Volume 83 1974 > Volume 83, No. 2 > Outrigger ages, by Edwin Doran Jnr., p 130-140
                                                                                             Previous | Next   

- 130

For more than 60 years the double outrigger canoe has been considered by most authorities to be the earliest craft used in the Pacific, although the sailing raft is sometimes suggested as a predecessor. It is the purpose of this paper to refute the conventional wisdom and to suggest that the double outrigger developed much later than the single outrigger canoe and double canoe. The culture history of Oceania is more easily understood with this assumption.

In order to present the argument, a few moments must be spent on the dynamics of sailing, because the way in which a boat is sailed may be as important as form in classifying boat types. To progress to windward sailboats must sail a zig-zag course because none can sail closer than about 45 degrees to the wind. Except for parts of the Indo-Pacific region, boats change from “zig” to “zag” by a procedure called “tacking”, during which the bow swings from the original course across the wind and over to the new course (another technique, “wearing”, is irrelevant here). A fundamentally different procedure is used in Micronesia and certain other areas. This technique, which will be called “shunting”, alternates ends of the boat as the bow, requires major relocation of mast and sail, and results from the desirability of keeping the outrigger float (or smaller hull of some double canoes) on the windward side. 1

In the entire area from Madagascar to Easter Island which is occupied by Austronesian-speaking people there are only three types of multi-hulled water craft as defined by hull form: double canoes, single outrigger canoes, and double outrigger canoes (Fig. 1). 2

Double outrigger canoes are always tacked, but single outriggers and double canoes may be either tacked or shunted. The five mathematically possible combinations of hull form and tacking technique may be resolved into three basic cultural groups: Polynesian craft are either double canoes or single outriggers, have some variant of the Oceanic sprit type of sail stepped well forward of amidships, and are tacked head to wind. Micro-

- 131
Hull and sail types in the Pacific

nesian craft are single outrigger canoes (with the exception of double canoes in Fiji and vicinity) which use Oceanic lateen sails and shunt from course to course when going to windward. Indonesian craft are double outriggers with rectangular lug sails which tack head to wind. The craft of Melanesia are mostly single outriggers sailed by the shunting technique, but using mat sails of the rectangular lug type. They are considered to be a transitional mixture of Micronesian and Indonesian influences.


The twentieth century discussion of the relative antiquity of double and single outrigger canoes in a dozen papers and monographs can be - 132 reduced to manageable proportions by presenting a chart (Fig. 2). Hornell's examination of the evidence convinced him that Madagascar double outrigger canoes were older than single outriggers. 3 The judgement was based on historical records indicating a decline in the numbers of double outriggers from about 1600 to 1850 and on a typological transition between the two kinds of boats. He later repeated the same arguments, 4 and, to my knowledge, no one has questioned them since. However, the historical evidence is not compelling, and his typology is as convincing in reverse as the way he presented it. On the Madagascar evidence alone either double or single outriggers could be the older.

The modern argument for greater age of double outriggers in the Pacific continues a hypothesis advanced by Friederici and earlier

Summary of the arguments on the relative ages of double and single outrigger canoes
- 133

scholars. 5 In 1920 6 Haddon pursued the argument, basing his idea on an early record of double outriggers at Boro Bodur (c. A.D. 800), the alleged peripheral location of double outrigger “relicts” in the overall outrigger distribution, and on the inferior sailing quality of double outriggers. In the same year, Hornell took the opposite position, that double outriggers were of lesser age, on the basis of the central distribution of double outriggers and on their superior seaworthiness. 7 He repeated the distributional argument in 1923, 8 then 13 years later disproved allegations that double outriggers had been found in peripheral locations in the Marquesas, Easter Island, and Palau. 9

Sometime between 1923 and 1938 Hornell abandoned his position and accepted that of Haddon in their joint volume of 1938. 10 In his final magnum opus, Water Transport, he again accepted Haddon's position that double outrigger canoes were older. 11 Finally Casson 12 simply repeats what has become the conventional wisdom of the last 35 years.

I shall attempt to prove that Hornell was correct in his 1920 and 1923 assertions of the lesser age of double outriggers. The central distribution of double outriggers, historical knowledge of culture thrusts, and recent archaeological evidence all point in this direction. The demonstrated seaworthiness of double outrigger canoes aids in the argument.


The distribution of canoe types in the Pacific c. 1750 is shown in Figure 3 (see App. 1 for sources). Typical Polynesian craft are found to the east. All canoes from Palau and the Marianas south-east to Fiji are typically Micronesian. To the west, in Indonesia, the Philippines, and as far east as Torres Strait, are Indonesian craft.

Culture thrusts known from good evidence are shown by arrows. Indonesian movements to the east along both the south and north coasts of New Guinea are known historically, 13 as is the changeover to the Micronesian canoe type in New Caledonia, and the Ellice, Tokelau, and Samoa Islands. 14 An earlier thrust from the New Hebrides north into Micronesia, thence west in the Carolines is suggested strongly by linguistic evidence. 15 The long-postulated movements to New Zealand, Hawaii, and Easter Island from somewhere in the Society-Marquesas area are now more firmly fixed archaeologically by both radio-carbon dating and similarities in culture complexes. 16

Other thrusts can be hypothesised on a more speculative basis. The

- 134
Distribution of canoe types in the Pacific

use of shunting as a sailing technique in Aitutaki, Manihiki, and the Tuamotus 17 suggests eastward movement of this idea, albeit in unsophisticated form. The single outrigger canoes of Melanesia are all shunted from course to course using Micronesian technique but nearly all have rectangular sails of probable Indonesian origin. 18 Double outriggers in Nissan and the Louisiades and the complexly built mon canoe of the Solomons which is similar to boats of the Moluccas also suggest transmission eastward of ideas from Indonesia. 19

A number of isolated boat types which are dissimilar to those around them are hypothesised as relicts of former distributions. The double canoes of Mailu in south-east New Guinea 20 and the double canoe models of Nanusa 21 and Truk 22 are probably relicts, especially when one notes that double canoes, presently or historically, are widely - 135 distributed in India, Indo-China, and China. 23 The Oceanic lateen sails of the Trobriands and at least eight occurrences of single outrigger canoes in Indonesia 24 also suggest a formerly much wider distribution of the Micronesian type. This assertion alone may be weak but when one notes the use of shunting in Ceylon by single outrigger canoes 25 and the suggested use of shunting by Madagascar single outriggers 26 a more substantial hypothesis appears.

On the basis of distributions of types and with cautious use of the Age-Area hypothesis one can suggest that the peripheral double canoes and single outriggers of Polynesia are the oldest type. Somewhat less peripheral and therefore younger is the Micronesian type of shunting single outrigger, not forgetting its analogues in Ceylon and Madagascar. Most central and presumably the latest in time is the double outrigger of Indonesia. Its probable late extension to Madagascar need not negate this assertion.

Hypothesised time perspective for Pacific canoes along a profile from Indo-China to the Marquesas. (See Appendix 2 for sources numbered above.)
- 136

When known and likely culture thrusts are added to the distributional picture, and when atypical craft are hypothesised as relicts, a considerably firmer hypothesis of successive culture waves eastward into the Pacific emerges. The east to west “backlash” in Micronesia seems intermediate in age. Additional confirmation is derived from archaeological data of the last 20 years.

For a number of Pacific Islands, the earliest known radiocarbon date of occupation is shown in Figure 4 with a time profile from Indo-china to the Marquesas drawn beneath. (Sources from which Fig. 4 was constructed are provided in Appendix 2). 27 A most interesting development is the recognition of a pottery type, Lapita, which has been found in the last two decades from New Britain to Fiji, with earliest dates everywhere about 1000 B.C. Obsidian from the Talasea Peninsula of New Britain was traded at the same time east to the Santa Cruz Islands. Close relationships between Lapita and Kalanay pottery in the Philippines and the Sa-huynh pottery in Indo-China have been suggested though later challenged by Golson. 28 Lapita ware has been excavated from the 1000-500 B.C. level in Tonga. 29 and together with undecorated ceramics from Samoa and the Marquesas it demonstrates an early and rapid push eastward of pottery-using people, among whom the art was lost later in the more eastern islands.

Upon the facts above I have overlaid a canoe hypothesis which fits all the evidence currently known. Modern distributions are shown in solid tint along the 1800 line, and the one firm prehistoric boat date, for double outriggers at Boro Bodur about A.D. 800, has been used to sketch a time-longitude area for Indonesian craft. The dashed time-space lines outlining Micronesian and Polynesian craft are only reasoned conjecture, but one point is clear. Haddon's use in 1920 of the Boro Bodur date to argue for greater antiquity of double outriggers is invalid. There was ample time for Polynesian and Micronesian craft to have preceded double outriggers in South-east Asia and Indonesia.

The upward slant of all lines from older in west to younger in east seems unlikely to be questioned because everyone, including Heyerdahl, agrees that most Oceanic traits came from South-east Asia and must, therefore, have been earlier in South-east Asia. The progressively slower rates of eastward movement, indicated by greater slant of lines, can be explained tentatively by increasing cultural resistances through time. The earliest push, about 1000 or 1500 B.C., into presumably unoccupied parts of the Pacific, was surprisingly rapid. Replacement of accustomed Polynesian craft by the faster Micronesian type occurred more slowly. Double outriggers moved even more slowly, perhaps because of Buddhist or Moslem traits that accompanied the boat type.

- 137

Assertions by Haddon and Hornell about the seaworthiness of various canoe types appear to be very largely just that—without evidence. Although Haddon's static descriptions are excellent, he seems to have known very little about sailing. Hornell was more knowledgeable about seamanship and the sailing characteristics of the boats he describes. The fact that the latter originally thought double outrigger canoes were seaworthy and later changed his mind is more likely a reflection of deference to Haddon than his real feeling on the matter.

To deal with a complex technological matter in brief space, I can do little beyond making my own unequivocal assertion that double outrigger canoes are at least as seaworthy, if not more so, than single outriggers and double canoes. Modern basnig in the Philippines, great 70- to 100-foot double outriggers, have been observed at sea proceeding quite comfortably in 12- to 15-foot waves. 30 Contrary to assertions about weakness, my personal sailing experience with both modern trimarans and Philippine vintas indicates that the lee float does not break off when it is submerged by waves. Although the static stability of all multi-hulled craft is demonstrably similar, the dynamic stability of double outriggers is greater. The lee float produces hydrodynamic lift, and, additionally, as the float submerges the capsizing rate is slowed by several crucial seconds which permit corrective action to be taken. Finally, both from historical documents and from my measurements of the two types, it can be stated that double outriggers can be as fast as or faster than single outriggers. 31


On the basis of the arguments presented above, it is hypothesised that Polynesian craft are the oldest in the Pacific, that Micronesian craft are younger, and that Indonesian double outrigger craft are still younger. These age relationships fit logically into the widely accepted pattern of successive culture waves, of people or traits or both, from west to east in the Pacific. The relative boat chronology suggested here should be of assistance in future efforts to work out details of Pacific culture history.


The degree to which a map such as Figure 3 should be documented is a scholarly puzzle. To cite each reference (for example 39 different pages in Hornell 1936, and 40 different pages in Haddon 1937) is cumbersome and would add undesirable clutter to the map. On the other hand, the reader should not be asked to accept a distribution on faith alone. This quandary is solved here by listing the references from which the map was compiled. The author retains the detailed list of citations and will supply - 138 it to any person desiring specific documentation on any part of the distribution.

  • Sources:
  • Advanced Research Projects Agency 1962
  • Haddon 1920
  • Haddon 1937
  • Haddon and Hornell 1938
  • Hornell 1920c
  • Hornell 1936
  • Ling 1970
  • Suder 1930
  • 1. Golson 1968
  • 2. Ambrose & Green 1972
  • 3. Barbetti & Allen 1972
  • 4. Howard 1967
  • 5. Davidson et al, 1967
  • 6. Green 1968
  • 7. Sinoto 1968
  • 8. Groube 1968
  • 9. Suggs 1960
  • 10. Green 1967
  • 11. Hornell 1920c
  • 12. Solheim 1964
  • ADVANCED RESEARCH PROJECTS AGENCY, 1962. The Junk Blue Book: A Handbook of Junks of South Vietnam. Washington, Department of Defense.
  • AMBROSE, W. R., and R. C. GREEN, 1972. “First Millennium B.C. Transport of Obsidian from New Britain to the Solomon Islands.” Nature, 237:31.
  • BARBETTI, M., and H. ALLEN, 1972. “Prehistoric Man at Lake Mungo, Australia, by 32,000 Years BP.” Nature, 240:46-8.
  • CASSON, Lionel, 1964. Illustrated History of Ships and Boats. Garden City N.Y., Doubleday.
  • DAVIDSON, J. M. et al, 1967. “Additional Radiocarbon Dates for Western Polynesia.” Journal of the Polynesian Society, 76:223-30.
  • DORAN, Edwin Jr., 1972. “Wa, Vinta, and Trimaran.” Journal of the Polynesian Society, 81:144-59.
  • GLADWIN, Thomas, 1970. East is a Big Bird. Cambridge, Harvard University Press.
  • GOLSON, Jack, 1968. “Archaeological Prospects for Melanesia”, in Yawata, & Y. H. Sinoto (eds.), Prehistoric Culture in Oceania. Honolulu, Bishop Museum Press, pp. 3-14.
  • —— 1972. “Both Sides of the Wallace Line: New Guinea, Australia, Island Melanesia and Asian Prehistory”, in N. Barnard (ed.), Early Chinese Art and Its Possible Influence in the Pacific Basin. New York, International Arts Press, Vol. 3, pp. 533-95.
- 139
  • GRACE, George W., 1964. “The Linguistic Evidence”, in K. C. Chang, G. W. Grace, and W. G. Solheim II, “Movement of the Malayo-Polynesians: 1500 BC to AD 500.” Current Anthropology, 5:361-8.
  • GREAT BRITAIN NAVAL INTELLIGENCE DIV., 1944. Netherlands East Indies, Vol. II. Geographical Handbook Series, B. R. 518 A. London.
  • GREEN, R. C., 1967. “The Immediate Origins of the Polynesians”, in G. A. Highland, et al (eds.), Polynesian Culture History. Honolulu Bishop Museum Special Publication 56, pp. 215-40.
  • —— 1968. “West Polynesian Prehistory”, in I. Yawata, & Y. H. Sinoto (eds.), Prehistoric Culture in Oceania. Honolulu, Bishop Museum Press, pp. 99-109.
  • GREEN, Roger and Marion KELLY (eds.), 1970. Studies in Oceanic Culture History, Vol. 1. Pacific Anthropological Records No. 11. Honolulu, Bernice P. Bishop Museum.
  • —— 1971. Studies in Oceanic Culture History, Vol. 2. Pacific Anthropological Records No. 12. Honolulu, Bernice P. Bishop Museum.
  • GROUBE, Leslie M., 1968. “Research in New Zealand Prehistory since 1956”, in I. Yawata, &. Y. H. Sinoto (eds.), Prehistoric Culture in Oceania. Honolulu, Bishop Museum Press, pp. 141-49.
  • —— 1971. “Tonga, Lapita Pottery, and Polynesian Origins.” Journal of the Polynesian Society, 80(3):278-316.
  • HADDON, A. C., 1920. “The Outriggers of Indonsian Canoes.” Journal of the Royal Anthropological Institute, 50:69-134.
  • —— 1937. The Canoes of Melanesia, Queensland, and New Guinea, in A. C. Haddon and James Hornell, Canoes of Oceania, Vol. 2. Honolulu, B.P. Bishop Museum Special Publication 28.
  • HADDON, A. C. and James HORNELL, 1938. Definition of Terms, General Survey and Conclusions in Haddon, A. C. and James Hornell, Canoes of Oceania Vol. 3, Honolulu. B.P. Bishop Museum Special Publication 29.
  • HORNELL, James, 1920a. “Les Pirogues à Balancier de Madagascar et de l'Afrique Orientale.” La Geographie, 34:1-23.
  • —— 1920b. “The Common Origin of the Outrigger Canoes of Madagascar and East Africa.” Man, 20:233-7.
  • —— 1920c. “The Outrigger Canoes of Indonesia” Madras Fisheries Bulletin, 12:43-114.
  • —— 1923. “The Origins and Ethnological Significance of Indian Boat Designs.” Memoirs, Asiatic Society of Bengal, 7:139-256.
  • —— 1934. “Indonesian Influence on East African Culture.” Journal of the Royal Anthropological Institute, 64:305-32.
  • —— 1936. The Canoes of Polynesia, Fiji, and Micronesia, in A. C. Haddon and James Hornell, Canoes of Oceania, Vol. 1. Honolulu, B.P. Bishop Museum Special Publication 27.
  • —— 1943. “The Fishing and Coastal Craft of Ceylon”. Mariners Mirror, 29:40-53.
  • —— 1946. Water Transport, Origins & Early Evolution. Cambridge University Press. Reimpressed 1970 by David and Charles.
  • HOWARD, Alan, 1967. “Polynesian Origins and Migrations”, in G. A. Highland et al (eds.), Polynesian Culture History. Honolulu, B.P. Bishop Museum Special Publication 56, pp. 45-101.
  • LEWTHWAITE, Gordon R., 1967. “Geographic Knowledge of the Pacific Peoples”, in Herman R. Friis (ed.), The Pacific Basin. American Geographical Society Special Publication No. 38, pp. 57-86.
- 140
  • LING, Shung-sheng, 1970. A Study of the Raft, Outrigger, Double, and Deck Canoes of Ancient China, The Pacific, and the Indian Oceans. Nankang, Taipei, Taiwan, Institute of Ethnology, Academia Sinica, Monograph No. 16.
  • ROMMY, Capt. C., 1970. Personnal communication, May 4.
  • SINOTO, Yosihiko H., 1968. “Position of the Marquesas Islands in East Polynesian Prehistory”, in Yawata, I. & Y. H. Sinoto (eds.), Prehistoric Culture in Oceania. Honolulu, Bishop Museum Press, pp. 111-8.
  • SOLHEIM, Wilhem G. II, 1964. “Pottery and the Malayo-Polynesians”, in K. C. Chang, G. W. Grace, and W. G. Solheim II, “Movement of the Malayo-Polynesians: 1500 BC to AD 500.” Current Anthropology, 5:360, 376-84.
  • SUDER, Dr. Hans, 1930. Vom Einbaum und Floss zum Schiff. Veröff. des Inst. für Meereskunde, Neue Folge, B. Historisch-volkswirtschaftliche Reihe. Heft 7. Berlin, E. S. Mittler u. Sohn.
  • SUGGS, Robert C., 1960. The Island Civilizations of Polynesia. New York, New American Library.
1   See Gladwin 1970:96-118 for details and excellent photographs of this manoeuvre.
2   See Haddon 1920; Haddon 1937; Haddon and Hornell 1938; Hornell 1920a and 1920b; and Hornell 1936 for the basic descriptions of these types.
3   Hornell 1920b:136-7.
4   Hornell 1934:320.
5   See Howard 1967 for an excellent summary of work on Polynesian origins, including canoes.
6   Haddon 1920:122.
7   Hornell 1920c:104-11.
8   Hornell 1923:228.
9   Hornell 1936:30, 100, 437.
10   Haddon and Hornell 1938:19.
11   Hornell 1946:269.
12   Casson 1964:208.
13   Great Britain 1944:60-5; Lewthwaite 1967:58-9.
14   Haddon & Hornell 1938:41-2.
15   Grace 1964:367.
16   Sinoto 1968:116, 117.
17   Hornell 1936:86, 167, 187, 188.
18   Haddon 1937.
19   Haddon 1937:108, 117, 254.
20   Haddon 1937:231.
21   Haddon 1920:77.
22   Hornell 1936:410.
23   Suder 1930:75; Ling 1970:165.
24   Haddon 1920:78, 97, 98, 116; Hornell 1920c:77, 89.
25   Hornell 1943:42.
26   Hornell 1920a:10.
27   More recent findings supporting the chronological conclusions of Figure 4 are now available in Green and Kelly 1970 and 1971. See particularly the articles by Bellwood, and Sinoto (1970), and by Shutler, Garanger and Golson (1971).
28   Golson 1972.
29   Groube 1971.
30   Capt. Rommy 1970.
31   Doran 1972:154.