Volume 109 2000 > Volume 109, No. 3 > Shorter communication: The conical clan in Micronesia: The Marshall Islands, by Per Hage, p 295-310
THE CONICAL CLAN IN MICRONESIA: THE MARSHALL ISLANDS
In an analysis of rank and the conical clan in the Caroline Islands, Petersen (1999) argues that social stratification in Micronesia has been greatly exaggerated by some anthropologists, most recently by Hage (1998a) and Hage and Harary (1996). According to Petersen the ideology of genealogical seniority does not match the reality of chiefly rank and succession. Petersen believes that the structure of the conical clan is inherently ambiguous because genealogical seniority is crosscut by age and generation. He concludes that Micronesian societies, past and present, are “simultaneously stratified and unstratified” (1999:396) and that the conical clan operates in pretty much the same way in all Micronesian societies. He attributes the misconstrual of Micronesian social organisation to the influence of Murdock's (1948) reconstruction of Micronesian sociopolitical history and to the reliance of Hage (1998a) and Hage and Harary (1996) on limited source materials. Petersen's article contains a number of errors, misunderstandings and unsupported generalisations, especially with regard to Marshallese society. It is inimical to the comparative study of Micronesian and Oceanic societies, and to a theoretical understanding of the conical clan in general.
Social stratification in the Marshall Islands
In a typological, as opposed to a historical linguistic, reconstruction of sociopolitical history in Micronesia, Murdock (1948:16) described the development of a “feudal political organization of unstable petty states [with] stratification into several differentiated social classes”. Such states were found in the Marshall Islands. Murdock identified weaker forms of stratification in the atolls of the Carolines, a “germinal” form in Truk and a more elaborated stable form in Yap. A number of post-Second World War ethnographers, without any reference to Murdock's evolutionary scheme, described Micronesian societies as having descent groups based on genealogical seniority—in today's terminology the conical clan—with - 296 hereditary distinctions between chiefs and commoners, e.g., Alkire (1965), Burrows and Spiro (1957), Mason (1947, 1954), Riesenberg (1968) and Spoehr (1949). Mason and Spoehr described Marshallese society as stratified into three distinct social classes based on three types of matrilineages: bwij-in-iroij ‘royal lineages’, bwij-in-bwirak ‘noble lineages’ and bwij-in-kajur ‘commoner lineages’. Some clans consisted only of commoner lineages while others consisted of all three types. They came to this conclusion on the basis of ethnographic evidence, traditional history and numerous early accounts, such as those contained in the Ergebnisse der Südsee Expedition 1908-10 (Thilenius 1913-38). Spoehr's characterisation of traditional Marshallese society, like that of Murdock, was essentially based on Erdland's comprehensive ethnography published in 1914, translated in part in 1942 in connection with the Navy Pacific Island Handbook Project. It is worth quoting Spoehr's summary in extenso:
In former times, the social distinctions that underlay this class system were very real indeed. The paramount chief was possessed of autocratic powers that were shared to a lesser extent by the nobility. The paramount chief and his nobles were the leaders in war and in sailing expeditions. They controlled the land and the fruits thereof. They provided the primary leadership of the community, and in turn enjoyed the privilege of being fed and supported by the commoners.
Of the noble class, the paramount chief himself was accorded the greatest respect. His position involved the hereditary acquisition of magical power, somewhat similar to Polynesian mana. He was approached only in the most deferential manner; in his presence persons walked stooped over, or moved on their knees. Of all the members of the community he was supposed to command the best information on the affairs of the Marshallese world. In recompense for his inherited responsibility, he received the best of the food produced on the land or caught in the sea. He lived in the most favored location. His lineage had its own cemetery. No restrictions were placed on the number of his wives and he had access to all commoner women. Over his people he exercised autocratic powers.
The commoners were the workers of the land, the fishermen, the sailors, and the ordinary fighting men. With the possible exception of the alabs—the heads of the commoner class were the workers (ri-jerbal) in every sense of the word. Their tribute supported the nobility. Their houses were built in less favorable parts of the island. Nor were they permitted the distinctive tatooing and the finer dress of the nobility (Spoehr 1949:77, also quoted by Oliver 1989:995 in his encyclopaedic account of Oceanic societies).
Spoehr (1949:77) expressed some reservations about the “despotic nature of chiefly rule” but otherwise did not doubt the reality of a class system in - 297 traditional Marshallese society. There is no reason to believe that Erdland, who lived for many years in the Marshalls during the German administration (1885-1914), and who spoke Marshallese and wrote a grammar and dictionary of the language (Erdland 1906), was deceived about Marshallese society nor that Mason and Spoehr were less than scrupulous in their reading and interpretation of earlier accounts of Marshallese society.
According to Petersen (1999:371), the “archetype of the Micronesian leader is one of gentle authority”. This is not the picture of Pohnpeian and Marshallese chiefs presented by Riesenberg (1968) and Erdland (1912). “Aggrandisement and avarice of [Pohnpeian] chiefs seem to have had few checks… a high chief might confiscate any article he wished simply out of greed, not necessarily as an act of punishment. Such an act was known as kuhl” (Riesenberg 1968:63-64). Erdland (1912:559) describes specific instances of the Harthertzigkeit (hardheartedness) of Marshallese chiefs toward their Untertanen (subjects). Not for nothing were Marshallese chiefs compared to frigate birds—“large tropical sea birds (genus Fregata) with extremely long wings and tail and a hooked beak: it commonly robs other birds of their prey” (Webster's New World Dictionary).
Structure of the Conical Clan
According to Petersen (1999:387), “There is a fundamental contradiction in the abstract principles of conical clanship”, because genealogical seniority is crosscut by generation and age. In fact, there is no such contradiction. The conical clan is based on a relation of unigeniture, ultimogeniture as in Kachin society (Leach 1954), and primogeniture as in many Oceanic societies. Unigeniture overrides generation and, theoretically, assigns every individual a unique rank. Not all anthropologists have understood this point. White (1959), for example, tried unsuccessfully to construct an abstract model of the conical clan that combines genealogical seniority and generation. The result was self-contradictory as shown in Hage and Harary (1996:108). Many anthropologists who have written about the conical clan, Fried (1967) and Sahlins (1958) for example, have not given a clear and consistent description of rank, and have used misleading shorthand expressions such as “distance from the founding ancestor or from the senior line of descent” (see Hage and Harary 1996:108).
In the patrilineal variant of the conical clan, as found in Tonga (Gifford 1929, Sahlins 1958), genealogical rank is traced exclusively through males: a father outranks a son, an elder brother outranks a younger brother, the male descendants of an elder brother outrank the male descendants of a younger brother, and the sons of an elder brother outrank the father's younger brother. The rank relation is a complete order (Hage and Harary 1996). 1 - 298 Every younger brother is a potential founder of a junior line of descent. Rank and succession are theoretically identical and succession is vertical. Thus: “In nearly all cases the succession of the Tu'i Tonga was from father to son, and it was the eldest son of the moheofo [principle wife] who succeeded. Normally there was no discussion about who should succeed; it was automatic” (Bott 1982:99). 2.
In the matrilineal variant of the conical clan, as found in the Marshalls, Pohnpei and some atolls of the Carolines, rank is traced exclusively through females: a woman outranks her daughter, an elder sister outranks a younger sister, the female descendants of an elder sister outrank the female descendants of a younger sister, and the daughters of an elder sister outrank the mother's younger sister. Women determine rank, but men (normally) succeed to the chieftainship. Petersen's (1999:387) hypothetical case in which an elder sister's younger sons are said to be genealogically superior to her younger brother is completely beside the point since it is the relative age of sisters that counts. In the Marshalls, the ideal successor is the eldest son of the eldest sister. If this were always the case succession would be vertical but the Marshallese conical clan has a lateral inflection: elder brothers are, ideally, succeeded by younger brothers before the title passes to the eldest sister's sons. Lateral succession is not a problem in the matrilineal variant of the conical clan because only sisters can found new descent lines.
As evidence for ambiguity in the structure of the conical clan Petersen cites Kiste's (1974) work on the exiled Bikini people.
When the head of a senior lineage was a member of a generation below that of the head of a junior lineage (in such instances, the head of the senior lineage was usually the classificatory maternal nephew of the head of the junior lineage), the principles did not specify either as being superior in rank, and there were no rules determining succession (Kiste 1974:52-3).
Bikini is a relatively isolated atoll in the poorer northern Marshalls. According to Mason (1954), Bikini never had a paramount chief of its own. In Arno, a larger atoll in the richer southern Marshalls, there may have been some uncertainty about rank within the commoner lineage but “within the aristocratic lineage, there is an expectation that males of junior branches [but of higher generation] might never succeed to the chieftaincy” (Rynkiewich 1972:64). Kiste, incidentally, characterised Marshallese society as highly stratified:
In the pre-contact era, most of the atolls in the Marshalls were divided among the realms of several paramount chiefs. Each of the chiefs was the head of a - 299 chiefly matrilineage, and he and his lineage mates comprised the highest ranking strata of a privileged social class. (The class system is best described by Mason 1947.) The amount of power and influence possessed by a paramount chief was in direct proportion to the number of atolls and islands in his domain and the number of kajur under his authority. Kajur was the class of commoners to which the majority of islanders belonged, and the term also denoted the power or strength of a chief.
A paramount chief maintained control over his realm by his success and reputation as a war leader. The several chiefs warred against one another as each attempted to extend his domain, and conflicts within the same chiefly lineage were commonplace. The extent of the respective domains of the several chiefs varied constantly with the fortunes of war (Kiste 1974:20).
In this connection, Petersen cites with approval Peoples' (1993) study of political evolution in Micronesia, but Peoples similarly described not only the Marshall Islands, but also Pohnpei, Kosrae and the northern Gilberts (Kiribati) as highly stratified:
Yet some low island residents of eastern Micronesia evolved levels of complexity comparable to those of Pohnpei and Kosrae. The southern Marshalls were linked into several multi-island polities headed by powerful land-claiming and tribute-taking hereditary chiefs. Commoners worked the land under the supervision of the heads of their matrilineages, and the paramount chief visited the islands of his domain periodically to collect tribute and supervise affairs (Mason 1954; Rynkiewich 1972; Spoehr 1949). Although somewhat different in structure, the low islands of northern Kiribati also developed complex political systems (Lambert 1966, 1978) (Peoples 1993:7).
A clear ethnographic example of matrilineal succession, and also of rivalry between junior and senior lines of the conical clan, is given in Spoehr's (1949:82-85) monograph on Majuro Atoll in the Ratak chain of the Marshalls. Just before it came under German control in 1885, Majuro had a single paramount chief, Lerok (Fig. 1). Lerok had two maternal nephews, Rimi, the son of an elder sister, and Jebrik, the son of a younger sister. According to Marshallese custom the chieftainship should have passed to Rimi, the elder sister's son, but Lerok decided before his death to divide the atoll between his two nephews. After Lerok's death Rimi and Jebrik struggled for control. Rimi was succeeded by his sister's son Kaibuki. Jebrik and Kaibuki fought the last “war” on Majuro but the division between the two lines remained. Kaibuki was succeeded by his younger brother, Kaibuki Larewa, followed by his next younger brother Kaibuki Mourjel. The latter was succeeded by his sister's son Reli who was succeeded by his younger - 300 brother Langlan, the last chief in an unbroken line of succession from Lerok. Langlan will be succeeded by his sister's son Aisea. Theoretically, Aisea's elder brother should succeed but he is a cripple. In the other senior line of descent, Jebrik was succeeded by his brother Jokane who was followed by another brother Lukuner (relative ages not specified). Lukuner was succeeded by his sister's son Lukutwerak who died in 1919 without an heir, ending this line. Twenty years later Lukutwerak's side selected Jitiam, the chief of a junior lineage, to succeed. “Jitiam himself is not in the direct line of succession from Lerok. For this reason, Langlan's followers are inclined to say that there really is only one paramount chief on Majuro, namely, Langlan” (Spoehr 1949:85).
Family Tree. elder sister, Lerok, younger sister, Rimi, Jebrik, Jokane, Lukuner, Kaibuki Mourjel, Kaibuki Larewa, Kaibuki, Lukutwerak, Langlan, Reli, Jitiam, Aisea, Figure 1: Chiefly succession in Majuro (from Spoehr 1949)
Not all cases of chiefly succession are so tidy. In the 19th and early years of the 20th centuries, the Ralik chain of the Marshall Islands was dominated by the royal Bwij-in-Iroij lineage. Succession to the title of paramount chief as reconstructed by Mason (n.d.) is shown in Figure 2 (see also Hage 1998b). The legendary founding ancestress of the principal Ralik- 301
Family Tree. Bwij-in-Iroij lineage, Erroja lineage, Lijjeleijel, Liwatoinmour, Jemaliut, Iroij, Litarao, Letalju, Leom, Lailiju, Lanini, Liwaju, Limijwa, Lamarein, Loj, Libokean, Lamari, Kaibuke, Lajutok, Loeak, Litalaur, Kabua, Leit, Rijjino, Figure 2: Chiefly succession in the Ralik chain of the Marshalls (based on data in Mason n.d.).
clans was a woman named Liwatoinmour. Her daughter, Iroji, was the ancestress of the Bwij-in-Iroij lineage. Iroij's son, Letalju, was the first paramount chief. Around 1800 Letalju was challenged, unsuccessfully, by his generationally senior, but genealogically junior “uncle”, Jemaliut. Letalju was succeeded by his sister's younger son Lanini, who evidently pre-empted the position of his elder brother by his outstanding service to his uncle and his unusual talent for leadership. We note with respect to Petersen's hypothetical example of structural ambiguity in the conical clan that Letalju was evidently the younger brother of his sister Litarao. This did not imply that Letalju was outranked by his elder sister's sons. On the contrary, the latter were the designated successors of the former. Lanini's rule does, however, illustrate the possibility of vertical conflict in the conical clan. Lanini married the niece of a rival paramount chief of another royal lineage and persuaded her to enflame a rivalry between her brothers and their mother's brother. The sister's sons assassinated their mother's brother and were in turn eliminated by their mother's brother's allies ending this royal lineage (Mason n.d.). Avuncucide may be a possibility in the matrilineal conical clan while patricide seems less likely in the patrilineal variant. Lanini's sister's son and logical successor, Lamarein, may have died before Lanini, so the title skipped a generation. Lanini's sister, Leom, had two daughters—Liwaju (elder) and Limijwa (younger). The title should have passed to Liwaju's sons but Liwaju raised her children in the northern islands away from the centre of action. In the event, Limijwa's eldest son by her second marriage, Kaibuke, succeeded. Kaibuke's logical successor was his younger brother Lajutok who either died before he could assume the title or, more probably, chose to align himself and live with his wife's clan (not a bad move for a younger brother especially since his elder brother, Kaibuke, ruled for almost 30 years from about 1840 to the late 1860s). Kaibuke was evidently succeeded by his eldest sister's son Loeak, who was continually challenged by his younger sister's son Kabua (illustrating the Marshallese proverb, “Children of brothers love one another, children of sisters fear one another” [Erdland 1914]). Loeak died in 1904 followed by Kabua in 1910. Loeak's sister's son, Rijjino, died before he could succeed his uncle. Kabua was succeeded by his brother Leit. There were no other matrilineal heirs so the royal Bwij-in-Iroij lineage came to an end with Leit's death in 1914.
Figure 2 illustrates an important but often neglected correlate of the conical clan. In many societies, a fully functioning conical clan system is associated with matrilateral cross-cousin marriage. In Kachin such marriages are hypogamous with women marrying down (Leach 1954). In Tonga, such marriages are hypergamous with women and tribute moving up as in marriages between sisters of the genealogically junior line of secular chiefs, - 303 the Tu'i Kanokupolu, and men of the genealogically senior line of sacred chiefs, the Tu'i Tonga (Bott 1982, Gifford 1929). In both societies matrilateral marriage is mainly a chiefly affair, important for establishing or maintaining relations of political alliance. In Island Networks (Hage and Harary1996), we hypothesised that matrilateral cross-cousin marriage once existed in Marshallese society. The hypothesis was confirmed by data in Mason's (n.d.) reconstruction of title succession in the Ralik chain (Hage 1998b). The dotted lines in Fig. 2 show some of the hypergamous matrilateral marriages between men of the royal Bwij-in-Iroij lineage and women of noble (bwirak) Erroja lineages. 3. Such marriages were a foundation of chiefly power. As Erdland wrote, “through marriages and inheritance each iroij has one or another buirak who pays tribute to him. The subchief puts his subjects at the high chief's disposal for free labour, fights for him, and supports him in the more important matters” (1914:73).
The conical clan provides a framework of social stratification with major implications for land tenure, economic exchange and chiefly sanctity. As Goodenough (1959:257) wrote in a review of Sahlins (1958), “the model of the ‘ramage’ [conical clan] enables us to see clearly the basic structural design shared by many Polynesian societies” and now, we would add, by many other Oceanic societies. Of course genealogical seniority may be overridden or contested on the grounds of generation, age and competence; heads of junior lineages may attempt to usurp the position of heads of senior lineages; and those who are successful may rearrange genealogies to validate their status. Who would argue with that? The point is that these contingencies are interpreted in terms of the framework of the conical clan. Anthropologists who have written about the conical clan have usually emphasised these points, e.g., Bott (1982), Gifford (1929), Hage (1998), Leach (1954). 4
Variations in Social Stratification
Conical clan societies vary with respect to the degree of social stratification. Not all Micronesian societies are like Pohnpei or the atolls of the Carolines as Petersen seems to imply. In Island Networks, we tried to describe these variations from an historical point of view. We hypothesised, on the basis of historical linguistic and comparative ethnographic evidence (Hayden 1978, Pawley 1982, Stillfried 1953), that Proto Nuclear Micronesian society, like Proto Polynesian society (Kirch 1984), Proto Oceanic society (Hage 1999), and probably Proto Malayo-Polynesian society (Bellwood1995) 5 was based on the conical clan. According to Petersen (1999:373), individuals who wish to succeed to titles in Micronesian systems “must engage in competitive giving”. This was true of Pohnpeian society (Riesenberg 1968), but not of the more stratified Marshallese society. In - 304 some atolls of the Carolines, Ifaluk for example, genealogical seniority distinguished chiefs from commoners but not chiefly from commoner classes (Burrows and Spiro 1957). In the isolated island of Nauru the conical clan was found, but without its characteristic functions in land tenure, and without significant chiefly privileges and prerogatives (Wedgwood 1936). In the northern Gilberts (Kiribati), the conical clan became patrilineal in form and operated on a smaller scale than elsewhere in Micronesia and Polynesia (Lambert 1978). In Truk, the conical clan disappeared entirely leaving matrilineal clans without distinctions between senior and junior lines of descent, with headship based on age rather than seniority (Murdock and Goodenough 1947). What Murdock (1948) saw in Truk was not the evolution of a class society but rather the devolution of a stratified Nuclear Micronesian society. Outside of Micronesia, in Hawai'i, traces of the conical clan remained in the “royal cum cosmological genealogies which, beginning in divine sources and proceeding patrilaterally through senior and cadet branches, fix the dynastic relations between the several islands” (Sahlins 1985:20). But here genealogical seniority became only one among several “arguments” in laying claim to high status, other arguments being alternative bilateral tracings of descent, descent from primary or secondary spouses, and the number of lines through which descent could be traced (Sahlins 1985, Linnekin 1990).
Independent Discoveries of the Conical Clan
The concept of the conical clan has proven elusive in the history of kinship studies. Petersen notes that MacLeod (1931) anticipated Kirchhoff's discovery of the conical clan, made in 1935, but not published until 1955 and later, and more accessibly, in 1959 (in Fried). As Service (1985) points out, there have been many independent discoveries of the conical clan including Fustel de Coulange's (1864) Greco-Roman gens, Firth's (1936) Polynesian ramage, Leach's (1954) Kachin gumsa lineage and Oberg's (1955) lowland South American chiefdom. We would add to this list Gifford's (1929) Tongan ha'a lineage and Mason's (1954) Marshallese clan. Gifford's and Leach's contributions are especially important because they connected the conical clan with matrilateral cross-cousin marriage. Mason's contribution is notable because he devised a numerical code that gives an exact definition of genealogical rank. This code is isomorphic to the mathematical model of a depth-first search tree (Hage and Harary 1996). The conical clan is also known in Oceanic anthropology as a “pyramidal descent group” in Fiji (Sayes 1984), and as a “descent group structured by genealogical seniority” in New Caledonia (Douglas 1979). A common nomenclature would have been helpful.- 305
Petersen cites Tikopia as an ethnographic case of an Oceanic society in which succession does not always, or even most of the time, follow a rule of primogeniture. Tikopia (like Bikini which he also cites) is a weakly stratified society. Thus, when Leach looked for a Polynesian exemplar of the conical clan he noted a formal but not a very substantial resemblance between the Tikopia clan and the Kachin gumsa lineage:
What makes the Kachins particularly interesting from an anthropological point of view is that they have a society which is simultaneously segmentary and class stratified. In most types of lineage system that have so far been described in any detail, the process of lineage segmentation leads to a ‘balanced opposition’ between the resulting segments rather than to a status ranking, superior and inferior. For this reason, among others, the interesting typology of political systems which Fortes and Evans-Pritchard have suggested for African societies would not cover Kachin gumsa society.
The Tikopia, as described by Firth, have indeed what may be considered a “pure” lineage system associated with notions of a class hierarchy, but here the whole scale of social activities is on such a minute scale that analogy is not very useful. I think that there are plenty of societies in the world of the Kachin gumsa type, but it so happens that social anthropologists have not yet got round to looking at them. That makes it all the more difficult for me to achieve lucidity (Leach 1954:159).
If Leach had read Gifford's account of the conical clan in Tonga he would have found a more suitable analogy to the Kachin gumsa lineage and to the Kachin system of chiefly matrilateral cross-cousin marriage. And if Mason had read Gifford's account of the Tongan lineage instead of Evans-Pritchard's (1940) account of the Nuer segmentary lineage when he tried to achieve lucidity about the Marshallese clan, he would have recognised a common type of descent group in eastern Micronesia and western Polynesia. If the presence of the conical clan had been recognised in New Caledonia and Fiji as well, Oceanic anthropologists would have had a unifying structure analogous to the archaeologist's Lapita complex.
Near the conclusion of his article Petersen criticises Hage (1998a, and by implication Hage and Harary 1996) and Mosko (1998) for their reliance upon limited source materials in their respective analyses of Micronesian political process. This hurts. Speaking for Hage and Harary, our aim was to give a comparative analysis of descent and marriage alliance in traditional Nuclear Micronesian-speaking societies. For that purpose we relied on several dozen sources, far more than those cited in Petersen's article. In - 306 Petersen's characterisation of Marshallese social stratification there seems to be an undue reliance on Kiste's (1974) Bikini monograph. Other important sources, such as Spoehr (1949), Mason (1947, 1954), Erdland (1914), Krämer and Nevermann (1938), Tobin (1967) and Rynkiewich (1972), are not mentioned at all. By not using these sources Petersen creates a picture of traditional Marshallese society that earlier generations of anthropologists would not recognise. By characterising all Micronesian societies as “simultaneously stratified and non-stratified” he ignores significant variations in social stratification. By confusing genealogical seniority and generation he misinterprets the structure of the conical clan. The downside of Murdock's real influence on Micronesian, and Oceanic, anthropology was not his illusions about social stratification, but rather his typological (non-linguistic) reconstruction of Proto Malayo-Polynesian, Micronesian and Oceanic society as unstratified and Hawaiian in type (Hage and Harary 1996), his lack of interest in the difference between ranked and egalitarian types of descent groups (Service 1985), and his treatment of marriage alliance as a secondary aspect of social structure (Hage and Harary 1996). The upside, of course, was his focus on the fundamental importance of kinship systems in Oceanic societies. It may well be that modern political process in Micronesia emphasises “'democratic, participatory and egalitarian aspects” as Petersen (1999:371) has experienced, but that is no reason to dismiss the accumulated ethnographic record.- 307
- 310 Page is blank
1 In Tonga, sisters outrank brothers not on the grounds of genealogical seniority but as a consequnce of hypergamy (Hage and Harary 1996).
2 “It is said that Tu'i Tonga Talatama… did not have a son, and was succeeded by his brother Tala-'i-Ha'apepe. But succession was always supposed to go from father to son in the Tu'i Tonga line, and so before Tala-'i-Ha'apepe could succeed, a fictitious king was made of wood, Nui Tamatou, Nui Tamatou then ‘died’, and then Tala-'i-Ha'apepe, his ‘son’ was declared Tu'i Tonga” (Bott 1982:94)
3 These unions included marriage with the mother's brother's widow as shown on Figure 2 for Lanini. Such marriages are common in matrilineal chiefdoms practising generalised exchange (Hage and Harary 1996, Hage 1998b, Rosman and Rubel 1981). In the Trobriands the successor to the district chieftainship inherits his mother's brother's wives (Malinowski 1932)
4 In Tonga, Bott emphasises the inportance of ability and political support by the mother's relatives in the succession of the “working king”, the Tu'i Kanokupolu. She also notes instances of lateral succession.
5 In an effort to provide a balanced view, Petersen (1999:377-79) accepts historical linguistic evidence for “some fundamental aspects of social stratification or rank” in Proto Malayo-Polynesian society, but he also agrees with Sutton's (1990) rejection of historical linguistic reconstructions of terms for hereditary leadership. A strong argument against Sutton is given by Green (1994) in the case of Proto Polynesian *qariki ‘chief’