A version of this paper was presented at the "South Seas Symposium, Easter Island in the Pacific Context," Albuquerque, New Mexico August 6-9, 1997. It was subsequently published in the Proceeding of the Fourth International Conference on Easter Island and East Polynesia in 1998.

VENERATED ANCESTORS; STONE SCULPTURES OF THE MARQUESAS ISLANDS, FRENCH POLYNESIA

Sidsel (Cecilia) Millerstrom
Archaeological Research Facility (ARF)
University of California, Berkeley
Berkeley, CA 94720, USA

Stone and wood sculptures of the Marquesas Islands have been described, admired, photographed, sketched, and painted since the arrival of the European to the archipelago in the 1790s . They have not, however, been systematically documented and analyzed until recently. In this essay I present a summery of metric information and empirical observation generated from data collected between 1984 and 1989 while working on the "Marquesan Rock Art Project". The project was initiated by Maeva Navarro, former director of Département d'Archéologie du Centre Polynésien des Sciences Humaines "Te Anavaharau" (presently known as Service de la Culture et du Patrimoine), Tahiti. All fieldwork was directed by Edmundo Edwards, then chief archaeologist with the department, in collaboration with the author. This paper presents some of the results of field survey of eighty-one tiki, documented since 1985. We worked in seventeen valleys on five of the presently six inhabited islands (Nuku Hiva, Ua Pou, Ua Huka, Hiva Oa and Tahuata). Architectural association, gender, stylistic variation, sculptural material, and applied decoration will be discussed. A few statues on Hiva Oa and Tahuata remain undocumented. American archaeologist Ralph Linton of the 1920 Bayard Dominick Expedition mentioned several statues that we so far have been unable to locate (e.g. Linton 1925:142,144-145,147,149. See also Linton 1925:photo XI C). (Figure 1. The author with one of the tiki at paepae Paeke, Taipi Valley, Nuku Hiva)
Fenua Enata Terre des Hommes (Land of the Men) as the Marquesan now wish their islands to be named , are known for their archeological art, especially the tiki. In general the word tiki refers to all anthropomorphic depiction on mediums such as stone, wood, cloth, and human skin (Dordillon 1931:384). It also refers to the large stone and wood statues (Handy 1927:121f, 1923:265). In local folklore the celebrated Tiki was the human creator and a legendary incestuous god both in tradition and religion (Suggs 1966:144). But the images do not represent Tiki, the human creator; according to historic writings the tiki depicted deified ancestors (Tautain 1887:673-678; Handy 1923:236; Delmas 1927).
The majority of the images, 65.5 percent, are found on the island of Hiva Oa. In general, sacred sites are larger and more variable on Hiva Oa than on any other island (Linton 1923:446, 1925:37). The reason for this is unclear. Few statues are left on Fatu Hiva and Tahuata, but several existed in the past (e.g. Von den Steinen 1928(2):75; Linton 1925:179, 180-180, 185, Plate XIC))
Forty-three tiki were documented in situ. The remaining thirty-eight tiki have been relocated and the archaeological associations are questionable. However, general provenience of some of these have been traced through earlier publications (e.g. Ferdon 1965; Linton (1925:Plate XI,E). But we will never know how many statues were relocated in the past. For instance, Steward, a missionary who visited the islands in 1828, described how members of the Haapa destroyed a temple (me'ae) in Taiohae (Nuku Hiva) and stole the idols during a skirmish between them and the Teiis tribe (Stewart 1832:149).

Classification
Based on morphology, 4 main categories are distinguished: I. Statues carved in full round; II. Modified boulders; III. Rectangular blocks; IV. Double figures. One seated figure and two "stretched out figures" constitute the "other" category.
Considering that no systematic survey of the statues has been undertaken, we were interested in knowing how prevalent the so-called "classic tiki" was. We thus examined the most common visual characteristics such as facial characteristics, protruding stomach, arm placement and the bent knee stance.
The head, because it was considered tapu (sacred, forbidden) and the seat of mana (supernatural power), was the most important part of the statue, and it is carved with careful attention to the smallest details. It may include intricate facial features and tattoos and headbands, while the rest of the body is only outlined. Typically, the size of the head is 1/3 or more of the body's height. It is usually resting directly on square shoulders lacking a neck. Round large and bulging eyes are often encircled with a 1-2 cm wide raised rims under high arched eyebrows (Figure 2). Sometimes the eyes have a raised or incised curved line that bisects them from the outer part of the eye to the inner corner of each eye. Some tiki have circular indented pits indicating the pupil .
A broad flat nose with wide nostrils fills out most of the center of the face. The outlines of the nose go upwards and connect with the eyebrows. High eyebrows reach down on the outer side of the head and link with the ears - resembling the stem of eye-glasses.
A long, wide, and oval mouth, sometimes with a protruding tongue or even teeth with great canines, covers the lower part of the face. Breasts and nipples are often depicted. Some of the images have the arms separated from the torso, while on others the arms are barely indicated.
Hands with fingers are resting on a protruding abdomen. Ritual knowledge and oral tradition were believed to be held in the stomach (Tautain 1887), thus hands placed on a protruding belly, may have been a way to protect these memories (Thornton 1989).
Wide rounded hips, short stubby and flexed legs rest on a pedestal. Some images have the legs only indicated while other sculptures are cut off below the waist. Raised circular knobs often indicate ankles. Only a few images depict toes. The back is often carved with spine and buttocks even when the statues are fitted into a stone platform.
Large round eyes, arms placed over the abdomen and a bent-knee stand are major common tiki attributes. For unknown reasons a protruding stomach is less prevalent. We have thus validated the "Classic Tiki
Stylistically the Marquesan tiki are remarkable similar obviously following certain social rules. However, numerous variations exist suggesting that each tiki not only symbolized an important deceased ancestor but may in fact, represented a specific progenitor.
The tiki vary in size from 13.7 to 250 centimeters above the ground. The majority of the statues, or 50.6%, measure between 50 to 100 centimeters (Figure 3). To estimate the total height of the tiki is impossible without excavation. Some statues have a pointed extension below the pedestal in order to secure it in the ground.

Manufacturing material
Fifty-eight percent of the statues are carved of a volcanic tuff; the red variety, or 38.27%, was especially preferred. Traces of red pigment were noticed by Linton in protected parts of a statue on me'ae Iopona, Hiva Oa (Linton 1925:162); we noticed pigment on an image in Punaei Valley, Hiva Oa (Millerstrom and Edwards ms 1985). This has led us to believe that basaltic stone statues were periodically painted red as red was considered a sacred color and was associated with high status throughout Polynesia (e.g. Handy 1927; Burrow 1938; Firth 1967; Lee 1992).
Several red tuff quarries are noted in the islands (Linton 1925:8-9), though undoubtedly there are others unknown to us. A red tuff quarry are located in Puamau (Hiva Oa) near me'ae Iopona. In 1920, Linton observed carved outlines of unfinished tiki (1925:74,). Only a quarry at the northeast bay of Ha'ata'ive'a (Suggs 1961:66) are known on Nuku Hiva. The red tuff horizontal stratum at Ha'ata'ive'a is from 5 to 8 m (7.62 m) thick (Figure 4). Horizontal outlines with undercuts of rectangular slabs, probably removed to carve tiki are still visible on the northern side (Suggs 1961:66; Millerstrom ms 1992). A yellowish tuff strata in the same area show three vertical niches in the cliff face where slabs for statues were removed (Millerstrom ms 1992) .

Decoration
Carved headdresses (hei) are the most common form of decoration, but wreath around the necks, tattoos, short mantles, and hair-knots on each side of the head also occur. Some tiki have drilled circular perforation in the earlobes for placement of earplugs. Takai'i, at me'ae Iopona, Puamau Valley (Hiva Oa), the largest Marquesan image, has tattoos on the chest and legs, what appears to be a hair knot at the back, and some 27 vertical grooves in the eyes, emphasizing their size (Millerstrom and Edwards ms 1985). Tattoos, when they occur, are placed on the sides of the mouth, chest, or on the thighs. Linton (1925:74) mentions ornamental grooves, circles and chevron patterns on the abdomen and buttocks on some images, but these are no longer visible.
We do not know if the neck wreath represents plant or marine material. Perforated shell tabs, pearl shell pendant, fish, porpoise and whale teeth are found archaeologically (Suggs 1961:134; Rolett and Conte 1995:215). In fact, whale teeth were so treasured that from AD 1100 onwards imitation whale teeth were carved from the lips of Cassis shell (von den Steinen 1928(2):22; Suggs 1961:136). Carved wreaths of perishable plant material are difficult to ascertain archaeologically. Petard (1996:82), a French biologist, reports that garlands were strung from the pandanus (pandanus tectorius) fruit to decorate the body during feasts. The Polynesians, according to the various parts the tree and the changing colors of the ripening fruit, recognized several varieties of the pandanus. In the Marquesas the red fruit ha'a kua, was reserved for hei to decorate the tiki and hence tapu for commoners (Petard 1986:84).
Tiki were dressed in bark (tapa) loincloth (Dordillon in Handy 1923:238). These were ceremonially made. Stewart (1832) describes old inspirational priests (tuhuka) beating loincloth for the gods. The act was so sacred that the priests could only eat in the evening (Steward 1832; Handy 1923:238). Offerings of human victims, animals, fruit and vegetables were placed in front of the statue or hanging in nearby trees (Stewart 1832:223; Handy 1923:238). Several of the images we have documented have a flat head that may have been a space to place decoration such as a crown of shells, teeth, bones or plant material.
Gender variation
It is generally assumed that all the tiki represent men. This is not the case. The majority - 72.8 percent - are genderless. Of the remaining statues 19.8 percent depict male while 7.4 percent depict females. For example, at paepae Paeke, Taipivai, 4 of the 12 tiki represent females and 4 are males. The sex of the remainder is un-determined. Linton, however, writes that all the images represent females (1925:116).
Documentation of the sculptures show that the size of the male genitalia are not exaggerated nor is there evidence that they have been deliberately mutilated as some literature suggests. It is possible that some of the male statues were re-carved from male to female; but at present we have no clear evidence to support this speculation.

Archaeological association
The majority of the images are associated with shrines, chief's platforms or temples located within public tribal complexes (tohua). In Nuku Hiva the statues are placed on or in the walls of platforms usually located on one end of the ceremonial plaza. Some are directly associated with chief's house platforms. A chief's house is sometimes referred to as a me'ae in the historic literature. It is conceivable that a chief's house became so imbued with mana that after his death the site became tapu. This is the case in Tikopia, a Polynesian Outlier (Kirch 1996). The sculptures then, appear to have been part of the upper-class stratum, and a visual display of status and political authority.
The age of the images is uncertain, as none has been dated directly. It appears, however, that the statues were developed in the late prehistoric period and that they continued to be carved into the late historic times. According to informants in 1880s who could recite approximately 25 generations, von den Steinen (1928(2):84-86), calculated that Iopona me'ae, with numerous sculptures, was constructed circa AD 1700-1750. Excavation at the same site by Heyerdahl (1965:123-150), leader of the Norwegian Archaeological Expedition in 1956, yielded three significantly earlier radiocarbon (uncalibrated) dates for the site. Ferdon Jr. (1965:117-121), also a member of the Norwegian Archaeological Expedition, collected early radiocarbon dates at Paeke, Taipivai. While these dates indicate the occupation of the sites, is uncertain if the dates reflect the age of the sculptures.
Suggs archaeological work on Nuku Hiva in the late 1950s, wrote that sculptures are most often associated with megalithic structures (paepae), which developed after approximately AD 1400 (Suggs 1961:155). Suggs furthermore, opined that sculptures of volcanic tuff or basalt were first made around AD 1600-1700 century (Suggs 1961). This suggested time-period for the sculptures is more in line with Von den Steinen's genealogy account discussed above.
It appears, however, that the conventional image face was first developed in the petroglyphs and later extended in the sculptures (Suggs 1961; Millerstrom 1997). Stone slabs with a mata komoe or face motif, so common in the Marquesan petroglyphs repertoire (Millerstrom 1997), was excavated in the lowest level of a fire pit by Suggs on Ha'ata'ive'a beach, Nuku Hiva. The valley was first occupied around AD 1100 (Suggs 1961:155).
*Mata, a Proto-Polynesian taxeme refers to eye and face in Tonga, Samoa, the Cooks, Easter Island, and among the Maori (Walsh and Biggs 1966). Maka is a Hawaiian cognate. Other glosses for mata includes genealogy, tribe or status lineage (Millerstrom 1997). These words still retained in Polynesian dialects indicate that the face motif are connected with concepts that go far back in prehistory (Millerstrom and Allen 1994).
The pervasive mata komoe motif expressed on the surface in most media e.g. images on stones and wood, on calabashes, canoes, tattoo, house-posts, bone carvings and so forth identify and intensify the ideology of a group of people (Figure 4). A pervasive art system also promote and reinforce social solidarity in order to maintain a belief systems that bolstered the political position of the hereditary chief, priests or warriors. Pervasive decorative organization experiences a much less rate of change than for example an opposing partitive art systems (DeBoer 1991:157). This theoretical perspective has important implications regarding changes in the Marquesan art system and how these changes reflect social transformation. (Millerstom 1997).
This brief essay present metric data rather than offer adequate interpretation. Several issues remain to be examined. For example, although the Marquesan tiki are remarkably homogeneous, attribute variations occur. Are these stylistic variations due to regional or chronological differences or are the differences due to the type of material used? Furthermore, is it possible to identify the sex of the 72.8 percent genderless statues by isolating specific characteristics of the male and female images? These are only some of the questions we will address in the future.

Acknowledgments
I wish to acknowledge Maeva Navarro, former director of C.P.S.H. (now Service de la Culture et du Patrimoine) Tahiti, for initiating the Marquesan Rock Art Project, for allowing us to conduct research in the Marquesas, and for her support and encouragement. I also thank the people of the Marquesan archipelago who graciously contributed our fieldwork.

REFERENCES

Burrows, E. G.
1938 Western Polynesia; a Study in Cultural Differentiation. Etnologiska Studier. : Walter Kaudern.

DeBoer, W. R.
1991 The Decorative Burden: Design, Medium, and Change. In Ceramic Ethnoarchaeology. W. A. Longacre, ed.. Tucson: The University of Arizona Press.

Delmas, Siméon
1927 La Religion ou la Paganisme des Iles Marquisiens. Paris, France.

Dordillon, Mgr René I.
1931 Grammaire et Dictionnaire de la Langue des Iles Marquises. Paris.

Ferdon, Edwin N.
1965 Surface Architecture of the Site of Paeke, Taipi Valley, Nuku Hiva. In Reports of the Norwegian Archaeological Expedition to Easter Island and the East Pacific, Vol. II . Thor Heyerdahl and E.N. Ferdon (eds.). PP 117-122. School of American Research and Kon-Tiki Museum, Monograph No 24, pt 2. Santa Fe, New Mexico.

Firth, R.
1967 The Work of the Gods in Tikopia. Humanities Press. New York.

Handy, E.S. Craighill
1923 The Native Culture in the Marquesas. Honolulu: Bernice P. Bishop Museum Bulletin 9.

Handy, E.S. Craighill
1927 Polynesian Religion. Honolulu, Hawaii: Bernice P. Bishop Museum Bulletin 34.

Heyerdahl, Thor
1965 The Statues of the Oipona Me'ae, with a Comparative Analysis of Possibly Related Stone Monuments. In Reports of the Norwegian Archaeological Expedition to Easter Island and the East Pacific, Vol. II . Thor Heyerdahl and E.N. Ferdon (eds.). PP 123-151. School of American Research and Kon-Tiki Museum, Monograph No 24, pt 2. Santa Fe, New Mexico.

Kirch, Patrick V.
1996 Tikopia Social space Revisited. Prepared for Pacific Culture History: Essays in Honor of Roger C. Green. In Press.

Lee, Georgia
1992 The Rock Art of Easter Island, Symbols of Power, Prayers to the Gods. Los Angeles: The Institute of Archaeology, University of California, Los Angeles. Monumenta Archaeologica 17.

Linton, Ralph
1923 The Material Culture of the Marquesas Islands. Memoirs of the Bernice P. Bishop Museum Vol. VIII (5). Honolulu.

Linton, Ralph
1925 Archaeology of the Marquesas Islands. Honolulu: Bernice P. Bishop Museum Bulletin 23. Honolulu.

Millerstrom, Sidsel
1992 Manuscript in possession of the author.

Millerstrom, Sidsel
1997 Carved and Painted Rock Images in the Marquesas Islands, French Polynesia. Archeology in Oceania. 32:181-196.

Millerstrom, Sidsel and Edmundo Edwards
1985 Manuscripts in possession of the authors.

Millerstrom, Sidsel and Trica Allen
1994 Carved Images and Punctured Skins: Rock Carvings and Tattoos in the Marquesas Islands. Paper presented at the 1994 International Rock Art Congress, Flagstaff, Arizona. Publication in progress.

Petard, Paul
1986 Plantes Utiles de Polynésie. Raau Tahiti. Papeete.: Haere Po No Tahiti.

Rolett, Barry V. and Eric Conte
1995 Renewed Investigation of the Ha'atuatua Dune (Nuku Hiva, Marquesas Islands): A Key Site in Polynesian Prehistory. Journal of the Polynesian Society 104(2):195-228.

Steinen, Karl von den
1928 Die Marquesaner und ihre Kunst. Plastik Vol. II. Berlin.

Stewart, C. S.
1832 A Visit to the South Seas in the U. S. Ship "Vincennes," during the years 1829 and 1830. London: Fisher, Son and Jackson, Newgate St.

Suggs, Robert. C
1961 The Archeology of Nuku Hiva, Marquesas Islands, French Polynesia. Anthropological Papers of the American Museum of Natural History. Vol. 49(1). New York.

Suggs, Robert C.
1966 Marquesan Sexual Behavior: An Anthropological Study of Polynesian Practices. New York: Harcourt, Brace & World, Inc.

Tautain, M. le Dr.
1887 Constructions et Monuments des Marquises. L'Anthropologie, Paris, France Tome VIII(1):538-558.

Thornton, Agathe
1989 Some Reflections on Traditional Maori Carving. The Journal of the Polynesian Society 98(2):147-166.

Walsh, D.S. and B. Biggs
1966 Proto-Polynesian Word List I. Linguistic Society of New Zealand.

---

For å komme til vår forkortede og oversatte versjon, klikk her.
For våre artikler om statuene i sydhavet - klikk her.
For å komme til våre artikler om pyramider og steinstrukturer i sydhavet, klikk her.

---