Volume 105 1996 > Volume 105, No. 2 > The first-order anthropomorphic gods of Polynesia, by Jeff Marck, p 217-258
                                                                                             Previous | Next   

- 217

The Polynesian race of the Eastern Pacific has an elaborate system of Cosmogony, which aims at explaining how the heavens were created and sustained, how gods and men came to be, how their own islands arose; but the details thereof vary much as given by the wise men in the various groups….(Fraser 1892:165)


This paper will suggest reconstructions of names and roles for the first-order anthropomorphic gods of Polynesians at various stages in prehistory: in the community of Proto-Polynesian (PPN) speakers, the community of Proto-Nuclear Polynesian (PNP) speakers, and the community of Proto-Central Eastern Polynesian (PCE) speakers. Here I shall define “first-order anthropomorphic gods” as those who were the first clearly anthropomorphic descendants of a Primordial Pair, a term which is explained below.

Early students of comparative Polynesian cosmogony were fascinated by the recurrence of cosmic entities, anthropomorphic gods and their deeds throughout Polynesia but were without a method to attempt reconstruction of beliefs to earlier points in time. There was no phylogeny of Polynesian linguistics. There was no phylogeny of Polynesian cultures. The single major attempt at reconstructing an overall cosmogony (Handy 1927) has seldom been cited except to note isolated factual details concerning individual traditions. It was not convincing, even at the time, and Handy's kind of work has been abandoned as speculation.

A second major comparative work soon followed Handy. Williamson (1933a) quite thoroughly laid out local traditions without attempting reconstruction although he did comment in his second volume (Williamson 1933b) on what may have been involved in the past. His speculation about the past took a different form from Handy's (1927). Where Handy had highly specific notions of an overall ancestral system and a later introduction of a “cult of Taŋaloa”, Williamson was more interested in smaller segments of the overall system and how they may have existed individually in the past. Burrows (1938:115–22) recognised basic differences in creation traditions between Western and Eastern Polynesia and believed many of the eastern traditions had developed locally. However - 218 this point was made in the context of a broad discussion of Polynesian cultural characteristics and he did not go into much detail.

Hiroa (1938a) considered the role of “Hawaiki” in the traditions of what we now call the Central Eastern Polynesian linguistic group, but this is only one of many themes or features common to many Polynesian cosmogonies that one might select for special attention. In this paper I have chosen the question of concepts involving the first order of anthropomorphic gods. I report elsewhere on another topic: the prehistory of naming associated with what I call the Primordial Pair 1 (Marck forthcoming).

I am well versed neither in Polynesian cosmogonic traditions nor in theories of how cosmogonic traditions should be compared and reconstructed. Nevertheless, I have found it striking that many of the distributions of traditions that confounded earlier students of the situation are remarkable in that they are found only within certain Polynesian linguistic subgroups as they are defined today. While Polynesian cosmogonic traditions may have differentiated according to a different phylogeny from Polynesian linguistic traditions, this report and the others (Marck forthcoming, n.d.) show that there is reason to believe that similar histories of differentiation were involved, that is, that a similar phylogeny obtains. This is not meant to be a definitive statement concerning the prehistory of Polynesian cosmogonies so much as it is an initial attempt to suggest what we might reasonably propose by comparison with the phylogeny of linguistics.


There is no narrowly specified method for this kind of work in the history of Polynesian studies although the constraints I shall propose resemble what Clark (1976) labels the “distributional method” of comparative linguistics when considering the antiquity of grammatical traits in Polynesian languages. They also follow in the tradition of Hiroa (1938a), Burrows (1938) and modern social theorists who believe the phylogeny of linguistics is of interest to the overall problem of cultural differentiation in Polynesia. What follows is one linguist's examination of a different order of cultural material in light of the linguistic phylogeny for the area in light of linguistic form-meaning reconstructions. It is not the only “correct” linguistic interpretation of the overall problem. There could easily be others.

Two basic ideas are being utilised:

  • 1. If a feature of the cosmogonies is universal or occurs in Tonga and Eastern Polynesia the feature will be suggested to have occurred amongst the beliefs of the PPN speakers. (Agreements between Tongan and
- 219
  • Samoan will be considered possible borrowings unless there is agreement from Eastern Polynesia as well). An example of a Tongan-Samoan-Eastern Polynesian agreement is the belief that the sky was close to the earth at the time of creation and that this caused great inconvenience to early anthropomorphic gods and people and that an early act of the early gods was to raise the sky into its present position 2 This would seem to have been a belief of the PPN speakers.
  • 2. Similarly, if a feature occurs in two widely separated groups not otherwise known to have borrowed from each other, the feature will be suggested to have occurred in the community of speakers of their common proto-language rather than to be borrowed. Examples are the naming of the male of the Primordial Pair as Papa-adj. in both Samoa and the Marquesas and the common Nuclear Polynesian (but not Tongan) belief that *Maaui-Tiki-Tiki-A-Talaŋa was the child of *Taŋaloa.

Certainly this is a better point of departure than the general framework of Handy 3 (1927) which was not specified to be the comparative method of social anthropology 4 as it was practiced at about that time but resembled it in many ways. The present work is not an evolutionary, psychological or philosophical attempt at reconstruction. It is a distributional attempt at reconstruction drawing upon the linguistic phylogeny of the area.

Since linguists work with languages, which seem all to change with time, there is, perhaps, a bias in that I assume that creation traditions have all changed with time. Perhaps some are more conservative than others. Some are certainly more innovative than others. But the result presented here assumes change on the part of all groups and this may be less true of some groups than others. Nevertheless, we are seeking out things which seem to have changed the least in the most number of groups so the bias may be imagined more than it is real.

The phylogeny of Polynesian linguistics that I assume is given in Figure 1. This involves some modifications of the received subgrouping (Pawley 1966, 1967) and Green (1966) based on work by Wilson (1985) and myself (Marck 1993, forthcoming2, in preparation).


Handy (1927) and especially Williamson (1933a, 1933b) have already been mentioned as secondary sources containing much useful material. Craig (1989) is a recent comparative dictionary of Polynesian mythology which does not attempt reconstructions. Handy, Williamson and Craig

- 220
Figure 1: Polynesian Subgrouping (after Pawley 1966, 1967, Green 1966, Wilson 1985, Marck 1993, forthcoming2, in preparation)
- 221

largely agree on what the most valuable primary material is, up to the date of their own publication. In fact it is an agreement upon what extant material has any value at all and the materials concerned added up to only about 300 titles by the time of Craig (1989). Many of those sources are very short, and those concerning cosmogony (rather than religion or general mythology) are very few indeed. Amongst these many will not be mentioned here because they seem traditions that have developed locally having neither actants nor elements in common with other groups. Some are not mentioned because they have actant names in common with other groups but attribute deeds to them not known from other localities. Thus there is nothing of comparative utility in them. Others are not mentioned because they are peripheral to the present topic. There is a need to remain focused. There is much that must await further study.

There is a general problem with specific cosmogonic materials lacking in otherwise comprehensive works. Firth (1967) on Tikopia, for instance, concerns religious practices. Religious practices in Polynesia seldom had much to do with the first-order anthropomorphic gods. Rather, ancestors were commonly worshipped as they were genealogically closer to the gods that first created or procreated people and these ancestors were seen as capable of interceding with these and other gods on behalf of living people. The earliest gods were distant and not the subject of much ritual. Thus a work describing religious practices need not deal with cosmogonic notions and many of them do not.

Firth's (1961) earlier work on Tikopian oral history and traditions addresses the issue of cosmogony but there we encounter a problem common to what little we know of the Polynesian Outliers' cosmogonies: they resemble little the traditions of Tonga or Samoa. This is similar to what Geraghty (1993:344 fn. 3) has noted in reference to the absence of the Tongan and Samoan Pulotu ‘paradise’ word and notion amongst the Outlier groups.

While cosmogony was neglected in favour of descriptions of religious life in some localities, another source of neglect was a preoccupation with general mythology. For example, von Steinen (1898, 1899, 1925-1928, 1933, 1934, cf. also Langridge and Terrell 1988) collected a great deal of Marquesan mythology and oral history but little of the cosmogony is mentioned in those materials. From a short passage in Christian (1895:1878) supported by a few paragraphs from Handy (1923:244-5) we find that there are many similarities between the basic Marquesan conceptualisation of creation with Tonga, Samoa and especially New Zealand Maori. But those few paragraphs represent all we know.

- 222

There is sometimes a problem with loss or transformation of cultural memory under Christian influence and the absence of interested European parties who wrote them down before their loss or transformation. This seems to have affected cosmogonic traditions more than general mythology. An example is the common source for Tahitian traditions (Henry 1928). Henry's traditions were collected after a period of contact with Christians and it is questionable how much Henry's detailed cosmogony pre-dated European times (cf. Barrère 1966, Marck forthcoming). Yet there is no more comprehensive earlier record of Tahitian traditions and we must search for elements of the past in a highly transformed set of traditions.

Similarly, Fraser (1892) appears to have been told a Christianised Samoan tradition at about a time when traditions more in keeping with the general Polynesian pattern were being collected (e.g., Turner 1861, 1884).

Barrère (1966) suggests some criteria for distinguishing authentic Polynesian traditions from those transformed by Christianity and specifically addresses the question of whether a would-be supreme being from around the Tahiti area and New Zealand (variously lo, Iho, 'Iho or Kiho) is a post-Christian development. Williamson (1933a, 1933b) barely mentions those traditions even though many were known by the time of his work, so possibly he, too, considered them suspect. In addition to Barrère's arguments we can note that New Zealand Maori lo lacks both the consonants of the would-be cognates around Tahiti, sounds which are otherwise almost never lost in New Zealand Maori (cf. Marck in preparation) in loans or directly inherited words 5 The absence of consonants strongly indicates that different ancient words were involved or that any borrowing was through the agencies of non-Polynesian speakers 6

In addition to the cautions Barrère (1966) mentions we can suggest that Christian influence is to be suspected when there is a pre-existing supreme being at all and when early acts of such supreme beings were creation of the land and waters and heavens. Comparative evidence suggests (Marck forthcoming) that ancient Polynesian cosmogonies conceived of the primordial condition 7 as one in which the land, sea and sky pre-existed, the sky was hugging the earth, and the primordial period was interrupted by actions of cosmic beings 8 or the Primordial Pair, none of which had anthropomorphic forms or incarnations except the Primordial Pair of the New Zealand Maori and Hawai'i, Rangi and Papa, and Waakea and Papa, respectively. In both cases they were the descendants of cosmic beings without anthropomorphic form. Pre-existing male anthropomorphic creators are found in Fraser's (1892) Samoan tradition and around the Tahiti area and in both cases reflexes of PPN *Taŋaloa are involved. However the Samoan tradition has few other elements or - 223 actants known from elsewhere. Neither do the traditions of Ta'aroa/ Tangaroa as creator god around Tahiti have elements or actants known from beyond that area… not even amongst the New Zealand Maori with which Tahiti is closely linked linguistically or amongst Hawaiians who may have borrowed some elements of their cosmogony from early Tahitic. Thus some traditions cited by Handy (1927), Williamson (1933a, 1933b) or Craig (1989) are not mentioned here because they show elements which have no counterparts outside a local area and these elements sometimes seem suspiciously similar to Judeo-Christian cosmogonic concepts.


In the main, Polynesian cosmogonies 9 attribute the end of the primordial period and the beginning of the world as we know it to the union of what I call a Primordial Pair, one male and one female. This is true of Tongan, Samoan, Marquesan, New Zealand Maori, and some Cook Island and Tuamotuan traditions. Traditions which differ, such as Tahitian, other Tuamotuan, some Southern Cook Islands and Hawaiian, would seem to have innovated out of the basic pattern seen in those mentioned before. Little is known of Rapanui (Easter Island) or Polynesian Outlier cosmogonic traditions and it is the traditions of the previously named groups with which we will be most concerned.

Numerous patterns are to be observed including: (1) The Primordial Pair give rise to cosmic beings who mate and are the forebears of the first-order anthropomorphic gods (Tonga and Samoa). (2.a) The Primordial Pair are the offspring of cosmic beings and direct parents of the first-order anthropomorphic gods (New Zealand Maori). (2.b) The Primordial Pair are direct parents of the first-order anthropomorphic gods but are, themselves, not known to descend from cosmic beings (the Marquesas, Tongareva). (3) Cosmic beings are unknown, the Primordial Pair are forgotten and only certain first-order anthropomorphic gods are remembered (Mangareva). The Primordial Pair are named as in Table 1 which includes reconstructions for some of the proto-languages.

As can be seen, no PPN reconstruction is made. This is due to the absence of agreement between Tongan and Eastern Polynesian 10 whose relationships are shown in Figure 1. The reconstructions in Table 1 are suggested on the basis of agreements between Samoan, Marquesan, New Zealand Maori and certain Cook Island and Tuamotuan traditions. Traditions around Tahiti, some of the Tuamotus, some of the Southern Cooks and Hawai'i seem highly transformed (Marck forthcoming, cf. Hiroa 1938a: 150, Barrère 1966) out of a pattern common to those seen in the traditions of those five first named - 224 groups. The Tuamotus and Cooks remained in contact with Tahiti after the divergence of New Zealand Maori from those other three and it is suggested here and in other works (Marck forthcoming, n.d.) that some of the Tuamotus and some Cook groups borrowed various transformations which seem to have emanated from or passed through the Tahiti area. Tongarevan (Penrhyn) traditions seem connected to Tahitian traditions at a time postdating the divergence of New Zealand Maori and preceding the move around Tahiti towards the loss of the Primordial Pair memory and the elevation of Tangaroa to a pre-existing being. Hawaiian traditions have some resemblances to Tahitic beliefs as they may have existed at about the time of the divergence of New Zealand Maori.

  Male Female
Tongan Limu Kele
PNP *Papa-adj. Papa-adj.
Samoan Papa-Tuu Papa-'Ele
PEP/PCE *Papa-adj. *Papa-adj.
PMQ *Papa-adj. *Papa-adj.
N. Marquesas Papa-'Uka Papa-'A'o
S. Marquesas Papa-'Una Papa-'A'o
Hawaiian Waakea 11 Papa
PTA *Aatea *Papa(-adj.)
Tahitian2 Tumu Papa-Raha-Raha
  Papa-Tuu-'Oi Aatea
  Aatea Ha'a-Hotu
Tuamotuan Tumu 12 Papa2
  Vaatea Hotu
Rarotongan Tumu Papa
  Vaatea Papa
Mangaian Vaatea/Avaatea Papa
Tongareva Aatea Haka-Hotu
New Zealand Maori Rangi-Aatea Papa
Table 1: The Primordial Pair and Possible Remnants of the Primordial Pair in Polynesian Cosmogonies. Source: Marck (forthcoming).

There is a basically different conceptualisation and naming of the Primordial Pair in Tongan as opposed to Nuclear Polynesian groups. Traditions external to Polynesia seem to have detailed agreements with neither. In - 225 Tonga the Primordial Pair are Limu ‘Seaweed’ and Kele ‘Soil Sediment’ while in Samoa and Central Eastern Polynesia they are earthly, atmospheric or heavenly rock or strata and have many cognate names, especially Papa which is always the female except in some obviously local developments in and near the Societies.

Papa, most basically, seems to have meant ‘flat rock’ in PNP as regards this appellation (Marck forthcoming), or at least that is what the Samoan evidence indicates. In PCE the “rock” word with modifiers can be reconstructed for both the male and the female but there was a sense of Papa-the-female being the earth itself and Papa-the-male being a stratum above the earth. Still they retained their old names in PCE with the “flat rock” word as the base. But the common word papa also had a sense of ‘flat surface’ so a “stratum” sense may be the older sense and Samoan may have lost it.

The primordial condition involved a sky that was hugging close to the earth and one of the early acts of the early gods was to raise the sky to create space for standing upright, the brightness of day and other modern conveniences 13. Three strata were distinguished: the earth, the sky and the space between them. There is no named deification of this space between them known from Tonga. Samoans name the deification of this space Vale-Vale-Noa. This is a similar concept to but not a name cognate with PCE *Aatea. *Aatea was the mate of Papa amongst some Tahitic groups and the Hawaiians while a child of Papa to the Marquesans. It is only amongst the New Zealand Maori that Rangi 14, the sky itself, was the actual name of the male mate of the female Papa. The Marquesans had an alternate name for the male, 'Ani-Motua ‘Sky Father’, but this was not the name used in the formal recitation of creation traditions. The common Tahitic and Hawaiian *Aatea was the deified space between the sky and earth so possibly the notion of the sky itself as the male developed independently in the Marquesas and amongst the New Zealand Maori (Marck forthcoming).

In the present work we shall consider the question of who the offspring of the Primordial Pair were according to PPN, PNP and PCE speakers.

5.1 Proto-Polynesian and Proto-Nuclear Polynesian

The first-order anthropomorphic gods of Tonga were half-brothers by the same father, a cosmic being descended from the Primordial Pair, and mothers also descended from the Primordial Pair 15 (Williamson 1933a:10-1). The brothers were Hiku-Le'o, Tangaloa and Maui. 16 Tangaloa governed the sky, Maui governed Lolo-Fonua (‘land or country below’) and Hiku-Le'o governed Pulotu (‘paradise’ cf. Geraghty 1993).

- 226

Craig (1989:28), Handy (1927:15) and Williamson (1933a:3) all cite Turner (1861, 1884) as their primary source on Samoan creation traditions. Creation begins with a genealogy of unions between cosmic beings such as “Nothing”, then “Fragrance”, then “Dust” and others which gradually merge with the genealogies of people (Turner 1884:3-9). In the first union mentioned in the genealogies of people, Papa-Tuu ‘Standing Rocks (male)’ and Papa-'Ele ‘Earth/Soil Rocks (female)’ pair, produce cosmic beings and seven generations later Tagaloa is born (Turner 1884:4). 17

While Turner's (1884:4) genealogical table gives no siblings of Tagaloa-the- originator-of-men, the next page mentions another child of the same father (“Cloudless heavens”, a cosmic being): Tagaloa-the-dweller-in-lands, who is said to have a different mother (“The eighth heavens” rather than “Spread out heavens”). Tagaloa-the originator-of-man is said to have sired Tagaloa-of-the-heavens, while Tagaloa-the-dweller-in-lands is said to have sired Tagaloa-the-explorer, who in turn sired Vale-Vale-Noa (‘space’ deified/ personified). 18 Thus Turner (1884:3-9) mentions siblings or half siblings named Tagaloa as the first anthropomorphic gods and does not indicate that they had siblings not named Tagaloa. But the source is a genealogy which would not mention ancestors other than those in a direct line back to Papa-Tuu and Papa-'Ele.

What we know of the ancient Samoan pattern cannot be shown to be similar to Tongan or much of Central Eastern Polynesia, where the union of the Primordial Pair resulted in groups of siblings. While we might suspect the “group of siblings” theme was a PPN or PNP feature, we cannot demonstrate this by comparison of the Tongan half-siblings with CE as only PPN *Taŋaloa can be reconstructed with the “first-order anthropomorphic god” status. Tonga's Maui is not a first-order anthropomorphic god in Nuclear Polynesian (cf. Marck n.d.) and Tonga's Havea-Hiku-Le'o is known elsewhere only from Samoa, where he seems not to have been of the order of Tagaloa and may, on distributional grounds, be a borrowing. In fact, Samoan Si'u-Leo is “Said to come from Tonga” (Turner 1884:52).

Elsewhere (Marck n.d.) I suggest that as early as Nuclear Polynesian times, there may have been a tradition of *Taŋaloa cuckolding *Talaŋa, the husband of *Maaui-Tiki-Tiki-A-Talaŋa's mother, by way of siring *Maaui-Tiki-Tiki-A-Talaŋa. But that is a different story. The point to be made here is that *Maaui-Tiki-Tiki is not a first-order anthropomorphic god in Nuclear Polynesian, as in Tonga, so his position at the PPN level is indeterminate. The first-order subgroups do not agree on this matter and there is no external evidence in agreement with either subgroup. In fact, Polynesian cosmogonies seem particularly Polynesian; no similarities of actants and deeds are known other than possible borrowings from Polynesian Outliers around Melanesia.

- 227

After their birth, in both Tonga and Samoa, *Taŋaloa and *Maaui occupy themselves with bringing the world as we know it into existence. They fish up islands, *Maaui raises the sky, *Taŋaloa sends worms or maggots to earth that become people and *Maaui goes into the bowels of the earth and obtains fire for gods and people 19 (Williamson 1933a:2-11, 41-2, 47-58, 1933b:184-91). The story of people coming from worms or maggots is particularly Tongan and Samoan and is not known from Central Eastern Polynesia, so this is a possible borrowing between Tonga and Samoa. The sky-raising story is also told around Central Eastern Polynesia and *Maaui is the most consistent hero of these traditions, so we have good reason to suspect that such was the belief amongst PPN, PNP, PEP, PCE, PMQ and PTA speakers (Marck n.d.). There is a specific agreement between Tongan, Samoan and Hawaiian traditions that *Maaui raised the sky in return for a drink of water from a mortal woman who asked him to do so. By our present method, this belief is then reconstructed to the cosmogony of PPN, PNP, PEP, and PCE speakers. Whether it should be reconstructed to PMQ or PTA is not clear as Hawaiians may have inherited the belief directly as members of Marquesic or may have borrowed the belief under later Tahitic influence.

The only other Samoan gods whose cognate figures have first-order anthropomorphic status elsewhere are Logo and Tuu, who are discussed under the PCE reconstructions *Roŋo and *Tuu.

There are numerous other cognate god names between Tongan, Samoan and CE groups, e.g., Samoan Luu and PCE *Ruu are cognate, Tongan Sinilau, Samoan Tinilau, and PCE *Tinirau are cognate and Tongan Hina, Samoan Sina and PCE *Hina are cognate. But these were not anthropomorphic gods of the first order in any of the traditions and detailed comparison of those figures was not attempted for the present work.

Reconstruction of additional first-order anthropomorphic gods to the PPN level is frustrated by a lack of further agreements between Tongan and Nuclear Polynesian. Reconstruction of additional first-order anthropomorphic gods to the PNP level is frustrated by a lack of further agreements between Samoan and Central Eastern Polynesian.

5.2 Proto-Central Eastern Polynesian Traditions

The reconstruction of beliefs to the PCE level proceeds most conveniently by the comparison of Marquesan and New Zealand Maori traditions which have much in common compared to highly transformed traditions from in and around the Societies or those of Hawai'i which seem as much Tahitic as they are Marquesic. When members of two first-order subgroups agree in form and meaning, one can reconstruct the agreement to the common protolanguage if borrowing or convergence are implausible. Marquesan and New - 228 Zealand Maori traditions agree on the PCE first-order anthropomorphic gods in Table 2. An additional agreement with Marquesan is taken from Mangaian in the case of PCE *Toŋa-Fiti and Tongarevan in the case of PCE *Mauri.

To the reconstructions of Table 2 we can add PCE *Taŋaroa ‘first-order anthropomorphic god’ by comparison of New Zealand Maori, Tongarevan and some Tuamotuan traditions with Tongan and Samoan. PPN *Taŋaloa, while elevated in and around Tahiti to a pre-existing being, was apparently demoted in Marquesan. He is not a first-order anthropomorphic god in the Marquesas but was relatively important to the Mangarevans (cf. Hiroa 1938b:16, 17-8, 19, 111). The possible reconstruction of *Aatea to the male of the Primordial Pair in PTA (Marck forthcoming), the possible borrowing of that status into Hawaiian (Marck forthcoming) and the position of Aatea as a first-order anthropomorphic god to Marquesans (Christian 1895:188, Handy 1923:244) suggests a high status for *Aatea amongst PCE speakers but one that is indeterminate. He did not necessarily become the male of the Primordial Pair in Proto-Tahitic (PTA) due to being a first-order anthropomorphic god in PCE, although this seems a possible scenario.

PCE Marquesan Maori Mangaia Tongareva
*(H)aumia 20 Aumia Haumia    
*Mauri Moui 21     Mauri
*Roŋo 'Ono Rongo Rongo Rongo
*Taane Taane Taane Taane Taane
*Toŋa-Fiti Tono-Fiti   Tonga-'Iti  
*Tuu Tuu Tuu    
Table 2: PCE first-order anthropomorphic gods as deduced from the comparison of Marquesan to New Zealand Maori and Cook Island traditions. Sources: Marquesan: Christian (1895:187-8) and Handy (1923:244-5); New Zealand Maori: Best (1925:746), Grey (1885:1-11) and Tregear(1904:462); Mangaian: Gill (1876:9-12), Tongarevan: Hiroa (1932:85).

It is difficult to speak comparatively of Polynesian cosmogony unless it is broken down into its components. Thus we are presently speaking - 229 of the first-order anthropomorphic gods and set aside the Primordial Pair and PPN *Maaui-Tiki-Tiki-A-Talaŋa for discussion elsewhere (Marck forthcoming and n.d.). Similarly, it is difficult to speak of the first-order anthropomorphic gods of Central Polynesia all at once as they are different from place to place and even when they are the same, their deeds and roles sometimes differ between localities. Thus it is convenient first to present their outline of origin for the major localities and then return to consider the individual reconstructions of Table 2 along with those for PCE *Tarŋaroa and *Aatea.

We will note here, and refer the reader to Hiroa (1938a), that Central Eastern Polynesian creation traditions are rarely autochthonous. Creation is remembered as having occurred in PCE *Hawaiki. Traditions through the Cooks, the Societies and the Tuamotus have specific memories of the creation of various islands of Samoa, Tonga and Fiji, and Marquesan genealogies name these same islands as ancestors. Such memories may be vaguer amongst the New Zealand Maori and Hawaiians but the present point is that they, too, remember creation as occurring elsewhere (in *Hawaiki) and think of people coming to New Zealand and Hawai'i after creation and the beginnings of human existence elsewhere.

5.2.1 Marquesic Traditions The Marquesas

Williamson (1933a:15) and Craig (1989:29) mention only two primary sources concerning the end of the primordial era and the origin of the first-order anthropomorphic gods: Christian (1895:187-8) and Handy (1923:244-5). Even Handy (1927) provides no further information on this specific question. In both Christian and Handy Papa-'Una/'Uka, 22 the ‘level above’ or ‘world above’(male) unites with Papa- 'A 'o, the ‘level below’ or ‘world below’ resulting in the birth of the first-order anthropomorphic gods. Christian names a sibling group of twelve, Handy nineteen. Those whose names have cognates outside the Marquesas are: Aatea, Taane, Tono-Fiti/ Toko-Hiti, Tiki, Aumia, Moui, Tuu and 'Ono (PCE *Rongo). Of the preceding, only Aatea, Taane and Toko-Hiti are mentioned by Christian, the rest coming from Handy. The others mentioned by Christian and Handy have no cognates in other Polynesian traditions, so far as I am aware.

Christian (1895:187) tells us that at first these siblings lived in a subterranean cave, longing for light and that it was Aatea who broke them out of the cave by stamping his foot through the earth to open space. Thus, we might presume, his name which translates freely as ‘clear space’ or ‘clear illuminated space’. Christian's tradition then goes on to speak of the lands over which Aatea and his eleven siblings came to hold dominion.

- 230

Handy's tradition does not mention the “breaking out” memory and focuses more upon the areas of human affairs over which the more important siblings held sway. Aatea is said to be “progenitor of all natives with Atanua his wife… Tu is a legendary character and patron of war. His name does not appear in the genealogies”. 'Ono (PCE *Rongo) “was of no importance in the actual worship - at least not under this name.” “Tane was of little importance in the Marquesas. His name appears in legend and chants, but not in genealogies” (Handy 1923:245).

“Tana-oa” (Tana'oa (PCE *Taŋaroa)) is mentioned at the end of these materials (Handy 1923:245) but recall that neither Christian nor Handy mention him as a member of the first-order siblings: “Tana-oa is mentioned elsewhere as a god of the wind and sea and patron of fishing.” Handy (1923:245) credits Tono-Fiti with thrusting apart the “level above” and “level beneath” but this seems unique in Polynesia where it was more often *Maaui who did so (cf. Marck n.d.) and this was the belief in Mangareva (Williamson 1933a:44).

The island-raising stories of the Marquesas and Central Polynesian groups in general concern Maui 23 (Williamson 1933a:35-6) and not *Tarŋaloa as in Tonga and Samoa. In general, the position of the Marquesas' Tana'oa/ Taka'oa is diminished compared to other Polynesian. Williamson (1933a:20-1) gives a translation of a Marquesan tradition from Fornander (1878:214-8) in which Tana'oa represents something like primal darkness and is defeated by Aatea. Elsewhere Williamson (1933b:235-6) notes the general dearth of materials concerning Tana'oa for the Marquesas and wonders if he is a recent introduction.

Cosmic beings, such as we have seen in Tonga and Samoa, are not mentioned in Handy's (1923:244-5) account for Marquesan but something of the sort seems part of Christian's genealogies (cf. 1895:191) which follow his cosmogonic tradition. There, such entities as “Po” (poo ‘darkness, night’) and “Ao” (ao ‘light, day’) are named as are various Western Polynesian and Fijian localities: “Havaii”, “Vevau”, “Fiti”, “Fiti-tapu” and “Tona-Tapu”. Modern people stand in a genealogical line between the Primordial Pair, their immediate offspring Aatea, these cosmic beings or lands (descendants of Aatea), and then (pre)historical people (descendants of the cosmic beings or lands). Mangareva

Hiroa (1938b:20-96) relates the traditional history of Mangareva. Only the first few pages have much to do with our current topic and relate quite a different kind of story from the autochthonous traditions of Tonga and Samoa or Central Eastern Polynesian traditions, which view creation - 231 as occurring in PCE *Hawaiki (Savai 'i in Samoa). In the tradition related by Hiroa (1938b:20-l) more or less mortal people are the first to arrive in Mangareva and have come from the Marquesas. Named Miru and Moa, they return to the Marquesas. Other people come to Mangareva and Tagaroa-Huru-Papa figures in this tradition. Other gods with cognate names elsewhere are not mentioned.

In later materials Hiroa (1938b:418-25) relates details of the primary gods. Aatea, Tagaroa and a few others not cognate with first-order anthropomorphic gods elsewhere. They have origins “so remote that they are termed gods without beginning” (Hiroa 1938b:418). There is no Primordial Pair or sibling group parented by them. Some of the gods commonly found as first-order anthropomorphic gods elsewhere are said to be the children of Tagaroa, including Tuu and Rogo (Hiroa 1938b:422).

Other sources on the cosmogonic traditions of Mangareva seem not to be well known, although Smith (1918:115-31) mentions that “Maui-matavaru” fished up the islands of Mangareva (p. 131) and that “Tiki” was the first man and husband of “Ina” (p. 129). Hawai'i

Hawaiian cosmogonic traditions are highly idiosyncratic. There are many competing traditions and Beckwith (1970:42-6) relates five, two of which are from Fornander (1916-1920) which seem the least influenced by Christianity. But there, too, is a suspicious trinity involving Kaane, Kuu and Lono (PCE *Taane, *Tuu and *Roŋo) and Beckwith (1970:46) acknowledges a Biblical “coloring” in all of them. In the longer of the two Fornander traditions:

The three gods Kane, Ku, Lono come out of the night (po) and create three heavens to dwell in, the uppermost for Kane, the next below for Ku, and the lowest for Lono… Next they make the earth to rest their feet upon… Kane then makes sun, moon, and stars, and places them in the empty space between heaven and earth… Next an image of man is formed out of red earth… A law is given him but he breaks the law and is then known as Kane-la'a-(kah)uli, “a god who fell because of the law.”

In the original garden… [is found] A tapu tree, sacred apples which cause death if eaten by strangers…(Beckwith 1970:43)

This does not seem the place to look for remnants of aboriginal Polynesian traditions. Craig (1989:101) has Kaane emerging from “the eternal poo (darkness)” as does Beckwith (1970:42-3), his source, but both ignore a passage in Fornander (1916-1917:17-8) more in keeping - 232 with the general Central Eastern Polynesian pattern that specifically has Kane and Kanaloa as offspring of Waakea and Papa, as they would seem to have been to PTA speakers. Due to this neglected passage, I think we can dismiss much of what has passed for Hawaiian cosmogonic tradition in Craig (1989) and Beckwith (1970) as late developments under the influence of Christianity. Rather we see in the passage from Fornander cited above the last dying gasp of an aboriginal cosmogony which cascaded towards Christian motifs soon after the arrival of Europeans.

To ask whether this small bit of unblemished tradition is more Tahitic or more Marquesic is probably not very meaningful. The Hawaiian Primordial Pair is more in keeping with Tahitic and the Hawaiian memory is one of Waakea and Papa coming from Tahiti, Kaane and Kanaloa apparently being born in Hawai'i or at least conducting their major deeds in Hawai'i. The memory of Kanaloa being a child of the Primordial Pair does not have a counterpart in the Marquesas but it does to some extent in Mangareva. So the evidence points, to some degree, towards a Proto-Marquesic first-order anthropomorphic god *Taŋaroa. Hawaiian Kanaloa may come from that tradition, the Tahitic tradition, or both (because they did not conflict).

Waakea and Papa are said to be thought of as people in the Hawaiian traditions (Beckwith 1970:294) but their offspring, some of the Hawaiian islands themselves, suggest a magical sort of people. The passage from Fornander mentioned above calls into doubt the notion that they were thought of as people rather than gods before Christian influence (as they were the parents of the gods Kaane and Kanaloa).

Hawaiian Waakea and Papa may be a revision of early Marquesic beliefs in the sense that Marquesan Aatea is considered the ancestor of all people but not the spouse of Papa or it may be a Tahitic influence as the Primordial Pair in PTA seems to have been *Aatea and *Papa. As Hawaiian Waakea and Papa are specifically associated with Tahiti rather than the Marquesas in Hawaiian traditions, it seems most reasonable to suspect that the origins of the tradition are Tahitic. 24

5.2.2 Tahitic Traditions New Zealand Maori

We will consider New Zealand Maori cosmogonic traditions first. By comparison with evidence external to Tahitic, they seem the most conservative Tahitic traditions. It is also the case that they have been recorded in greater detail than for many other Polynesian groups. The cosmogony of the New Zealand Maori is quite clear: cosmic beings began - 233 evolving and their ultimate progeny, Rangi ‘Father Sky’ or ‘Sky Father’ and Papa ‘Mother Earth’ or ‘Earth Mother’, are the direct parents of the first-order anthropomorphic gods. There are competing traditions that Tangaroa was the spouse of Papa and that Rangi absconded with her (c.f. Tregear 1904:462, Biggs n.d.) but the union of Papa amongst Nuclear Polynesian groups is normally with a terrestrial, atmospheric or sky stratum and not an anthropomorphic god. Consequently, I have argued elsewhere (Marck forthcoming) that the Tangaroa-Papa tradition seems a local tradition.

There are many sources for the Rangi-Papa tradition (e.g., Grey 1885, Shortland 1856, 1882, Clark 1896, Tregear 1904, Cowan 1910, 1930a, 1930b, Best 1924, 1925, Andersen 1928) which seem to come from a limited array of primary sources (Biggs n.d.). A passage from Cowan (1910) seems most useful for our present purpose:

Rangi and Papa, the Sky-Father and Earth-Mother, were the parents of the following deities, who are the chief gods of the Polynesians and Maoris:

  • Rongo (God of Cultivations).
  • Tane (God of Man, also Forests and Birds).
  • Tangaroa (God of the Ocean and Fish).
  • Tawhiri-matea (God of the Wind and Storms).
  • Haumia (God of Fern-root and Uncultivated Foods).
  • Ruai-moko (God of Volcanoes and Earthquakes).
  • Tu-mata-uenga (God of Man and of War).

… Following upon the begetting of their seven children (there are many others mentioned in legends and genealogies, but the foregoing are the principal and deified ones), came the separation of Heaven and Earth… It was Tane-mahuta who forced his parents apart by standing on his head and thrusting Rangi upwards with his feet. (Cowan 1910:105)

Amongst Cowan's list of the most important first-order anthropomorphic gods we find all the gods with cognate names in other localities so we will not mention the others. As can be observed, there are very basic similarities with the Marquesan cosmogony, the main differences being in the naming of the Primordial Male, the inclusion of Tangaroa as amongst the first-order progeny, and the role of Taane instead of Marquesan Tono-Fiti or Mangarevan Maui in lifting the sky. These events are conceived of as occurring in Hawaiki in both the Marquesas and amongst the New Zealand Maori. The discovery and settlement of New Zealand is viewed as being the work of mortal people.

- 234 Cook Islands

The reconstruction of *Aatea as the Primordial Male for PTA (Marck forthcoming) is based on agreements between certain Cook Island and Tuamotuan traditions and the Rangi ~ Rangi-Aatea from New Zealand Maori. It is not supported by evidence external to Tahitic. Still it is these same Tuamotuan and Cook Islands that tend to have the most resemblances to Marquesan and New Zealand Maori in respect of the first-order anthropomorphic gods, so I will present first what seem the most conservative Cook Island traditions and then those that seem more influenced by transformed Tahitian traditions. Tongareva (Penrhyn)

These Northern Cook Island people remember creation as being the result of the union between Aatea and Haka-Hotu. 25 Their children were Taane, Tangaroa, Te Kapua, Mauri, Rongo-Nui, Tahaki, Te Porou-Rangi, Te Tou, Maru, Haka-Peka, and Putahi-Aitu (Hiroa 1932:85). Hiroa (1932) does not mention a sky-raising story or the fishing up of the islands nor do I know of other sources on such matters for Tongareva although Gill (1876:48) relates a Mangaian myth that Tongareva was raised by Vaatea while he was fishing. This seems unusual in two respects. Firstly, it is not common for a member of the Primordial Pair to be assigned anthropomorphic acts and secondly PCE *Aatea is not mentioned elsewhere in Polynesia in the context of the fishing up of the islands. Mangaia

Gill (1876:1-22) relates an outline of creation from Mangaia and Williamson (1933a:11-4) summarises it. The creation tradition is quite different from other Polynesian, even other Tahitic ones, but a number of familiar figures are present. First there is a complex notion of the universe resembling a coconut with a root extending downward and with a single hole in the top:

At various depths, in the interior of the coconut, which Gill calls Avaiki, was a series of floorings or lands, one above another, and communicating with each other… …in the lowest depth of Avaiki, where the sides of the shell nearly met, lived a woman, a demon of flesh and blood… She was the great mother…

This great mother, being desirous of offspring, plucked a piece of her right side, and it became a human being - the first man - Avatea or Vatea (Noon). - 235 This being, the father of gods and men, was half man and half fish…(Williamson 1933a:12-3)

The name of Vaatea's mother (Vari-Ma-Te-Takere) is not cognate with any creation name that I know of outside Mangaia. Vaatea eventually weds Papa (which Williamson translates as ‘Foundation’):

Vatea and Papa had five sons [all gods]. The first two, Tangaroa and Rongo, were twins; they were the first beings of perfect human form, having no second shape; the third was Tonga-iti, incarnate in the white and black spotted lizards; the fourth was Tangiia; and the fifth was Tane-papa-kai (Tane-piler-up-of-food). Rongo had three grandsons - Rangi and two others who dragged up the island of Mangaia from Avaiki to the light of day…(Williamson 1933a:14)

So here we see a familiar set of relationships between the Primordial Pair and some of the early gods. Hiroa relates:

In Mangaia there is no exploratory period, for the sons of the well known god, Rongo, drew the island out of an under-seas spiritual Avaiki to its present material position, with themselves upon it. With them commences the settlement period which thus links directly with the mythical period… As the period becomes more remote, details disappear, and the pedigree becomes a single list of names. The single list merges into the mythical period, from which it is often not clearly defined. (Hiroa 1932:16)

Ruu and Maaui are associated with the raising of the sky in Mangaia (Gill 1876:58-60, Williamson 1933a:43). Rarotonga

Williamson (1933a:14-15) relates two Rarotongan versions of creation. The first:

A genealogy of the royal (Makea of Karika) family of Rarotonga commences with references to Papa (the earth), which grew, became beautiful, budded, became mature, had duration, and became a parent; and its child, regarded apparently as a spirit or human being. This child married Ina [this is the same as Sina of Samoa and Hina of Tahiti], the daughter of the god Rongo, and it was from this first marriage that the Makea or Karika family's ancestry was traced.

(Williamson 1933a:14)

- 236

The second:

According to another Rarotongan account of the opening up of Papa (the earth), it was said that a person or being called Te Tumu [tumu means the ‘root’, ‘origin’, ‘source’ or ‘foundation’ of a thing] took Papa to be his wife and had by her three children - Te Uira, Te Aa, and Te Kinakina… After this Papa gave birth to the gods Rongo, Tane, Ruanuku, Tu, and Tangaroa. We are then told of a descendant of Te Tumu who begat the god Atea and others…(Williamson 1933a:14)

Overall, the Rarotongan legends are somewhat unusual, individually, and as a group and are not always consistent with each other. Whereas there is no male ancestor of Tumu in the tradition related by Williamson above, there are some brief passages from Gill (1911:134-6) which indicate that Tumu was the child of Aatea and Papa-Roa-I-Te-Itinga, while Tumu's wife was Papa-I-Te-Opunga. Tumu is Papa's spouse in Tahiti as we will see presently, something we did not see in Tongareva or Mangaia or for the New Zealand Maori. Thus some Rarotongan traditions blend with some seemingly recent Tahitian traditions in a way that Tongarevan and Mangaian traditions do not. Maaui is associated with the raising of the sky in Rarotonga (Smith 1899:70-2) rather than Taane as amongst the New Zealand Maori and Caillot's (1932) Tuamotuan tradition. Tahiti

Since the cosmogonies of the New Zealand Maori, some of the Tuamotus and much of the Cooks have so much in common with that of the Marquesas, it will be appreciated that the very different concepts concerning creation in Tahiti, for which similarities are to be found in only some of the Tuamotus and Rarotonga, must be local developments, as Hiroa (1938a:150) suggested.

The outstanding feature of Tahitian cosmogonies is the elevation of Ta'aroa (PCE *Taŋaroa) to a pre-existing being and supreme creator of the world as we know it today. Barrère (1966:104) attributes this to an adoption of Ra'iatean beliefs. The assertion is unreferenced but Williamson (1933a:379), also in an unreferenced assertion, also places the “centre of the Tangaroa-Oro cult” in Ra'iatea.

Henry (1928) is the common source for Tahitian cosmogonic traditions (cf. Handy 1927, 26 Craig 1989) and there are a smattering of others (cf. Williamson 1933a:34, fn. 4; 35, fn. 1-6, Ellis 27 1829a, 1829b). Henry's (1928:336-4) traditions were collected in 1822, 1824, and 1833. Barrère (1966:104-7) provides an overview, mainly in the context of questioning whether the belief in Ihoiho was ancient.

- 237

Ellis (1829b:42-3) was interested in what he said was a “Hindu” notion that creation began with an egg. Try as he might, he could not find people who would relate similar notions to him in Tahiti at that time. In Henry (1928:336-44) is the specific notion of Ta'aroa as a pre-existing being, living out the primordial eternity inside a kind of egg shell out of which he finally broke and fashioned the world as we know it today. We can raise the possibility that Ellis' asking after an egg tradition motivated this Tahitian innovation. Ellis (1829a:5) first embarked for Tahiti in 1816 and last left those islands on 31st December 1822 (1829b:574). The “egg” traditions were first collected in 1822 and later so it is possible that Ellis made inquiries about an egg tradition and that this inspired further revision of cosmogonies that were already seeing some European notions creeping in (cf. Barrère 1966:104). In any event, the egg notion is found in Polynesia and outside the Societies only in certain of the Tuamotus and cannot be shown to be of general antiquity around Polynesia.

Within these highly transformed traditions the search for remnants of a Primordial Pair and first-order anthropomorphic gods of the Marquesan or New Zealand Maori variety is not clearly fruitful (cf. Marck forthcoming). We shall not repeat those materials here but simply note that of the three possible remnants of the Primordial Pair in Henry's (1928) traditions (see Table 1), none produce a sibling group of first-order anthropomorphic gods in the typical fashion of Central Eastern Polynesian speaking groups' traditions.

As elsewhere, the sky is said to be held down against the earth. In at least one Tahitian tradition it is an octopus which holds down the sky. Williamson, citing Tyerman (1831a:526) relates:

There was a belief in the Society island of Ra'iatea that the sky originally lay flat on the face of the earth and ocean, being held down by the “legs” (? tentacles) of a huge cuttle-fish; but Maui dived down to the bottom of the sea and dismembered the cuttle-fish, whereupon the sky flew up, became convex, resting on the horizon and having the vertical sun as its keystone. (Williamson 1933a:42)

Ellis (1829:43) relates a tradition that “at the first the heavens joined the earth, and were only separated by the teva plant, draconitum pollyphillum, till their god, Ruu, lifted up the heavens from the earth” and Henry (1928:409-13) relates a similar story involving both Ruu and Maaui. Williamson (1933a:34-5 citing Hale 1846:23, 25, Moerenhout 1837:450, Young 1898:109, Forster 1778:159, 541, Smith 1903:239, Ellis 1829a:167, Cuzent 1872:44, and Bovis 1863:274) relates that in the - 238 Societies Maaui is credited with raising the islands from the sea. These are the Society Islands sources of which I am aware for traditions concerning the raising of the sky and the fishing up of the islands. It is consistent with the common notion of Central Eastern Polynesian speaking groups that *Maaui was responsible for these deeds. The Tuamotus

The Tuamotus are numerous and cover a vast area. As the traditions to be related here come from only a few of those islands, it is important not to view the known Tuamotuan traditions as indicative of the archipelago as a whole.

Craig (1989:29) mentions only Henry's (1928:347-9) Tuamotuan creation tradition which does not give the home island of Pairore “the chief and regent” from which it was collected (Henry 1928:347, fn. 30). Young (1919) also mentions Pairore but does not mention the area of the Tuamotus from which he came. Williamson (1933a:15-6) mentions Tuamotuan creation traditions collected by Montiton (1874) from the islands of Fangatau and Takoto and another collected by Smith (1903:221-42) for which the island source is not mentioned. Caillot (1932:37-57) is another primary source and some of those traditions were recorded on Makemo and Hao, though it is not clear if those to be mentioned here were.

The account of Paiore collected by X. Caillet 28 in 1890 and published by Henry (1928:347-9) is notable in that it contains elements of the transformed Tahitian traditions as well as elements that had not undergone such extreme transformations. In that tradition Tumu and Papa, not Tangaroa, were contained inside the primordial egg which burst open “and produced three layers, one below propping two above. Upon the lowest layer remained Te-Tumu and Te-Papa, who created man, animals, and plants” (Henry 1928:347). In this account, Tangaroa is born to other early anthropomorphic gods or godlike people (neither with cognate names in Polynesia) and an initial raising of the sky occurs before the birth of Tangaroa. Tangaroa's realm is darkness and the nether lands, which is reminiscent of Marquesan. 29 The only Tangaroa mentioned is the father of the only Maaui mentioned, much in the mould of Nuclear Polynesian traditions as a whole. Another Tuamotuan tradition follows in Henry (1928:349-53) but no date or home island of the narrator is given. It concerns Aatea, Taane and Maaui, giving parents of Taane that are not reported elsewhere and not counting any of the three as siblings or descendants of Tumu and Papa and there is nothing further in that tradition to link with the “group of siblings” theme in the conceptualisation of the first-order anthropomorphic gods amongst some other Central Eastern Polynesian speaking groups.

- 239

The traditions mentioned by Williamson (1933a: 15-16) are covered in two paragraphs, the Moniton tradition including an account of a battle between Aatea and Taane (mentioned also in Henry 1928:349-53) and of Taane's raising the sky. The Smith (1903) tradition is highly abbreviated in Williamson (1933a: 16) but, as usual, very true to the source. There is no sibling group in the typical Central Eastern Polynesian pattern in the primary source.

The Tuamotu tradition of Taane raising the sky, as amongst the New Zealand Maori, is a quite mysterious similarity and may be due to independent developments or some kind of borrowing between certain of the Tuamotus and the New Zealand Maori, or Tuamotuans having a part in the settlement of New Zealand.

In a very short passage (Caillot 1932:50-1), Aatea is mentioned as sleeping with Hotu, the result being the sibling group “Ru, Pigao, Tope, Pepe, Tane, Tagaroa, Titi, Tiki, Ruanuku, Maui, Gaohe, Vaerua”. Calliot (1932:51) mentions that Vaatea, Taane and Tagaroa are “trois formes d' une même divinité” so it would seem that older traditions, based on the Primordial Pair-sibling group pattern, were current but being transformed in light of Christian notions.

5.3.1 *Aatea

PCE *Aatea ‘space deified’, MQA Aatea ‘space deified/personified and first-order anthropomorphic god’, HAW Waakea, PTA *Aatea, Var.: *Awaatea, *Waatea, TUA Aatea, Vaatea, Tongarevan Aatea, Mangaian Vaatea ‘space deified and Primordial Male’, MAO Rangi-Aatea ‘sky deified and Primordial Male’ TAH Aatea ‘space deified and father of Taane

Biggs (1994): PPN *qaho-qatea, PCE *awatea ‘midday’, PPN *qaho, PCE *ao ‘day’, PPN *qaatea, PCE * aatea ‘clear, unobstructed’

The PCE name *Aatea denoted the space between the sky and the earth which was compressed in the primordial state by the sky's hugging close to the earth. Reconstruction of a precise status to the PCE level is not possible because of a basically different conceptualisation of *Aatea's role in creation between the two CE subgroups and an absence of external cognate names or even a similar type of entity with similar deeds. 30 In the instance of Marquesic, we have a clear statement of his origin for the Marquesas, an origin so ancient it was forgotten in Mangareva and one that seems influenced by Tahitic in Hawai'i. In the instance of Tahitic, we have agreement - 240 amongst many of the Cooks and at least one Tuamotuan tradition that *Aatea was the Primordial Male while New Zealand Maori traditions conceptualise creation in a similar way but have largely forgotten the *Aatea name except in the Rangi-variant Rangi-Aatea. Tahitian has a basically transformed conceptualisation of creation and this is shared to some extent with some of the Tuamotus and some of the Southern Cooks. On the basis of various Cook Island agreements with New Zealand Maori and fragments of certain Tuamotuan traditions, we can reconstruct: PTA *Aatea ‘space deified and Primordial Male’ with some confidence.

Marquesan Aatea is a first-order anthropomorphic god in the sibling group who were the children of Papa- 'Uka and Papa- 'A 'o, the Primordial Pair in the widespread Polynesian pattern. From within the earth, he kicked through to open space so that he and his siblings could reside on the earth's surface; hence, apparently, his name. All Marquesans trace their ancestry to him and his wife Ata-Nua.

A rather transformed Hawaiian set of traditions agrees to some extent with the PTA Primordial Pair of *Aatea and *Papa in that Waakea and Papa were great creator gods in Hawai'i, responsible for the origin of many of the islands, and the parents of Kaane and Kanaloa. This fits with Tahitic notions but Waakea and Papa are remembered as the progenitors of all Hawaiians, very much as the Marquesans remember Aatea and Ata-Nua as their ancestors. Tahitic groups do not trace their ancestry to PTA *Aatea except through one of the first-order anthropomorphic gods, so the Hawaiian memory of being descended from Waakea may be Marquesic.

It is not clear which group has changed what, since PCE times, and two possibilities can be imagined without much difficulty, both with two possible explanations for the situation in Hawai'i:

  • 1. That there was a Primordial Pair named *Aatea and *Papa in PCE and that this continued into PMQ and PTA but was modified in MQA where *Aatea came to be amongst the first-order offspring of the Primordial Pair. In this instance Hawaiian Waakea and Papa could be either from direct inheritance of PMQ speakers' beliefs or from indirect inheritance (a borrowing) from TA.
  • 2. That there was a PCE Primordial Pair Papa-adj, and Papa-adj, and that *Aatea was amongst their first-order anthropomorphic offspring. In this instance PTA speakers would be seen to have elevated him to the status of the Primordial Male while MQA simply did not change. The Hawaiians' Waakea and Papa would then be either a borrowing from Tahitic or a local transformation of earlier Marquesic notions.
- 241

The first solution is not very satisfying and was not adopted in Marck (forthcoming) as it requires a renaming, in Marquesan, of the Primordial Male as Papa-adj, and the claim that the similarity to the naming of Samoa's Primordial Male, also Papa-adj., is an independent development. Linguists, perhaps, come to such problems with an aversion to independent development solutions, or at least the present linguist has, and the solution taken elsewhere (Marck forthcoming) is the second stated above, because it does not require independent similar developments in Samoan and Marquesan. By this solution, the Waakea-Papa pair of the Hawaiians could be a Tahitic loan or independent local development of Waakea as Papa's spouse. The Tahitic loan hypothesis is appealing because the Hawaiians remember them as coming from Tahiti. The hypothesis that has this as an Hawaiian transformation of old Marquesic beliefs is appealing because it is a particularly Marquesan notion that Aatea is the ancestor of all people and this is shared with Hawaiian traditions but is not Tahitic other than in some Cook Island localities (see Cook Island section above). *Taane is more commonly the ancestor of people in the Tuamotus, Societies and amongst the New Zealand Maori.

Again, the position of *Aatea at the PCE level is indeterminate by our method but he may have been one of the first-order anthropomorphic gods and is mentioned here for that reason.

5.3.2 *(H)aumia

PCE *Haumia, MQA Aumia ‘first-order anthropomorphic god’, MAO Haumia ‘god of fern root, first-order anthropomorphic god’

This is a marginal reconstruction with only two cognates which agree imperfectly. I have not yet identified common vocabulary from which the name may have been taken. Haumia is well known from New Zealand Maori traditions as a first-order anthropomorphic god who was the god of the fern root, an important source of starch to the New Zealand Maori as taro and breadfruit do not flourish in their southern environment. Handy (1923:244) mentions Aumia among the siblings of Aatea, Taane, Tuu, 'Ono and others but does not mention his (or her) realm and other primary and secondary sources do not mention a Marquesan Aumia at all. If they are cognate, the New Zealand Maori Haumia agreement with Marquesan Aumia is irregular in that New Zealand Maori has irregularly inserted an initial consonant or Marquesan has irregularly lost one. It is also possible that Handy did not transcribe an initial h (although we otherwise know he heard and transcribed that sound in that environment in that language).

- 242

It is possible that this is a chance resemblance since I have not found it reported for other Polynesian peoples but the possibility of a chance resemblance would be about one in four thousand and nothing near that number of comparisons were made in the present work. Possibly it is a memory, on the part of Marquesans, of a name from a time when fern root was an important food around central Eastern Polynesia, as it may have been at the time of initial settlement before breadfruit and taro were abundantly available.

Craig (1989:53) equates New Zealand Maori Haumia-Tiki-Tiki with Hawaiian Haumea, Marquesan Haumei, and Tuamotuan Faumea and calls New Zealand Haumia-Tiki-Tiki a “goddess” in the “Haumea” entry while calling Haumia-Tiki-Tiki a “god” and son of Rangi and Papa in the “Haumia-Tikitiki” entry. This linguist's inclination is to suggest that at least three names are involved in those and an associated entry: the Haumea of Hawai 'i and the Faumea of the Tuamotus, which seem regularly cognate names but share nothing of the same deeds as related in Craig (1989:38,53); the Haumei of the Marquesas, whose form is regularly cognate with none of the others and seems (Craig 1989:53) to share no deeds with the others; and the Haumia-Tiki-Tiki of the New Zealand Maori, which is probably not cognate with the others but possibly cognate with the Marquesan Aumia mentioned above.

5.3.3 *Mauri

PCE *Mauri, MQA Moui, Tongarevan Mauri ‘first-order anthropomorphic god’

Biggs (1994): PPN *maquri, PCE *mauri ‘life, alive’

This reconstruction is almost as obscure as the previous one except that more groups remember PCE *Mauri as a god, though not a first-order anthropomorphic god. We would expect **Mou'i 31 for Marquesan but the source (Handy 1923) did not record Marquesan glottal stops. Neither Handy (1923:244) for Marquesan nor Hiroa (1932:85) for Tongarevan describe a realm or deeds for this god, only that he was amongst the sibling group first born to the Primordial Pair. We must note the reconstruction as weak. It is quite possible both the Marquesans and Tongarevans raised a more junior god to first-order status as many siblings of the main gods are mentioned (but not realms or deeds) and they are rarely cognate between groups (for groups that have the “group of siblings” theme).

- 243
5.3.4 Roŋo

PCE, PMQ, PTA *Roŋo, MQA 'Ono/'Oko, TUA Rogo, Rarotongan, Tongarevan, Mangaian, MAO Rongo ‘first-order anthropomorphic god’, TAH Ro'o, HAW Lono ‘one of the earliest gods’

Biggs (1994): PPN, PCE, PMQ, PTA *roŋo ‘to hear’

PCE *Roŋo, probably PCE *Roŋo-Nui. is remembered as a messenger god in Tahiti, a role in which the Samoan god with a partially cognate name, Logo-Noa. is also found (Fraser 1892:265), although not as a first-order anthropomorphic god as he often is in Central Eastern Polynesian groups. In Samoa he was the messenger of Tagaloa (Craig 1989:142) and in Tahiti the messenger of Taane (Henry 1928:369-71). He was often a god of food in Tahitic traditions, e.g., Tongareva (Hiroa 1932:87), Mangaia (Gill 1876:11-2) and amongst New Zealand Maori (Tregear 1904:462) and this was also true of Mangareva (Hiroa 1938b:422) so possibly he was the god of food amongst PCE speakers. Little is known of him from the Marquesas (Handy 1923:245) and he is not mentioned as a god of food in the Tuamotus and Tahiti in sources I have seen. Still, the agreement of Mangarevan and Cooks/ New Zealand is enough to establish his status as god of food in PCE if this is not a borrowing on the part of Mangarevan. There is an interesting passage in Gill (1876:10-4) which states that Rongo and Tangaroa were twins and the first born of Vaatea and Papa in Mangaia. “Tangaroa should have been born first, but gave precedence to his brother Rongo” (Gill 1876:10). There followed a great deal of competition between the two as to who would have dominion over what. The passage is significant mainly because it says something about the birth order of the first-order anthropomorphic gods, something which is otherwise left to the imagination in Central Eastern Polynesian accounts of these gods (unless we take their order in being named to be the birth order.).

Ro 'o in Tahiti and Lono in Hawai'i are early gods, as early as most others in these highly transformed traditions, but they are not known to be offspring of the Primordial Pair in those localities.

5.3.5 Taane

PCE, PMQ, PTA *Taane, MQA Taane, HAW Kaane, TUA, Tongarevan, Mangaian, Rarotongan, MAO Taane, ‘first-order anthropomorphic god’ TAH Taane ‘early god’

Biggs (1994) PPN *taqane, PCE, PMQ, PTA *taane ‘male’

- 244

The evidence for PCE, PMQ, PTA *Taane ‘first-order anthropomorphic god’ is abundant and unequivocal. He is a child of the Primordial Pair in at least one tradition from the Marquesas, the Tuamotus, New Zealand Maori, Rarotonga, Mangaia and Tongareva 32 and he is found as a member of a sibling group including other such gods as seen in Table 2 along with PCE *Tanŋroa in Tahitic traditions and Hawai'i and PCE *Aatea in the Marquesas. Many books could be written about *Taane in Tahitic traditions. Our purpose here is complete with the simple demonstration of his status in PCE and some mention of his status in PMQ and PTA. He is not known from Tonga or Samoa.

There are basic differences between the Marquesan Taane and that of Tahitic accounts. I will comment upon Kaane in Hawai'i after mentioning some of Taane's characteristics in Marquesan and Tahitic.

Tregear (1904:453) wonders that the “sublime Trinity-worship” of Taane, Rongo and Tuu is not found in New Zealand when it was conducted “with such solemn ritual and embodied in such magnificent hymns… in the Hawaiian Islands and the Marquesas.” I am aware of some motive for his question in the case of Hawai'i but have found nothing to support his assertion in the case of the Marquesas. Handy relates that:

Tane was of little importance in the Marquesas. His name appears in legend and chants, but not in genealogies… The concepts of a male principle and light belong to Atea rather than to Tane here, though Tane was regarded as a “light” god in the sense that he was believed to have had light skin and hair and to have been the ancestor of the white race. (Handy 1923:245)

He is described as “of little importance” by Handy but first-order anthropomorphic gods are often described as such (cf. Craig 1989:99 on Samoa's Tagaloa) because they are not associated with much religious ritual. A single Marquesan tradition of which I am aware credits the raising of the sky to Taane in association with Aatea and a similar story is known from the Tuamotus (Williamson 1933a:26-7, cf. also Caillot 1932:60) but another credits Marquesan Tono-Fiti with this act (Handy 1923:245), while closely related Mangarevan has memories of Maaui having done so (Williamson 1933a:44). This seems the older pattern, given the agreement with Tongan, Samoan, Hawaiian and some Cook Island traditions on the matter. Hiroa (1938b:424) mentions a rather marginal status for Taane in Mangareva.

Taane in Tahitic, on the other hand, is normally quite prominent. The Tahitian traditions are generally idiosyncratic and knowledge of Tuamotuan traditions quite limited so we shall turn to the Cooks for our first look at Taane in Tahitic. Starting in Tongareva, Hiroa (1932:85-6) notes that Taane - 245 was a child of Aatea and Haka-Hotu and an active god at the time of creation. Taane, or a Taane incarnation, is in one of three main lines of descent in the genealogies (Hiroa 1932:17-9). Other than this, nothing is said of him in that work.

Gill (1876:11) relates that Taane was a principal god of Mangaia along with Tangiia, a Mangaian first-order anthropomorphic god for whom I know of no cognates in Polynesia. 33 Gill does not expand upon his assertion that Taane was a principal god of Mangaia. There is no description of deeds attributed to him apart from a tradition of his marrying “Ina” ('Ina from PCE *Hina) (Gill 1876:107-14). However, outside Tahitic, *Hina is not generally associated with *Taane so much as with *Taŋaloa and *Maaui and the story line has no elements which seem similar to those other traditions. It would seem to be Rongo rather than Taane to whom the “original” Mangaians trace their ancestry (Gill 1876:17).

Taane in Rarotonga is mentioned briefly by Savage (1919:57-61). We find there only the mention of his being amongst the first-order anthropomorphic gods along with Rongo, Tuu, Tangaroa and Rua-Nuku.

Taane is relatively well known from New Zealand Maori sources. He separated Rangi and Papa (Grey 1885:l-3,Tregear 1904:461,Best 1924:37-8, 1925:745), adorned “the breast of heaven (his father, Rangi) with stars” (Tregear 1904:435, cf. 461; cf. Best 1924:39) and covered the earth (his mother, Papa) with trees (Andersen 1928:403-4), was the god of forests (Grey 1885:2, Tregear 1904:439), made the first woman (Andersen 1928:407, Cowan 1930a:8), made the first man (Tregear 1904:464) or was the “Father of Man” (Cowan 1930b:44)or “God of Man” (Cowan 1910:105), and fought battles with other gods (Grey 1885:5, Tregear 1904:478-9). Altogether, there is a great deal of content against which we can begin to evaluate the prehistory of traditions around Tahiti and in Hawai'i.

We will consider Tuamotuan traditions briefly before going on to Tahitian. The material I will cite is from Caillot (1932), which has been mentioned in the summary of Tuamotuan sources as more conservative in some respects than others. Recall that in those materials Taane is amongst a sibling group of first-order anthropomorphic gods in the broad Central Polynesian pattern, children of Vaatea and Hotu in this instance (Caillot 1932:50-1). In a following passage (Caillot 1932:60) a number of things are attributed to Taane. Roughly translated, they include the raising of the sky, the covering of the barren earth (with plant life), holding the status of the principal force of life in all beings, fathering universal life, responding to needs of the heart, and arbitrating the destiny of people's lives (while living but not after their death).

- 246

In Tahiti there is no sibling group of first-order anthropomorphic gods nor a clear Primordial Pair in the broader Polynesian pattern. Nevertheless, Taane is the sole child of Aatea and Papa, who are the Primordial Pair elsewhere in Tahitic even though their genders are reversed in the tradition related by Henry (1928:364-9). 34 Henry's (1928:353-71) traditions have a great deal to say about Taane, though little of it resembles what has been seen for New Zealand Maori's Taane or that of the Tuamotus. Aside from Ta'aroa, who is elevated to a pre-existing creator god, Taane figures most prominently in these traditions. Often it is adulation reminiscent of Judeo-Christian eulogising about the glory of their deity or a setting out of the world in a fashion more similar to the Old Testament order than other Polynesian memories of creation. Nothing is very similar to other Tahitic accounts as regards Taane except his high status.

The question of what Hawaiian memories of Kaane are most similar to is muddled by 1) the transformation of Hawaiian traditions, much of it apparently after European contact, 2) the possible transformation of Tahitian traditions after early borrowings by Hawaiians, and 3) the dearth of materials from the Marquesas.

Beckwith (1970:60-80) relates a number of traditions concerning Kaane in which Kanaloa is normally involved somehow. In one (Beckwith 1970:60-1) Kaane and Kanaloa are said to draw men in the earth and Kaane's lives while Kanaloa's does not. Beckwith's rendition of this tradition varies somewhat from the source (Fornander 1919-1920:267-8) but the point here is that Kaane is associated with the origin of people. This is a Tahitic belief but not a Marquesan belief (where Aatea is the progenitor of people). There is a fight between Kaane and Kanaloa (Fornander 1919-1920:268) in which Kaane prevails. It was Aatea that fought off Tana'oa in the Marquesas (Fornander 1878:214-218), where Tana'oa was thought of as a god of darkness, as was Hawai'i's Kanaloa. This is something which happens nowhere in Tahitic, or, apparently, in Mangareva. But like the Marquesan Tana'oa, the Hawaiian Kanaloa is still god of the sea. So there are a few resemblances to Marquesan in Hawai'i as concerns Tana'oa and Kanaloa but the prominent role of Kaane cannot be identified as having a Marquesic source and seems most similar to Tahitic Taane.

While the role of Kaane in Hawai'i resembles that of *Taane in Tahitic, there seem to be no specific resemblances between Kaane traditions from Hawai'i and *Taane traditions from Tahitic localities. The activities of Kaane and Kanaloa generally seem to take place in Hawai'i itself, rather than *Hawaiki as in Tahitic traditions, so there is little to compare other than prominence of the role itself, which could have occurred independently.

Finally, the Hawaiians remember Maaui lifting the sky (Beckwith - 247 1970:230,379) rather than Kaane. This could be either a Marquesic memory or one from around Tahiti. Taane took that role in the Tuamotus and New Zealand (but not in parts of the Cooks and Tahiti, which still remember Maaui as having done so (cf. Marck n.d.)).

Thus we find no compelling reason to suggest Hawaiian Kaane's high status is a loan from early Tahitic, a retention from early Marquesic or an independent development. Superficially there is the greater resemblance to Tahitic accounts but there is an absence of specific elements in the conceptualisation and deeds of the Hawaiians' Kaane other than the notion of the creation of the first man which could be independent developments as the names of the men created are not cognate between Hawaiian and Tahitic. With the borrowing of the Tahitic pronunciation for “priest” into Hawaiian 35 and the possible borrowing of the Hawaiian names for the Primordial Pair from Tahitic (Marck forthcoming) we might be inclined to view the role of Kaane as a further Tahitic influence. But we cannot demonstrate such by our method.

PCE *Taane may have had a high status that was displaced by Aatea in the Marquesas or he may have been the relatively obscure god that he was in the Marquesas and been elevated to greater prominence in many of the Tahitic groups. In any event, he seems to have been among the first-order anthropomorphic siblings born to the Primordial Pair in the cosmogony of PCE speakers.

5.3.6 *Taŋaroa

PPN, PNP *Taŋaloa, PCE, PMQ, PTA *Taŋaroa, TON *Tangaloa, SAM Tagaloa, MQA Taka'oa/Tana'oa, MVA Tagaroa, HAW Kanaloa, TUA Tagaroa, Tongarevan, Mangaian, Rarotongan, MAO Tangaroa ‘first-order anthropomorphic god’, TAH Ta'aroa ‘pre-existing creator god’

Biggs (1994) PPN *Taŋaloa ‘a principal god of the pantheon’, PPN *taŋa ‘1. bag, 2. drive into enclosure, surround’, PPN *loa ‘long’

It is tempting to reconstruct “senior first-order anthropomorphic god” for PCE *Taŋaroa but the “senior” turns out to be difficult to support from within CE. While it is true that Tangaloa was the senior of three half-sibling first-order anthropomorphic gods in Tonga, the only anthropomorphic god of first rank in Samoa and elevated to pre-existing creator god in Tahiti, Central Eastern Polynesian traditions other than Tahitian do not generally specify that he was senior. Gill's (1876) account for Mangaia, mentioned above in the context of PCE *Roŋo, specifically has Tangaroa and Rongo as twins and first born of the Mangaian sibling group of first-order anthropo- - 248 morphic gods. But he seems not to be otherwise specifically mentioned as the eldest around Central Eastern Polynesian groups. We might suspect that the Tahitians took him to be their pre-existing being because he was the most senior of the siblings, but except for the Mangaian account, there is nothing in materials internal to Central Eastern Polynesian to suggest that he was most senior. We might also mention that he was the only god or man common to Central Eastern Polynesian traditions other than Tiki to be remembered by the Rapanui people. Tangaloa is said to have died on Rapanui in those traditions (Englert 1970).

In Central Eastern Polynesian groups, unlike Tonga and Samoa, *Taŋaroa seems never to be associated with the fishing up of the islands. Caillot (1932:60) mentions him as finishing the sky-raising work of Taane, but I know of no other mention of him in connection with the sky-raising story in Polynesia except Fraser's (1892) suspicious Samoan tradition.

He is the god of craftsmen in the Tuamotus (Caillot 1932:60-1) so given the agreement with Tonga (Collocott 1921:153) on this matter, we can reconstruct that status to PPN, PNP, PCE, and PTA. He is the god of the sea amongst the New Zealand Maori (Tregear 1904:462), as in the Marquesas (Handy 1923:245), so we can reconstruct this status to PCE, PMQ and PTA.

His loss of rank in the Marquesas is curious. There he is remembered as a thirty-ninth generation offspring of the Primordial Pair, but not so in Mangareva, where he is one of the gods so ancient in the genealogies that previous ancestors are unknown and their origin is unspecified. The diminution of Kanaloa in Hawai'i may be recent and independent of Marquesan. As mentioned previously, Beckwith (1970) mentions only the “trinity” of Kaane, Kuu and Lono. However, a neglected but more authentically Polynesian sort of tradition can be found in Fornander (1916-1917:18), which has Kanaloa a full sibling of Kaane born of Waakea and Papa. Beckwith (1970:cf. 60, 62) remarks more than once that Kaane and Kanaloa are normally mentioned in the same breath. Her sources extend far beyond Fornander (1878-1885, 1916-1920), and are often in the context of propitiatory recitations. This would seem to indicate that Kaane and Kanaloa held similar sway over human affairs before Kanaloa's apparent demotion in the Kaane-Kuu-Lono “trinity” in the historic period.

5.3.7 Tonga-Fiti

PCE *Toŋ-Fiti, MQA Tono-Fiti, Mangaian Tonga-'Iti ‘first-order anthropomorphic god’, Rarotongan Tonga-'Iti ‘early god’

Biggs (1994): PPN, PNP, PCE, PMQ, PTA *tonga ‘south, south wind’, - 249 Marck: PPN, PNP, PCE, PMQ, PTA *Toŋa ‘Tonga’, PPN, PNP, PCE, PMQ, PTA *Fiti ‘Fiji’

This is a marginal reconstruction. The second vowel of the Marquesan form does not agree with that of the others and the glottal stops of Mangaia and Rarotonga are assumed (the sources, Gill 1876:9-12 and Savage 1919:58, do not mark glottal stops). The only deeds attributed to any of them is the lifting of the sky by Tono-Fiti in the Marquesas (Henry 1923:245) and that of Tongo-'Iti parenting the children who “opened up” Papa, the earth, in Rarotonga (Savage 1919:58). Langridge and Terrell (1988:194-7) also recount a tale about Marquesan Tono-Fiti but not one which comments on cosmogony. Christian (1895:187) relates that his domain was “Havaiki” in the Marquesas, i.e., the afterlife or paradise.

While the reconstruction is marginal, it allows us to recall the topic of the extent to which Tonga and Fiji, in addition to *Hawaiki, were remembered around central Eastern Polynesia. It seems that Marquesans, Tuamotuans, Tahitians and Cook Islanders all remembered the Tonga/Fiji area, more or less. In the Marquesas the islands of “Tona-Tapu”, “Vevau”, “Fiti”, “Fiti-tapu” and “Havaii” are remembered as ancestors in genealogical recitations in the ninth to eleventh generations after the Primordial Pair. Borabora (“Pai-o-Poapoa”) is remembered as an ancestor in the twenty-fifth generation after the Primordial Pair (Christian 1895:191). Specific Marquesan memories of these localities as real-world geographical places can be seen in Langridge and Terrell (1988, cf. 104-9).

If the Marquesans remember these places it should not be surprising that Tuamotuans and people progressively closer to Samoa/Tonga/Fiji might also do so and such references are found from time to time in the cosmogonic materials of at least the Tuamotus, Societies and Cooks. I will not present them here but simply want to make the point that if there is any further reluctance to accept Hiroa's (1938a) assertion that the Central Eastern Polynesian memory of *Hawaiki refers specifically to Savai'i in Samoa, how then can we dismiss specific associated memories of Fiji and Tongan islands as well?

5.3.8 Tuu

PNP *Tu(q)u ‘primary god of war’, PCE, PMQ, PTA *Tuu, MQA, MAO Tuu, TUA Tuu ‘first-order anthropomorphic god; primary god of war’, TAH Tuu ‘early god and artisan’, HAW Kuu ‘early god of growth, rain, forests’, MVA Tuu ‘god of breadfruit’

- 250

Biggs (1994): PPN, PNP *tuqu, PCE, PMQ, PTA *tuu ‘stand, be upright’

There can be little doubt that PCE *Tuu was a first-order anthropomorphic god and that his status as god of war continued one held in PNP, though in this case we can be less certain about his parentage. He was the primary god of war but not a child of the Primordial Pair in Samoa (Turner 1884:61). In Hawai'i Kuu was part of what had become a trinity by the early years of European contact (Beckwith 1970) but his position was not generally so illustrious elsewhere. Specific deeds are not associated with him in many localities although he had a major role in the argument with his siblings about raising the sky in the case of the New Zealand Maori (Grey 1885:1-7). In the Marquesas he is mentioned simply as a first-order anthropomorphic god, patron of war and as absent from the genealogies (Handy 1923:245). In Mangareva he was patron of breadfruit rather than war (Hiroa 1938b:422). He is not mentioned as among the children of the Primordial Pair in Tongareva (Hiroa 1932:85) or Mangaia (Gill 1876:1-22) or Rarotonga (Savage 1919:57-61). Caillot (1932:62-4) mentions Tuu as a first-order anthropomorphic god for one Tuamotuan locality and a Tuu incarnation as god of the sea but does not mention him as god of war in that locality.


We have examined a topic that is, perhaps, more meaningful for Central Eastern Polynesian groups than for Tonga and especially Samoa. We have reconstructed a sibling group of first-order anthropomorphic gods, children of the Primordial Pair, for Proto-Central Eastern Polynesia. The sibling group, or at least three half-brothers, is also found in Tonga but not with names cognate with what we have reconstructed for PCE other than in the case of PCE *Taŋaroa. Samoa's Tagaloa is also a first-order anthropomorphic god but seems to have had no siblings not also named Tagaloa. Thus the sibling group may be independent developments in Tongan and PCE cosmogonies.

PCE *Roŋo and *Tuu, part of the sibling group in PCE, are clearly cognate with Samoan Logo and Tuu and reconstruct to PNP but not as first-order anthropomorphic gods. PCE *Taane is the final member of the sibling group about whom we can express a high level of confidence, occurring in both MQ and TA with widely agreeing attributes mainly in TA. PCE *(H)aumia, *Mauri and *Tonga-Fiti are rather marginal reconstructions but allowed by our method.

The reconstruction of *Aatea as a first-order anthropomorphic god in PCE is supported only by MQ, TA suggesting that he was the male of the - 251 Primordial Pair. If we take Hawaiian traditions to be mainly Marquesic, then we need to reconstruct PCE *Aatea as the Primordial Male. But Hawaiians have clearly borrowed their pronunciation of the “priest” word (kahuna) from Tahitic, and they only remember coming from Tahiti. Furthermore, reconstruction of *Aatea as the Primordial Male in PCE displaces an appealing agreement between Samoan Papa-adj, and Marquesan Papa-adj. for the Primordial Male from which agreement PCE *Papa-adj. ‘Primordial Male’ is reconstructed (Marck forthcoming).

We have not made a great deal of progress in suggesting whether Hawaiian traditions seem more Marquesic or more Tahitic. Mostly they are simply Hawaiian and show no special resemblances to either Marquesic or Tahitic. But we have identified a passage from Fornander (1916-1917:18) which contains an Hawaiian tradition more in the mainstream of Central Eastern Polynesian cosmogonic traditions than those summarised by Beckwith (1970) and Craig (1989).

The outstanding characteristic of this kind of work is its multiplicity of facets. There are many actants. There are many localities. There are competing traditions within those localities. I have here and elsewhere (Marck forthcoming, n.d.) tried to reduce the problem to a manageable level by choosing small questions about which we have some hope of giving comprehensive treatment, or at least of mentioning those Polynesian traditions or motifs from traditions which occur in more than one locality.

The availability of encyclopaedic secondary sources, especially Williamson (1933a, 1933b) and Craig (1989) has greatly facilitated the work. Not all primary sources were available during the course of this work and not all primary sources at hand were searched as exhaustively as they might have been. On the other hand, relevant material on the present topic tends to be limited to a very few paragraphs or pages in each source except the shortest sources such as chants, genealogies or individual legends.

The purpose of this and related presentations (Marck forthcoming, n.d.) was to revive hope that work on such topics might be fruitful. The distributions that confounded earlier students of the situation, especially Handy (1927) and Williamson (1933a, 1933b), can now be seen in a different light given the current phylogeny of linguistics for the area. It is now clear that certain distributions of cosmogonic notions, actants and their deeds occur exclusively within certain linguistic subgroups and that languages and cosmogonies may have differentiated along similar lines during the prehistory of the area.

In applying the method proposed here I have felt least secure about the notion that there was a unified Proto-Tahitic cosmogony in the same sense - 252 that there was a unified Proto-Tahitic language. Certain elements of Cook Island beliefs seem quite conservative compared to developments common to Tuamotuan, the Societies and the New Zealand Maori. An example is the central status of Taane in the latter groups, while the Cooks have rather less interest in Taane and some, along with Tahitian, retain a belief that Maaui and not Taane raised the sky. These traditions were (and are) maintained mainly by families and their political networks. Our modelling will probably approximate reality more closely if it assumes a certain amount of diversity in the past. Such was not done here but it was the purpose of the present work to identify central themes in subgroups and in accomplishing this other interesting material was neglected or beyond the scope of the present narrow topic.


Frequently used language and subgroup names are often cited using the conventions familiar from Polynesian linguistics:

1. Subgroups:

  • CE Central Eastern Polynesian
  • MQ Marquesic
  • PN Polynesian
  • EP Eastern Polynesian
  • NP Nuclear Polynesian
  • TA Tahitic

2. Proto-languages of these subgroups have P prefixed to the subgroup name, e.g.:

  • PNP Proto-Nuclear Polynesian

3. Other languages:

  • HAW Hawaiian
  • MQA Marquesan
  • MAO New Zealand Maori
  • TUA Tuamotuan
  • TAH Tahitian

This work was done while on an Australian National University Ph.D. scholarship.

- 253 Page of endnotes

- 254 Page of endnotes

- 255
  • Andersen, J. C., 1928. Myths and Legends of the Polynesians. London: George G. Harrap and Co.Ltd.
  • Barrère, D., 1966. Revisions and adulterations in Polynesian creation myths, in G. A. Highland (ed.), Polynesian Culture History: essays in honor of Kenneth P. Emory. Honolulu: Bishop Museum, pp. 103-19.
  • Beckwith, M. W., 1970(1940). Hawaiian Mythology. Honolulu: University of Hawai'i Press. (1940 Yale University Press.)
  • Best, E., 1924. Maori Religion and Mythology. Wellington: The Polynesian Society.
  • —— 1925. Tuhoe: The Children of the Mist, Vol 1. New Plymouth: Thomas Avery.
  • Biggs, B., 1994. Comparative Polynesian Lexicon (POLLEX) Computer data base, Department of Maori Studies, University of Auckland.
  • —— n.d. The position of Tangaroa in Maori mythology, (ms.).
  • Bovis, L. de, 1863. Etat de la Société Taitienne, à l'arrivée des Européens (sp). Annuaire des Etablissements Française de l'Océanie, Pape'ete.
  • Burrows, Edwin G., 1938. Western Polynesia: a study of cultural differentiation. Etnologiska Studier (Goteborg), Volume 7.
  • Caillot, A.-C. E., 1932. Histoire des Réligions de l'Archipel Paumotu. Paris: Librairie Ernest LeRoux.
- 256
  • Christian, F. W., 1895. Notes on the Marquesans. Journal of the Polynesian Society, 4:187-202.
  • Clark, K., 1896. Maori Tales and Legends. London: D. Nutt.
  • Clark, R., 1976. Aspects of Proto-Polynesian Syntax. Te Reo Monographs, Auckland: Linguistic Society of New Zealand.
  • Collocott, E. E. V., 1921. Notes on Tongan religion. Journal of the Polynesian Society, 30:152-63, 227-40.
  • Cowan, J., 1910. The Maoris of New Zealand. Christchurch: Whitcombe and Tombs.
  • —— 1930a. Legends of the Maori, Volume 1: mythology, traditional history, folk-lore and poetry. Wellington: Harry H. Tombs.
  • —— 1930b. The Maori Yesterday and Today. Auckland: Whitcombe and Tombs.
  • Craig, R. D., 1989. Dictionary of Polynesian Mythology. New York: Greenwood Press.
  • Cuzent, G., 1872. Voyage aux Iles Gambier (Archipel de Mangareva). Paris: V. Masson.
  • Ellis, W., 1829a, 1829b. Polynesian Researches, 2 Vols. London: Dawsons of Pall Mall.
  • Emory, K. P., 1938. The Tahitian account of creation by Mare. Journal of the Polynesian Society, 47:45-63.
  • Englert, S., 1970. Island at the Centre of the World. London: Robert Hale.
  • Evans-Pritchard, E. E., 1965. The comparative method in social anthropology. The Position of Women in Primitive Societies and Other Essays in Social Anthropology. London: Faber and Faber. pp. 13-36.
  • Firth, R., 1961. History and Traditions of Tikopia. Wellington: The Polynesian Society.
  • —— 1967. The Work of the Gods in Tikopia, 2 Vols. London: Athlone.
  • Fornander, A., 1878, 1880, 1885. An Account of the Polynesian Race: its origin and migration and the ancient history of the Hawaiian people to the time of Kamehameha I, 3 Vols. London: Trubner.
  • —— 1916-1917, 1918-1919, 1919-1920. Fornander Collection of Hawaiian antiquities and Folklore, translated and edited by Thomas G. Thrum, 3 Vols. Bishop Museum Memoirs 4-6, Honolulu: Bishop Museum Press.
  • Forster, J. R., 1778. Observations Made During a Voyage Round the World. London: G. Robinson.
  • Fraser, J., 1892. The Samoan story of creation - A ‘Tala’. Journal of the Polynesian Society, 1:164-89.
  • Geraghty, P., 1993. Pulotu, Polynesian homeland. Journal of the Polynesian Society, 102:343-84.
  • Gifford, E. W., 1924. Tongan Myths and Tales. Bishop Museum Bulletin No. 8, Honolulu: Bishop Museum Press.
  • Gill, W. W., 1876. Life in the South Seas. London: Henry S. King.
  • —— 1911. Extracts from papers of the late Rev. W. Wyatt Gill, LL.D. Journal of the Polynesian Society, 20:116-51, 189-223.
  • Green, R. C., 1966. Linguistic subgrouping within Polynesia: the implications for prehistoric settlement. Journal of the Polynesian Society, 75:6-38.
- 257
  • —— 1988. Subgrouping of the Rapanui language of Easter Island in Polynesian and its implications for East Polynesian prehistory, in C. Cristino, P. Vargas, R. Izaurieta and R. Budd (eds), First International Congress, Easter Island and East Polynesia, Volume 1: Archaeology. Isla de Pascua (Easter Island), Universidad de Chile, Facultad de Arquitectura y Urbinismo, Instituto de Estudios. pp. 37-57.
  • Grey, G., 1885(1855). Polynesian Mythology and Ancient Traditional History. Auckland: H. Brent, Evening Star Office. 2nd edition.
  • Hale, H., 1971(1846). United States Exploring Expedition during the Years 1838-1842. New York: G.P. Putnam.
  • Handy, E. S. C., 1923. Native Culture in the Marquesas. Bishop Museum Bulletin 9, Honolulu: Bishop Museum.
  • —— 1927. Polynesian Religion. Bishop Museum Bulletin 34. Honolulu: Bishop Museum.
  • Henry, T., 1928. Ancient Tahiti. Bishop Museum Bulletin No. 48. Honolulu: Bishop Museum.
  • Hiroa, Te Rangi, 1932. Ethnology of Tongareva. Bishop Museum Bulletin No. 9. Honolulu: Bishop Museum.
  • —— 1938a. Vikings of the Sunrise. Philadelphia: J.B. Lippincott Company.
  • —— 1938b. Ethnology of Mangareva. Bishop Museum Bulletin 157. Hawaii: Bishop Museum.
  • Krämer, A. F., 1902. Die Samoa-Inseln, 2 Vols. Stuttgart: E. Schweitzerbart.
  • Langridge, M. and J. Terrell, 1988. Steinen's Marquesan Myths. Canberra: Target Oceania and The Journal of Pacific History.
  • Marck, J., 1993. Eastern Polynesian subgrouping today. Paper presented at the First Conference on Oceanic Linguistics, University of the South Pacific, Vila, Vanuatu.
  • —— forthcoming. Was there an early Polynesian “Sky Father?” Journal of Pacific History.
  • —— forthcoming2. A revision to the standard theory of Polynesian linguistic subgrouping and its culture history implications, in R. Blench and M. Spriggs (eds), Proceedings of the World Archaeology Conference. London and New York: Routledge.
  • —— in preparation. Polynesian language and culture history. PhD thesis, Department of Linguistics, Research School of Pacific and Asian Studies, Australian National University.
  • —— n.d. Deconstructing Polynesian religion and mythology, paper presented at the conference on Deconstructing the Island Group, Department of Pacific and Asian History, Research School of Pacific and Asian Studies, Australian National University, Canberra, December, 1994.
  • Moerenhout, J. A., 1837. Voyages aux iles du Grand Ocean, 2 Vols. Paris: A. Bertrand.
  • Montiton, A., 1874. Les Paumotous. Les Missions Catholiques, 6.
  • Pawley, A., 1966. Polynesian languages: a subgrouping based on shared innovations in morphology. Journal of the Polynesian Society, 75:39-64.
- 258
  • —— 1967. The relationships of Polynesian Outlier languages. Journal of the Polynesian Society, 76:259-96.
  • Savage, S., 1919. A history of Taniia-Nui. Journal of the Polynesian Society, 28:57-64.
  • Shortland, E., 1856. Traditions and Superstitions of the New Zealanders. London: Longmans.
  • —— 1882. Maori Religion and Mythology. London: Longmans Green.
  • Smith, S. P., 1899,1918-1921. History and traditions of Rarotonga. Journal of the Polynesian Society, 8, 27-30. (Five issues carried materials under this title. It is convenient to cite those articles by year and page).
  • —— 1903. Some Paumotu Chants. Journal of the Polynesian Society, 12:221-42.
  • —— 1918. Notes on the Mangareva, or Gambier Group of islands, Eastern Polynesia. Journal of the Polynesian Society, 27:115-31.
  • Tregear, E., 1904. The Maori Race. Wanganui: A.D. Willis.
  • Turner, G., 1861. Nineteen Years in Polynesia. London: J. Snow.
  • —— 1884. Samoa a Hundred Years Ago and Long Before. London: Macmillan and Co.
  • Tyerman, D., 1831a, 1831b. Journal of Voyages and Travels by the Rev. Daniel Tyerman and George Bennet Esq., 2 Vols. London: Westley and Davis.
  • Steinen, K. von den, 1898. Reise nach den Marquesas-Inseln. Verhandlungen der Gesellschaft für Erdkunde zu Berlin 25.
  • —— 1899. Ein Marquesanisch Sarg. Ethnologisches Notizblatt. Herausgegeben von der Direktion des königlichen Museums für Völkerkunde in Berlin, Bd II, Heft I. Berlin: Museum für Völkerkunde.
  • —— 1925-1928. Die Marquesaner und ihre Kunst: Studien iiber die Entwicklung primitiver Südseeornamentik, 3 Vols. (sp) Berlin: D. Reimer.
  • —— 1933. Marquesanische mythen. Zeitschrift für Ethnologie, 65:1-44, 326-73.
  • —— 1934. Marquesanische mythen. Zeitschrift für Ethnologie, 66:191-240.
  • Williamson, R. W., 1933a, 1933b. Religious and Cosmic Beliefs in Central Polynesia, 2 Vols. Cambridge: Cambridge University Press.
  • Wilson, W., 1985. Evidence for an Outlier source for the Proto-Eastern Polynesian pronominal system. Oceanic Linguistics, 27(1:2):85-l33.
  • Young, J. L., 1898. The origin of the name Tahiti. Journal of the Polynesian Society, 7:109-10.
  • —— 1919. The Paumotu conception of the heavens and of creation. Journal of the Polynesian Society, 28:209-21.
1   The parents of the first-order anthropomorphic gods or their direct ancestors in all Polynesian traditions but Rapanui, some divergent traditions centred in Tahiti, and the Outliers for which we generally have no information.
2   The Polynesians generally thought of the sky as solid and fixed above the earth in a dome whose edges came to rest upon the earth out to sea beyond the known world.
3   So far as I know Handy's (1927) framework was his own and seems undisciplined speculation at this point in time. It may have been within the framework of comparative mythology and comparative cosmogony as it was being conducted for other culture areas at the time but I am not familiar with that literature. We can, however, broadly characterise the intellectual atmosphere as one in which there was little knowledge and appreciation of the time depth separating the various cultures of the earth and a consequent expectation that links would eventually be established between many or most of them if only there were enough information. This lead him to assume that an Asian origin for Tangaloa would necessarily present itself to the trenchant investigator and of course he found one (Handy 1927:325).
4   Which Evans-Pritchard (1965) so solemnly dismissed.
5   See also Emory (1938:48-52) on the equivalence of New Zealand Maori Io or Iho and Tahitian 'Iho- 'Iho.
6   We can, for instance, speculate about the role of speakers of an “h-less” English dialect passing the Tahitian word and notion to the New Zealand Maori.
7   Or at least the condition at the time of the Primordial Pair.
8   “Cosmic beings” is a term applied by many of the late nineteenth century and early twentieth century descriptive and comparative sources to a recurring element in Polynesian cosmogonic traditions. These are the earliest “beings” in creation. They reproduce leading eventually to the Primordial Pair which have form but are not attributed with conciousness or deeds except in the cases of New Zealand Maori and Hawai'i. The cosmic beings lack deeds, conciousness, and sometimes even form. It might be appropriate to speak of them as personifications in sense of “embodiment” but not in the sense of anthropomorphic form as they have no such form in any instance.
9   Cosmogony - “a theory or story of the genesis or origination of the universe”. Cosmology - “the branch of philosophy that concerns itself with the origin and general structure of the universe, its parts, elements, and laws, esp. with such characteristics as space, time, and causality”. Definitions from The Macquarie Dictionary, 2nd Revision, Sydney: The Macquarie Library Pty Ltd, 1991. This article examines Polynesian traditions concerning the genesis of the cosmos more than it is concerned with their philosophical interpretation and is, therefore, a study of cosmogony rather than cosmology.
10   While Tongan Kele agrees with Samoan 'Ele there is the possibility of a borrowing and no reconstruction was made.
11   Here suggested to be a loan from Early Tahitian.
12   Not the Primordial Pair but hold a similar position.
13   Often it is *Maaui who accomplishes this feat and the specific reconstruction of *Maaui having done so for a drink of water in the cosmogony of PPN speakers is possible through agreements between Tongan, Samoan and Hawaiian as to that specific element or motif (Marck n.d.).
14   And Rangi-Aatea is a variant of his name. Thus the PTA reconstruction *Aatea ‘Primordial Male’.
15   Several Tongan creation traditions are known and differ somewhat, one having Limu and Kele in sexual union producing the rock Toui-A-Futuna from which burst forth cosmic beings (Williamson 1933a:9-ll and Craig 1989:28-9), while another (Collocott 1921:152-3) has Limu and Kele passing by Toui-A-Futuna which is a rock in the sea and magically causing it to burst forth with cosmic beings. Gifford (1924:19) relates a third, where Limu and Kele produce a male and a female child which mate and begin creation. All agree that Limu and Kele are the beginning and Hiku-Le'o, Tangaloa and Maui the first-order anthropomorphic gods that ultimately emerge. Gifford (1924:14) mentions other minor variations.
16   PPN *Maaui is reconstructed (Biggs 1994). The “a” is short in Tongan but not Niuean.
17   Craig (1989:28), citing Turner (1884:3-5,10) and Krämer (1902:7), would seem to have “power demons Saolevao and Saveasi'uleo” born directly to Papa-Tuu and Papa-'Ele. This is not supported from Turner (1884:3-5,10) and Craig's own wider reading of Krämer has Saolevao (or Salevao) as a child of Tua-Faile-Matagi and Papa-Tea (Craig 1989:242 citing Krämer 1902:8, 23, 75, 79-80, 105, 115) and Savea-Si'u-Leo as a child of Taufa and Alao (Craig 1989:243 citing Krämer 1902:104-8 and Turner 1884:259). Craig's (1989:28) characterisation therefore seems an abbreviation for his dictionary purposes and cannot be taken literally.
18   The personification or deification of space is common in Central Eastern Polynesia but the name, PCE *Aatea, is not cognate with Samoan Vale-Vale-Noa. See above in main text.
19   It is a particularly Nuclear Polynesian tradition that *Maaui obtains fire from PNP *Mafuike (Marck n.d.). It is also Maui who obtains fire in Tonga but a name similar to *Mafuike is not known from Tonga.
20   The agreement between MQA and MAO is imperfect. If the reconstruction is correct, MQA has irregularly lost the initial consonant. The MQA source (Handy 1923) is not otherwise known to have omitted initial h- in his transcriptions, MQA is not otherwise known to have lost an initial *h- nor is MAO otherwise known to have inserted one (Marck in preparation) so PCE *(H)aumia is the proper reconstruction.
21   We would expect **Mou'i but the single source (Handy 1923:244) did not record Marquesan glottal stops.
22   Marquesan has northern and southern dialects, the northern dialect reflecting PCE *ŋ as k and the southern dialect reflecting PCE *ŋ as n. Both reflect PCE *r as glottal stop and PCE *f as f or h.
23   The first vowel is short in Marquesan.
24   There are formal reasons for regarding Hawaiian mythology and religion as influenced by Tahitic in addition to the observation that Hawaiians remember only coming from Tahiti and replacing the menehune people (small people) on arrival. Most striking is the Hawaiian word kahuna ‘priest’ which is a borrowing from Tahitic. Marquesic continued PPN *tufuŋa where the first vowel was “u”. Only Tahitic shows the unexpected first vowel “a”. Thus the Hawaiian pronunciation is clearly borrowed from Tahitic.
25   We can reconstruct *Faka-Hotu or *Haka-Hotu on the basis of a Tuamotuan Hotu, Tahitian Ha'a-Hotu and Tongarevan Haka-Hotu. But the era to which it is reconstructed is called “Post-Tahitic” which refers to developments after the time of Proto-Tahitic in which language the Primordial Female was certainly *Papa. Thus the Tongarevan name is probably taken from the Tahiti area at a time after the divergence of New Zealand Maori and after Hawaiian borrowings from Tahitic.
26   Handy (1927) and Henry (1928) were both published by the Bishop Museum and Handy had access to Henry's manuscript while his book was in preparation.
27   Ellis' source was Orsmond which was ultimately published by Henry (1928) (Niel Gunson, personal communication).
28   Note that there was a X. Caillet and a A.-C.E. Caillot who both worked on matters concerning the Tuamotus and that they were different individuals.
29   And Hawaiian traditions after some years of contact. I am, however, inclined to believe that the Hawaiian demotion of Kanaloa is a post-European development. See section on Hawai'i.
30   Save Samoan Vale-Vale-Noa, mentioned in the section on PPN and PNP. But there is no similarity in his deeds or role to that of *Aatea in Central Eastern Polynesian groups.
31   “**” is used here to indicate an expected but unattested form.
32   He is even a child of Papa-Tuu-'Oi and Aatea in Henry's (1928) traditions… the only member of the common sibling group not simply conjured forth by Ta'aroa in these highly transformed traditions.
33   Although Gill (1876:19,23) mentions a mortal chief by that name from Rarotonga and another (possibly the same person) from Tahiti.
34   Craig (1989:101-2) misstates the sex of Aatea and Papa-Tuu-'Oi, which he has as male and female, respectively. His source (Henry 1928:364-369) clearly has them as female and male, respectively. While Aatea later changes gender (Henry 1928:372-6) and has more children as a male, none are cognate with first-order anthropomorphic gods elsewhere. Possibly Craig has glossed over this Tahitian idiosyncrasy for brevity. It is noteworthy that the female Tahitian Aatea changes gender by exchanging natures with Fa'a-Hotu, also a woman but with a masculine nature, and that Hotu is the name of the male Aatea's wife in Caillot's (1932) Tuamotuan materials (cf. Marck forthcoming). Haka-Hotu is the female spouse of the male Aatea in Tongareva as well (see Table 1).
35   HAW kahuna < PTA *kahuŋa not PMQ *tuhuŋa.