collection of stone artefacts. No radiocarbon dates have been made, and the hypothetical date of settlement is an extrapolation from theoretical models of the migrations of the first Polynesians. Prior to the arrival of Europeans canoes were the sole means of communication between the islands. However the descriptions they left us require careful reading. Once Raivavae accepted Christianity it closed in on itself. The Tahitian language Bible became the centre of daily life and in consequence Tahitian has come to replace the local language throughout most of French Polynesia although Raivavae has been able to maintain its traditions and language.

Conclusion Canoe construction is learned only through on the job apprenticeship. There are no drawings, or builder’s models. Rather each builder carries a ‘theoretical’ model in his head as a reference. The work is carried out empirically, the builder working primarily ‘by eye’ to lay out the lines and trim the shape with the adze. The construction sequence presented here is for a canoe whose bottom is made of two (nearly identical) pieces. However, canoes are made just as often with a one piece bottom. The size, length and construction of the latter are otherwise identical. There is only one form or style of canoe on Raivavae, but two modes of construction. It may be that the two-piece bottom with a vertical joint is used to give greater flexibility to the structure. It is certainly not a reflection of the availability of timbers long enough for one piece bottoms, nor of the difficulties of transport of such timbers. Storage is given special care by the owners who adapt the supports to the environment because the longevity of the canoe depends primarily on the care given during periods of non-use. There are three complementary modes of propulsion: punting, paddling and since the early 1960’s outboard motors although the latter are still uncommon. Although sailing is unknown on Raivavae today, the results of efforts to sail Raivavae canoes on Moorea may presage a new future, and who knows, perhaps the return of sails to the lagoon. Except on Rimatara where small canoes are still built collectively, the rest of the Australs have lost their traditional canoes. Unsophisticated canoes are still built on Tubuai but the skill of the builders does not rival that of Raivavae, although paradoxically barely 20 years ago, dozens of sailing canoes furrowed Tubuai’s lagoon while on Raivavae as early as the 1920’s no living islander had seen a sailing canoe. Figure 10: Paddling a canoe. Drawing © Alexis Nguyen-Thé. Punting a canoe. Drawing © Alexis Nguyen-Thé. Canoe under power. Drawing © Alexis Nguyen-Thé.

The Raivavae canoes are vestiges of a nearly forgotten maritime heritage. The prehistory of the island is still unknown. Archaeological investigations have been

119

ROBERT VECCELLA In the Tuamotus, canoes were swept aside in the 19th century by the trading schooners that criss-crossed the Pacific stopping at the smallest islands. This disappearance was also due to the shortage of timber on the atolls. On Raivavae their voluntary isolation led to the loss of sailing but the generous forest and protected lagoon fostered the continuation of canoes. They are much more than simple working craft. They are part of community life and contribute to group cohesion such as other communal activities (taro cultivation, sports, etc.). Amédée told us that on attaining 16 years, he received as a gift his canoe. Not a model but a real canoe to go fishing in. This gift seems to have been a ritual to mark the passage of a boy into adulthood. Another custom was that in the past carpenters were priests, who guarded knowledge. Their training was long and initiatory. Today tāmuta are members of a lineage, those who guard this ancient knowledge. An individual may only enter therein by birth or marriage. Today modernity is catching up with the Raivavae, but I wager that this will not cause the last of these ancient outrigger canoes to disappear, and with them the knowledge of their construction, which is already competing successfully against modern aluminium and plywood boats.

published in 1848. Publications de la Société des Océanistes 1, Musée de l’Homme, Paris. Hornell, J. 1936 “Austral Islands and Rapa”. In A.C. Haddon and J. Hornell, Canoes of Oceania, vol. 1, “The canoes of Polynesia, Fiji and Micronesia”. Special Publications 27, 28 and 29, pp. 146-156. Reprint in 1975. Bishop Museum Press, Honolulu. Jacquier, H. 1955 “Notes sur les pirogues actuelles de l'île de Raivavae”. Bulletin de la Société des Etudes Océaniennes 9 (113): 494-496. Lemaitre, Y. 1973 Lexique du Tahitien contemporain. ORSTOM, Paris. Marshall, D. 1961 Ra'ivavae, an expedition to the most fascinating and mysterious island in Polynesia. Library of Congress catalog card number 60-13545, Doubleday Company Inc., Garden City, New York. Simutoga, P. C., 1992 Technologie traditionnelle à Wallis, Essai de sauvegarde de la mémoire collective des charpentiers wallisiens (tufuga) du district de Hihifo. Publications de la Société des Océanistes 44, Musée de l'Homme, Paris. Stimson, J. F. 1928-1931 Preliminary report of the Cornelius Crane expedition of the Bishop Museum to the Austral Islands, Associate in Polynesian Linguistics. Bernice P. Bishop Museum of Honolulu (3rd Version) and Special Editor Webster's New International Dictionary (Second Edition) for Polynesia, Melanesia and Micronesia, Microfilm catalog 121, Reel #5, 1970, Peabody Essex Museum, Salem Massachusetts. Stokes, J. F. G., [n.d.] 1929-1930 Notes manuscrites on Raivavae, 141 p. Bishop Museum, Honolulu, Hawaii.

References Académie tahitienne 1999 Dictionnaire Tahitien-Français, Fa'atoro Parau Tahiti-Farani, Fare Vana'a. Papeete. Branagan, C. (ed.) 2001 Science in a Sea of Commerce. The Journal of a South Seas Trading Venture (1825-27) by Samuel Stutchbury. University of Sydney, Darlington. Brunor, M. A. 1969 Arts and Crafts Of The Austral Islands. A Special Exhibition, 17 December 1968 to 30 April 1969. Peabody Museum of Salem, Massachusetts. Printed by The Anthoensen Press, Portland, Maine. Corney, B. G. 1913-1919 The Quest and Occupation of Tahiti by Emissaries of Spain during the Years 17721776. translated into English and compiled for Hakluyt Society ser. 2, vols. 32 (1913), 36 (1915), 43 (1919), London. Edwards, E. 1993 The archaeological survey of the island of Raivavae Austral Island, French Polynesia. Inventaire Archéologique de Polynésie française, Easter-Chili, Département Archéologie. Centre Polynésie des Sciences Humaines. Unpublished. Henry, T. 2000 Tahiti aux temps anciens. Translated from English by Bertrand Jaunez. Originally

120

is characterized by a distinctive style of stamped pottery and is called the Lapita culture. Beginning about 3,200 cal BP, the Lapita peoples rapidly expanded into Remote Oceania, establishing settlements in the proximal Santa Cruz Islands and to the south in Vanuatu, New Caledonia and the Loyalty Islands (Sand 2000). Another branch of the expansion moved eastward, colonizing the islands of Fiji and Tonga and Samoa, in West Polynesia. Most major archipelagos in both branches of the expansion likely had initial settlements by 2,750 cal BP at the latest (Sand 2000; Sand et al. 2002).

Chapter VII Simulating Island discovery during the Lapita expansion by Chris Avis*, Álvaro Montenegro* and Andrew Weaver* The Colonization of Oceania

The archaeological record of Pacific settlement implies that, following the Lapita expansion, there was a pause in eastward expansion lasting for between 500-1,700years (Irwin 1998; Anderson 2003; Spriggs and Anderson 1993), with recent evidence supporting a longer pause (Hunt and Lipo 2006). During this interval, there was movement into central-eastern Micronesia (likely from the Solomon and Vanuatu groups) and into Niue, Rotuma and Pukapuka in the Northern Cooks (Intoh 1997; Anderson 2003), although the early dates for presence in Pukapuka, and hence its colonization during the pause, have been questioned (Di Piazza 2005).

In the present view, Oceanic settlement occurred as a result of a series of waves of rapid expansion interspersed by long pauses. Humans first arrived in Australia, Tasmania and New Guinea between 60,000 - 40,000 calendar years before present (cal BP, Stringer 2000), when sea levels were lower and these three landmasses were connected, forming a continent called Sahul. Sahul was never directly connected to South-East Asia during the Pleistocene, indicating that the initial settlers of this region must have had at least rudimentary maritime technology. The islands of the Bismarck Archipelago, lying to the north-east of New Guinea and separated from Sahul by a narrow deep water trench, were reached by at least 35,000 cal BP (Kirch 1997, 2000; Spriggs 1997). The Solomon Islands, extending to the south-east of the Bismarcks, were also reached from Sahul or the Bismarcks during the Pleistocene. The remainder of Oceania, including the Santa Cruz and Vanuatu island groups, was not settled until much later, starting at about 3,200 cal BP.

Navigation and Maritime Technology Prehistoric remains of Oceanic boats are extremely scarce and those that have been found are very fragmentary (McGrail 2001). European explorers who visited the region beginning in the 16th Century observed the use of log rafts, single-hulled canoes (with and without outriggers) and double-hulled canoes. Of these, the double-hulled canoes seem best suited for lengthy sea voyages as they could be quite large, more than capable of carrying enough cargo to survive a long sea voyage or to establish a founding settlement on a new island (McGrail 2001; Sharp 1963). However, Anderson et al. (2006) argue on the basis of linguistic evidence that double hulled canoes were developed only after the colonization of western Polynesia had been accomplished and that early stages of colonization were accomplished using outrigger canoes and rafts.

It is unclear why the 350 km crossing from the southern Solomons east into the Santa Cruz Islands presented such a formidable barrier to the early explorers and settlers of this region. One possible explanation is that early explorers could have reached from mainland South-East Asia to the Solomon Islands by means of a series of intervisible water crossings wherein the destination island became visible before the departure island had sunk below the horizon (Irwin 1992). Thus, the main island groups of Near Oceania could have been discovered as a result of cautious exploration by early seafarers who always remained within sight of land. In contrast, the islands of the Santa Cruz and Vanuatu groups were never connected to the Solomons via intervisible crossings, even at times of greatly reduced sea level.

The navigational skills of Pacific peoples have been extensively debated and there seems to be little doubt that repeated contacts occurred between widely separated island groups in Melanesia and Polynesia at some point in prehistory, within so-called interaction spheres. Europeans at the time of contact reported that mariners knew directions to distant islands and in some cases had sophisticated traditional navigation methods based on celestial observations. Further, x-ray fluorescence analysis has demonstrated the transfer of basaltic adze material over thousands of kilometers (Weisler 1997, 1998), helping to quantify the extent of interaction spheres. Presumably settlement of newly discovered islands would also have required some form of navigation

At about 3,500 cal BP, the archaeological record of Near Oceania reveals the appearance of a new culture occupying coastal areas of the region (Hurles et al. 2003; Kirch 2000) with the earliest sites found in the islands of the Bismarck Archipelago. This new, likely alien culture *

School of Earth and Ocean Sciences, University of Victoria, Victoria BC, Canada.

121

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

or at least directed sailing for colonists to arrive at their destination. But all of this behaviour would have occurred after islands had been located and says nothing about the means by which they were first discovered.

navigators who exploited sustained periods of winds from other directions (Finney et al. 1989). Anderson et al. (2006) have argued that maritime technology at the time of colonization was not sufficiently advanced to allow crews to make any significant progress against the winds, supporting the latter suggestion. Voyages down the direction of the dominant winds would be considerably riskier since the return trip home could not be assured. Part of the appeal of this theory is that the first waves of Oceanic expansion, through Melanesia and from western to central-eastern Polynesia, were eastward and against the easterly trade winds. Later, when technology and sailing abilities improved, explorers would sail across and then down the prevailing winds on their outbound voyages, allowing them to settle the most distant islands of Oceania.

At present, the favored view of Pacific exploration (best stated by Irwin 1992) is that Pacific settlement occurred as a result of a systematic exploration of the Pacific by skilled voyagers. These mariners might have refined their voyaging skills in the safe ‘voyaging nursery’ of the intervisible islands of Near Oceania where there are predictable seasonal reversals of wind and current. Voyages of exploration into Remote Oceania would then initially move against the direction of the prevailing winds so that explorers could rely on a rapid return voyage home by sailing downwind. Outbound voyages could be made by tacking against the winds, or by astute

Figure 1: The islands of Near Oceania and the Lapita expansion. Islands are arranged according to conventional geographic groupings. The coastlines are identified from the GLOBE 30” DEM dataset.

122

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

In an earlier view, best advanced by Sharp (1963) Oceanic explorers lacked the navigational knowledge to complete two-way voyages of exploration. A vessel’s course could have been crudely estimated by observing its motion relative to the wind. This ‘dead-reckoning’ technique could not account for the effects of ocean currents and lateral drift, which might introduce significant errors. Latitude could have been gauged from the stars, but no measurement of longitude was possible and, lacking this, voyagers would have no means of accounting for longitudinal errors in course. With a sufficient understanding of regional waters, the reckoning system might be revised to counter such errors, but such knowledge would not be available to those who dared explore uncharted waters, making accurate navigation home impossible. For these reasons, Sharp (1963) argues that one-way voyages, either intentional or accidental, must have played an important role in the settlement of Oceania.

Scope of Work Here we use a computer model to test two aspects of these different hypotheses: discovery of new islands by drifting vessels and by directly downwind sailing. The former is one method by which the one-way voyaging proposed by Sharp (1963) may have occurred. Downwind sailing is a simple model of deliberate exploration in which crews intentionally orient their boats to move with the wind for ease of sailing. Such crossings can be either with or against the prevailing winds depending on the orientation of the winds experienced locally by the boat. We focus our study on crossings associated with the Lapita expansion (Fig. 1), paying particular attention to how seasonal and inter annual environmental variability could have influenced the crossings from Near Oceania to islands further to the east.

Figure 2: Top Left: Average surface winds in the area of interest during the Austral summer. The data come from montly long-term means from the NCEP/NCAR reanalysis project (Kistler et al. 2001). A 10 m/s vector is included for reference. Bottom left: Average winter surface winds. Top Right: Austral summer surface currents. The data come from MITgcm ECCO experiment (Stammer et al. 2002). A 50 cm/s vector is included for reference Bottom Right: Winter surface currents.

123

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

more intermittently east of Santa Cruz. To the east of Fiji they become much more episodic and occur in shorter bursts. For example, Finney et al. (1989) cite historic records of monsoon winds which extended as far east as the Societies and Southern Cooks and state that, on rare occasions, such winds may even reach as far east as the Marquesas. Thus, while the mean winter wind field provides an indication of the region where there are reliable monsoon winds, the actual eastward extent of such winds in a particular year may be much greater.

Climate of the Study Area Climate at the Time of Colonization Surface wind and current data which accurately reflect conditions at the time of initial colonization are needed to drive the simulations devised in this work. Reconstructions of paleoclimate conditions in the southern Pacific (Markgraf et al. 1992) indicate that the atmospheric circulation across the Pacific was essentially the same as in the present day climate by 3,000 cal BP. Since the Lapita expansion began just prior to this time, we make the assumption that the present day climate is a reasonable approximation to that at the time of colonization.

A larger-scale phenomenon which can also cause extended periods of anomalous westerlies within Melanesia and Polynesia is El Niño / the Southern Oscillation (ENSO), the dominant source of inter annual variability in the equatorial Pacific. In the mean state, the easterly trade winds make up the bottom arm of the zonal Walker circulation cell which spans the tropical Pacific. Accumulation of warm surface water in the western Pacific by these winds warms the air and results in a region of rising air and low surface air pressure; in the east, cold, upwelling water from depth cools the surface air and leads to a region of subsidence and high surface air pressure.

Surface Winds The surface winds in the study area show distinct annual fluctuations. During the Austral winter, the wind field is dominated by the strong south-east trade winds (Fig. 2). The circulation is more complicated during the Austral summer (Fig. 2). During this time, the north-east trades strengthen and the south-east trades weaken. In addition, the surface winds reveal the influence of the Australian monsoon circulation associated with a low pressure centre which extends from northern Australia eastward into Melanesia as a result of heating of southern hemisphere land masses during the summer months. This complicated circulation results in a band of northwesterly surface winds which generally extends into the Lapita region from the south-east coast of New Guinea through the Santa Cruz islands and partway to Fiji.

Westerly winds aloft form the remaining branch of the circulation. The zonal sea level pressure (SLP) gradient reinforces the easterly trades, demonstrating the coupling between atmosphere and ocean that is a key aspect of this circulation. El Niño events are identified by a strong and sustained increase in the sea surface temperature (SST) of the eastern equatorial Pacific. This influences the surface air temperature across the basin and results in a reduction of the zonal SLP gradient, in turn reducing the strength of the easterly trade winds. During very strong El Niño events, the surface winds of the western-central Pacific may reverse as the pool of warm surface water sloshes eastward. La Niña episodes are associated with negative SST anomalies in the eastern Pacific resulting in an increase in the zonal SLP gradient and the strength of the Walker circulation and the trade winds.

Surface Currents Mean surface ocean currents are largely driven by windstress. In the study area, the dominant current is the broad, westward flowing South Equatorial Current (SEC). During the Austral winter, this current extends from the equator to about 15° S (Fig.2). In this season, the North Equatorial Counter Current (NECC), a narrow continuity current, flows eastward in a band from the equator to about 5° N. In Austral summer (Fig. 2), the NECC shifts to higher latitudes and another east-ward flowing current, the South Equatorial Counter Current (SECC), forms in response to the surface winds of the summer monsoon. The SECC typically forms in a narrow band from 5° S to 10° S and extends from the northern coast of New Guinea through the Solomon Islands to the northeast of Samoa.

In the present climate, ENSO is irregular, displaying variation in the amplitude and frequency of El Niño / La Niña episodes. Successive events of the same phase generally occur every 3-5 years, setting a time scale for ENSO-related climatic variability. Analysis of paleoproxies from regions of the world strongly affected by this phenomenon indicates enhanced environmental variability starting about 5,000 cal BP (Markgraf and Diaz 2000). This suggests that around this time, ENSO had begun to behave as it does in the present climate, although the frequency of occurrence of El Niño conditions seems to have varied since the mid-Holocene (Moy et al. 2002).

Interannual Variability Wind reversals associated with the Australian monsoon occur predictably within the western Lapita region, from New Guinea to the Santa Cruz region, where they dominate the summer wind field, and occur progressively

124

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

2002) so these are neglected in the simulation. Due to the asymmetry of floating objects, Vwinddrift is not parallel to the local wind velocity but deviates by a leeway angle, α. Experimental work on different types of drifting vessels has shown that vessels of the same class will tend to drift to the right or to the left of the surface wind with roughly equal probabilities (Hackett et al. 2006), and the exact direction is determined by small differences in the form of the vessel and the direction of the surface wind and waves (Fig. 3). We adopt Vwinddrift and α values empirically determined by the US Coast Guard (USCG 2002). Both values are kept fixed for all simulations and were based on the drift of a type of vessel which resembles a large canoe with a basic canopy, similar to the type of craft which is thought may have been in use at the time of Oceanic settlement (McGrail 2001). The direction of the deflection caused by the leeway angle (33 degrees either to the right or to the left of the wind) is randomly selected for each vessel every 24 hours, both directions having the same probability of occurrence.

Model Methodology Experiments were performed using a variant on a computer drift model developed by Montenegro et al. (2006). The model uses surface wind and current data derived from numerical models adjusted by observed weather data. It was originally designed to study prehistoric ocean crossings into the Americas and was subsequently used to investigate the introduction of the sweet potato into Polynesia (Montenegro et al. 2007). Wind and Current Data The surface wind data used in the simulation are daily mean winds that came from the National Center for Environmental Prediction / National Center for Atmospheric Research (NCEP/NCAR) reanalysis product. The NCEP/NCAR data have a horizontal resolution of approximately 1.9° (Lat.) x 1.875° (Long.) and were provided by the NOAA-CIRES Climate Diagnostics Center located in Boulder, Colorado. Ocean current data were taken from the Estimating the Circulation and Climate of the Ocean (ECCO) experiment, which, as with the winds, are a combination of modelled and measured data. The ECCO data utilized are 10-day velocity means and have variable horizontal resolution, ranging from 0.3° x 1.0° at the equator to 1° x1° at high latitudes with each point representing the motion of the upper 5 m of the water column. A total of 13 years of wind and current data were selected from January, 1993 through to December, 2005. A 3dimensional interpolation was performed on both data sets so that wind and current data had identical temporal and spatial resolution. The final data used in the simulation have a 1-day time step with the same horizontal resolution as the ECCO data. The selected weather data extend from 40.5° N to 59.5° S and from 139.5° E to 290.5° E, encompassing all of Oceania and including marginal areas. For the experiments the wind data were interpolated to the current spatial resolution.

Figure 3: The velocity components of a drifting vessel. The net velocity may be decomposed into wind and current induced components. The wind-drift is generally not parallel to the surface wind, but deviates by a leeway angle α. Vectors in the bottom left corner indicate the direction of the hypothetical surface wind and current vectors.

Modelling Drift The velocity of a drifting object, Vdrift, is given by: Vdrift = Vcurrent + Vwinddrift + Vwavedrift Vcurrent is the current drift velocity due to the surface current. Vwinddrift is the component of vessel drift caused by the direct effect of the wind on the vessel. The final term, Vwavedrift, accounts for the wave force on the vessel and the Stokes’ drift associated with the swell/wave currents. Studies have shown that wave forces are negligibly small for objects less than 50 m in length (Hackett et al. 2006) and that the Stokes’ drift is small compared with current-and wind-induced drift (USCG

Earlier Simulations The role of drift voyages in Pacific colonization was addressed in a pioneering and influential computer simulation by Levison et al. (1973), who used wind and current data which had been collected sporadically between 1855 to 1938 and 1854 to 1952, respectively, by merchant and naval ships traversing their study area. Monthly wind observations and quarterly current

125

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

readings were binned into a 5° by 5° grid, though the number of measurements per cell was highly variable across the study area, with cells lying along major shipping routes being much better sampled.

Experiment Design Two sets of experiments were conducted: one where vessel displacement is based on the USCG parameters (defined from now on as drifting) and a second where boat displacement is based on the parameters adopted by Levison et al. (1973) (which we define as directly downwind sailing). In both cases, boats start from the centre of selected cells from our weather data grid and move under constant wind and current conditions over a 1 day timestep. A one-day timestep was selected because, given the characteristic winds and currents in the study area, vessels would generally cross no more than one or two grid cells per day.

Lacking a sequential time series of data, Levison et al. (1973) adopted a stochastic approach to determine the weather conditions experienced by their drifters. One drawback of their simulation was that it did not capture the temporal autocorrelation of winds and current velocities1 nor the correlation between wind and current. The nature of Levison et al. (1973)’s data did not allow for a study of the influence of interannual variability on their drift voyages. Furthermore, the coarse spatial resolution of Levison et al. (1973)’s data did not capture small-scale spatial variability which might have affected the trajectories and lengths of their simulated voyages.

Drifting and directly downwind sailing boats were launched starting every 7 days from January 1993 to December 20042. This value was selected based on the computational efficiency of the simulation code in order to produce a set of results in a reasonable length of time. Based on sample drifts from a subset of the full set of launching sites, it was found that simulation results were not significantly different when a higher frequency of vessel launches was employed, thus validating the selected launch frequency. Voyages were allowed to last for up to 90 days, which was taken as a conservative estimate of the length of time that a vessel and crew could survive at sea and is significantly shorter than the length of the longest historical drift voyages with survivors (Montenegro et al. 2006). Vessels that did not encounter land before the maximum voyage length were identified as being ‘lost at sea’3.

The vessel movement parameters adopted by Levison et al. (1973) were based on historical accounts and on data from replica Polynesian canoes. In the context of Equation 1, Vwinddrift is in the same direction as the wind, neglecting any cross-wind drift, while Vcurrent is parallel to the surface current. Levison et al. (1973)’s Vwinddrift is generally between 0.15 and 0.3 of the surface wind speed, comparable to the performance of replica Polynesian canoes sailing with the wind (Finney 1977). A comparison of the wind-and current-induced drift response under the USCG and Levison movement schemes, demonstrates that wind induced speed of vessels under the Coast Guard scheme is much lower that the wind induced speed of craft moving under Levison’s scheme. Thus, while Levison et al. (1973) intended to model drift, the parameters they adopted are in fact more representative of a vessel sailing directly downwind.

Definition of Island Groups To identify island groups which would serve as launching and target sites, a high resolution map of land-masses in the Pacific Ocean was generated using the Global Land One-kilometer Base Elevation (GLOBE) Digital Elevation Map (DEM). The GLOBE model is an independently peer-reviewed DEM distributed by NOAA’s National Geophysical Data Center with a latitude-longitude grid spacing of 30”, which is roughly equivalent to a spacing of 1 kilometer at the equator (Hastings and Dunbar 1999). Coastlines were defined as points in the dataset having elevations between 0 and 50 m.

The Levison et al. (1973) model and data were used by Irwin et al. (1990) in simulations which compared the success of different intentional voyaging strategies in the exploration of the Pacific. More recently, Evans (1999) developed a voyaging program which attempted to assess the influence of simulated weather and vessel performance characteristics on a navigated voyage to a known destination in Polynesia. Both of these works model decision-making by vessel crews and are difficult to directly compare with our experiments so they will not be further addressed.

Experienced mariners are able to deduce the presence of land by looking for signs such as the presence of island clouds, the deflection of swell, seabirds, logs and floating vegetation. 30 nautical miles is often taken as a 1

Winds in the real world are not completely random: the direction that the wind blows on a particular day influences the following day’s wind direction probability distribution. Winds can blow from the same direction for several days and these spells would not be accurately simulated by stochastic weather generation schemes.

2

Experiments starting in the latter half of 2004 could last into the following year which is why data from 2005 were needed. 3 This does not imply that crews couldn’t survive beyond this time, it simply means the vessels hadn’t reached land by the end of the simulation.

126

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

conservative estimate of the distance from which Pacific seafarers could detect land, with tall islands visible from even further away (Dodd 1972; Frankel 1962). We adopted 30 nautical miles as a measure of the detection distance, or sighting radius, for the shoreline of islands and extended sighting circles having this radius around each shoreline point. Closely spaced groups of islands form extensive island screens as sighting circles from adjacent islands intersect, as shown by Fig. 4 in the case of the Fiji group. The resolution of the simulation is ultimately set by that of the weather data so that island groups were represented using cells from the weather grid. Grid cells whose centres lay within the sighting circles of an island were selected to represent an island or island group. The data spatial resolution is such that this procedure generates a good approximation of island sighting screens (Fig. 4, bottom). Figure 5 shows the islands and archipelagos of the Lapita region and nearby sites in Micronesia and Melanesia as represented by grid cells. The Bismarck, Solomon and Tonga groups all consist of one large ‘main’ island screen and several smaller, numbered screens representing outlying islands in the group. The centres of the selected grid boxes are the starting points for the simulated voyages and these cells also serve as target boxes for islands or island groups. Voyages end when a vessel’s trajectory intersects cells of an island group other than the departure group or if it intersects the departure group after 6 or more days at sea. We opted to invoke these criteria because it may take a few days for vessels to cross the larger island groups; during this time the boats would be within sight of land and not lost at sea or out exploring. It is assumed the crew of drifting vessels would attempt to make landfall upon sighting land after being lost. Likewise, explorers stop when they reach land, to investigate a new island or to return to populated areas after an unsuccessful exploration outing. We also defined as targets some areas of eastern New Guinea and eastern Australia. Voyages can also end if the boats move out of the study area, and to quantify the number of such vessels, this is represented in the simulation by an ‘out of bounds’ target box.

Figure 4: Top: Shorelines of the islands of the Fiji group (white lines) surrounded by 30 nautical miles sighting circles (black area). Bottom: Grey rectangles represent the model grid cells which, in the simulations, correspond to the Fiji group.

Contact Probabilities The parameter used to describe contact between island groups is the contact probability. To understand how this quantity is calculated consider the following example. Suppose one considers crossings from group A to group B where A consists of 10 grid cells and B consists of 5 grid cells. On a given launch day, 10 voyages start from A and are tracked by the simulation. Each of these 10 trajectories is then checked to see if they intersect any of the 5 target cells from group B. If a boat’s trajectory intersects a target cell, a ‘hit’ is registered and the boat is stopped. There can be a maximum of 10 hits in group B from this set of launches. The code then moves on to the

next launch day under consideration and adds the number of hits in group B for this launch day to the total number of hits and so forth. In general, the contact probability from A to B is then calculated as follows: ∗ 100% Total#of Hits toB (# of Launch Sites in A)(# of Launch Days) Crossing probabilities do not reflect the differing size of the launch groups and this is important to recognize when

127

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

considering the results. For example, crossings to a target from cells representing an individual, isolated island and a large archipelago might have the same contact probability, but over time one would expect a greater number of total launches (and thus successful crossings) from the larger group.

all 12 years) as well as mean seasonal probabilities and seasonal probabilities for launches which began in specific years. The seasons were defined based on conditions in the southern hemisphere as follows: summer - December, January, February; fall - March, April, May; winter - June, July, August and spring September, October, November. In addition to contact probabilities, voyages were also analyzed according to their minimum and mean crossing times.

Results will be presented in terms of overall probabilities (based on launches which occurred in all months and in

Figure 5: The islands of the Lapita region in terms of simulation grid cells. Islands which were settled during the Lapita expansion are shaded in grey; different shades of grey have been added to distinguish adjacent grid cells that are part of different island groups. White cells are Oceanic islands which are thought to have been settled after the Lapita expansion. Black grid cells represent Australia and New Guinea, which served as target sites but not launching sites. “Loyalty Group” includes New Caledonia. Results Drift and Sailing Results

so differences are almost entirely due to differing response to surface winds.

The number of possible crossings into the eastern Lapita region is much greater for the directly downwind sailing experiments than for the drifters. Directly downwind sailing crossings also tend to have higher probabilities and shorter mean crossing times than the drifting only results. Both the USCG and Levison parameterizations have a very similar response to surface currents (whose effects are generally much less important than the wind),

Figure 6 shows sample trajectories for the same launch day from Solomon-3 with sailing trajectories in grey and drifters in black. The mean step size for the sailing boats is considerably greater than for the drifters, indicating that these vessels move much further in an average 24hour timestep. The drift trajectories appear somewhat jagged as a result of the drifters changing their orientation

128

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION with respect to the driving wind4. Both types of vessels initially follow a track due east. North of Santa Cruz, the sailing ships move southeast, eventually reaching Fiji. The drifters veer more sharply to the south, arriving in Santa Cruz. The different tracks taken reflect the relative speeds of the vessels; by the time the drifters had reached the vicinity of the Santa Cruz group, the winds had a more southerly component than those experienced by the sailors.

Santa Cruz, to Fiji/Rotuma and to Tonga/Samoa. We aim to identify launch sites from which these groups can be reached and the conditions under which these crossings are probable. Results indicate that crossings from the west to Futuna and Wallis are also possible but these islands do not offer critical connections to other island groups and so are omitted to maintain focus on the larger groups. In this exercise, the possible crossing pathways for both drifting and downwind sailing proved to be sufficiently different that separate tables were warranted.

The differences in crossing parameters between both sets of simulations can be understood in terms of the relative ‘efficiency’ of the different movement schemes. This also explains why a significantly higher percentage of drift voyages end up back in the departure group. Drifters have a lower chance of clearing their home group on favorable winds before the wind changes direction, pushing them back. This is especially true of departures from large island groups such as Fiji or the Solomons since they present a much larger area to intercept returning drifts.

Crossings to Santa Cruz Crossings from the Bismarcks and the Solomons to Santa Cruz easily exceed the threshold crossing probability when considering the mean seasonal results for summer and fall both in drifting (Table 3) and directly downwind sailing (Table 4) schemes. Tikopia, Vanuatu and the Loyalty group may be easily reached starting from Santa Cruz (Table 2). Crossings to Fiji and Rotuma

Overall and Summer Results

Drift crossings to Fiji and Rotuma do not exceed the 5% probability cutoff  in the overall and mean seasonal results. However, drift crossing probabilities are significantly higher when considering drifts which occur in the summer months of specific years (Table 5). Conditions in summer 2002 and 2003 are particularly favorable for reaching this area from the Solomons and Santa Cruz, though mean crossing times are on the order of 60-70 days, significantly longer than the drifts to the Santa Cruz group from Near Oceania.

Only crossings which occurred with probabilities greater than 5% in either the drift or sailing schemes are presented and discussed. This is consistent with Levison et al. (1973) who considered this probability to be a reasonable threshold for crossings to be likely to occur. Considering only crossings which exceeded this threshold did not decrease the overall area that was reached in either movement scheme. However, certain key crossings exceed the probability threshold only in certain seasons or years, reflecting the influence of seasonal and interannual variability on the experiments.

In the downwind sailing experiments, Fiji can be reached from the western Lapita region under mean summer conditions with probabilities as high as about 9% (Table 6). The crossing probabilities in summer months vary from year to year and additional crossings are possible when considering specific years. For instance, Rotuma can be reached in the sailing scheme from Santa Cruz, Solomon-1, Solomon-2, Solomon-3 and Tikopia if we consider launches that occur in the summer months of particular years (Table 8). Crossing probabilities to Fiji are also significantly higher than the average summer values in specific years. For instance, the crossings from Santa Cruz and Tikopia occur with a probability of 28% and 25%, respectively, in summer 2002. In the case of both drift and sailing, Fiji can be easily reached from Rotuma in the summer months.

The overall results, which take into account voyages which began in all seasons, confirm the expected prevalence of westward crossings, but also indicate a large number of crossings to the east. The overall crossing parameters for launches from the major groups in the Lapita region are displayed in Table 1. The importance of seasonal variability is clear if only voyages that started during the Austral summer are considered (Table 2). During this season, the probability of occurrence and number of eastward crossings increase while the duration of these trips decreases. Crossings in the Lapita expansion We now focus on specific crossings that would allow the easternmost islands of the Lapita region to be reached from the inter-visible islands of Near Oceania. For this we focus our attention on three key eastward crossings: to 4

Recall the drift direction is randomly assigned to the right or left of the wind on a given day, always deviating by the constant leeway angle.

129

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

Figure 6: Sample trajectories for vessels using the USCG (black) and Levison (grey) parameters. Drifts start from the centre of the three grid cells making up the Solomon -3 launch area. The selected drifts come from the summer 1993 set of launches. USCG vessels all arrive in the Santa Cruz group after 22-24 days; Levison vessels all reach the Fiji group after 19-28 days. Crossings to Tonga and Samoa

could achieve these crossings only with contact probabilities of less than 1%.

Crossings to Tonga and Samoa from the west are not viable under the mean annual or mean seasonal conditions in either scheme and only occur for launches in specific summers. In the case of drift, these eastern targets can be reached only from Rotuma (Table 7), while directly downwind sailing experiments allow for crossings from the western Lapita sites as well as Fiji and Rotuma (Table 8). The overall contact probability table (Table 1) demonstrates that Tonga can be readily reached from Samoa, but the reverse crossing requires specific conditions.

Discussion In the simulations, the entire Lapita region can be reached by drift and sail from Near Oceania if a 5% probability of occurrence cut-off is adopted. A crossing probability of 5% implies that, on average, 1 out of 20 vessels launched will accomplish the crossing under the conditions for which the probability was generated. As anticipated, given the direction of mean winds and currents, westward crossings are most probable overall and have enhanced probabilities in the Austral winter and spring, when the trades are strongest and most consistent. The ease of such westward journeys would allow those who had found new land to easily drift or sail to populated areas in the west with news of their discovery. Whether they could return to their departure site would depend on the navigational skill of the voyagers.

Beyond the Lapita expansion Island groups east of the Lapita region were also modelled in our simulation. From Samoa, sailing and drifting vessels can reach Niue, a nearby island that is thought to have been settled after the initial Lapita dispersal. However, we found that no additional crossings to the east from the easternmost Lapita sites exceeded our probability cut-off under any conditions. The nearest such groups, the Northern and Southern Cooks, could be reached from Tonga and Samoa by the directly downwind sailing vessels with probabilities as high as about 3.5% in specific years, though the drifting vessels

Since westward movement can be achieved with apparent ease, we focus our discussion on the case of crossings to the east. Figure 7 summarizes the key simulation results, demonstrating the sections of the Lapita region reached under particular conditions, assuming that all discoverers ultimately originated in the Bismarcks or the Solomons.

130

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

Figure 7: Summary of simulation results. The shaded areas indicate the different areas which are reached by drift and downwind sailing under the specified conditions. Santa Cruz to the Loyalty Group

shows the opposite seasonal signal with this crossing most probable in the winter and spring. While the influence of the monsoon winds becomes more limited further to the east, these winds occur in a predictable manner in the vicinity of this crossing. These figures demonstrate how this strong seasonality could be exploited in facilitating interaction between islands in Near Oceania and those further east, once their presence was recognized.

Crossings to Santa Cruz from the Solomon Islands are probable in an overall sense, but probabilities are heightened in the summer and fall under the influence of the Australian monsoon (Tables 3, 4). Arrivals in this group are most likely from the outlying Solomon groups (Solomon-3, in particular) since these are separated from the large island screens of the main Solomon and Bismarck groups. Vessels starting from these larger groups generally return there allowing for a week at sea, such is the influence of the broad island screens.

From the Santa Cruz group, vessels can drift or sail to Vanuatu and from there reach the Loyalty Islands and New Caledonia; these crossings appear in the overall table but, again, are more probable in the summer and fall. In all simulations Vanuatu and the Loyalty group were ‘dead ends’ in terms of eastward exploration as launches from these sites reach no new islands under any conditions. An interesting set of results from this western portion of the study area are pathways between Remote Oceania and Australia. Australia is reached with relatively high overall contact probabilities from Vanuatu (drifting: 9.9%, sailing: 8.8%) and the Loyalty group(drifting: 15.5 %, sailing: 24.8 %). Remote Oceania

The great increase in the likelihood of crossings to Santa Cruz in the summer months is reflective of the strong and predictable seasonality of the winds in this part of Melanesia. Figure 8 shows a time series of the monthly crossing probabilities from Solomon-3 to Santa Cruz, revealing a distinct seasonal signal. Probabilities peak around January-February, reaching values as high as 80 100 % in the drift model and 60 - 75 % in the sailing model. Figure 9 shows a time series for the westward crossing from Santa Cruz to the main Solomon group and

131

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

Drift

Launch Site Target Site Bismarck -Main Bismarck -1 Bismarck -2 Bismarck -Main Other Lost at Sea

Downwind sailing

Cross.%

Min Time

Mean Time

Cross.%

Min Time

Mean Time

9.1 % 3.1 % 70.5 % 8.7 % 8.6 %

2 2 7 -

7.7 10.2 8.2 -

9.5 % 5.9 % 43.5 % 31.8 % 9.3 %

2 2 7 -

5.1 5.2 8.0 -

59.0 % 2.0 % 0.2 % 2.5 % 32.7 % 2.5 % 1.1 %

7 18 25 15 11 -

7.6 29.2 38.1 26.4 30.4 -

16.0 % 12.9 % 5.6 % 6.3 % 40.5 % 17.1 % 1.6 %

7 6 7 5 5 -

9.7 12.2 15.8 11.5 14.3 -

15.5 % 0.3 % 54.5 % 3.0 % 0.2 % 26.5 %

23 32 7 27 -

48.8 43.6 7.7 45.5 -

24.8 % 10.9 % 8.5 % 27.9 % 6.1 % 21.8 %

7 9 7 7 -

21.8 17.1 8.8 17.4 -

2.4 % 7.1 % 3.7 % 34.0 % 3.3 % 0.8 % 7.5 % 9.2 % 4.0 % 22 % 4.2 % 1.8 %

14 9 16 7 31 55 4 6 10 7 -

36.8 20.1 27.6 7.7 49.1 71.2 16.0 16.6 36.2 13.5 -

5.3 % 5.2 % 5.5 % 5.7 % 6.6 % 6.6 % 4.6 % 6.6 % 6.5 % 17.9 % 26.7 % 2.8 %

6 4 6 7 12 20 2 3 4 3 -

17.9 11.3 13.7 10.5 21.4 31.7 7.5 9.7 16.4 7.4 -

31.2 % 7.3 % 5.2 % 36.4 % 7.2 % 8.1 % 4.6 %

7 6 8 4 3 -

9.7 13.4 18.7 12.4 28.0 -

6.8 % 10.5 % 3.1 % 40.2 % 12.0 % 25.4 % 2.0 %

7 3 5 3 2 -

11.6 6.0 13.2 6.5 11.8 -

12.9 % 1.5 % 73.1 % 8.8 % 3.7 %

2 5 7 -

13.8 24.6 7.8 -

31.8 % 5.1 % 36.1 % 23.8 % 3.2 %

2 2 7 -

6.8 11.1 8.0 -

60.5 % 6.1 % 25.2 % 3.8 % 3.2 % 1.2 %

4 2 7 24 -

10.8 8.6 8.3 49.9 -

64.8 % 4.1 % 5.9 % 7.9 % 16.1 % 1.2 %

2 2 7 7 -

5.0 5.6 9.3 20.6 -

9.9 % 4.8 % 7.8 % 9.8 % 7.6 % 3.4 % 23.8 % 0.3 % 32.6 %

35 24 2 25 9 7 7 -

57.6 42.6 12.8 45.4 17.5 17.7 7.9 -

8.8 % 19.5 % 6.0 % 13.4 % 12.9 % 16.5 % 8.5 % 3.3 % 11.1 %

14 7 2 8 4 3 7 -

31.0 16.5 9.4 20.4 8.4 10.1 11.1 -

Fiji Fiji Santa Cruz Solomon -Main Tikopia Vanuatu Other Lost at Sea

Loyalty Australia Bismarck -Main Loyalty New Guinea Other Lost at Sea

Samoa Fiji Futuna Rotuma Samoa Santa Cruz Solomon -2 Tonga -1 Tonga -2 Tonga -Main Wallis Other Lost at Sea

Santa Cruz Santa Cruz Solomon -3 Solomon -4 Solomon -Main Vanuatu Other Lost at Sea

Solomon -Main Bismarck -Main Santa Cruz Solomon -Main Other Lost at Sea

Tonga -Main Fiji Tonga -3 Tonga -Main Vanuatu Other Lost at Sea

Vanuatu Australia Bismarck -Main Loyalty New Guinea Solomon -4 Solomon -Main Vanuatu Other Lost at Sea

Table 1: Overall crossing parameters for major island groups under the drifting and directly downwind sailing. Cross. %, contact probability; Min. Time, fastest crossing time, in days; Mean Time, average crossing time, also in days. Crossing parameters are based on all simulated voyages and the selected launching groups are the largest groups in the study area. Crossings to targets which have a probability of 5% or greater in either scheme are singled out, crossings with less than this percentage fall into the other category. The lost at sea category represents the percentage of vessels that have not encountered land after 90 days.

132

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

USCG Parameters

Launch Site

Hits%

Min Time

Mean Time

Hits %

Min Time

6.0 % 75.8 % 1.8 % 6.1 % 4.6 % 5.7 %

2 7 9 2 -

9.5 7.9 32.4 19.2 -

11.2 % 53.9 % 5.0 % 11.7 % 15.5 % 2.7%

2 7 4 2 -

4.8 7.6 13.5 7.7 -

68.3 % 3.8 % 23.9 % 1.4 % 2.6 %

7 26 15 -

8.0 54.6 41.1 -

27.1 % 6.6 % 42.0 % 22.1 % 2.2 %

7 9 6 -

11.4 21.3 18.4 -

15.9 % 56.6 % 0.4 % 0.2 % 26.9 %

24 7 44 -

50.6 7.7 59.5 -

42.6 % 11.3 % 7.8 % 10.56 % 27.8 %

7 7 13 -

24.2 10.2 32.0 -

4.7% 3.2 % 2.9 % 38.9 % 15.6 % 12.4 % 12.6 % 6.4 % 1.3 % 2.0 %

14 13 11 7 4 8 10 7 -

38.9 24.1 24.1 8.1 16.7 18.3 37.7 16.9 -

12.0 % 5.5 % 5.2 % 10.4 % 9.5 % 11.7 % 16.8 % 12.4 % 15.4 % 1.1 %

6 5 3 7 3 4 6 3 -

16.3 11.6 12.1 10.8 7.6 9.3 15.6 8.9 -

0.3 % 48.1 % 6.5 % 10.6 % 20.6 % 5.0 % 8.9 %

59 7 6 3 3 -

74.7 9.5 15.0 16.5 31.2 -

9.0 % 11.4 % 13.8 % 8.5 % 30.0 % 24.6 % 2.7 %

7 7 3 2 2 -

22.0 11.6 9.5 6.6 13.3 -

1.7 % 4.1 % 5.0 % 4.0 % 80.6 % 0.9 % 1.2 % 2.5 %

3 5 2 2 7 13 -

16.0 23.9 9.2 15.5 8.1 36.6 -

8.3 % 13.6 % 7.5 % 8.4 % 40.0 % 7.2 % 12.8 % 2.2 %

2 2 2 2 7 3 -

10.3 10.4 5.4 9.2 8.5 14.9 -

44.8 % 10.0 % 34.1 % 3.5 % 4.9 % 2.7 %

5 3 7 27 -

11.1 10.1 8.8 56.0 -

58.6 % 6.1 % 9.7 % 13.1 % 10.0 % 2.5 %

2 2 7 9 -

5.3 5.9 9.2 23.1 -

9.3 % 12.4 % 2.2 % 3.0 % 1.3 % 42.7 % 0.6 % 28.5 %

40 3 48 12 11 7 -

70.2 16.4 73.1 25.2 37.4 8.4 -

17.3 % 14.0 % 5.1 % 8.1 % 12.5 % 19.1 % 9.6 %

16 2 12 5 4 7 -

35.4 11.8 38.2 14.9 18.0 12.6 -

Target Site

Bismarck- Main Bismarck-2 Bismarck- Main Solomon-2 Solomon- Min Oter Lost at Sea

Levison Parameters Mean Time

Fiji Fiji Loyalty Vanuatu Other Lost at sea

Loyalty Austrlalia Loyalty New Guinea Other Lost at Sea

Samoa Fiji Futuna Niue Samoa Tonga-1 Tonga-2 Tonga-Main Wallis Other Lost at Sea

Santa Cruz Fiji Santa Cruz Solomon-Main Tikopia Vanuatu Other Lost at Sea

Solomon- Main Bismarck- Main Santa Cruz Solomon-3 Solomon-4 Solomon-Main Vanuatu Other Lost at Sea

Tonga-Main Fiji Tonga-3 Tonga-Main Vanuatu Other Lost at Sea

Vanuatu Australia Loyalty New Guinea Solomon-4 Solomon-Main Vanuatu Other Lost at Sea

Table 2: Like Table1, but only for voyages that start during the Austral summer (December, January, February). Table reproduced from Avis et al. (2007).

133

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

Drifting Season

Hit %

Min Time

Mean Time

Santa Cruz

Summer

6.6 %

25

50.0

Solomon-2 Santa Cruz

Summer

12.4 %

23

48.5

Overall Summer Fall

17.6 % 46.9 % 20.3 %

5 5 5

20.8 20.4 20.5

Launch Site Target Site

Solomon-1

Solomon-3 Santa Cruz Santa Cruz Santa Cruz

Table 3: Drift crossings into the Santa Cruz island group from sites in Near Oceania. Crossings have a higher probability in the summer months when the frequency of occurrence of winds having a westerly component is higher. Downwind sailing Launch Site Target Site

Solomon-1 Santa Cruz

Season

Hit %

Min Time

Mean Time

Summer

7.8 %

7

19.5

Overall Summe Fallr

7.6 % 18.4 % 7.3 %

5 5 6

23.0 16.9 24.2

Overall Summer Fall

11.3 % 30.5 % 11.5 %

3 3 3

10.0 9.3 9.8

Overall Summer Fall

5.1 % 13.6 % 5.8 %

2 2 4

11.1 10.4 9.8

Solomon-2 Santa Cruz Santa Cruz Santa Cruz Solomon-3 Santa Cruz Santa Cruz Santa Cruz

Solomon-Main Santa Cruz Santa Cruz Santa Cruz

Table 4: Crossings into the Santa Cruz island group from sites in Near Oceania under the Levison (sailing) parameters. Drifting

Launch Site

Year

Hit %

Min Time

Mean Time

2003

5.6 %

42

62.7

2002

14.3 %

64

66.0

2002 2002 2003

10.7 % 17.9 % 8.3 %

80 55 64

84.2 65.3 69.3

Target Site

Santa Cruz Rotuma

Solomon-1 Rotuma

Solomon-2 Fiji Rotuma

Table 5: Drift crossings into Fiji and Rotuma. Drift pathways into these groups are probable when considering crossings that take place in the summer months of particular years.

134

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

Downwind sailing

Launch Site Target Site

Santa Cruz Fiji

Solomon-2 Fiji

Solomon-3 Fiji

Tikopia Fiji

Season

Hits %

Min Time

Mean Time

Summer

9.0 %

7

22.0

Summer

5.4 %

16

31.7

Summer

8.0 %

9

22.7

Summer

6.6 %

8

18.3

Table 6: Directly downwind sailing crossings into Fiji. Crossings are viable under mean summer conditions and occur with mean crossings times that are 1/3-1/2 of those of the drifters.

Drifting

Launch Site Target Site

Year

Hit %

Min Time

Mean Time

1993 1993

12.5 % 16.7 %

44 33

49.3 47.3

1993

6.3 %

48

48.0

Rotuma Tonga -1 Tonga -2

Tonga -2 Samoa

Table 7: Drifting crossings to Tonga and Samoa. Crossings are probable in particular summers, though all pathways from the west require Rotuma as an intermediate stepping stone. Additional crossings to Tonga and Samoa are possible using Wallis and Futuna as stepping stones but these must be reached via Rotuma. was peopled beginning about 3,500 years ago, and it is curious that there appears to be little evidence for prehistoric interactions between Remote Oceania and Australia given the apparent ease with which such crossings could occur.

2002 from Solomon-2, but here the mean crossing time is nearly three months, significantly longer than in the sailing experiments. Sailing boats reach Rotuma from Solomon-1 in 1998, Solomon-2 in 1996, 2002 and 2003, Solomon-3 in1993 and Tikopia in 2003 (Table 8), while drifters can reach the island from Santa Cruz in 2003, Solomon-1 in 2002 and Solomon-2 in 2002 and 2003 (Table 5). Again, the number of probable crossings in the drift scheme is greatly reduced over the sailing connections. Fiji can be easily reached from Rotuma in both sets of simulations, presenting another pathway by which it could have been discovered. We have included Rotuma as a launching site in these simulations since it is the only point from which Tonga and Samoa may be reached by drift. Its inclusion is, however, problematic since Rotuma is thought to have been first settled after the initial Lapita expansion (Anderson 2003). Thus, the role of this island in the actual expansion is questionable.

Fiji and Rotuma Fiji is reached from Santa Cruz, Solomon-2, Solomon-3 and Tikopia with probabilities ranging from 5.4 % to 8.9 % in the sailing experiments when mean summer results are considered (Table 6) and with significantly higher probabilities in particular summers. Fiji is nearly 1,500 kilometers from the nearest western launch point, yet these crossings have a mean time of a month or less. These crossings are not possible by drift in mean summer conditions, the first indication of the more limited set of contact pathways open to drifters. Sailing crossings further east are also not probable under mean summer conditions. Drifts directly to Fiji occur only in summer,

135

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

Downwind sailing Year

Hits %

1993 2002 1997 1999 2001

6.8 % 6.0 % 5.1 % 8.6 % 6.3 %

16 3 4 3 5

20.5 13.9 5.1 8.3 9.5

1993 2004 1993 1994 2003 1993 1996 1999 2000 2002

12.5 % 7.1 % 8.3 % 7.1 % 7.1 % 8.3 % 7.1 % 19.0 % 9.5 % 9.5 %

12 15 10 8 37 21 9 5 8 9

16.3 20.7 16.0 15.0 39.7 29.0 10.3 8.5 16.0 12.3

Samoa Tonga-1 Tonga-2 Tonga-Main

1995 2003 1995 2003 2003 2002

9.7 % 12.8 % 5.1 % 5.1 % 5.1 % 6.1 %

24 11 33 39 28 26

33.1 24.6 47.8 50.5 31.3 32.9

Solomon-1 Rotuma Tonga-1 Tonga-Main

1998 1993 2002

7.1 % 12.5 % 7.1 %

21 32 34

21.0 32.0 34.0

Samoa Tonga-2 Tonga-Main

1996 2002 2003 1995 1993 1993

7.1 % 7.1 % 10.7 % 6.0 % 8.3 % 6.3 %

24 18 28 47 28 39

24.5 27.8 35.9 52.6 32.5 39.3

Solomon-3 Rotuma Tonga-Main

1993 2002

12.5 % 7.1 %

16 30

19.0 39.0

Tikopia Rotuma Tonga-3

2003 2002

17.9 % 7.1 %

17 24

20.6 24.0

1996 2003 2004

7.1 % 14.3 % 7.1 %

25 3 26

25.0 4.0 26.0

1993 1996 2003 2004

18.8 % 7.1 % 10.7 % 10.7 %

10 16 6 22

15.0 21.0 17.7 24.0

Launch Site Target Site

Fiji Samoa Tonga-3 Tonga-Main

Rotuma Samoa Tonga-2 Tonga-Main

Santa Cruz Rotuma

Solomon-2 Rotuma

Tonga-1 Samoa

Tonga-2 Samoa

Min Time

Mean Time

Table 8: Downwind sailing crossings to Rotuma, Tonga and Samoa. Crossings are probable in particular summers. A number of pathways to the easternmost groups are possible, including some direct connections from the western Lapita region. Tonga and Samoa

particular summers as voyages to this group from the west are not probable under mean summer conditions, even from Tonga and Rotuma and even allowing Futuna and Wallis as intermediate launching points. Under

To allow for pathways to Tonga and Samoa in West Polynesia, we must again consider crossings that occur in

136

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

specific conditions, sailing boats from Fiji, Rotuma, Santa Cruz and the Solomons reach Samoa; Tonga is reached from all preceding sites as well as from Tikopia (Table 8). Sailing crossings from Samoa to Tonga occur with high overall probabilities. The reverse crossing is possible, too, but only in particular years.

for drifts as far as the Loyalty group and sailing connections as far as Fiji. Further movement eastward is possible in certain years but not in the general case. From these eastern sites, the island of Niue may be reached, but further eastward crossings, by drift or by sail, occur only in specific years and with reduced crossing probabilities which fail to exceed the 5% threshold imposed on the simulation. These results suggest that the swath of ocean east of Tonga and Samoa presents a sort of bottleneck to eastward expansion by the simple exploration strategies we modelled. This is a significant finding since, with the exception of Niue, the area reached by the vessels in our simulation coincides with the area settled during the rapid Lapita dispersal, which was followed by a great pause in eastward movement.

Drifting vessels reach Tonga only from Rotuma (Table 7). Drifts directly to Samoa from western sites are not probable, but the group can be reached from Tonga in summer, 1993. Thus, while a number of sailing pathways to western Polynesia are possible, there is a single drift path to Tonga and Samoa, one which must necessarily include Rotuma as an intermediate point. The experiments demonstrate pathways to all the island groups in the Lapita region by drift and by sail from initial launches in the Solomons and Bismarcks. The region of influence of the mean monsoonal winds allows

Figure 8: Time series of monthly crossing probabilities from the Solomon-3 launch group to the Santa Cruz group. Probabilities for the drifting are shown in the left panel and for the sailing in the right panel. These plots show a distinct seasonal signal: crossings occur with high probabilities in the Austral summer and not at all in the winter. Survival at Sea

show that a majority of simulated vessels encounter land within three months at sea, reflecting the influence of the extensive screens of islands in the equatorial Pacific which can ‘catch’ drifting or sailing vessels. So, these results imply that, if survival at sea for 3 months is accepted, the number of fatalities may not be large even in voyages of drift or undirected exploration.

The proper way to interpret the ‘lost at sea’ category in the results is not as a measurement of fatalities, but as representation of the percentage of vessels which did not reach land within 90 days. Drifts and sailing voyages which start from islands lying in the band stretching from the equator to 20° S, both within Melanesia and Polynesia, have low ‘lost at sea’ percentages of 25% or less (considering overall results), with the majority having percentages of 10 % or lower. These low values

A number of factors affect survival including availability of food and water, precipitation received en route (which supplies drinking water), the nature of wind and

137

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

temperature extremes encountered and so on. In an attempt to address this issue, Levison et al. (1973) allowed for variable voyage lengths: individual voyages were randomly assigned maximum lengths following a survival probability table based on anecdotal reports of drifts. Survival at sea is a complicated matter and to model it, given the unknowns, would require a number of additional and arbitrary assumptions, and we chose not to model it at all.

stores of food and water to help them withstand lengthy voyages, so survival at sea for extended periods is much easier to accept in this case. In the Levison et al. (1973) simulation, vessels stood a 50% chance of capsizing upon encountering winds of force 9 on the Beaufort scale or greater (> 20.6 m/s). We make no attempt to model the sea worthiness of vessels and assume that boats withstand all weather that they encounter. Since nothing firm is known about maritime technology at the time of Pacific settlement (McGrail 2001), there is no reason to choose any particular value as the maximum wind speed that could be withstood by boats, and doing so involves making further assumptions in the simulations. We do note however that force 9 winds and higher occur within the study area only infrequently.

We acknowledge that mariners, especially drifters, could expire before three months was up though it should be noted that there are a number of historical accounts of drifts with survivors which lasted for at least half a year (Montenegro et al. 2006). Furthermore, since the downwind sailing experiments represent intentional exploration, crews of such ships would surely bring

Figure 9: Time series of monthly crossing probabilities from Santa Cruz to the main Solomon target. The plots show a seasonal signal opposite to that in Figure 8 with crossings most probable in the Austral winter. between 3 and 14 and the number of hurricanes/typhoons per year has varied between 1 and 12 (UKMet 2007). Frequencies generally increase in strong El Niño years, associated with positive anomalies in sea surface temperature in the western Pacific. Individual systems only affect a small portion of the basin, and Irwin (1992) argues that the frequency of occurrence of cyclones is sufficiently small that the odds of a particular canoe encountering one are low.

Such intense winds are often associated with cyclones which typically originate east of the Solomons at around 8° S, generally between November and March (Irwin 1992). Thus the summer months, when eastward movement is most probable, coincide with the cyclone season in the Lapita region. Tropical storms are cyclonic disturbances having peak sustained wind speeds between 17 m/s and 32 m/s and hurricanes or typhoons are systems having maximum sustained wind speeds in excess of 32 m/s. Since 1988, the number of tropical storms per year in the eastern Australia basin5 has ranged

A Comparison with Levison et al. (1973) While Levison et al. (1973) focused most of their analysis on crossings within Polynesia, they did test some

5

The eastern Australian basin is defined as the area between the equator and 40◦ S and between 135° E and 225° E.

138

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

crossings in the Lapita region. In agreement with our results, the authors found that westward drifts were quite feasible with a large number of crossings having probabilities of 10 - 20 % or higher. Drifts to the east, north and south occurred with low probabilities and the key crossings to Fiji and Rotuma and to central Polynesia occurred with negligible probabilities, if at all. This led the authors to conclude that “the four hundred-mile expanse of ocean between the Solomons and New Hebrides (Vanuatu) to the west and Fiji to the east presents a formidable barrier to eastward drifts”. Levison et al. (1973)’s work was an important early application of computer models to anthropology and its main results, demonstrating the apparent implausibility of eastward crossings in a direction that is against the prevailing trade winds, are similar to our overall results for sites not strongly influenced by monsoonal winds, as is the recognition that westward drifts are highly probable. However, their model cannot account for the influence of any kind of interannual variability nor sustained periods of anomalous winds in particular months, both of which could be important aspects of a strategic exploration strategy which exploits such unique conditions. In this respect, our results differ significantly from those of Levison et al. (1973), especially with regards to crossings into Fiji/Tonga/Samoa. Such pathways are possible both using Levison et al. (1973)’s own “drift” parameters and more conservative estimates of drift from the US Coast Guard.

the stochastic nature of the phenomenon makes it difficult to predict which crossings are possible in a particular episode. That some eastward crossings also occur in La Ni˜na years, when easterlies would generally be expected to be more intense, further reflects the complicated nature of the interannual variability in the Lapita region.

The Influence of El Niño

Conclusions

There were five periods of sustained El Niño conditions during the simulations (1993, 1994-95, 1997-98, 2002-03 and 2004-05). Of these, the most pronounced is the 199798 El Niño, well known as the most intense such event on record. While the majority of paths that are probable in specific years occur in El Niño years, a few eastward crossings occur in 1996 and 1999, under La Niña conditions. Furthermore, the paths that are possible in one El Niño are different in another and there is no apparent correlation between the number of viable pathways in the Lapita region and the strength of the El Niño. For instance, the number of crossings possible in winter 1993 exceeds those in 1997-1998 despite the latter being a much more intense event.

That the Lapita peoples and their descendants were capable seafarers is not in question. The extensive interaction spheres and exchange of artifacts between widely separated islands and archipelagos imply they could cross long stretches of open water with apparent ease, perhaps employing relatively sophisticated navigational skills. But sailing and navigating in a sea of known islands where mean currents and winds are understood is a different matter than exploring an unknown ocean for the first time. Our simulation results suggest that the discovery of islands in the Lapita region was possible without advanced exploration schemes. Contact pathways from Near Oceania to Tonga and Samoa are possible both through drift and directly downwind sailing. More advanced exploration strategies could have been employed, but they are not necessary under the assumptions of our model.

ENSO in the Holocene According to one proxy reconstruction, the frequency of occurrence of anomalous conditions associated with the El Niño appears to have varied since the mid-Holocene, with the number of moderate-to-strong El Niño events per 100 years increasing from about 5 at about 4,000 years ago to over 25 between 1,000 and 2,000 years ago (Moy et al. 2002). Anderson et al. (2006) have proposed that prehistoric Pacific watercraft were not advanced enough to permit sailing against the trades and observe that El Niño related wind reversals could facilitate eastward movement. The present view is that Oceanic colonization was highly episodic and the authors observe that colonization episodes either coincide with or slightly lag behind periods where there was an increase in the frequency of El Niño events. Our results indicate that important eastward crossings in the Lapita region and Polynesia have higher probabilities in El Niño years, and lend support to Anderson et al. (2006)’s hypothesis. However, caution must be exercised until a larger number of long high-resolution El Niño proxies are available.

More simulations using a longer weather data set that captured additional intense El Niños might be used to look for a correlation between event strength and crossings within central Polynesia. Sampling the influence of additional El Niños might also reveal new crossings, possibly including a stronger connection between Tonga and Samoa and the Cooks than we were able to demonstrate. Overall, these results suggest that climate fluctuations associated with El Niño may play an important role in facilitating crossings to the east, though

Even if the Lapita people were very skilled mariners, it is unreasonable to say that they never got blown off course or lost at sea, and such events could be one way by which drift voyages might occur. We do not hold that drifters can account for the settlement of new islands, but they could explain their discovery. From the drift ‘lost at sea’ results, such drifters stand a very good chance of reaching

139

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

land, accepting that the crews can survive at sea for 3 months. The majority of drifters return back to the area from which they were lost but a sizable percentage move eastward. Having encountered new land, the crews of such boats could wait for easterly winds and attempt to sail or drift west to inform others of their discovery. The knowledge of land to the east could then serve as motivation for deliberate voyages of exploration and, later, settlement. That many key drift crossings in the study region occur only in particular years when the wind field is different from the mean state fits well with this scenario since it seems more plausible for vessels to lose their way when faced with unusual weather.

that would be involved. We suspect, though, that the fundamental results of this simulation would also apply to models incorporating more advanced exploration strategies: eastward courses would be fastest and easiest to maintain in summer months and El Niño conditions when the frequency of occurrence of easterly winds decreased. The simulations point to the importance of considering climate variability in models of Pacific exploration, a factor which could not be addressed in Levison et al. (1973)’s influential work. While our experiments are doubtless an over simplification of the actual means by which the Pacific was explored, they are more sensitive to this variability than more advanced exploration models might be. It is sometimes said that the settlers of the Pacific perceived their world not as a barrier, but a highway. To extend this metaphor, we suggest that the ocean can indeed be seen as a highway, but one with an ever shifting roadmap.

Irwin (1992)’s safe exploration model suggests that the Lapita region and central-eastern Polynesia were probed by vessels sailing against the direction of the dominant trade winds. Downwind sailing is an element of this model, but in Irwin (1992)’s view, crews on outbound voyages would only use this strategy if progress to the east could be made. In our simulations, vessels sail directly downwind no matter which way the wind blows and since the simulation does not prevent boats from leaving when easterlies blow, it is not a pure test of this aspect of Irwin (1992)’s theory. However, a comparison of the Austral summer crossing probabilities with corresponding results from the overall table gives an indication of how strategic exploration that exploits westerlies could increase the frequency of contact to the east.

References Anderson, A. 2003 “Initial human dispersal in Remote Oceania: pattern and explanation”. In: Sand, C. (ed.), Pacific Archaeology: Assessments and Prospects. Les Cahiers de l’Archéologie en Nouvelle Calédonie 15: 71-84. Service des Musées et du Patrimoine de NouvelleCalédonie, Nouméa. Anderson, A., J. Chappell, M. Gagan and R. Grove 2006 “Prehistoric maritime migration in the Pacific islands: an hypothesis of ENSO forcing”. The Holocene 16 (1): 1-6. Avis, C., A. Montenegro and A. Weaver 2007 “The discovery of Western Oceania: a new perspective”. Journal of Island and Coastal Archaeology 2: 197-209. Di Piazza, A. 2005 “Excavations at Avarua (rak-1): A late archaeological assemblage from a pearl shell workshop on Rakahanga, northern Cook Islands”. People and Culture in Oceania 20: 69-88. Dodd, E. 1972 Polynesian Seafaring. Dodd, Mead, New York. Evans, B. 1999 Simulating Polynesian double-hulled canoe voyaging: combining remotely-sensed and experimental data. M.A. thesis, University of Hawai’i. Finney, B. 1977 “Voyaging canoes and the settlement of Polynesia”. Science 196 (4296): 1277-1285.

The directly downwind sailing experiments imply a greater number of pathways in the Lapita region, many of which occur with significantly higher probabilities than in the drift case. One would therefore anticipate a faster rate of island discovery of new islands in the sailing case because of these heightened probabilities. Directly downwind sailing is the simplest exploration strategy to test and involves the fewest assumptions, requiring only that the crew opt to sail with the wind and are able to make landfall when islands are sighted. The boats used by explorers and settlers could almost certainly sail, not just directly downwind, but also at various angles to the wind. Indeed, this would be necessary if crews were to steer towards land once islands were sighted. This steering ability would increase the possible courses in the Lapita region over those represented by our simulations. Modelling directed sailing makes simulations more difficult to interpret and to justify as decision-making must be modelled in order to choose a course to be sailed and the means by which the crew would attempt to maintain it. Thus, while eastward contact probabilities might increase (and crossing times decrease) under a more advanced exploration scheme, we opted not to test such models given the added complexity and assumptions

140

SIMULATING ISLAND DISCOVERY DURING THE LAPITA EXPANSION

Finney, B., P. Frost, R. Rhodes, N. Thompson 1989 “Wait for the west wind”. The Journal of the Polynesian Society 98 (3): 261–302. Frankel, J. 1962 Polynesian Navigation. Journal of the Inst. of Navigation 9: 35-47. Hackett, B., O. Breivik and C. Wettre 2006 “Forecasting the drift of objects and substances in the ocean”. In E. Chassignet and J. Verron (eds.), Ocean Weather Forecasting: 507-523. Springer, Berlin. Hastings, D. and P. Dunbar 1999 The Global Land One-kilometer Base Elevation (GLOBE) Digital Elevation Model, version 1.0. Tech. rep., National Oceanic and Atmospheric Administration, National Geophysical Data Center, Boulder, Colorado. Hunt, T. and C. Lipo 2006 “Late colonization of Easter Island”. Science 311 (5767): 1603–1606. Hurles, M., E. Matisoo-Smith, R. Gray, and D. Penny 2003 “Untangling Oceanic settlement: the edge of the knowable”. Trends in Ecology and Evolution 18 (10): 531-540. Intoh, M. 1997 “Human dispersals into Micronesia”. Anthropological Science 105 (1): 15–28. Irwin, G. 1992 The Prehistoric Exploration and Colonization of the Pacific. Cambridge University Press, Cambridge. 1998 “The colonisation of the Pacific plate: chronological, navigational and social issues”. The Journal of the Polynesian Society 107 (2): 111–143. Irwin, G., S. Bickler, and P. Quirke 1990 “Voyaging by canoe and computer: experiments in the settlement of the Pacific Ocean”. Antiquity 64 (242): 34–50. Kirch, P. 1997 The Lapita Peoples: Ancestors of the Oceanic World. Blackwell Publishers, Cambridge, Massachusetts. 2000 On the Road of the Winds: An Archaeological History of the Pacific Islands before European Contact. University of California Press, Berkeley. Kistler, R., E. Kalnay, W. Collins, S. Suranjana, G. White, J. Woollen, M. Chelliah, W. Ebisuzaki, M. Kanamitsu, V. Kousky, H. van den Dool, R. Jenne and M. Fiorino 2001 “The NCEP/NCAR 50 year Reanalysis: Monthly means CD-ROM and documentation. Bulletin of the American Meteorological Society 82(2), 247–267.

Levison, M., R. Ward, and J. Webb 1973 The Settlement of Polynesia: a Computer Simulation. The University of Minnesota Press, Minneapolis. Markgraf, V. and H. Diaz 2000 “The past ENSO record: a synthesis”. In: H. Diaz and V., Markgraf (eds.), El Niño and the Southern Oscillation: Multiscale Variability and Global and Regional Impacts: 465-488. Cambridge University Press, Cambridge. Markgraf, V., J. Dodson, A. Kershaw, M. McGlone and N. Nicholls 1992 “Evolution of late Pleistocene and Holocene climates in the circum-South Pacific land areas”. Climate Dynamics 6 (3-4): 193-211. McGrail, S. 2001 Boats of the World: From the Stone Age to Medieval Times. Oxford University Press, New York. Montenegro, Á., C. Avis, and A. Weaver 2007 “Modelling the pre-historic arrival of the sweet potato in Polynesia”. Journal of Anthropological Science (in press) Montenegro, Á., R. Hetherington, M. Eby and A. Weaver 2006 “Modelling pre-historic transoceanic crossings into the Americas”. Quaternary Science Reviews 25 (11-12): 1323-1338. Moy, C., G. Seltzer, D. Rodbell, and D. Anderson 2002 “Variability of El Niño/Southern Oscillation activity at millennial timescales during the Holocene epoch”. Nature 420 (6912): 162165. Sand, C. 2000 “La datation du premier peuplement de Wallis et Futuna: contribution à la définition de la chronologie Lapita en Polynésie occidentale”. Journal de la Société des Océanistes 111 (2): 165-172. Sand, C., J. Bolé, J.and A. Ouetcho 2002 “Site LPO023 of Kurin: Characteristics of a Lapita settlement in the Loyalty Islands (New Caledonia) ”. Asian Perspectives 41 (1): 129147. Sharp, A. 1963 Ancient Voyagers in Polynesia. Angus and Robertson Ltd., London. Spriggs, M. 1997 The Island Melanesians. Blackwell Publishers, Cambridge, Massachusetts. Spriggs, M. and A. Anderson 1993 “Late colonization of East Polynesia. Antiquity 67 (255): 200-217.

141

CHRIS AVIS, ÁLVARO MONTENEGRO AND ANDREW WEAVER

Stammer, D., C. Wunsch, R. Giering, C. Eckert, P. Heimbach, J. Marotzke, A. Adcroft, C. Hill and J. Marshall 2002 “The global ocean circulation during 19921997 estimated from ocean observations and a general circulation model”. Journal of Geophysical Research 107 (C9): 3118-3146. Stringer, C. 2000 “Coasting out of Africa”. Nature 405 (6782): 24-26. UKMet 2007 United Kingdom Meteorology Office: Eastern Australian Basin tropical cyclone activity. Http://www.metoffice.gov.uk/weather/tropic alcyclone. Last accessed: February 2nd, 2007. USCG 2002 U.S. Coast Guard addendum to the United States National Search and Rescue supplement to the International Aeronautical and Maritime Search and Rescue manual. Second edition. Tech. rep., United States Coast Guard, comdtinst m16130. Weisler, M. 1997 Prehistoric long distance exchange in Oceania. New Zealand Archaeological Association, Monograph 21, Auckland. 1998 “Hard evidence for prehistoric interaction in Polynesia”. Current Anthropology 39 (4): 521-532.

142

Statistical analysis of simulated voyages incorporating synoptic weather data and experimental archaeologicallyderived sailing parameters shows that this new methodology produces results that are statistically significantly different from simulated voyages calculated using the methodologies of previous studies. In addition, courselines created using this new simulation appear similar to those of actual experimental archaeological voyages. With this in mind, courselines produced using this study should prove useful in answering questions about the best times of departures in terms of seasonal variations and with respect to larger scale phenomena such as the El Niño Southern Oscillation (ENSO) for voyages between specific islands and archipelagos. This in turn will aid in answering larger questions on the timing and method of settlement of Polynesia. More immediately though, results of this research may prove valuable to the planning of several experimental voyages planned in the near future.

Chapter VIII Simulating Polynesian Double-Hulled Canoe Voyaging. Combining Digital and Experimental Data to Prepare for a Voyage to Rapa Nui (Easter Island) by Bradley M. Evans∗ The first important use of computers to answer questions about Polynesian voyaging was in the early 1970's. In their study, Levison, Ward and Webb rejected random drift voyaging as a viable method for the settlement of Polynesia in favor of a model that called for systematic exploration by early Polynesians. A continuing series of experimental voyages using traditional navigational methods has also helped to illuminate issues of the settlement of Polynesia. These efforts have produced better data on the sailing capabilities of Polynesian canoes. The wind and current data available to previous computer simulations were sparse in many areas.

This paper is in four sections. The first section provides a brief history of the theories of the settlement of Polynesia. The next reviews the computer simulations and experimental voyages carried out in the last two and a half decades. The third explains the current simulation with respect to the methodologies of previous computer simulations. Lastly, the results of this new methodology along with directions for future research are presented.

Today, hourly readings of weather conditions across the Pacific are available from satellites. The simulation described here incorporates these new datasets, and produces courselines that are statistically different from and more accurately reflect the results of experimental voyages than do those of previous computer simulations. Future refinements of this simulation can aid in planning courselines for upcoming experimental archaeological voyages. This new methodology will also help in answering issues of the best times of year for voyaging, the effect of events such as the El Niño Southern Oscillation (ENSO), and the strategies used to colonize and explore the Pacific Basin.

Settlement of Polynesia The Polynesian Triangle is an area of 20 million square kilometers stretching across much of the South Pacific Ocean (Fig. 1). Within it, Polynesian voyagers visited hundreds of islands and colonized more than a dozen archipelagos. While archaeological dating, linguistics, and artifact seriation studies have provided insight into the settlement of Polynesia, the actual mechanism of the Polynesian dispersal has remained a subject of scientific curiosity since the recognition of a Polynesian people more than two centuries ago.

Before the late 1960’s, the settlement of Polynesia was investigated primarily by way of archaeology, ethnography and linguistics. Since that time, researchers have added two methods: computer simulation and experimental archaeology. Until now, these two methods had not been used in combination. This paper updates the classic computer simulations of Polynesian voyaging by incorporating data on sailing parameters from experimental archaeology, and remotely-sensed, synoptic weather data that was not available to earlier researchers. The aim of the project discussed is to determine if such a simulation will produce simulated courselines that are more similar to those of actual experimental voyages than courselines produce by the methods of previous studies.

As early as the 1770's, Captain James Cook realized the close kinship of the languages of island peoples across the Pacific Basin and questioned how it was that this race had come to spread itself across thousands of kilometers of ocean (Finney, 1994b). By this time long-distance, interarchipelago voyaging and colonization had ceased throughout Polynesia and the canoes constructed at the time were intended for fishing, warfare and other relatively near-shore and short-distance needs. Despite this, oral traditions told of lengthy voyages between farflung archipelagoes such as Tahiti, in the Society Islands, and Hawai‘i (Finney 1994b).

∗ Anthropology Department, Bernice P. Bishop Museum, 1525 Bernice St., Honolulu, 96817

143

BRADLEY M. EVANS

Figure 1: The Polynesian Triangle colonization from the Americas would resurface a century later and prove difficult to remove from the public’s consciousness.

Subsequent speculation on the origin of the Polynesians ranged from the sensible to the fantastic. While some Western explorers and scientists believed that the Polynesians had colonized their islands through intentional navigation, others preferred the hypothesis that the island’s original settlers were the storm–tossed castaways of unfortunate fishing trips. Still others thought that the Polynesians were the descendants of mountain dwelling tribes that retreated to high ground when a lost Pacific continent sank beneath the waves (Bellwood 1979).

By the early twentieth century, informed mainly by linguistic evidence and oral history, a few prominent Pacific anthropologists saw the Polynesians as the “Vikings of the Pacific.” They held that the ancestors of modern Polynesians had migrated from Indonesia, sailing confidently over vast tracts of ocean as they moved eastward across the Pacific, colonizing islands, and maintaining lines of communication between island groups thousands of kilometers apart (Buck 1938).

In the nineteenth century, amateur anthropologists, primarily missionaries, espoused a variety of theories of Polynesian settlement based on both linguistics and ethnography (Bellwood, 1979). These studies are of limited value for the most part because the authors were blind to the enormous changes that contact with the West had wrought upon Polynesian society. In addition, many of the theories appear to be either tortured attempts to fit Polynesian prehistory into Christian dogma or to satisfy the author’s fantastic theories of the ultimate origins of the Polynesian people, quite often in ancient Egypt. Some pointed out that the prevailing tradewinds would make exploration to the east difficult and suggested an American origin. Though dismissed by most anthropologists today, this idea of a westward

Thor Heyerdahl and the voyage of the Kon Tiki challenged this notion in 1947. This renowned trip involved a more than three month drift voyage in a balsa wood raft from off the coast of Peru that reached eastern Polynesia (Heyerdahl, 1968). He believed that the predominant tradewinds and currents made an eastward colonization of Polynesia improbable, and suggested an American origin for the Polynesian people. To strengthen his arguments, Heyerdahl also pointed out botanical introductions to Polynesia that he claimed were of South American origin. While most his examples strained credulity, the sweet potato, found throughout much of

144

SIMULATING POLYNESIAN DOUBLE-HULLED CANOE VOYAGING

voyages from a number of islands and South America, and as a result, rejected unintentional drift voyaging as a method of settlement of much of Polynesia. Using synthesized weather data collected from British ships from the mid-nineteenth through mid-twentieth century, Levison et al. simulated canoes that drifted before the wind and along with currents, not attempting to navigate or hold a course. Islands were considered to lie at the center of a circular sighting-radius. The size of this circle was determined by the size and elevation of the island and reflects the distance at which voyagers would have been able to detect the island’s presence. If a drifting canoe entered an island’s sighting-radius before food and water were exhausted, the voyage was considered successful, otherwise the crew perished at sea.

Polynesia, is undisputedly originally a South American cultigen (Bellwood 1979). Heyerdahl’s thesis that Polynesia had been settled by drift voyages from the east was not new. While he was, and is, ridiculed by the scientific community for his questionable methods and interpretations, his voyage had addressed one on the continuing enigmas of the settlement of Polynesia: across the Polynesian triangle the prevailing tradewinds blow from the east (Bellwood 1979). As any sailor knows, tacking for long periods into the wind makes for long, difficult, uncomfortable, and possibly dangerous voyages. The double-hulled canoes of the Polynesians, while quite seaworthy, probably could not hold a course much closer than 75 degrees off the wind (Finney 1994b). The practical effect of this is that if the destination of a Polynesian canoe lay directly upwind, tacking at 75° off of the wind means an addition of nearly four miles for every mile made good toward the destination. Thus, a voyage of a few hundred kilometers as the crow flies becomes a bumpy, exhausting, and possibly fatal trip of a few thousand kilometers. Perhaps short voyages against the wind might be concluded successfully, but Heyerdahl and others believed that a consistent eastward expansion from Southeast Asia was blocked by the predominant easterly tradewinds.

Reverse experiments were also conducted to find the starting points of drift voyages that would have ended on specific islands. Levison et al. stated that in the case of Rapa Nui, the chance of drift voyaging from either the Americas or East Polynesia was virtually nil. They went on to rank the “spontaneous creation” of the population of Rapa Nui as more likely than a drift voyage regardless of the point of origin (Levison et al. 1973). Irwin is the most active researcher in his use of computers to test theories of the strategic exploration and colonization of Polynesia (Irwin et al. 1990; Irwin 1992; Irwin 1998). Using a portion of the wind speed and direction dataset of Levison et al., Irwin has simulated the settlement of Polynesia testing a variety of voyaging strategies, then compared the results with the archaeological record and linguistics evidence. He made the case that the most logical exploration and colonization strategy used refined navigational techniques intended to maximize the crews’ safety (Irwin 1992).

Further skepticism was aimed at the navigational techniques used by Polynesian voyagers. The lack of methods to accurately determine longitude and estimate the effect of current was seen by some to preclude the possibility of any intentional voyages over a few hundred kilometers. This point was driven home by Sharp (1963). His view was that Polynesia was settled mainly by oneway drift voyages, reasoning that any canoe that had sailed for more than a few days would find itself irretrievably lost. Some ethnological research on noninstrument navigation seemed to support Sharp’s claims (Akerblom 1968).

Irwin hypothesized that Polynesian voyagers explored eastward against the prevailing tradewinds for strategic reasons: if an expedition were unsuccessful in finding new lands, it was a simple matter to return to the island of departure, relying upon the tradewinds for a quick passage home. The settlement pattern created using Irwin’s simulation roughly approximates that suggested by radiocarbon data, though important questions about the timing of certain periods of colonization remain (Irwin 1992).

By the late 1960’s, researchers were seeking newer and more rigorous methods to answer questions of Polynesian origins. The two most contentious assertions of Sharp and Heyerdahl, namely, the incompetence of traditional Polynesian voyaging techniques and the difficulty of colonizing Polynesia from west to east against the prevailing tradewinds, have been tested by a number of different researchers since the late 1960’s using computer simulations and experimental archaeology. The results of their studies have drastically altered views on the settlement of Polynesia.

Though Irwin and Levison et al.’s studies were pivotal in reevaluating the possibility of intentional navigation, it has been the reconstruction of double-hulled voyaging canoes and an attempt to recover non-instrument navigation techniques that has gained wide public exposure. These experimental archaeological studies have helped to rebut skeptics of both the capabilities of noninstrument navigation and the seaworthiness of Polynesian voyaging canoes. In addition, the meticulous record keeping of many voyages provides more realistic sailing parameters for the modeling of Polynesian

Previous Investigations: Computer Simulations and Experimental Archaeology The first important use of computer simulation in this area was by Levison, Ward, and Webb (1973). As part of their study Levison et al., simulated over 100,000 drift

145

BRADLEY M. EVANS

Marquesas in some 40 days”. Heyerdahl’s voyage of the Kon Tiki, by contrast, took over 100 days to drift the same distance1 (Finney, 1994a). Larger scale phenomena such as the El Niño Southern Oscillation (ENSO) have also been theorized to have played a part in the search and settlement strategies of Polynesian voyagers (Finney 1985).

voyaging by computer simulations than was previously possible. Early on in the experimental archaeology effort, investigation of traditional non-instrument navigation and the reconstruction of Polynesian canoes were carried out separately. Lewis employed traditional Micronesian noninstrument navigation in 1966 on a voyage of several thousand kilometers through the South Pacific (Lewis 1994) by interpreting clues from winds, waves, stars and biological indicators. Since that time, continuing research by Lewis and others (Gladwin 1968; Thomas 1987) has shown that, despite a lack of a method for longitudinal reckoning, non-instrument navigation techniques are accurate and quite applicable to voyages of hundreds or even thousands of kilometers across open ocean. The studies have found that, despite Sharp’s assessment, errors in navigation are not cumulative over the course of a voyage. Instead, navigators constantly refine and recalculate their position as new data becomes available, quite often with stunningly accurate results (Lewis 1994).

For both cultural and scientific reasons, construction of double-hulled replicas and the undertaking of traditional– style voyages is continuing throughout the Polynesian community. In previous voyages, the Höküle‘a has traveled to Aotearoa (New Zealand) and the newer Hawai‘iloa, along with Höküle‘a has traveled to and from Hawai‘i and central Polynesia. PVS completed the navigation of the Polynesian triangle with a trip to Rapa Nui in 1999. The voyage was broken into stages with the last one, the 2000 km from Pitcairn Island to Rapa Nui, being the longest and most difficult. Due to the easterly course and the tiny size of Rapa Nui, the navigators of this voyage consider it the most challenging passage yet attempted in their 20 years of experimental voyaging (Nainoa Thompson and Pi‘ikea Miller 1998).

In 1967, Finney built and tested the performance of the Nälehia, a reconstruction of a Hawaiian double-hulled canoe (Finney, 1977). The Nälehia was modeled on a near-shore Hawaiian canoe described by a nineteenth century French naval officer. While not a true voyaging canoe, the Nälehia was instrumental in providing data on the sailing capabilities and seaworthiness of Polynesian canoes in general (Finney, 1977).

Overcoming Limitations Previously, the applicability of the results of previous computer simulations in the planning of specific experimental voyages has been limited by three factors. The first factor is the poor quality of the weather data, second is the technique used to model the voyaging environment, and third is the lack of sailing parameters based upon experimental voyages.

In 1973, the Polynesian Voyaging Society (PVS) was founded in Hawai‘i with the goal of building and sailing a working replica of a traditional double-hulled voyaging canoe. The PVS’s first voyaging canoe, Höküle‘a, has since completed voyages of up to 3,200 kilometers between the many archipelagos of Polynesia using methods approximating traditional non-instrument navigation. These voyages and those of other reconstructed Polynesian canoes have provided a wealth of data of the sailing characteristics of those vessels, and also insights into the capabilities and strategies of noninstrument navigation.

Data Quality The simulations of Levison et al. (1973) and Irwin (1992) were based upon readings collected by British merchant and naval vessels between 1855 and 1938. The roughly rectangular study area from 150° east to 90° west, and 50° south to 35° north, was divided into 392 cells of 5° by 5° (Fig. 2). Readings of wind direction were along sixteen compass points (22.5 degrees between directions) and their speed was categorized in terms of the Beaufort scale, a commonly used classification of wind force (readings from 0 to 9). Current readings were classified into the same sixteen directions and six arbitrary levels. The wind and current data were not systematically recorded either temporally or spatially. The readings were taken at various times of the day, month and year by different crew members of different ships using different measurement techniques. In addition, the data were collected primarily along shipping lines, so readings were

Finney et al. (1989) reason these sailing characteristics make it unlikely that long, tacking voyages against the tradewinds were undertaken. Instead, Polynesian navigators likely used short periods of westerlies that punctuate breakdowns in the tradewinds to travel eastward. This “wait for the west wind” strategy has been used in experimental voyages and is also supported by ethnographic and oral histories (Finney 1988). As a result of the continuing series of experimental voyages, Finney (1994a) believes the sweet potato was introduced to Pacific Basin by two-way voyages of Polynesia navigators using this “wait for the west wind” strategy. He points out that double-hulled canoes, such as the Höküle‘a “barring mishaps…could probably cross the 4000 or so miles from Peru to the Tuamotus or the

1

Finney (1994a) and others have pointed out that the voyage of the Kon Tiki was not actually begun on the coast of Peru. Instead, the raft was towed out to sea nearly 80 km in order to avoid near-shore ocean currents and increase the chance of the experiment’s success.

146

SIMULATING POLYNESIAN DOUBLE-HULLED CANOE VOYAGING

much denser in certain areas, while large sections of the study area contain very sparse data. Readings for the

entire study area comprised a database of 800,000 records.

Figure 2: The study areas of Levison et al. (1973) and Irwin (1992) and this simulation. Heavy gray line is the limits of the previous studies, while the black grid marks the 5° by 5° cells used to create the synthesized weather conditions. The dotted line represents the boundary of this simulation. Note that the grid spacing for synoptic data used in this simulation is 2.5° by 2°, though the cells are not shown in the figure. with wind vector data from European Center for MediumRange Weather Forecasts (ECMWF) analyses (Atlas et al. 1996). The combined dataset was created and made available by researchers at the Data Assimilation Office (DAO) of NASA’s Goddard Space Flight Center, who also tested the resulting dataset against buoy and ship readings for accuracy.

The weather database used in this simulation is a portion of a worldwide dataset and ranges from 160° to 100° west and 4° and 32° south, with a grid spacing of 2.5° longitude by 2° latitude (Fig. 2). Readings are every six hours from 7/11/87 through 12/31/962. Wind direction and speed are measured to one decimal point of degrees and meters per second, respectively. These readings comprise a database of more than 4.2 million records. The dataset is based on readings gathered by Special Sensor Microwave Imager (SSM/I) instruments aboard three Defense Meteorological Satellite Program (DMSP) satellites collecting data since 1987 (Atlas et al. 1996).

Weather Modeling Techniques Due to the sporadic nature of the collection of the data, Levison et al. (1973) and Irwin (1992) were obliged to aggregate the data and create synthesized, or randomized, weather. Probabilities were determined for wind direction and speed on a monthly basis (Table 1). For example, if a simulated canoe found itself in cell 386c in the month of August, there would be a 7.1% chance of southerly winds. Those southerly winds would blow at force 2 on the Beaufort scale 55.6% of the time. Ocean current speed and direction probabilities were determined on a quarterly basis instead of monthly due to even sparser data availability. Irwin makes no mention of the effect of currents in his simulations. In addition, his experiments

The surface of the ocean reflects a variety of wavelengths of light as wind moves over it. The surface’s roughness is an artifact of surface wind speed, and can be measured using passive microwave detectors. Though the SSM/I instruments had provided large, high-resolution datasets, their applicability was limited by lack of data on wind direction. The SSM/I wind speed data were combined 2 Two months of data are missing from the database: December 1987 and May 1992.

147

BRADLEY M. EVANS

Sailing Parameters

included only the January and July matrices to reflect the entire of the summer and winter seasons. During the simulated voyages in Levison et al. (1973) and Irwin (1992), probability matrices were consulted once a day and canoes were moved according to rudimentary sailing parameters.

Levison et al. (1973) modeled the daily movement of canoes on a simplified relationship between wind speed and canoe speed (Table 2). Data for this table were derived from a variety of sources, including drift voyages of lifeboats from World War II. Results of experimental sailing voyages of reconstructed double-hulled voyaging canoes had not yet been published at the time of their study, though some preliminary data from experimental voyages did inform Levison et al.’s (1973) sailing parameters. The speed of simulated canoes in Irwin’s study is a constant that does not vary with wind speed or direction. The operator of the computer set the speed before beginning the simulation, and it remains unchanged until the end of the simulated voyage.

Cell 386c, August Wind Direction

Wind Force (Beaufort scale) 0 1

NNE NE ENE E ESE SE SSE S SSW SW WSW W WNW NW NNW N Total

2 1 1 1

1 3 5 1 1 1 2 1 1

3 1 1 2 3 2 2

4 3 3 3

5

6 1

7

8

9

1

This simulation uses sailing parameters from Finney (1977) that are based on the actual sailing performance of Nälehia. The speed of the canoe is expressed as a ratio of the true wind speed at a given course off the wind (Fig. 3). While the Nälehia is not a deepwater canoe, the sailing characteristics are probably similar to those of the Höküle‘a in moderate tradewind seas and are probably conservative (Ben R. Finney, personal communication 1998).

1 1 1

2 3

3 3 1 1 1 3 2 6 2 1 3 3 2 1 3 2 1 2 6 4 1 1 3 4 5 1 2 1 2 126 observations

1 1 2

Wind Speed Distance traveled (Beaufort scale) per day (Nautical miles) 0 0 1 12 2 24 3 48 4 72 5 108 6 144 7 168 8 144 9+ 108

1

Table 1: A typical synthesized wind speed and direction matrix for August for a single grid cell. Adapted from Levison et al. (1973). The weather dataset used in this simulation is less a synthesized model of the weather across the study area at a given time, than a snapshot. While interpolations and measurement error are present in the database, checking the readings against ship and buoy readings show the dataset to be a fair representation of wind conditions (Atlas et al. 1996)

Table 2: Distance traveled per day as a function of wind speed. Adapted from Levison et al. (1973). Synthesized versus Synoptic Weather Methodologies

Current is unavailable presently in useful and detailed form, so as in Irwin’s study, it is omitted. This may not be of great importance since currents in the study area are often light (Ben R. Finney, personal communication 1998.). In addition, the effect of current is more important in transequatorial voyages, from Tahiti to Hawai‘i for example, where currents are perpendicular to course and have a larger impact on navigation and planning. Nevertheless, future studies would benefit from the inclusion of a similarly detailed current dataset.

In general, this simulation is straightforward. A canoe sets off on a predetermined course on a direct bearing, or rhumbline, from Pitcairn Island to Rapa Nui. Wind speed and direction determine the distance traveled in six hours based upon the performance of the Nälehia as quantified by Finney (1977). When faced with contrary winds the simulated canoe chooses the optimal tack, that is, the bearing that provides the most distance made good toward the final destination. If forced to move off-course, a new bearing to Rapa Nui is calculated and that becomes the desired course. As in previous simulations, Rapa Nui

148

SIMULATING POLYNESIAN DOUBLE-HULLED CANOE VOYAGING

is considered to lie within a circular sighting-radius of approximately 50 kilometers. When a simulated canoe enters this circle, the voyage is considered successful and ends.

Discussion The two methodologies produce sets of courses that are significantly different. Four courseline variables were compared between the synoptic and synthetic methodologies: length of trip, number of course changes necessitated by contrary winds, and the furthest latitude reached to the north and south during the voyage. Taking the two samples as a whole, the four variables show significant differences when t-tests are performed on variable pairs. In all four cases, the probability is well below .05 (Table 3). Each variable shows a significant difference between the synoptic and synthetic methodologies. The synoptic methodology shows much wider variability. The coefficients of variation for each of the variables using the synoptic methodology are from two to five times as great as those using the random weather methodology (Table 4).

The weather database was applied in two ways; first, using a synthesized methodology similar to that of earlier studies; and second, using synoptic weather with voyages experiencing actual weather at six-hour intervals. When using synthesized weather, winds are applied on a monthly basis as in Levison et al. (1973), so if a voyage begins in January, winds are taken at random from any January hour, day and year in the entire ten year database, for that coordinate. Synoptic weather is applied sequentially on a six-hourly basis. A variety of details are kept for each voyage and stored in a database. These details include length of the trip, number of course changes, and the points furthest south and north reached on the voyage. The latter two parameters are recorded to find which of the two methodologies will produce courses that vary more from the reference rhumbline course. Additional parameters such as waiting a predetermined period for sustained westerlies, taking down sails above a certain wind speed, destruction of the canoe in winds over a certain speed, and navigator error are built into the simulation and can be manipulated by the user. While these parameters can be used to create still more databases available for analysis, discussion of these parameters will occur in other forums.

Synoptic Synthesized Mean Mean t-value Length (Days) 18.204 17.453 10.298 Standard (3.816) (2.066) Deviation Course Changes 43.807 Standard (22.528) Deviation

p 0

40.553 (11.276)

7.691

0

28.081

27.492

29.897

0

(1.094)

(0.430)

Northernmost (Degrees) Standard Deviation

24.054

24.653

-25.692

0

(1.3)

(0.492)

N

3351

3600

Southernmost (Degrees) Standard Deviation

Table 3: T-tests of courseline attributes created using synoptic and synthesized weather methodologies. These results can best be interpreted by comparing the courses produced by both methodologies to actual courses of experimental voyages (Fig. 4). Synthesized, random weather allows the simulated canoe to stay relatively close to the original reference course since there are no periods of sustained winds from a single direction to consistently push the canoe off-course (Fig. 5). Courses created using the synoptic weather methodology are more similar in many respects to the tracks of actual experimental voyages (Fig. 6). The most

Figure 3: Sailing parameter curve used in this simulation. Canoe speeds (gray line) are calculated as a ratio of wind speed at given courses off the wind. Data are derived from the performance of Nälehia. Adapted from Finney (1977).

149

BRADLEY M. EVANS

apparent similarity is the shape of the course. A veering wind shift is obvious in both figures 4 and 6 as lowpressure systems move from west to east south of the canoe. The easterly tradewinds are disrupted and the canoe responds by making course corrections as the wind appears to move counterclockwise around the compass.

Length (days) Course changes Southernmost point (Degrees) Northernmost point (Degrees)

Seasonal variations in the length of simulated voyages are evident, as are yearly variations. Westerly winds occur more often during the Austral Summer (December March), thus one would expect faster voyages to occur in this period. In fact simulated voyages using synoptic weather show just the opposite: the winter month voyages (June through September) are generally faster, with fewer course changes on average, and remain closer to the reference course than voyages made in the summer months (Fig. 7). This is due to the nature of those westerlies. Summer westerlies occur during the hurricane season and are blustery creating rough seas that actually impede canoe travel. Winter westerlies, while rarer, are safer and faster. This winter strategy was used successfully on the long easterly voyage of the Hokule’a from Samoa to Tahiti in 1986 (Finney, et al. 1988).

Synoptic methodology, Coefficient of variation 0.21

Synthesized methodology, Coefficient of variation 0.12

0.51

0.28

0.04

0.01

0.05

0.02

Table 4: Coefficients of variation for courseline attributes created using synoptic and synthesized methodologies.

Figure 4: The course of the Höküle‘a, from September 21 to October 9, 1999. The circles are the sighting radii for Pitcairn (left) and Rapa Nui (right). Horizontal line is 25° South latitude, vertical lines represent 130° and 110° degrees west. Small numbers along course are the day numbers. Adapted from Finney (2001).

150

SIMULATING POLYNESIAN DOUBLE-HULLED CANOE VOYAGING

Yearly variations are not as great as the seasonal variations and are not as easy to interpret (Fig. 8). The blustery and dangerous westerlies common in the Austral Summer are even more pronounced during ENSO events. Even still, some have speculated on the possible effects of ENSO on the settlement of Polynesia, especially with regards to the apparent early colonization of the Marquesas Islands vis a vis the more westerly Society Islands (Finney 1985). Unfortunately, the database used

in this simulation does not cover any major ENSO event. The 1992-1993 winter was a mild ENSO, and simulated voyages in that period are among the lowest on average, but more years of data are needed for comparison before this simulation methodology might provide more conclusive results.

Figure 5: A typical courseline generated using the stochastic wind model and sailing parameters from Levison et al. (1973). The straight line is the rhumbline reference course. The circles are the sighting radii for Pitcairn (left) and Rapa Nui (right). Horizontal line is 25° South latitude, vertical lines represent 130° and 110° degrees west. Small numbers along course are the day numbers. Lines intersecting the courseline represent wind direction and strength. simulation may not suffer from these drawbacks. Additionally, an exciting feature of this simulation is the opportunity to compare actual weather and sailing data compiled during the upcoming experimental voyage with simulated voyages in order to both test its utility and further to refine the simulation.

Conclusion This simulation represents a new step in the use of computer simulation in investigating issues of Polynesian settlement. Unlike earlier computer simulations, this study uses detailed synoptic wind data and sailing parameters derived from experimental voyages that were unavailable to previous researchers.

The simulation in turn can then be further refined to help answer broader questions of Polynesian settlement. For instance, the similarity of courselines created using synoptic weather to those of actual voyages suggests that the exploration of strategies like waiting for the west wind may be more effectively modeled within this newer methodology. More basically, this newer methodology may prove helpful in reconciling the archaeological record with new perspectives on the strategies and methods of Polynesian voyaging.

One of the common complaints about output from computer modeling in anthropology, and archaeology in particular, is the lack of utility. This springs from what Aldenderfer calls “the inherent flaws and fundamental ‘noisiness’ of archaeological data (Aldenderfer 1991). The inclusion of detailed, remotely-sensed, synoptic weather data and sailing parameters derived directly from experimental archaeology hold out the promise that this

151

BRADLEY M. EVANS

Figure 6: A simulated voyage beginning on September 22, 1994 using the sequential weather model and the sailing parameters of Finney (1977). The circles are the sighting radii for Pitcairn (left) and Rapa Nui (right). Horizontal line is 25° South latitude, vertical lines represent 130° and 110° degrees west. Small numbers along course are the day numbers. Lines intersecting the courseline represent wind direction and strength.

Figure 7: Length of voyage from Pitcairn to Rapa Nui, using synoptic weather methodology, by month.

152

SIMULATING POLYNESIAN DOUBLE-HULLED CANOE VOYAGING

Figure 8: Length of voyage from Pitcairn to Rapa Nui using synoptic weather methodology, by year.

Atlas, R., Hoffman, R.N., Bloom, S.C., Jusem, J.C., and Ardizzone, J. 1996 “A Multiyear Global Surface Wind Velocity Dataset Using SSM/I Wind Observations”. Bulletin of the American Meteorological Society 77 (5): 869-82. Bellwood, P. 1979 Man's conquest of the Pacific: The Prehistory of Southeast Asia and Oceania. Oxford University Press, New York. Buck, P.H. 1938 Vikings of the Pacific. University of Chicago Press, Chicago. Evans, B.M. 1999 Simulating Polynesian Double-Hulled Canoe Voyaging: Combining RemotelySensed and Experimental Data. Unpublished Master, Honolulu. Finney, B. 1977 “Voyaging Canoes and the Settlement of Polynesia”. Science 196 (4296): 1277-85. Finney, B. 1985 “Anomalous Westerlies, El Niño, and the Colonization of Polynesia”. American Anthropologist 87: 9-26.

Acknowledgments. The Data Assimilation Office (DAO) of the NASA Goddard Space Flight Center graciously provided the synoptic weather data that made this project possible, and special thanks go to Joseph V. Ardizzone who helped to pare down the worldwide dataset to a useable size and form. I am grateful to the Bernice P. Bishop Museum for its support of this research. I would also like to thank Matthew McGranaghan and Shannon Patrick McPherron for comments on drafts of this manuscript.

References Akerblom, K. 1968 Astronomy and Navigation in Polynesia and Micronesia: A Survey. Monograph Series, Publication No. 14. The Ethnographical Museum, Stockholm. Aldenderfer, M. 1991 “The Analytical Engine: Computer Simulation and Archaeological Research”. In M. Schiffer (ed.); Archaeological Method and Theory, Volume.

153

BRADLEY M. EVANS

1988

“Voyaging Against the Trades: A Report of an Experimental Canoe Voyage from Samoa to Tahiti”. American Anthropologist 90: 401-405. 1991 “Myth, Experiment, and the Reinvention of Polynesian Voyaging”. American Anthropologist 93: 383-404. 1994a “Polynesian Voyagers to the New World”. In Man and Culture in Polynesia 10: 1-13. 1994b Voyage of Rediscovery: A Cultural Odyssey Through Polynesia. University of California Press, Berkeley. 2001 “Voyage to Polynesia's Land's End”. Antiquity 75: 172-181. Finney, B., P. Frost, R. Rhodes, and N. Thompson 1989 “Wait for the West Wind”. Journal of the Polynesian Society 98: 261-302. Gladwin, T. 1968 East is a Big Bird: Navigation and Logic on Puluwat Atoll. Harvard University Press, Cambridge. Heyerdahl, T. 1968 Sea Routes to Polynesia. George Allen and Unwin: London. Irwin, G. 1992 The Prehistoric Exploration and Colonisation of the Pacific. Cambridge University Press, Cambridge. 1998 “The Colonisation of the Pacific Plate: Chronological, Navigational, and Social Issues”. Journal of the Polynesian Society 107 (2): 111-43. Irwin, G., S. Bickler and P. Quirke 1990 “Voyaging by Canoe and Computer: Experiments in the Settlement of the Pacific”. Antiquity 64: 34-50. Levison, M., R.G. Ward and J.W. Webb 1973 The Settlement of Polynesia: A Computer Simulation. University of Minnesota Press, Minneapolis. Lewis, D. 1994 We, the Navigators: The Ancient Art of Landfinding in the Pacific. University of Hawai‘i Press, Honolulu Sharp, A. 1963 Ancient Voyagers in Polynesia. Angus and Robertson, Sydney. Thomas, S. 1987 The Last Navigator. Henry Holt and Company, New York.

154

BAR S1802 2008

Canoes of the Grand Ocean Edited by

Anne Di Piazza and Erik Pearthree DI PIAZZA & PEARTHREE (Eds) CANOES OF THE GRAND OCEAN

B A R

BAR International Series 1802 2008